Defluoridation of Water Using Al-Mg-Ca Ternary Metal Oxide-Coated Sand in Adsorption Column Study

Abstract

Defluoridation of water was investigated in an adsorption column study using Al-Mg-Ca-coated sand (AMCCS), a ternary metal oxide adsorbent with eco-friendly components that were shown to be effective for water defluoridation, in a batch adsorption study. A packed column of the AMCCS sorbent was evaluated as function of column flow rate, solution type, and sorbent recyclability. Adsorption column experiments included two column flow rates of 2 mL/min and 10 mL/min using two different solutions: deionized water and a synthetic solution representative of groundwater. Greater fluoride column adsorption capacity was obtained at the lower flow rate for both solutions, mainly due to longer contact times between solution and AMCCS sorbent. Adsorption of fluoride occurred through physical adsorption, which followed the Langmuir adsorption model and second-order kinetics for deionized water and synthetic solution. A lower AMCCS column fluoride adsorption capacity was observed for the synthetic solution due to the competition from adsorption of other ions in the synthetic solution, whereas fluoride adsorption by the AMCCS column was influenced by interphase mass transfer to a lesser extent using the synthetic solution than deionized water. The re-coating of spent AMCCS sorbent in the adsorption column resulted in effective recycling and reuse of the AMCCS adsorption column for both deionized water and the synthetic solution, rendering the AMCCS adsorption column a recyclable and sustainable flow through water defluoridation system.

1. Introduction

Fluoride-related health issues are considered serious due to its ability to circulate throughout the body via the bloodstream following exposure through sources such as drinking water, food, or other pathways. Prolonged fluoride intake can result in serious conditions, including bone fluorosis, paralysis, osteoporosis, and damage to vital organs such as the kidneys, liver, and brain. The severity of these effects is determined by fluoride concentration and exposure duration, making fluoride toxicity a critical area of research [1,2]. Fluoride contamination occurs via two principal pathways: natural processes, including geochemical reactions and volcanic emissions, and anthropogenic activities such as those associated with ceramics, aluminum, iron, steel, fertilizer production, and thermal power plants [3,4,5,6,7,8].

Various fluoride standards have been established globally for protection of public water supply and health. The potable water maximum contaminant level (MCL) for fluoride has been set at 4 mg/L by the U.S. Environmental Protection Agency [9]. While the World Health Organization (WHO) has recommended a limit of 1.5 mg/L of fluoride in drinking water [10,11], the U.S. Public Health Service recommends a lower drinking water limit of 0.7 mg/L of fluoride as the optimal fluoride concentration [12]. There are various reasons for how these limits are determined, leading to variations between regulatory guidelines. These include the sources of fluoride exposure, individual age, lifestyle habits, and daily water consumption, among others [13].

Consequently, developing and implementing effective fluoride removal technologies is crucial for meeting these standards. Common methods include adsorption, ion exchange, coagulation–precipitation, membrane separation, bioremediation, and electrocoagulation [14,15,16,17]. Each method varies in effectiveness, depending on local conditions such as water chemistry, infrastructure availability, cost considerations, and target populations. For example, coagulation–precipitation is suitable for high-fluoride waters (>100 mg/L) but is less effective at lower concentrations and may produce toxic aluminum fluoride complexes or elevate sulfate levels [18]. Membrane separation is highly efficient but often removes essential minerals and requires costly post-treatment remineralization.

Among these methods, adsorption has emerged as one of the most promising and versatile techniques for fluoride removal, owing to its cost-effectiveness, operational simplicity, and adaptability to varying fluoride concentrations. The method is compatible with a wide range of materials, many of which are regenerable, and does not compromise water taste, odor, or clarity. However, challenges remain, including pH sensitivity, high production costs for certain materials, and the potential for adsorbent leaching [14,17]. Adsorbents can be categorized into metal oxides, biosorbents, organic carbons, natural zeolites, clays, and industrial by-products. Among these, metal oxide-based adsorbents have shown exceptional selectivity and efficiency, with commonly studied materials based on aluminum, iron, magnesium, calcium, copper, and titanium [11,19,20,21,22]. Aluminum-based adsorbents are particularly attractive due to their availability and low cost; however, their performance diminishes under neutral to alkaline pH conditions [23]. To overcome this limitation, researchers have combined aluminum with magnesium and calcium to create ternary oxide systems that offer broader pH stability, high fluoride affinity, and non-toxic properties [24,25,26,27,28,29]. These elements are also biocompatible and affordable, although they can raise the pH of treated water [21,22,30,31,32].

Recent advancements have focused on engineered adsorbents with enhanced surface characteristics, particularly metal oxide-coated sorbents, which are valued for their high surface area and suitability for packed-bed columns. Several studies have evaluated fluoride removal performance in adsorption column systems using different adsorbents, as well as different column flow rates and initial fluoride concentrations. Wei et al. [33] developed a calcium–iron mixed metal oxide (CCF), demonstrating defluoridation performance across pH 3–12, although regeneration was limited by CaF₂ precipitation. Similarly, Iwar et al. [34] synthesized a raffia palm shell-derived carbon–alumina composite, highlighting the potential of bio-based metal oxide materials. Anas et al. [35] evaluated activated alumina across batch, column, and pilot scales. Mani and Bhandari [36] investigated zirconium–copper complexed polyvinyl alcohol hydrogel beads, achieving a high column adsorption capacity despite a low BET surface area. Bakhta et al. [37] reported high adsorption capacity of fluoride using Al(OH)₃-loaded activated carbon. Teutli-Sequeira et al. [38] evaluated the fluoride performance of aluminum-modified zeolite, Nur et al. [39] assessed that of hydrous iron oxide packed with anthracite in adsorption columns, while Ye et al. [11] studied a magnesia–pullulan composite. Recent studies have explored hybrid approaches that couple adsorption with processes such as chemical precipitation, filtration, or magnetic separation to improve system efficiency. Zhao et al. [40] reviewed such systems and showed that integrating slaked lime precipitation, multimedia filtration, or magnetic separation with adsorption can enhance fluoride removal and reduce operational challenges in industrial settings.

Despite progress in fluoride removal technologies, continuous-flow column systems still face several operational challenges, including premature breakthroughs, competition from co-existing anions (e.g., bicarbonate, sulfate), mass transfer limitations, and diminished performance across regeneration cycles. For instance, Deng et al. [41] demonstrated that while a nanocomposite-based fixed-bed system maintained high fluoride removal efficiency in a multi-column series configuration, most of the equilibrium adsorbent capacity was not utilized. To overcome these limitations, research continues to emphasize the development of more robust sorbent formulations, surface functionalization techniques, and accurate predictive models, which can effectively simulate breakthrough behavior and optimize design parameters for large-scale column operations [35].

Previously, in a batch adsorption study for defluoridation of water [28], we demonstrated the effective defluoridation performance of a ternary hybrid metal oxide adsorbent with eco-friendly components, the aluminum–magnesium–calcium-coated sand (AMCCS). In the present study, we assessed the performance of the AMCCS sorbent for defluoridation of water in a flow-through column adsorption system, examining different flow rates and solutions, as well as investigating the effects of recoating the sorbent on its column performance. Column studies are valuable tools in the water treatment field, providing a practical procedure to predict sorbent behavior in real-world applications and to assess the potential for pilot-scale implementation.

2. Materials and Methods

2.1. Chemicals and Reagents

This study utilized various chemicals for sorbent preparation and experimental procedures, including sodium fluoride, magnesium chloride hexahydrate, aluminum chloride hexahydrate, sodium bicarbonate, sodium sulfate, and calcium chloride dihydrate, all of which had over 99% purity and were of ACS (American Chemical Society) grade. Clean silica sand (quartz) with particle size of 50–70 mesh (average particle size of 250 µm) was obtained from Sigma-Aldrich (St. Louis, MO, USA). All reagents were purchased from Fisher Scientific (Fairlawn, NJ, USA), while the deionized (DI) water with a resistivity greater than 8 MΩ was produced in the laboratory and subsequently used for preparation of all solutions.

2.2. Sorbent Preparation

The fabrication process of the AMCCS sorbent was previously developed and applied in a batch setup as part of our earlier study [28]. The preparation of the proposed sorbent began by weighing 40 g of silica sand, which was then transferred to a beaker and mixed with three coating agents—aluminum chloride, magnesium chloride, and calcium chloride—each with a molarity of 1 M. The coating solution had a volume of 100 mL, consisting primarily of aluminum chloride (85 mL), with the remaining 15 mL equally divided between calcium chloride (7.5 mL) and magnesium chloride (7.5 mL). A magnetic mixing plate was used to stir the mixture of coating solution–sand continuously. After 24 h of mixing, the solution was separated from the coated sand and the spent coating solution was then placed in HDPE plastic bottles for use in future applications. The coated sand was then placed in a drying oven at 110 °C for 24 h. The final step of the fabrication process involved the calcining of the dried coated sand in a furnace for 24 h at 220 °C.

2.3. Characterization of the AMCCS Sorbent

The surface characterization of the AMCCS sorbent was carried out using X-ray powder diffraction (XRD), transmission electron microscopy (TEM), and scanning electron microscopy with Electron Dispersive X-ray Spectroscopy (EDX). The XRD analysis was performed using the XRD Bruker D8 Discover System (Billerica, MA, USA) from 2θ = 10° to 80° at 40 kV and 30 mA with a Cu tube (1.5418 Å). The specific surface area of the AMCCS sorbent was investigated using Brunauer–Emmett–Teller (BET) method using an Accelerated Surface Area and Porosimeter system (Micromeritics Instrument Corporation, Norcross, GA, USA). SEM-EDX was performed to determine the elemental composition of the AMCCS sorbent surface using a TOPCN ABT-150S SEM with a JXA 840 EDX instrument (Tokyo, Japan). TEM was used to investigate the microstructure of the AMCCS sorbent with a JEOL JEM-3010 300 KV TEM instrument (Tokyo, Japan). The XRD and TEM analysis of the AMCCS sorbent (Figure S1) showed that the metal oxides on the surface of the AMCCS sorbent were mainly amorphous with negligible crystallinity, while the SEM-EDX analysis showed the presence of Al, Mg, and Ca on the surface of the AMCCS sorbent. The BET surface area of the AMCCS sorbent was determined to be 1.255 m²/g.

2.4. Column Adsorption Experiments

Two dynamic column adsorption experiments were carried out using glass columns (KONTES CHROMAFLEX™ 420830-3020, DKW Life Sciences, Vineland, NJ, USA). Each column was packed with 120 g of AMCCS sorbent, occupying about 42% of the total column volume, and exhibiting a porosity (Φ) of 0.41. Figure 1 shows the experimental setup for adsorption column system; additional operational parameters are listed in

Table 1. Characteristics of AMCCS adsorption column.

Column Parameter	Value
adsorption column material	borosilicate glass
height	30 cm
diameter	2.5 cm
volume	147.3 cm3
pore volume	61.3 cm3
flow rate	10 mL/min, 2 mL/min
linear velocity	2.04 cm/min, 0.41 cm/min
AMCCS bed height	13 cm
AMCCS bed volume	63.83 cm3
AMCCS bed pore volume	26.58 cm3

.

The experiment commenced by introducing an aqueous solution of 5 mg/L fluoride from the top of the adsorption column, with the eluent solution then traveling downwards through a packed bed of the AMCCS sorbent inside the adsorption column. Sampling of the column effluent was carried out at designated time periods. The defluoridation performance of the AMCCS adsorption column was assessed as function of column flow (2 mL/min and 10 mL/min), type of aqueous fluoride solution (DI water and synthetic solution), and sorbent recycling. The column operation continued until the effluent concentration of fluoride was above 4 mg/L, indicating near exhaustion of the adsorption column. The spent AMCCS sorbent was removed from the adsorption column, rinsed with DI water, and was then recycled. The influent and effluent concentrations of fluoride were determined using ion-selective fluoride electrode according to Standard Method 4500-F [42].

Two distinct feed solutions, DI water and synthetic solution with 5 mg/L of fluoride each, were prepared for the column adsorption study. The synthetic solution was formulated to resemble typical groundwater by preparing an aqueous solution containing several ions commonly present in groundwater. The synthetic solution was prepared by adding CaCl₂, NaHCO₃, and Na₂SO₄ to DI water, resulting in a solution containing 1 mM calcium, 2.5 mM bicarbonate, and 0.5 mM sulfate. The synthetic solution had an alkalinity of 125 mg/L as CaCO₃, a total hardness of 100 mg/L as CaCO₃, and a total dissolved solids (TDS) of 340 mg/L.

The adsorption breakthrough curves for the AMCCS adsorption column experiments were obtained by plotting (C_t/C₀) against elution time for each column flow rate, where C₀ was the fluoride level in influent, C_t was the fluoride level in effluent, and adsorbed C was the difference between the influent fluoride and effluent fluoride levels (C₀ − C_t). The cumulative fluoride mass (mg) removed by the AMCCS adsorption column (q_total) for column flow Q was calculated as follows:

$$q_{total}=\mathrm{Q}{\int }_{\mathrm{t}=0}^{\mathrm{t}=\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}(adsorbed C)\mathrm{d}\mathrm{t}$$

(1)

The AMCCS column adsorption capacity (q₀) for fluoride was calculated as

$$q_0={\displaystyle \frac{q_{total}}{\mathrm{M}}}$$

(2)

where M (kg) was the AMCCS sorbent mass present in the AMCCS adsorption column. The fluoride adsorption capacity of the AMCCS column at the breakthrough point was determined when the fluoride concentration reached a target of 1 mg/L of fluoride in the column effluent and (C_t/C₀) = 0.2 for 5 mg/L influent fluoride level. The fluoride adsorption capacity of the AMCCS column at near column exhaustion of the AMCCS sorbent was determined when the fluoride concentration reached about 4.5 mg/L of fluoride in the column effluent and (C_t/C₀) ~ 0.9 = (4.5 mg/L ÷ 5 mg/L).

2.5. Recycling of Column Adsorbent and Reuse Studies

To facilitate the recovery and reuse of AMCCS sorbents within the saturated column, a regeneration method involving the re-coating of spent sorbents was implemented. A total of 120 g of used AMCCS material was collected from the column, thoroughly rinsed with distilled water, and separated into three equal portions of 40 g each. Each portion of the spent AMCCS sorbent was placed in a beaker alongside 100 mL of the previously used coating solution and stirred continuously for 24 h using a magnetic stirrer and plate. Following the mixing process, the coating solution was decanted and saved for subsequent re-coating cycles, while the treated sorbents were placed in a drying oven at 105 °C for 24 h and then placed in a furnace at 220 °C for 24 h. All regenerated sorbent batches were consolidated and reloaded into the column for further reuse and recycling evaluations.

3. Results and Discussion

3.1. Adsorption Column Performance for DI Water

Figure 2 shows the adsorption breakthrough curves for adsorption of fluoride in the AMCCS adsorption column for the two column flow rates. The breakthrough point for the 10 mL/min column flow rate occurred at 620 min for DI water (C_t/C₀ = 0.2 and C_t = 1 mg/L), while the breakthrough point for the 2 mL/min column flow rate occurred at 3360 min for DI water (C_t/C₀ = 0.2 and C_t = 1 mg/L). The AMCCS column fluoride adsorption capacities at the breakthrough point were observed to be 251 mg/kg for the 10 mL/min column flow rate and 273 mg/kg for the 2 mL/min column flow rate.

A longer breakthrough time in the AMCCS adsorption column was observed for the lower column flow rate (2 mL/min) than for the higher column flow rate (10 mL/min). The breakthrough time increased significantly from 620 min (10.3 h) for the 10 mL/min column flow rate to 3360 min (56 h) for the 2 mL/min column flow rate. The later column breakthrough time for the 2 mL/min column flow rate was attributed to the longer residence time, which provided a longer contact time between the AMCCS sorbent and the fluoride solution inside the AMCCS adsorption column, whereas fluoride ions did not have sufficient contact time with the AMCCS sorbent for the 10 mL/min column flow rate, which resulted in their exit from the AMCCS adsorption column.

Other researchers have found that breakthrough time decreases as the flow rate increases. They attributed this to the reduced contact time between fluoride and adsorbent, leading to insufficient time for fluoride ions to diffuse into the adsorbent’s pores or surface area. As a result, fluoride ions leave the column before reaching equilibrium, reducing the efficiency of adsorption [37,43,44]. The breakthrough curve using DI water at a 10 mL/min flow rate for near exhaustion of the AMCCS adsorption column is presented in Figure 3. The breakthrough curve data show that a major increase in effluent fluoride concentration occurred from about 500 min to about 1000 min of column operation, while a lesser gradual increase in effluent fluoride concentration was observed after 1000 min of column operation, approaching greater than 90 percent exhaustion of the AMCCS adsorption column after 2000 min of column operation.

The uptake of fluoride (q_e) in the AMCCS adsorption column is also presented in Figure 3, where the steady uptake of fluoride occurred at a higher level of about 8.2 mg/kg for the initial 500 min of column operation. The fluoride uptake decreased significantly from 500 min to 1000 min of column operation, while slowly decreasing to about 1 mg/kg after 2000 min of column operation. The effluent pH for the AMCCS column is shown in Figure S2, showing an increase in effluent pH from 500 min to 1000 min, with a steady effluent pH after 1000 min of column operation.

3.2. Adsorption Column Performance for Synthetic Solution

Breakthrough curves for adsorption of fluoride in the AMCCS column using the synthetic solution at two column flow rates are shown in Figure 4, where the breakthrough points were observed at 300 min for 10 mL/min column flow and at 2280 min for 2 mL/min column flow. The column adsorption capacities at the breakthrough point were 118 mg/kg for the 10 mL/min column flow and 183 mg/kg for 2 mL/min column flow. The difference in fluoride adsorption capacity was mainly due to the column flow rate and contact time between the AMCCS sorbent and the synthetic solution inside the AMCCS column. Competing ions in the synthetic solution, such as bicarbonate, reduce fluoride removal by occupying the AMCCS sorbent’s active sites. The higher column flow rate of 10 mL/min resulted in a reduced contact time between the AMCCS sorbent and the fluoride solution, limiting fluoride’s ability to interact with the sorbent, allowing competing ions to take up the active sites more quickly. Conversely, the lower flow rate of 2 mL/min provided longer contact time, improving fluoride’s interaction with the sorbent and increasing its adsorption capacity. The treated volumes of synthetic solutions at breakthrough were 4560 mL for column flow of 2 mL/min and 3000 mL for the column flow of 10 mL/min, further demonstrating the impact of column flow rate on treatment capacity.

The fluoride breakthrough curve for 10 mL/min column flow rate using the synthetic solution is presented in Figure 5 for near 90 percent exhaustion of the AMCCS adsorption column. The breakthrough curve data show that a major increase in effluent fluoride concentration occurred from about 200 min to about 800 min of column operation, while a lesser gradual increase in effluent fluoride concentration was observed after 800 min of column operation, approaching greater than 90 percent exhaustion of the AMCCS adsorption column after 1400 min of column operation. The uptake of fluoride (q_e) in the AMCCS adsorption column is also presented in Figure 5, where the steady uptake of fluoride occurred at higher level of about 8 mg/kg for approximately the initial 200 min of column operation. The fluoride uptake decreased significantly after 200 min from about 8 mg/kg to about 2 mg/kg at 800 min of column operation, while slowly decreasing to about 1 mg/kg after 1400 min of column operation. Other research have found that fluoride breakthrough time decreased with increasing bicarbonate concentration in the synthetic solution, mainly due to competition between the fluoride anions and the bicarbonate anions for the available adsorption sites [11]. The effluent pH from the AMCCS column using the synthetic solution is shown in Figure S3. The effluent pH for the AMCCS column shows an increase in effluent pH within the initial 200 min of column operation, with a steady effluent pH after 400 min of column operation.

3.3. Adsorption Capacity of AMCCS Adsorption Column

The fluoride adsorption capacity of the AMCCS adsorption column for DI water was 401 mg/kg for 5 mg/L influent fluoride at 10 mL/min column flow. While adsorption capacity is commonly reported based on adsorbent mass, it can also be reported based on adsorbent area. The fluoride adsorption capacity of the AMCCS adsorption column, based on the surface area of the AMCCS sorbent, was obtained by dividing the AMCCS column fluoride adsorption capacity of 401 mg/kg by the AMCCS sorbent BET surface area of 1.255 m²/g to give a surface-normalized AMCCS column adsorption capacity of 319.5 µg/m². Table 2 shows the column adsorption capacities of several adsorbents using DI water, including the AMCCS adsorption column for comparison. The results shown in Table 2 for surface-normalized column adsorption capacities of several other adsorbents indicate that the surface-normalized fluoride adsorption capacity of the AMCCS adsorption column was comparable to other adsorption columns, while the smallest adsorbent surface area belonged to the AMCCS sorbent. This showed the effectiveness of the Al-Mg-Ca coating of the AMCCS sorbent for appreciable fluoride adsorption occurring in the AMCCS adsorption column.

The equilibrium adsorption of fluoride onto the AMCCS sorbent is shown in Table S1 to follow the Langmuir adsorption model [45] for DI water, the synthetic solution, and tap water (properties of Chicago tap water shown in Table S3); all three solutions had an initial fluoride concentration of 5 mg/L. As shown in Table S1, the equilibrium fluoride adsorption capacity of the AMCCS sorbent for the three solutions decreased as follows: DI water > synthetic solution > tap water. Based on the lower equilibrium fluoride adsorption capacity for tap water, the removal of fluoride from tap water in the AMCCS adsorption column would be expected to be lower than that for DI water and the synthetic solution. The adsorption of fluoride onto the AMCCS sorbent was shown to be physical adsorption for DI water, synthetic solution, and tap water according to the results for E (mean free energy of adsorption) obtained from the Dubinin–Radushkevich (D-R) adsorption equation [46], where the values of E for the three solutions were all less than 8 kJ/mole (Table S1).

Table 2. Comparison of the adsorption column performance in removal of fluoride.

Adsorbent Name	Fluoride in Column Influent (mg/L)	Column Flow Rate (mL/min)	Column Adsorption Capacity qm (mg/kg)	Adsorbent BET Surface Area (m2/g)	Surface-Normalized Column Adsorption Capacity (µg/m2)	Reference
Aluminum–Magnesium–Calcium-coated sand	5	10	401	1.255	319.5	Current work
Activated alumina	5	20	1450	250	5.8	[24]
(PCZH)-complexed PVA hydrogel bead Copper−zirconium	10	0.5	10,430	1.95	5349	[36]
Hydrous ferric oxide	30	20.5	6710	148	45.3	[39]
Kanuma mud	20	5	585	144	4.06	[43]
Al(OH)3@AC	10	7.5	41,840	20	2092	[37]
Aluminum-modified zeolite	10	1	3240	-	-	[38]
Red mud	5	5	1490	10.2	146.1	[47]
Magnesia–pullulan composite	10	16	16,600	32.89	504.7	[11]

Table 2. Comparison of the adsorption column performance in removal of fluoride.

Adsorbent Name	Fluoride in Column Influent (mg/L)	Column Flow Rate (mL/min)	Column Adsorption Capacity qm (mg/kg)	Adsorbent BET Surface Area (m2/g)	Surface-Normalized Column Adsorption Capacity (µg/m2)	Reference
Aluminum–Magnesium–Calcium-coated sand	5	10	401	1.255	319.5	Current work
Activated alumina	5	20	1450	250	5.8	[24]
(PCZH)-complexed PVA hydrogel bead Copper−zirconium	10	0.5	10,430	1.95	5349	[36]
Hydrous ferric oxide	30	20.5	6710	148	45.3	[39]
Kanuma mud	20	5	585	144	4.06	[43]
Al(OH)3@AC	10	7.5	41,840	20	2092	[37]
Aluminum-modified zeolite	10	1	3240	-	-	[38]
Red mud	5	5	1490	10.2	146.1	[47]
Magnesia–pullulan composite	10	16	16,600	32.89	504.7	[11]

The effect of AMCCS sorbent dosage was shown previously in a batch adsorption study [28] to increase the removal of fluoride. Based on the results obtained previously, greater removal of fluoride would be expected if a larger mass of AMCCS sorbent was used in the adsorption column for the same inlet fluoride concentration and column flow rate. The effect of fluoride concentration was shown [28] in a batch adsorption study to have resulted in greater adsorption capacity of the AMCCS sorbent at higher initial fluoride concentrations of up to 30 mg/L. Based on the results obtained previously, the fluoride adsorption capacity of the AMCCS adsorption column would be expected to increase for inlet fluoride concentrations higher than 5 mg/L at the same column flow rate.

The adsorption mechanism for adsorption of fluoride by the AMCCS sorbent was mainly due to electrostatic attraction between the anionic fluoride species and the positively charged surface of the AMCCS sorbent up to the PZC of the AMCCS sorbent (Figure S4), which had a pH_PZC of 10.4 obtained from zeta potential analysis of AMCCS sorbent (Figure S5). The column effluent pH values ranged from 4 to 8 for DI water (Figure S2) and ranged from 4.5 to 8.5 for synthetic solution (Figure S3). The adsorption of fluoride onto the active adsorption sites on the AMCCS sorbent surface occurred on the protonated sorbent surface below the pH_PZC of 10.4, where the sorbent surface was positively charged due to the presence of aluminum, magnesium, and calcium on the AMCCS sorbent surface, as shown by the SEM-EDX results (Figure S1).

For adsorption column operation, an empty bed contact time (EBCT) may be reported for the adsorption column based on column flow rate as follows:

column EBCT = column adsorbent bed volume ÷ column flow rate

(3)

The EBCT values for the AMCCS adsorption column were calculated as 6.4 min for the 10 mL/min column flow rate and as 31.9 min for the 2 mL/min column flow rate; these EBCT values were within the range of commonly applied EBCT values for adsorption column operations. The lower column flow rate (higher EBCT) provided for a longer contact time between the fluoride solution and the AMCCS sorbent; based on the results obtained for EBCT values of 6.4 min and 31.9 min, it would be expected to have greater removal of fluoride at EBCTs longer than 32 min (flow rates lower than 2 mL/min) and to have lesser removal of fluoride at EBCTs shorter than 6 min (flow rates higher than 10 mL/min).

As the fluoride solution flows through the column, the adsorption zone where most of the fluoride removal occurs progressively moves down through the adsorption bed, resulting in an increase in fluoride concentration in the column effluent over time. Benefield et al. [48] explored the adsorption zone in an adsorption column in terms of its formation and transport using the following parameters described in

Table 3. Column adsorption zone parameters.

Parameter	Description
Vs	solution volume processed from column breakthrough to column exhaustion (mL)
Qw	influent flow rate (mL/min)
tz = Vs ÷ Qw	time for fluoride adsorption zone to travel the length of its adsorption zone height after being established (min)
VE	solution volume processed until column exhaustion (mL)
tE = VE ÷ Qw	time needed for adsorption zone to establish itself and exit the bed (min)
hz	adsorption zone height (cm)
h	column bed depth (cm)
Uz = hz ÷ tz = h ÷ (tE − tf)	adsorption zone rate of travel within the column bed (cm/min)
tf = (tz)(1 − F) C0 C VB	time needed for initial formation of fluoride adsorption zone (min) influent fluoride (mg/L) effluent fluoride (mg/L) solution volume processed at breakthrough (mL)
VE Sz = ∫ (C0 − C)dV VB	removal of fluoride through the adsorption zone between breakthrough and column exhaustion (mg)
Smax = (C0)(VE − VB)	maximum amount of fluoride that can be removed through the adsorption zone between breakthrough and column exhaustion (mg)
F = Sz ÷ Smax	AMCCS sorbent fraction within the adsorption zone at breakthrough that still has the capacity to remove fluoride
(100%)[h + (F − 1) hz] ÷ h	percentage saturation of AMCCS sorbent at breakthrough

: t_z, t_E, t_f, U_Z, F, and percent saturation.

Table 4. Specification of AMCCS sorbent column operating parameters for 10 mL/min flow rate for DI water and synthetic solution.

Column Eluent	qtotal (mg)	qe (mg/kg)	EBCT (min)	tz (h)	tE (h)	SZ (mg)	Smax (mg)	F	tf (h)	Uz (cm/h)	hz (cm)	Saturation
DI water	48	401	6.38	21.7	32	17.95	65	0.276	15.7	0.797	17.3	3.8%
synthetic solution	31.5	263	6.38	15	20	17.34	45	0.385	9.23	1.207	18.1	14.4%

shows the results of the calculated adsorption column parameters for the AMCCS adsorption column according to the parameters from Table 3. The AMCCS sorbent column demonstrates better performance with DI water compared to the synthetic solution across most parameters. DI water shows a higher total adsorption capacity (q_total = 48 mg vs. 31.5 mg), higher adsorption per unit mass (q_e = 401 mg/kg vs. 263 mg/kg), and longer breakthrough (10.3 h versus 5 h) and exhaustion times (t_E = 32 h versus t_F = 20 h); these results are indicative of greater removal efficiency and a longer operational lifespan for DI water.

The fluoride adsorption zone traveled within the AMCCS column was at a higher velocity for the synthetic solution (U_z = 1.207 cm/h) than for DI water (U_z = 0.797 cm/h), resulting in a shorter timespan for the adsorption zone to travel through the AMCCS column (t_z = 15 h versus t_z = 21.7 h). The higher velocity and the shorter travel time of the fluoride adsorption zone resulted in a lower fluoride adsorption capacity and removal efficiency of the AMCCS adsorption column using the synthetic solution. Greater saturation of the AMCCS sorbent for fluoride occurred for the synthetic solution mainly due to competition from bicarbonate ion for the fluoride adsorption sites available on the AMCCS sorbent surface, resulting in a lower fluoride removal efficiency of the AMCCS adsorption column using the synthetic solution.

3.4. Column Adsorption Modeling

To assess the potential of the designed sorbent for industrial applications, the Thomas model was applied to analyze lab-scale column experiments and predict breakthrough curves. The Thomas model [49] assumes that adsorption follows the Langmuir adsorption equilibrium model [45] and second-order adsorption kinetics [50], while limited by interphase mass transfer with negligible axial dispersion. The Thomas model may be linearized using

$$\mathrm{l}\mathrm{n}\left({\displaystyle \frac{{\mathrm{C}}_0}{{\mathrm{C}}_{\mathrm{t}}}}-1\right)={\displaystyle \frac{{\mathrm{K}}_{\mathrm{T}\mathrm{h}}{\mathrm{q}}_0\mathrm{X}}{\mathrm{Q}}}-{\mathrm{K}}_{\mathrm{T}\mathrm{h}}{\mathrm{C}}_{0}t$$

(4)

where, for column flow Q, K_Th denotes the Thomas rate constant, q₀ denotes the column adsorption capacity, and X denotes the adsorbent mass. Since the Thomas model is derived based on second-order adsorption kinetics, the model results are limited by the effect of interphase mass transfer, leading to errors when applying the model to column adsorption.

The Thomas model was used to fit the experimental AMCCS column adsorption data for C/C₀ ratios above 0.01 (Figure 6).

Table 5. AMCCS adsorption column parameters using Thomas model.

AMCCS Column	Influent Fluoride C0 (mg/L)	Column Flow Rate = Q (mL/min)	AMCCS Sorbent Mass = X (g)	KTh (L/mg.min)	Calculated Adsorption Capacity = q0 (mg/kg)	Experimental Adsorption Capacity (mg/kg)	R2	% Error
DI water	5	10	120	7.2 × 10−4	528	401	0.794	24
Synthetic solution	5	10	120	1.02 × 10−3	282	263	0.800	7

shows the results for column adsorption parameters, K_Th and q₀ values (calculated adsorption capacity), obtained from the linear regression of the Thomas model linearized form for 5 mg/L influent fluoride at 10 mL/min column flow.

The AMCCS sorbent was shown to follow the Langmuir adsorption model (Table S1) and second-order kinetics for DI water (Table S2), while the AMCCS sorbent was shown to also follow the Langmuir adsorption model (Table S1) and second-order kinetics (Table S2) for the synthetic solution. As shown in Table 5, the experimental adsorption capacity value for the synthetic solution was more similar to the value predicted by the Thomas equation (q_o) than the corresponding values for DI water, with a 7 percent deviation for the synthetic solution and a 24 percent deviation for DI water.

The results show a better fit of the Thomas model for the synthetic solution after breakthrough of fluoride (Figure 6), and an overall less error in estimating the AMCCS column adsorption capacity for the synthetic solution (Table 5). The Thomas model results for DI water and the synthetic solution indicate that the adsorption of fluoride using the synthetic solution was affected by interphase mass transfer to a lesser extent. Adsorption of fluoride by the AMCCS sorbent in batch adsorption experiments was shown to be affected by interphase mass transfer for both DI water and the synthetic solution according to the results obtained from plots of q vs. t^1/2 (Figure S6) based on the Weber and Morris equation for intraparticle diffusion [51], where the effect of external mass transfer for DI water was more pronounced as shown by the steeper first segment of the plots for DI water versus the synthetic solution.

Based on the results that showed the adsorption of fluoride by AMCCS sorbent using tap water while also following the Langmuir adsorption equilibrium model and second-order kinetics (Tables S1 and S2), the column adsorption of fluoride using tap water would be expected to be influenced less by interphase mass transfer due to the lesser external mass transfer limitation observed in Figure S6 for tap water versus DI water and the synthetic solution (less steep slope of the first segment of plot for tap water).

3.5. Column Recycling and Reuse

The study of recycling and reusing exhausted sorbent is crucial, as it is an important aspect of designing a system that is practical, economical, and competitive in the market. The AMCCS adsorption column performance results using consecutive adsorption column cycles are presented as breakthrough curves in Figure 7. The AMCCS adsorption column performance results using consecutive adsorption column cycles are presented as fluoride uptake curves in Figure 8. The column study results for AMCCS sorbent recycling showed promising outcomes, demonstrating that the AMCCS column successfully removed fluoride from both solutions (Figure 7 and Figure 8) after re-coating the spent AMCCS sorbent, with fluoride removal efficiency comparable to its performance before re-coating. Additionally, there was no negative impact of the co-existing ions in the synthetic solution on the re-coating process and the column performance for the synthetic solution using sorbent recycling.

Table 6. Comparison of the AMCCS column performance using sorbent recycling for DI water and synthetic solution for 5 mg/L fluoride influent and 10 mL/min column flow.

AMCCS Sorbent Column (DI Water)	Adsorption Capacity (qm, mg/kg)	Breakthrough Point (min)	AMCCS Sorbent Column (Synthetic Solution)	Adsorption Capacity (qm, mg/kg)	Breakthrough Point (min)
Fresh Column	401	620	Fresh Column	263	300
Re-coated Column	424	740	Re-coated Column	289	380
2nd Re-coated Column	388	660

shows the results from the AMCCS column sorbent recycling experiments demonstrating the potential for reuse of the AMCCS column, with some performance variations across adsorption cycles. The recycling performance of the AMCCS column for DI water showed improvement in adsorption capacity from 401 mg/kg (fresh column sorbent) to 424 mg/kg (re-coated column sorbent), while extending the breakthrough point from 620 min to 740 min.

This suggests that re-coating effectively regenerated the AMCCS column sorbent, restoring and even slightly enhancing its fluoride adsorption capacity for DI water. After the second re-coating, the adsorption capacity was 388 mg/kg and the breakthrough point was 660 min. Using DI water, the average fluoride uptake using DI water for the three adsorption column runs was 404 mg/kg and the average breakthrough point for the three adsorption column runs was 673 min. Re-coating the sorbent enhanced the column’s performance when treating the synthetic solution, increasing the adsorption capacity from 263 mg/kg to 289 mg/kg and extending the breakthrough time from 300 to 380 min. Breakthrough curves support these findings, illustrating prolonged breakthrough times for re-coated columns relative to the original. Overall, the results suggest that recycling and reusing AMCCS sorbent in column configuration is a viable strategy. This contrasts with previous studies, which often reported an immediate decline in performance and breakthrough time following reuse [35,44].

The surface composition of the AMCCS sorbent before the column adsorption experiment was based on the results obtained from SEM-EDX analysis (Figure S1). The SEM-EDX surface analysis showed that there was 3.6 percent aluminum, 1 percent magnesium, and 1 percent calcium present on the surface of the AMCCS sorbent as components of the surface coating of the AMCCS sorbent. The surface composition of the regenerated AMCCS sorbent after recoating was expected to be similar to the composition of the fresh sorbent, resulting in similar removal of fluoride for both DI water and synthetic solution for the fresh AMCCS sorbent and for the regenerated AMCCS sorbent (as shown in Figure 7 and Figure 8).

The sustainability of the AMCCS sorbent was improved by its capacity to undergo multiple re-coating cycles, significantly reducing the frequency of sorbent replacement. This approach conserves raw materials and lessens the overall environmental impact. Unlike traditional regeneration methods, which often require high energy input and chemical use and produce wastewater that demands expensive treatment, re-coating offers a more economical and environmentally responsible alternative. Re-coating lowers operational costs and environmental strain by minimizing the need for transporting new sorbents and reducing wastewater generation. It also avoids the carbon emissions and resource consumption involved in manufacturing new sorbent materials, resulting in a reduced environmental footprint. These advantages make AMCCS sorbent a strong candidate for long-term application, delivering cost savings in material production, waste management, and logistics while supporting more sustainable water treatment practices.

4. Conclusions

Column adsorption of fluoride was carried out using the Al-Mg-Ca-coated sand (AMCCS) sorbent in packed column experiments. Two different column flow rates were applied for two different solutions of DI water and a synthetic solution, both with 5 mg/L of fluoride in the influent column flow. The synthetic solution resembled a typical groundwater solution containing calcium, bicarbonate, and sulfate. The column breakthrough point and the column adsorption capacity were assessed for 1 mg/L fluoride in the column effluent flow using the breakthrough curves obtained for adsorption column experiments. Longer breakthrough times were observed for the lower column flow rate using either DI water or the synthetic solution. Breakthrough times of 620 min for the 10 mL/min flow rate and 3360 min for the 2 mL/min flow rate were observed for DI water. Breakthrough times of 300 min for the 10 mL/min flow rate and 2280 min for the 2 mL/min flow rate were observed for the synthetic solution. The AMCCS column adsorption capacity at the breakthrough point for DI water was 251 mg/kg for column flow of 10 mL/min and 273 mg/kg for column flow of 2 mL/min. The AMCCS column adsorption capacity at the breakthrough point for the synthetic solution was 118 mg/kg for column flow of 10 mL/min and 183 mg/kg for column flow of 2 mL/min. Greater column adsorption capacity and longer breakthrough time were observed for DI water; the smaller adsorption capacity observed for the synthetic solution was caused by other ions competing with fluoride. The recycling and reuse of the AMCCS column was carried out through the re-coating of the spent AMCCS column sorbent in sequential column adsorption cycles. Successful recycling of the AMCCS column was achieved, where defluoridation with recycled AMCCS column sorbent was similar to defluoridation with fresh AMCCS column sorbent for both column eluents (DI water and the synthetic solution).