Ultrasound-Assisted Synthesis of Fe³⁺/Zr⁴⁺-Modified Layered Double Hydroxides for RSM-Optimized Fluoride Remediation: Structural Insights and Evaluation in Groundwater

Abstract

This study investigates the structure–performance relationship of Fe³⁺- and Zr⁴⁺-modified layered double hydroxides (LDHs) for fluoride removal from water. Mg–Al LDHs with different metal loadings (Zr0.05, Zr0.1, Fe0.8, and Fe1) were synthesized via ultrasound-assisted coprecipitation and characterized using XRD, SEM–EDS, FTIR, XPS, and N₂ physisorption. Among the synthesized materials, Zr0.05-LDH exhibited the highest adsorption performance. Response surface methodology identified adsorbent dosage as the most influential parameter, achieving a maximum fluoride removal efficiency of 98.17% under optimal conditions (pH ≈ 5, adsorbent dose of 0.88 g/L, and initial fluoride concentration of 12.6 mg/L). Zr0.05-LDH showed the largest specific surface area (261 m²/g) and a maximum adsorption capacity of 137 mg/g, as described by the Langmuir isotherm model. Kinetic studies indicated rapid adsorption, with equilibrium reached at approximately 180 min. Fluoride removal was governed primarily by inner-sphere complexation at Zr⁴⁺ and Fe³⁺ sites, accompanied by anion exchange and electrostatic interactions. The adsorbent retained 89% of its capacity after five regeneration cycles. Groundwater tests from Durango, Mexico, demonstrated effective fluoride reduction below Mexican and WHO guideline limits despite competing anions. These results demonstrate the potential of modified LDHs for fluoride-contaminated groundwater treatment.

1. Introduction

Fluoride contamination in drinking water is a major environmental and public health concern, particularly in regions of Asia, Africa, and Latin America, where groundwater is the primary source of drinking water [1,2]. Fluoride (F⁻), owing to its small ionic radius, high electronegativity, and chemical stability, can be readily mobilized under specific hydrogeochemical conditions. Its occurrence in groundwater is commonly associated with the dissolution of fluoride-bearing minerals such as fluorapatite, fluorite, and cryolite, as well as with anthropogenic sources including phosphate fertilizers, industrial emissions, and untreated wastewater discharges [3,4]. In arid and semi-arid regions, elevated pH, high alkalinity, and limited calcium availability enhance fluoride release by suppressing Ca²⁺ coprecipitation and promoting OH⁻ substitution within mineral structures [5]. Low concentrations of fluoride in drinking water may be beneficial for dental health. However, prolonged exposure can lead to dental and skeletal fluorosis. The World Health Organization (WHO) recommends a maximum fluoride concentration of 1.5 mg/L in drinking water [6]. Therefore, it is important to develop efficient defluoridation technologies [7].

Various technologies have been applied for fluoride removal, including chemical precipitation, coagulation, ion exchange, electrocoagulation, nanofiltration, and reverse osmosis [8,9]. Although effective, these methods are often limited by high operational costs, significant energy requirements, membrane fouling, and the generation of secondary waste streams [10]. As a result, adsorption has gained considerable attention due to its operational simplicity, cost-efficiency, and high removal efficiency [11,12]. However, its performance is strongly dependent on adsorbent properties, including structural stability, surface chemistry, and susceptibility to competing ions [13].

Among emerging adsorbents, layered double hydroxides (LDHs) have attracted significant attention because of their tunable composition, high anion-exchange capacity, and structural versatility [14,15]. However, pristine LDHs often exhibit limited adsorption capacity, low selectivity, and structural instability under variable aqueous conditions, which restricts their practical application [16]. Consequently, recent studies have focused on modifying LDHs through metal cation substitution to improve adsorption performance [17]. Iron-modified LDHs (Fe–LDHs) have shown enhanced fluoride adsorption because of increased surface charge density and strengthened electrostatic interactions and surface complexation mechanisms [16]. Similarly, zirconium-based materials are widely reported to exhibit a high affinity for fluoride, which may be associated with the formation of inner-sphere complexes involving Zr–F interactions [18].

The development of LDH-based adsorbents has often been guided by the assumption that increased metal incorporation results in enhanced adsorption performance. However, this assumption remains largely unverified and may be overly simplistic. Adsorption efficiency does not necessarily increase linearly with metal content but instead depends on the interplay among metal dispersion, structural organization, and active-site accessibility. These considerations highlight limitations in prevailing design strategies and emphasize the importance of mechanistic understanding for the rational development of high-performance adsorbents.

Nevertheless, the relationship between metal incorporation, structural organization, and adsorption performance remains insufficiently understood, particularly in LDH systems modified with high-valence cations such as Zr⁴⁺. Potential non-linear effects of metal loading, together with trade-offs among active site density, metal dispersion, and structural stability, are still not fully understood, which makes the rational design of optimized adsorbents more challenging.

Accordingly, this study systematically evaluates the effect of Fe³⁺ and Zr⁴⁺ incorporation on the structure–performance relationships of LDH-based adsorbents for aqueous fluoride removal. The adsorbents were synthesized and characterized, and response surface methodology (RSM) was applied to optimize the adsorption conditions and assess the factors affecting fluoride removal. The results showed that adsorption performance was more strongly related to the balance between structural crystallinity and metal dispersion than to the total amount of metal incorporated into the material. These findings suggest that the prepared materials are promising adsorbents for fluoride removal in water treatment applications.

2. Materials and Methods

2.1. Materials

Analytical-grade reagents were used for the synthesis of LDHs, including magnesium nitrate hexahydrate (Mg(NO₃)₂·6H₂O, Meyer, Mexico City, Mexico; 98%), aluminum nitrate nonahydrate (Al(NO₃)₃·9H₂O, Sigma-Aldrich, St. Louis, MO, USA; 99%), iron(III) nitrate nonahydrate (Fe(NO₃)₃·9H₂O, Sigma-Aldrich, St. Louis, USA; 98%), and zirconium oxynitrate hydrate (ZrO(NO₃)₂·H₂O, Sigma-Aldrich, St. Louis, USA; 98%). Sodium carbonate (Na₂CO₃, J.T. Baker, Phillipsburg, NJ, USA; 99%), sodium hydroxide (NaOH, 97%, Karal, Mexico City, Mexico; 97%), and hydrochloric acid (HCl, Fermont, Monterrey, Mexico; 37%) were used for pH adjustment. Fluoride stock solutions were prepared from sodium fluoride (NaF, Sigma-Aldrich, St. Louis, USA; 98.5%). Deionized water was used in all experiments.

2.2. Synthesis of Adsorbents (LDHs)

Layered double hydroxides (LDHs) with a nominal Mg/Al molar ratio of 3:1 were synthesized as reference materials. Metal-modified LDHs were prepared by incorporating Fe³⁺ (0.8 and 1 mol) into the Mg–Al framework while maintaining a total trivalent cation content of 1 mol. For Zr⁴⁺-modified materials (0.05 and 0.1 mol), the combined Al³⁺ + Zr⁴⁺ content was fixed at 1 mol to preserve charge balance within the layered structure.

All materials were synthesized by ultrasound-assisted coprecipitation. Briefly, a mixed metal nitrate solution was added dropwise to a 500 mL Erlenmeyer flask containing 50 mL of 0.5 mol/L Na₂CO₃ under vigorous stirring. Simultaneously, a 1 mol/L NaOH solution was added to maintain the pH at 10. The resulting suspension was subjected to ultrasonic irradiation (130 W, 20 kHz) for 30 min, followed by aging at 65 °C for 18 h under continuous stirring (170 rpm).

After aging, the precipitates were repeatedly washed with deionized water until neutral pH was achieved, then filtered and dried at 80 °C for 12 h. The obtained solids were designated as LDH, Fe0.8-LDH, Fe1-LDH, Zr0.05-LDH, and Zr0.1-LDH according to their composition. Prior to use, the materials were ground and passed through a 200-mesh sieve (~75 µm) to ensure a uniform particle size distribution, followed by calcination at 500 °C for 4 h to obtain the corresponding mixed metal oxides. Despite calcination, these materials are referred to as LDH-based materials due to their origin and their ability to undergo partial reconstruction in aqueous media via the memory effect.

2.3. Characterization of Adsorbents

The physicochemical properties of the synthesized adsorbents were characterized using complementary analytical techniques.

X-ray diffraction (XRD) patterns were collected using a Bruker D8 Advance diffractometer (Bruker, Karlsruhe, Germany) equipped with a Göbel mirror, Cu Kα radiation (λ = 1.5406 Å), and a NaI detector. Measurements were conducted over a 2θ range of 3–80°, with a step size of 0.02° and a scanning rate of 2°/min.

Surface chemical composition and oxidation states were analyzed by X-ray photoelectron spectroscopy (XPS) using a Thermo Scientific K-Alpha spectrometer (East Grinstead, UK). The instrument was operated under ultra-high vacuum conditions (2.4 × 10⁻⁷ mbar) with a monochromatic Al Kα source (1486.68 eV). The analyzed area had a diameter of approximately 400 μm.

Morphological characteristics were examined using a field-emission scanning electron microscope (FE-SEM, JEOL JSM-7800F Prime; Tokyo, Japan). The elemental composition and its distribution were analyzed by energy-dispersive X-ray spectroscopy (EDS, Bruker) coupled with the microscope.

The textural properties of the samples, including specific surface area, pore volume, and pore size distribution, were determined from N₂ adsorption–desorption isotherms using a Micromeritics ASAP 2020 analyzer (Norcross, GA, USA). Prior to analysis, the samples were degassed at 100 °C for 12 h. The specific surface area was calculated using the Brunauer–Emmett–Teller (BET) method.

The functional groups present on the adsorbents were identified by Fourier transform infrared (FTIR) spectroscopy. Spectra were recorded using a Shimadzu IRAffinity-1S spectrometer (Shimadzu Corp., Kyoto, Japan) equipped with a GS10800 attenuated total reflectance (ATR) accessory (Specac Ltd., Orpington, UK). Measurements were performed in ATR mode over the spectral range of 400–4000 cm⁻¹, with a resolution of 2 cm⁻¹ and 45 scans per sample.

2.4. Batch Adsorption Experiments

Batch adsorption experiments were conducted in Erlenmeyer flasks using orbital shaking at 150 rpm and 25 ± 5 °C. After adsorption, water samples were collected and filtered through a 0.45 μm syringe filter prior to fluoride analysis. The residual fluoride concentration was measured using an ion-selective electrode (Orion 9609BNWP, Thermo Fisher Scientific Inc., Waltham, MA, USA) after the addition of TISAB III. Each experimental condition was performed in duplicate to ensure reproducibility. The fluoride removal efficiency (%) and adsorption capacity (q_e, mg/g) were calculated using Equations (1) and (2).

$$Removal \left(\%\right)={\displaystyle \frac{(Co-Ce)}{Co}}\times 100$$

(1)

$$q_{e }(\mathrm{m}\mathrm{g}/\mathrm{g})={\displaystyle \frac{Co-Ce}{m}}\times V$$

(2)

where Co and Ce (mg/L) represent the initial and equilibrium fluoride concentrations, respectively, while m (g) and V (L) denote the mass of adsorbent and the solution volume.

Initial adsorption experiments were conducted using LDH as the reference material, and Fe1-LDH, Fe0.8-LDH, Zr0.1-LDH, and Zr0.05-LDH. In each test, 20 mg of adsorbent was dispersed in 20 mL of a fluoride solution (10 mg/L), corresponding to an adsorbent dosage of 1 g/L. Experiments were performed at the natural solution pH (≈6) with a contact time of 240 min.

2.4.1. Experimental Design (RSM)

Based on the adsorption experiment results, fluoride removal using Zr0.05-LDH was optimized using a Central Composite Design (CCD) within the Response Surface Methodology (RSM) framework. The independent variables were initial fluoride concentration, pH and adsorbent dosage, each evaluated at three levels (

Table 1. Independent variables and their levels used in the Response Surface Methodology (RSM) design for fluoride adsorption optimization.

Independent Factors	Levels
	−1	0	1
Initial fluoride concentration (mg/L)	5	10	15
pH	4	5	6
Adsorbent dose (g/L)	0.5	0.75	1.0

), while fluoride removal (%) was selected as the response variable. A total of 40 experimental runs, including duplicates, were conducted under constant orbital stirring (150 rpm) for 240 min at 25 ± 5 °C.

2.4.2. Adsorption Kinetics

Adsorption kinetic studies were conducted for all synthesized adsorbents. In each experiment, 500 mg of adsorbent was contacted with 500 mL of a fluoride solution (12 mg/L) at pH 5 in Erlenmeyer flasks. At predetermined time intervals (5, 10, 15, 20, 30, 45, 60, 120, 180, 240, 300, and 360 min), aliquots were collected and filtered using 0.45 μm syringe filters prior to fluoride analysis. The adsorption capacity at time t (qₜ, mg/g) was calculated from the residual fluoride concentration.

The kinetic data were analyzed using the pseudo-first order (Equation (3)), pseudo-second order (Equation (4)), and Elovich (Equation (5)) models [18,19]. Model parameters were obtained by nonlinear regression using Qtiplot software (version 0.9.8.3-3).

$$qt=qe \left(1-e^{-K1t}\right)$$

(3)

$$qt={\displaystyle \frac{K_2{q_e}^2 t}{(1+K_2{q}_{e}t)}}$$

(4)

$$qt={\displaystyle \frac{1}{\beta }}ln(1+\alpha \beta t)$$

(5)

where q_e is the adsorption capacity at equilibrium (mg/g), q_t is the adsorption capacity at time t (mg/g), t is the contact time (min), K₁ is the pseudo-first-order rate constant (1/min or min⁻¹), K₂ is the pseudo-second-order rate constant (g/mg·min or g mg⁻¹·min⁻¹)), α is the initial sorption rate (mg/g·min⁻¹), and β is the desorption constant (g/mg).

2.4.3. Adsorption Isotherms

Equilibrium adsorption experiments were conducted to determine the maximum fluoride adsorption capacity of the synthesized adsorbents. Fluoride solutions with initial concentrations ranging from 5 to 80 mg/L (5, 10, 20, 30, 40, 50, 60, and 80 mg/L) were prepared and contacted with the adsorbents at a dosage of 1.0 g/L and pH 5. The suspensions were agitated at 150 rpm for 5 h at room temperature to ensure equilibrium conditions.

The equilibrium adsorption data were analyzed using the Langmuir (Equation (6)) and Freundlich (Equation (7)) isotherm models [20,21]. Model parameters were obtained by non-linear regression using OriginPro 2021 software (OriginLab Corporation, Northampton, MA, USA).

$$q_e={\displaystyle \frac{q_m{K}_L{C}_e }{1+K_L{C}_e}}$$

(6)

$$qe=K_F{C_e}^{{\displaystyle \frac{1}{n}}}$$

(7)

where q_e is the adsorption capacity at equilibrium (mg/g), Ce is the equilibrium concentration of the adsorbate in solution (mg/L), q_m is the maximum adsorption capacity (mg/g), K_L is the Langmuir affinity or energy constant (L/mg), K_F is the Freundlich constant related to adsorption capacity [mg/g (L/mg)^1/n], and 1/n is the heterogeneity factor associated with adsorption intensity (dimensionless).

2.4.4. Regenerability and Recyclability

The regeneration and reusability of Zr0.05-LDH and Fe0.8-LDH were evaluated using 1.0 M NaOH as the desorbing agent. After fluoride saturation during the kinetic experiments, the adsorbents were recovered, dried at 105 °C for 12 h and weighed. The materials were then contacted with an appropriate volume of desorption solution to maintain a constant adsorbent dosage of 1 g/L under continuous stirring for 120 min.

Reusability was assessed over five consecutive adsorption–desorption cycles. In each cycle, adsorption was carried out using a fluoride solution with an initial concentration of 12 mg/L at pH 5. After desorption, the adsorbents were washed with deionized water, filtered and dried at 105 °C for 12 h. A calcination step at 500 °C for 4 h was subsequently applied to restore active sites before the next cycle.

2.4.5. Groundwater Adsorption Study

The adsorption performance of Zr0.05-LDH and Fe0.8-LDH was evaluated using groundwater collected from a well in the state of Durango, Mexico, where fluoride levels exceed the maximum permissible limit (1.0 mg/L) set by the Mexican standard NOM-127-SSA1-2021. Adsorption tests were conducted at a dose of 1.0 g/L and an initial pH of 5. The concentrations of sulfate (SO₄²⁻), chloride (Cl⁻), nitrate (NO₃⁻), phosphate (PO₄³⁻), fluoride (F⁻), and silica (SiO₂) were determined before and after treatment using a HACH DR1900 portable spectrophotometer (Hach, Loveland, CO, USA). Fluoride concentrations were specifically quantified using an ion-selective electrode (Thermo Scientific Orion, model 9609BNWP; Waltham, MA, USA).

2.4.6. Statistical Analysis

All adsorption experiments were conducted in duplicate (n = 2), and results are reported as mean ± standard deviation. Error bars representing standard deviations have been included in the experiment of preliminary, kinetic and isotherm adsorption, regeneration and groundwater application.

3. Results and Discussion

3.1. Characterization of Materials

3.1.1. X-Ray Diffraction (XRD) Analysis

The X-ray diffractograms of the adsorbent materials exhibited typical behavior of layered double hydroxides (LDHs) with a hydrotalcite-type structure, consistent with the crystallographic reference pattern COD 554–551 (Figure 1a). All samples showed characteristic reflections in the 2θ range of approximately 11–60°, which were indexed to the (003), (006), (009), (015), (018), (110) and (113) planes, in agreement with previous studies reported for the LDH phase [22,23]. Although the LDH structure was confirmed for all materials, variations in relative peak intensities and peak broadening were observed. These differences are attributed to reduced crystallinity and structural distortions induced by the incorporation of a third cation (Fe³⁺ or Zr⁴⁺) [24,25]. These changes are commonly associated with the partial isomorphic substitution of Al³⁺ by Zr⁴⁺ or Fe³⁺ within the brucite-like layers [26].

In Figure 1b, the diffractograms of the samples before and after adsorption, as well as the calcined Zr0.05 sample, are shown. After fluoride adsorption (Zr0.05-LDH + F⁻ and Fe0.8-LDH + F⁻), a partial reconstruction of the LDH phase was observed due to the memory effect, as evidenced by the reappearance together with attenuation and broadening of the (003), (006), and (009) reflections. This indicates that the layered structure was partially restored, although with reduced crystallinity. These results suggest that fluoride adsorption occurred without complete destruction of the LDH structure as shown in Figure 1b, but with some degree of structural modification, possibly due to surface interaction and/or partial intercalation processes. This behavior is consistent with previous studies reporting partial reconstruction of calcined LDHs upon anion adsorption [27].

The unit cell parameters a and c were calculated from the (110) and (003) reflections, where c = 3d₀₀₃ corresponds to the interlayer repeat distance and a = 2d₁₁₀ represents the average cation–cation distance within the brucite-like layers [27]. As summarized in

Table 2. Unit cell parameters and characteristic XRD reflection data for pristine LDH and the synthesized Zr- and Fe-modified adsorbents.

Unit Cell Parameters	(003)		(110)		c = 3d003	a = 2d110
Adsorbent	2θ	d (Å)	2θ	d (Å)
LDH	11.36	7.7805	60.49	1.5292	23.3415	3.0584
Zr0.1	11.25	7.8602	60.32	1.5332	23.5806	3.0664
Zr0.05	11.26	7.8295	60.36	1.5323	23.4885	3.0646
Fe1	11.41	7.7460	59.47	1.5530	23.2380	3.1060
Fe0.8	11.43	7.7388	59.71	1.5474	23.2164	3.0948

, incorporation of Zr⁴⁺ resulted in a slight increase in a parameter with increasing molar ratio, which is consistent with the larger ionic radius of Zr⁴⁺ (0.72 Å) compared with Al³⁺ (0.675 Å), confirming partial isomorphic substitution within the LDH structure [28,29]. In contrast, Fe³⁺-modified materials exhibited negligible variation in a parameter, while a modest increase in c at higher Fe³⁺ loadings suggests adjustments in interlayer spacing associated with cation incorporation [30].

3.1.2. Fourier Transform Infrared Spectroscopy (FTIR) Analysis

Fourier transform infrared (FTIR) spectroscopy was employed to identify surface functional groups and assess the structural integrity of the adsorbents before and after fluoride adsorption. As shown in Figure 1c, both pristine and fluoride-loaded samples exhibit the characteristic vibrational features of layered double hydroxides (LDHs), indicating that the layered framework remains intact after adsorption. A broad absorption band in the range of 3000–3633 cm⁻¹ is attributed to the stretching vibrations of hydroxyl groups (–OH), arising from both structural hydroxyls and interlayer water molecules. The band at approximately 2989 cm⁻¹ is associated with physically adsorbed water within the lamellar structure [31], while the signal near 1636 cm⁻¹ corresponds to the H–O–H bending vibration. The absorption band at 1407 cm⁻¹ corresponds to the symmetric stretching vibration of interlayer carbonate anions [14].

Following fluoride uptake, the FTIR profiles of both Zr0.05–LDH + F⁻ and Fe0.8–LDH + F⁻ still show the main features of the original LDH materials, indicating that the structure remains stable after fluoride adsorption. Bands in the 500–700 cm⁻¹ region are assigned to Zr–O and Zr–OH vibrations [32]. The FTIR band at 800 cm⁻¹ corresponds to lattice metal–oxygen (M–O) vibrations (M = Mg, Al, Fe, Zr) [26].

After fluoride adsorption, the decrease in the intensity of the –OH stretching band indicates that surface hydroxyl groups are involved in the fluoride adsorption process. This behavior is likely associated with an OH⁻/F⁻ ion-exchange mechanism, in which hydroxyl groups are partially replaced by fluoride ions [33]. Therefore, the results suggest that fluoride ions may interact with metal active sites on the calcined LDH surface, potentially involving inner-sphere complex formation. Together with the XPS results, these findings support a synergistic adsorption mechanism governed by hydroxyl exchange and metal–fluoride complex formation.

3.1.3. X-Ray Photoelectron Spectroscopy (XPS) Analysis

X-ray photoelectron spectroscopy (XPS) was used to investigate the surface chemical states of calcined Zr0.05-LDH and Fe0.8-LDH before and after fluoride adsorption (Figure 2a). After adsorption, distinct changes were observed in the F 1s, Zr 3d, Fe 2p, and O 1s regions, providing direct evidence of fluoride interaction with surface metal sites and functional groups [34].

The F 1s signal, which was absent in the pristine materials, appeared after fluoride exposure, with binding energies centered at 684–685 eV, characteristic of metal–fluoride (M–F, where M = Zr, Fe) bond formation, thus confirming direct fluoride complexation (Figure 2b,c). In calcined Zr0.05-LDH, an additional shoulder at higher binding energy (~688 eV) was detected and attributed to weakly coordinated or electrostatically retained fluoride species, indicating the coexistence of chemisorption and non-specific interactions [35].

The Zr 3d spectrum exhibited peaks at 182–185 eV, consistent with Zr⁴⁺ species (Figure 2d). After fluoride adsorption, a slight shift (~0.3 eV) toward lower binding energy was observed (Figure 2e). All spectra were calibrated using the C 1s peak (284.8 eV). The observed shift suggests that electrons may transfer from F⁻ to Zr⁴⁺, associated with the formation of stable Zr–F bonds and slight changes in the local electronic environment [36,37].

These results suggest the possible involvement of inner-sphere complexation during fluoride adsorption. In this mechanism, fluoride ions bind directly to metal sites on the surface of the calcined LDHs adsorbent by replacing surface hydroxyl (–OH) groups, resulting in the formation of metal–fluoride bonds (≡M–F). This interaction may also promote partial dehydration of fluoride ions and the formation of stable surface complexes with some covalent character. This process occurs at Lewis acid sites such as Zr⁴⁺ because of their high charge density and strong affinity for fluoride ions. Compared with outer-sphere interactions, inner-sphere complexation is generally associated with more stable and less reversible adsorption. In this context, it may contribute to the formation of stronger adsorption sites and to the overall adsorption performance observed in this system.

In the Fe 2p region, peaks at approximately 711 and 724 eV, together with their satellite features, confirmed the predominance of Fe³⁺ species (Figure 2h) [38]. After fluoride uptake, changes in the Fe³⁺/Fe²⁺ ratio were observed, which may suggest the possible involvement of surface redox processes in fluoride adsorption (Figure 2i). This behavior has been reported for Fe-based LDHs, where partial reduction of Fe³⁺ stabilizes Fe–F surface complexes [23,34,39]. However, given the inherent limitations of XPS quantification and potential fitting uncertainties, this interpretation should be considered as an initial contribution rather than a definitive mechanistic proof.

The O 1s spectra were deconvoluted into lattice oxygen (M–O, ~530 eV), surface hydroxyl groups (M–OH, ~531.5 eV), and adsorbed water (~533 eV) (Figure 2f,j). After fluoride adsorption, the relative contribution of M–OH decreased, while that of adsorbed H₂O increased, consistent with an OH⁻/F⁻ exchange mechanism in which displaced hydroxyl groups form water (Figure 2g,k) [35]. The surface chemistry of calcined LDHs is consistent with the FTIR results, which show attenuation of the –OH stretching bands. However, carbonate-related vibrations remain largely unchanged, indicating preservation of the LDH framework.

Detailed XPS analysis reveals that fluoride adsorption using calcined LDHs proceeds through a synergistic combination of mechanisms. These include initial OH⁻/F⁻ anion exchange, evidenced by the depletion of surface hydroxyl groups, together with direct surface complexation at zirconium and iron sites. This was confirmed by the appearance of distinct F 1s signals and associated binding energy shifts. In parallel, electrostatic attraction toward the positively charged calcined LDH layers provides an additional driving force for fluoride uptake [36,39]. In this system, chemisorption via stable M–F bond formation is the dominant pathway. Zirconium sites in calcined LDHs are suggested to play an important role through possible inner-sphere interactions with fluoride, while iron may contribute by enhancing redox flexibility and increasing the density of active sites [40,41].

The XRD and BET results further support that the structural integrity of the adsorbent is maintained after adsorption and demonstrate the complementary roles of Zr and Fe in controlling adsorption performance.

3.1.4. Specific Surface Area

Nitrogen physisorption analysis at 77 K (BET/BJH), as shown in

Table 3. Textural characteristics of Mg–Al LDH and their changes following Fe³⁺ and Zr⁴⁺ modification.

Adsorbent	Surface Area (m2/g)	Diameter (nm)	Volume (cm3/g)
LDH	218.0	26.53	1.44
Zr0.1-LDH	243.2	15.40	0.94
Zr0.05-LDH	261.0	16.00	1.04
Fe1-LDH	215.0	16.14	0.87
Fe0.8-LDH	233.5	18.29	1.067

, revealed that Zr0.05-LDH exhibited the highest specific surface area (261 m²/g). All metal-modified samples, except Fe1-LDH, showed higher surface areas than pristine LDH, confirming the beneficial effect of controlled metal incorporation on textural properties. An optimal Zr loading (0.05 mol) enhanced surface accessibility without significantly altering mesoporosity, as supported by scanning electron microscopy (SEM) observations.

In contrast, higher Zr⁴⁺ incorporation (0.1 mol) resulted in reduced pore volume and pore diameter, likely due to structural densification and limited Zr⁴⁺ dispersion within the LDH framework. This behavior is consistent with energy-dispersive X-ray spectroscopy (EDS) and X-ray diffraction (XRD) results, which indicate increased lattice distortion and reduced structural openness at higher Zr⁴⁺ contents [42].

For Fe-modified materials, Fe0.8-LDH exhibited a higher specific surface area than Fe1-LDH, highlighting the structural role of residual aluminum. Even low Al³⁺ incorporation improved structural stability and promoted a more homogeneous elemental distribution, as confirmed by SEM and EDS analyses. Conversely, Fe1-LDH showed greater heterogeneity and structural disorder, likely associated with iron-rich domains and partial formation of amorphous Fe₂O₃ phases, which reduce porosity and effective surface area [43].

These results indicate that metal composition and dispersion critically influence LDH structural organization and porosity, which in turn govern adsorption performance. In particular, the incorporation of Zr⁴⁺ and Fe³⁺ changes the surface and structural properties of the adsorbent. In this study, Zr⁴⁺ exhibited a high affinity for fluoride, which may be attributed to Lewis acid–base interactions that favor inner-sphere complexation. Fe³⁺ also affected the adsorption behavior by increasing the surface charge and creating more adsorption sites. However, excessive metal incorporation may alter the structure of calcined LDHs, leading to pore blockage, reduced porosity, and limited accessibility of active sites. Therefore, metal dispersion, surface accessibility, and structural integrity play important roles in calcined LDHs and influence fluoride removal performance, as observed for the Zr0.05–LDH sample.

3.1.5. Scanning Electron Microscopy (SEM), Energy-Dispersive X-Ray Spectroscopy (EDS), and Elemental Mapping

SEM micrographs (Figure 3) show that all synthesized materials retain the characteristic lamellar morphology of hydrotalcite-like structures, confirming the successful formation of calcined LDHs. Nevertheless, cationic modification induces distinct and systematic morphological variations. In Zr⁴⁺-modified materials, platelet size and thickness increase with increasing nominal Zr loading, consistent with the expansion of the lattice parameter determined by XRD. This behavior suggests interlamellar structural distortion induced by Zr⁴⁺ incorporation. In addition, the more compact aggregation of platelets at higher Zr⁴⁺ contents may be attributed to increased surface energy, promoting denser packing [29].

In contrast, Fe³⁺-modified materials exhibit a more heterogeneous morphology, characterized by less uniform platelet size and irregular stacking, indicative of less controlled nucleation and growth. Similar morphological features have been reported for Fe-doped LDH systems and are commonly associated with the partial formation of iron oxide species [44]. This interpretation is consistent with the appearance of additional FTIR bands in the 550–600 cm⁻¹ region, which are attributed to Fe–O vibrations.

EDS analysis supports the morphological observations and provides quantitative insight into metal incorporation. In Zr-modified materials, the detected Zr⁴⁺ content increased only from approximately 0.30 to 1.06 wt.% despite a twofold increase in nominal loading, indicating non-proportional incorporation and suggesting a substitution limit within the LDH lattice. Elemental mapping revealed a homogeneous Zr⁴⁺ distribution without localized enrichment, ruling out the formation of segregated crystalline Zr phases and suggesting the presence of highly dispersed species or amorphous domains (e.g., ZrO₂). This behavior is consistent with the reduced crystallinity observed in the XRD patterns [45].

For Fe-based materials, the higher Fe³⁺ content (approximately 18–19 wt.%) reflects greater incorporation efficiency. However, the binary Mg/Fe system shows less homogeneous elemental distribution, consistent with the presence of Fe-rich domains. In contrast, the Mg/Al/Fe system exhibits a more uniform distribution of elements, indicating that Al³⁺ plays a key structural role in stabilizing the layered framework and regulating cation dispersion [46].

Overall, these results demonstrate that chemical composition and cation distribution critically govern LDH structural order, substitution behavior, and the formation of secondary phases. These factors directly influence crystallinity, porosity, and ultimately, the adsorption performance of the materials.

3.2. Batch Experimentation of Fluoride Adsorption

3.2.1. Fluoride Adsorption Tests

The fluoride adsorption performance of LDH, Fe1-LDH, Fe0.8-LDH, Zr0.1-LDH, and Zr0.05-LDH was evaluated under identical conditions (Figure 4). Pristine LDH exhibited the lowest removal efficiency (~80%) with low standard deviation, highlighting the beneficial effect of metal modification. Zr0.05-LDH achieved the highest fluoride removal efficiency (~96–97%), outperforming Zr0.1-LDH and demonstrating a non-linear relationship between metal loading and adsorption performance.

This behavior indicates that fluoride uptake is governed not only by metal content, but also by the balance between active site availability and metal dispersion. Zirconium-based materials exhibit strong fluoride affinity through inner-sphere complexation (Zr–F); however, excessive Zr⁴⁺ incorporation may promote the formation of amorphous domains, limit site accessibility, and reduce adsorption efficiency [47].

Fe-modified materials also showed enhanced performance compared with pristine LDH, with Fe0.8-LDH outperforming Fe1-LDH. This result highlights the stabilizing role of Al³⁺, which promotes structural order and a more homogeneous distribution of active sites within the LDH framework [15]. Overall, these results demonstrate that adsorption performance is primarily dictated by structural organization and metal dispersion rather than by nominal metal composition alone.

3.2.2. Optimization of Fluoride Removal via RSM

Based on the adsorption results and structural characterization, calcined Zr0.05-LDH was selected for process optimization. Optimization was performed using Response Surface Methodology (RSM) with a Central Composite Design (CCD). The resulting quadratic model exhibited an acceptable predictive performance (R² = 0.83), indicating that a second-order polynomial adequately describes fluoride removal behavior within the studied experimental range. This moderate coefficient of determination is consistent with the physicochemical heterogeneity of calcined LDH-based adsorption systems, associated with non-uniform active site distribution and the coexistence of multiple adsorption mechanisms.

The low standard error of estimate (1.67) and mean absolute error (MAE = 0.83) indicate good agreement between experimental and predicted values. R² values above 0.70 are generally considered acceptable for systems with inherent variability [48]; therefore, an R² value of 0.83 can be regarded as satisfactory for optimization purposes. The resulting regression model is expressed by Equation (8), indicating that the model is reliable for predicting fluoride removal within the studied conditions. The remaining minor residual variability is likely attributable to the intrinsic heterogeneity of the LDH adsorption sites rather than poor model fitting.

$$Removal \left(\%\right)=95.94+0.52x_1+7.42x_2-0.42x_3+1.75x_2·x_3-3.61x_1^2-3.09x_2^2-1.57x_3^2$$

(8)

with the coded variables x₁ (concentration), x₂ (ads dose), and x₃ (pH).

ANOVA results indicate that adsorbent dosage is the dominant factor (p < 0.001), whereas the linear effects of initial concentration and pH are not statistically significant (p > 0.05), although their quadratic terms contribute to model curvature. This is also observed in the standardized Pareto chart (Figure 5a), in which the adsorbent dosage exhibits the largest standardized effect, exceeding the significance limit. In contrast, the effects of initial concentration and pH remain below the critical line, indicating a smaller statistical influence in the studied range. The significant quadratic terms suggest non-linear behavior, with fluoride removal reaching optimal values under specific conditions.

The lack-of-fit was significant (F = 29.31, p < 0.05), indicating that the quadratic model does not fully describe the response beyond the pure error, as shown in

Table 4. Analysis of Variance ANOVA for SQRT.

Source Model	Sum of Squares	Degree of Freedom	Mean Square	F Value	p Value
A: pH	7.30888	1	7.30888	0.24	0.6280
B: Adsorbent dose	1493.97	1	1493.97	49.02	0.0000
C: Initial concentration	4.87948	1	4.87948	0.16	0.6920
AA	375.899	1	375.899	12.33	0.0015
AB	10.0172	1	10.0172	0.33	0.5709
AC	9.5481	1	9.5481	0.31	0.5800
BB	275.65	1	275.65	9.04	0.0054
BC	48.7902	1	48.7902	1.60	0.2159
CC	70.7672	1	70.7672	2.32	0.1384
Blocks	26.9781	1	26.9781	0.89	0.3546
Lack of Fit Test	868.262	19	45.698	29.31	<0.0001
Pure error	15.5893	10	1.5589
Total (corr.)	3105.94	39

. This implies that additional curvature or unmodeled effects may be present within the explored factor space. The model is therefore used mainly to identify factor influence and guide optimization trends, while predictions are interpreted cautiously.

The residuals versus predicted values plot (Figure 5b) shows a random scatter of data points around zero, with no visible patterns, indicating good agreement between the experimental and predicted values. Although a slight clustering of residuals at higher predicted values can be observed, it remains within acceptable limits and does not compromise model validity. No significant outliers were detected, further supporting the reliability of the regression.

Although the lack-of-fit result suggests unexplained variability in the experimental data, the ANOVA, Pareto chart, and residual analyses indicate that the model remains statistically useful for process optimization, with adsorbent dosage identified as the primary factor controlling fluoride removal.

Fluoride was 0. 88 g/L, followed by a plateau at around 98%, indicating a site saturation regime in which further increases in available active sites do not enhance adsorption performance. The interaction between pH and initial fluoride concentration (Figure 6a) shows a decline in removal efficiency at pH values above 6, which is associated with reduced surface protonation and increased competition from OH⁻ ions. Under mildly acidic conditions (approximately pH 5), fluoride uptake is favored due to enhanced electrostatic attraction and inner-sphere complexation with Zr active sites [49].

In contrast, the weak interaction between adsorbent dosage and initial fluoride concentration (Figure 6b) suggests that, once enough adsorbent is available, the system becomes less sensitive to variations in fluoride concentration. The interaction between adsorbent dose versus pH (Figure 6c) indicates that the adsorbent dosage exerts a stronger influence than pH, as dosages above 0.85 g/L consistently achieve fluoride removal close to 98%. The optimal conditions (C₀ = 12.6 mg/L, dosage = 0.88 g/L, and pH = 4.9) resulted in a maximum fluoride removal of 98.17%, demonstrating that process performance is governed primarily by accessible active site density and surface chemistry rather than by nominal fluoride concentration alone.

3.2.3. Kinetic Study

Adsorption kinetics were evaluated using the pseudo-first-order (PFO), pseudo-second-order (PSO), and Elovich models (Figure 7). For all materials, a rapid uptake phase was observed during the initial stage (0–60 min), followed by a slower approach to equilibrium, which was reached after approximately 180–240 min.

The kinetic parameters are shown in

Table 5. Kinetic model parameters and statistical error analysis for fluoride adsorption on pristine and calcined Fe³⁺- and Zr⁴⁺-modified Mg–Al LDHs.

Model	Parameters	LDH	Zr0.1	Zr0.05	Fe0.8	Fe1
Pseudo-first order	qe (mg/g)	9.90	9.98	11.17	10.10	8.96
	K1(1/min or min−1)	0.06	0.02	0.02	0.01	0.03
	R2adj	0.98	0.95	0.99	0.87	0.90
	RMSE	0.42	0.68	0.29	0.99	0.79
Pseudo-second order	qe (mg/g)	11.05	11.53	12.17	11.71	9.96
	K2 (g/mg·min or mg−1·min−1)	0.00036	0.0021	0.0015	0.0015	0.0034
	R2adj	0.98	0.95	0.98	0.90	0.94
	RMSE	0.42	0.64	0.46	0.95	0.61
Elovich	α (g/mg·h)	0.08	0.51	0.34	0.36	2.4
	β (g/mg·h)	0.22	0.40	0.337	0.37	0.7
	R2adj	0.98	0.95	0.96	0.92	0.88
	RMSE	0.43	0.71	0.70	0.82	0.82

, the PSO model provided a satisfactory description of the kinetic behavior of the modified LDH systems, particularly for Fe0.8-LDH (R²_adj ≈ 0.984), suggesting that surface-controlled interactions play an important role in fluoride uptake. However, the comparable performance of different models in several cases indicates that no single kinetic model fully captures the complexity of the adsorption process. Notably, Zr0.05-LDH showed the best fit with the PFO model (R²_adj = 0.992; RMSE = 0.29), indicating that the initial adsorption stage is dominated by rapid uptake at readily accessible surface sites.

In contrast, the PSO model predicted a higher equilibrium adsorption capacity (qₑ = 12.17 mg/g) for Zr0.05-LDH and more accurately described the later stages of adsorption, suggesting the coexistence of multiple kinetic regimes [50]. This behavior likely reflects an initial fast interaction with surface hydroxyl groups, followed by slower processes involving specific metal-based active sites [50,51]. The Elovich model further supports the presence of surface heterogeneity across all materials, consistent with a distribution of adsorption energies induced by metal incorporation [52].

The kinetic performance followed the order Zr0.05-LDH > Fe0.8-LDH > Zr0.1-LDH > LDH > Fe1-LDH, indicating that adsorption kinetics are governed primarily by the accessibility and distribution of active sites rather than by their nominal concentration. These results reinforce the role of structural organization and metal dispersion in controlling adsorption dynamics.

As shown by the error bars in Figure 7 (standard deviation, n = 2), fluoride adsorption remains highly consistent across all samples. Slightly larger standard deviations were observed for some data points. This variation is expected given the rapid initial uptake, where even small differences in sampling time can introduce minor fluctuations.

3.2.4. Isotherm Study

The equilibrium adsorption behavior of fluoride onto LDH, Zr0.05-LDH, and Fe0.8-LDH was evaluated over an initial concentration range of 5–80 mg/L using the Langmuir and Freundlich isotherm models (Figure 8). The experimental data were fitted using nonlinear regression to avoid distortions associated with linearization, and the resulting parameters are summarized in

Table 6. Isotherm parameters and statistical error analysis for fluoride adsorption on pristine and metal-modified calcined Mg–Al LDHs (Zr0.05-LDH and Fe0.8-LDH).

Model	Parameters	LDH	Zr0.05	Fe0.8
Langmuir	qe (mg/g)	54.73	137.12	82.37
	R2	0.95	0.99	0.99
	R2adj	0.95	0.99	0.98
	KL (L/mg)	0.03	0.01	0.017
	RMSE	0.15	0.09	0.04
Freundlich	KF (mg/g)·(L/mg)1/n	4.43	2.04	2.93
	n	0.50	1.32	0.64
	R2	0.89	0.99	0.97
	R2adj	0.89	0.98	0.95
	RMSE	0.26	0.70	0.82

[53].

Among the evaluated models, the Langmuir isotherm showed the best fit, with adjusted coefficients of determination (R²_adj) of 0.992 for Zr0.05-LDH, 0.988 for Fe0.8-LDH, and 0.947 for pristine LDH. This behavior suggests adsorption on a relatively homogeneous distribution of energetically comparable active sites and is consistent with a predominantly monolayer adsorption process. Accordingly, fluoride uptake within the investigated concentration range is likely governed primarily by specific surface interactions rather than multilayer adsorption.

FTIR analysis shows a decrease in the intensity of –OH stretching bands after adsorption, indicating the participation of surface hydroxyl groups through an OH⁻/F⁻ exchange mechanism. In parallel, XPS revealed the appearance of F 1s signals (~685 eV) and the binding energy shifts in the Zr 3d and Fe 2p regions confirmed the formation of metal–fluoride (M–F) interactions and associated local electronic redistribution. These observations support a mechanism involving surface complexation at metal centers in addition to anion exchange, in agreement with the adsorption behavior inferred from the Langmuir model.

This behavior may also be related to the partial reconstruction of LDH-like structures upon contact with aqueous solutions, enabling the coexistence of oxide surfaces and hydroxylated layers that participate in adsorption through surface complexation and anion exchange. The maximum adsorption capacities followed the order Zr0.05-LDH (137 mg/g) > Fe0.8-LDH (82.4 mg/g) > pristine LDH (54.7 mg/g), highlighting the role of controlled metal incorporation in enhancing both active site density and accessibility [36].

In most cases, the standard deviation, as shown in Figure 8, was low compared to the corresponding adsorption capacities. However, since only duplicate experiments were performed (n = 2), the variability should be interpreted with some caution, and a higher number of replicates would provide greater statistical confidence.

As summarized in Table 7, the performance of Zr0.05-LDH is competitive with, and in some cases superior to, previously reported LDH-based adsorbents. Compared with other LDH systems such as MgAlZr and Mg/Fe/La materials (Table 6), Zr0.05-LDH exhibits comparable or higher adsorption capacity, which may be attributed to optimized Zr⁴⁺ incorporation that increased active-site accessibility while preserving structural integrity. These results place the developed material among the most effective LDH-based adsorbents reported for fluoride removal and suggest that metal dispersion plays a more important role than composition alone.

Although the Langmuir model provided the best overall fit, the Freundlich model also showed good agreement (R²_adj = 0.98), suggesting a degree of surface heterogeneity and the contribution of secondary mechanisms such as electrostatic attraction and OH⁻/F⁻ exchange [53,54]. Fluoride adsorption is best described as a combined process, where Langmuir-type behavior reflects dominant site-specific interactions, while spectroscopic evidence (FTIR and XPS) indicates the contribution of chemisorption through surface complexation [33].

Table 7. Comparative analysis of the maximum fluoride adsorption capacity (q_e) of the synthesized Zr0.05-LDH and Fe0.8-LDH relative to previously reported LDH-based adsorbents.

Adsorbent	qe (mg/g)	Reference
Mg/Al-LDH-Cl−	59.60	[21]
MgAlZr-LDH	51.92	[36]
CaAlZr-LDH	131.03	[55]
NiFe-LDH	23.49	[56]
Mg/Fe/La	60.00	[57]
Zr0.05-LDH	137.12	This study
Fe0.8-LDH	82.36	This study

Table 7. Comparative analysis of the maximum fluoride adsorption capacity (q_e) of the synthesized Zr0.05-LDH and Fe0.8-LDH relative to previously reported LDH-based adsorbents.

Adsorbent	qe (mg/g)	Reference
Mg/Al-LDH-Cl−	59.60	[21]
MgAlZr-LDH	51.92	[36]
CaAlZr-LDH	131.03	[55]
NiFe-LDH	23.49	[56]
Mg/Fe/La	60.00	[57]
Zr0.05-LDH	137.12	This study
Fe0.8-LDH	82.36	This study

3.2.5. Recyclability of Zr0.05-LDH and Fe0.8-LDH

Regenerability is a key factor determining the economic feasibility and operational stability of adsorbent materials. In this study, the recyclability of Zr0.05-LDH and Fe0.8-LDH was evaluated over consecutive adsorption–desorption cycles using 1.0 M NaOH as the regenerating agent. The use of NaOH is justified by the strong affinity of OH⁻ ions to replace fluoride in the LDH structure, facilitated by similar ionic radii and the high rehydration capacity of hydrotalcite materials [35,58].

As shown in Figure 9, both adsorbents exhibited high adsorption efficiency over the first five adsorption–desorption cycles, indicating good regeneration capability. The regeneration results show that Zr0.05-LDH maintains more stable performance with lower variability compared to Fe0.8-LDH across cycles. After five cycles, Zr0.05-LDH achieved an adsorption capacity of 9.0 mg/g (89% retention), while Fe0.8-LDH retained 8.4 mg/g (82%). However, a significant decrease in performance was observed in the sixth cycle, with a 31% decrease in removal efficiency for Zr0.05-LDH and 50% for Fe0.8-LDH.

After several adsorption–desorption cycles, the material showed a gradual loss in adsorption capacity. This behavior may be related to partial deactivation of active sites, pore saturation by residual species, and some degree of particle agglomeration. Fluoride removal is generally associated with chemisorption processes, potentially involving inner-sphere complex formation; under these conditions, some adsorbed species may remain strongly attached to the surface and may not be fully removed during desorption [59]. To mitigate these effects, optimization of regeneration conditions (e.g., NaOH concentration and contact time), incorporation of stabilizing supports or binders, and development of composite or granular forms of LDHs could improve structural stability and long-term reusability in practical applications.

As shown in Figure 9, the standard deviations indicate that the data are not equally consistent across all groups. Some data have a wider spread, meaning the values vary more and are less consistent, while others are more tightly grouped, indicating more precise and reliable results. This behavior suggests that certain conditions may lead to greater variability in the data.

These results confirm that Zr0.05-LDH exhibits enhanced durability and regeneration stability compared to Fe-modified systems, reinforcing its potential as a cost- effective and sustainable adsorbent for water treatment applications.

3.2.6. Evaluation Under Real-Water Conditions

Groundwater usually contains various dissolved ions that can interfere with fluoride adsorption by competing for the active sites of the adsorbent. To evaluate this effect under real conditions, groundwater from a well located in Durango, Mexico, was tested. The groundwater sample had a fluoride concentration of 2.99 mg/L, which is higher than the limit allowed by NOM-127-SSA1-2021 (1.0 mg/L). As shown in Figure 10, fluoride removal by the calcined LDHs remained relatively high in the presence of groundwater anions, which include Cl⁻ (3.0 mg/L), NO₃⁻ (1.0 mg/L), SO₄²⁻ (18 mg/L), PO₄³⁻ (0.8 mg/L), and SiO₂ (46.8 mg/L). This behavior differed from that obtained with synthetic fluoride solutions, where the removal efficiency was close to 99%. Part of this decrease may be related to competitive adsorption effects caused by the coexistence of other dissolved species in the groundwater matrix.

After treatment, some of these anions also showed partial removal. This indicates that the adsorbents can interact with several anionic species simultaneously and are therefore not completely selective toward fluoride under real groundwater conditions. Silica species may interact with surface metal centers or form surface complexes, further contributing to the decrease in fluoride selectivity [60]. While this multicomponent removal contributes to overall water-quality improvement, it also reduces fluoride removal efficiency due to competition for active sites.

Under groundwater conditions, fluoride removal efficiencies decreased to 69.42% for Zr0.05-LDH and 68.99% for Fe0.8-LDH, corresponding to adsorption capacities of 1.798 and 1.787 mg/g, respectively. This reduction is attributed to competitive adsorption effects, particularly from multivalent anions, whose higher charge density enhances electrostatic interactions and may partially displace fluoride from active sites [15,36].

This competitive behavior can be further rationalized by considering fundamental physicochemical parameters such as ionic charge density (z/r) and hydration energy. A mixture of anions such as SO₄²⁻ and PO₄³⁻ can interact directly with hydrotalcite-like materials because of their higher charge and polyatomic structure. It can be explained that the competition of fluoride for available adsorption sites is enhanced by the strong electrostatic attraction to the positively charged LDH layers, including their affinity for surface –OH groups [36]. Conversely, fluoride differs due to its monovalency, small ionic radius and high electronegativity which results in high charge density and strong hydration energy.

According to HSAB theory, F⁻ acts as a hard base and presents a strong binding with hard Lewis acid centers such as Zr⁴⁺. These results suggest the possible formation of inner-sphere Zr–F interactions [61]. Consequently, Zr0.05-LDH exhibits higher fluoride removal efficiency, adsorption capacity, and greater tolerance to multivalent interferents under groundwater conditions. This interpretation is consistent with previous reports [49,62], and is further supported by FTIR and XPS analyses, which confirm that fluoride adsorption is governed by synergistic surface complexation and ion-exchange mechanisms.

Overall, although the complexity of groundwater matrices reduces adsorption efficiency relative to idealized systems, both materials demonstrate promising performance. However, the lower and more consistent standard deviation values for Fe0.8-LDH suggest better stability in complex groundwater matrices, as reflected in Figure 10.

4. Conclusions

This study shows that calcined Fe³⁺- and Zr⁴⁺-modified LDHs can remove fluoride from groundwater. The adsorption performance was influenced not only by the amount of metal incorporated into the structure, but also by its dispersion and the structural characteristics of the material. Zr0.05-LDH exhibited the best balance between material structure and active-site accessibility. Therefore, these materials demonstrate high adsorption capacity (137 mg/g) and fluoride removal (98%) under the selected operating conditions.

The XPS and FTIR results suggest that fluoride adsorption mainly occurred through interaction with Zr⁴⁺ active sites, probably by the formation of inner-sphere surface complexes. Ion exchange between OH⁻ and F⁻, together with electrostatic interactions, also contributed to the adsorption process. The adsorption behavior indicated the presence of various types of active sites and interaction mechanisms on the calcined LDH adsorbent surface. Kinetic and isotherm analyses further suggest that adsorption proceeds through rapid surface interaction followed by chemisorption-dominated equilibrium.

Under real groundwater conditions, the presence of competing anions reduced fluoride removal efficiency due to competitive adsorption effects; however, both Zr0.05-LDH and Fe0.8-LDH reduced fluoride concentrations below regulatory limits. The superior performance of Zr0.05-LDH under these conditions highlights the role of specific chemical affinity between fluoride and Zr⁴⁺ centers in maintaining selectivity within complex aqueous matrices.

From a translational perspective, the high adsorption capacity and regeneration stability demonstrated by calcined Zr0.05-LDH (up to 89% capacity retention after five cycles) support its potential for practical water-treatment applications. For future work, it would be worthwhile to explore transforming powdered LDHs into pelletized or composite forms to improve their mechanical stability and make them more suitable for potential use in packed-bed column systems. Nevertheless, the present study was limited to batch-scale experiments using powdered adsorbents. Therefore, the results mainly contribute to a better understanding of adsorption behavior and material performance at the laboratory scale and can serve as a foundation for further studies aimed at practical water treatment applications.