Removal of HF via CaCl₂-Modified EAF Slag: A Waste-Derived Sorbent Approach

Abstract

This study evaluates CaCl₂-modified electric arc furnace (EAF) slag for fluoride removal from synthetic hydrofluoric acid (HF) wastewater. Adsorption performance was assessed under different particle sizes (850 μm–1.7 mm, 250–850 μm, and <250 μm), temperatures (25–45 °C), and initial pH values (2–11), using oxidized (EOS) and reduced (ERS) slags in raw and modified (C1, C2) forms. Characterization included isotherm modeling (Langmuir and Freundlich), X-ray diffraction (XRD), and inductively coupled plasma mass spectrometry (ICP-MS). The CaCl₂-modified slags (particularly EOS-C2 and ERS-C2) demonstrated stable performance under all conditions. ERS-C2 achieved the maximum adsorption capacity of 16.13 mg/g at 600 mg F⁻/L. EOS-C2 maintained capacities above 8.0 mg/g across pH 2–11, whereas unmodified slag showed a decline in performance above pH 5, with residual concentrations exceeding 250 mg F⁻/L and capacities dropping to 1.14–2.14 mg/g. XRD analysis indicated increased amorphization and enhancement of dicalcium silicate and brownmillerite phases after modification. Isotherm fitting showed better agreement with the Freundlich model, suggesting multilayer adsorption. Leaching tests confirmed that Cr, Cu, and As concentrations were within safe limits, while Pb and Cd were not detected. These results demonstrate the strong potential of CaCl₂-modified EAF slag as an efficient, pH-stable, and environmentally safe adsorbent for treating HF-containing industrial wastewater.

1. Introduction

As of 2024, the global semiconductor market is valued at approximately USD 665 billion, with Asia accounting for 50.94% of the total market [1]. The rapid growth of the semiconductor industry has led to increasing concerns regarding the environmental pollutants and wastewater generated during semiconductor fabrication. Semiconductor manufacturing involves multiple steps, including wafer production, oxidation, photolithography, etching, and deposition, which consume large volumes of chemicals and produce thousands of tons of wastewater. Particularly, wet etching commonly employs strong inorganic acids, such as nitric acid, phosphoric acid, and hydrofluoric acid (HF), generating highly acidic wastewater that poses significant environmental risks. Owing to the high toxicity of HF and its resistance to biological treatments, chemical methods must be used for its removal [2].

Fluoride-containing wastewater is typically treated via adsorption, coagulation with lime and alum, and filtration. However, at high fluoride concentrations, these methods require excessive input of chemicals and generate large volumes of sludge, leading to increased operational costs [3]. Among the available treatment methods, coagulation and adsorption are widely used because of their simplicity and ease of operation. Various studies have focused on the development of effective adsorbents. Traditional adsorbents include alumina; aluminum-, calcium-, and iron-based materials; as well as natural or waste-derived materials such as lignite, fine coke, and bituminous coal. In recent years, industrial by-products such as slag have been explored as alternative adsorbents for the removal of various contaminants [4,5,6].

Steelmaking slags have been reported to effectively adsorb heavy metals such as Cu²⁺, Cd²⁺, Pb²⁺, and Mn²⁺ [7,8,9] and also can remove phosphorus, nitrogen, ammonia, and dissolved organics from wastewater after acid, alkali, or thermal modification [10]. Kostura et al. demonstrated that blast furnace slag can remove phosphate from water [11], and other studies have reported the effective removal of dyes using similar materials [12].

Electric arc furnace (EAF) slag is generated during steel production from scrap metal and contains a high content of free lime (CaO), which causes volume expansion and disintegration when exposed to moisture. Consequently, EAF slag is more commonly used as a construction aggregate than as a cement precursor. Additionally, owing to the heterogeneous composition and potential heavy metal content of EAF slag, its applications are limited. Nevertheless, EAF slag is rich in CaO and Al₂O₃, which can react with fluoride ions to form insoluble compounds, making it a promising candidate material for HF wastewater treatment. However, its low porosity and specific surface area limit its adsorption performance [10].

To enhance its adsorption capacity, EAF slag was thermally modified to improve nitrogen and phosphorus removal in some studies [13,14], whereas other studies have used aluminum hydroxide to modify converter slag for the removal of ammonium, phosphate, and cadmium [15,16]. Despite these developments, few studies have explored the application of modified slags to HF wastewater treatment. Particularly, CaCl₂ was selected as a modifying agent in this study because Ca²⁺ ions can react with fluoride to form insoluble CaF₂, providing precipitation and surface activation effects.

Therefore, this study aimed to enhance the fluoride adsorption performance of EAF slag through CaCl₂ modification. The modified slags were applied to synthetic HF wastewater (300 mg/L) under varying particle size, pH, and temperature conditions. The adsorption characteristics were evaluated using isotherm experiments, structural changes were analyzed via X-ray diffraction (XRD), and the leaching behavior of heavy metals (Cr, Cu, As, Pb, and Cd) was assessed to confirm the environmental safety of the modified materials.

2. Materials and Methods

2.1. EAF Slag

The electric arc furnace (EAF) slag used in this study was provided by a domestic steelmaking company (Company A, Busan, Republic of Korea). Two types of slags were used: oxidized slag (Figure 1a) and reduced slag (Figure 1b). The oxidized and reduced slags were generated during the EAF process under different furnace conditions: oxidized slag was produced during the oxidizing refining stage (with abundant oxygen injection), while reduced slag was formed during the reducing refining stage (with the addition of reducing agents).

Typically, oxidized slag is generated in the early to intermediate stages of steelmaking, whereas reduced slag appears after oxidizing refining or at the final stage of the process. The chemical compositions of the slags were provided by Company A, and their major components are summarized in

Table 1. Composition of reduced and oxidized slags (wt%).

											(Unit: %)
Type		CaO	SiO2	MgO	Al2O3	Fe	MnO	P2O5	SO3	Cr2O3	TiO2	F2O3
EAF OS	Avg.	16.8	18.8	10.0	14.1	18.2	9.6	0.1	0.0	4.0	1.5	23.4
	Max.	20.0	20.6	12.8	18.5	27.0	15.8	0.2	0.0	5.2	1.8	34.7
	Min.	13.9	15.6	8.1	11.4	11.6	7.2	0.0	0.0	2.5	1.1	14.9
EAFRS	Avg.	31.9	29.1	17.4	9.9	2.5	4.2	0.0	0.0	0.4	0.8	3.2
	Max.	42.4	35.8	24.0	15.8	25.8	25.5	0.1	0.1	4.6	1.5	33.2
	Min.	16.0	16.1	7.4	6.3	0.5	0.7	0.0	0.0	0.0	0.4	0.7

.

After primary crushing, the slag samples were sieved using standard mesh sieves (80, 150, and 200 mesh) and classified into three particle size ranges of 850 μm–1.7 mm, 250–850 μm, and <250 μm, respectively.

The unmodified oxidized and reduced slags were designated as EOS-B and ERS-B, respectively. For chemical modification, 0.5 and 1.0 M CaCl₂ solutions were prepared. The slag (30 g) was added to each CaCl₂ solution (300 mL), and the mixture was stirred at 100 rpm for 6 h at 40 °C. After cooling to room temperature, the solid was separated using a GF/C filter and dried at 105 ± 0.5 °C for 12 h.

The slags treated with 0.5 M CaCl₂ solution were designated as EOS-C1 and ERS-C1, whereas those treated with 1.0 M CaCl₂ solution were designated as EOS-C2 and ERS-C2, respectively.

2.2. HF Wastewater Solution

In semiconductor manufacturing processes, hydrofluoric acid (HF) is widely used to remove metal impurities, organic contaminants, unwanted particles, and surface oxide films from wafers without damaging the substrate. As a result, a significant amount of HF-containing wastewater is generated.

Typically, HF wastewater accounts for approximately 60–70% of the total effluent from semiconductor fabrication, with a strongly acidic pH of 2–3. The representative characteristics of semiconductor wastewater in South Korea, as reported in previous studies, are summarized in

Table 2. Characteristics of semiconductor wastewater in South Korea.

				(Unit: mg/L)
Item	Jeon (2003) [17]	An (2010) [2]	Byeon (2005) [18]	An (2012) [19]
pH	3.37	2.8	2.3	2.3
F−	214.94	330.2	375~442	327
PO43-P	-	322.7	1000~1350	432
NH4+-N	-	137.8	-	176
NO3−	-	-	310~400	62
SO42−	-	-	350~400	1200
CODCr	56.5	-	-	561

[2,17,18,19].

To simulate the effluents typically generated from semiconductor manufacturing processes, synthetic HF-containing wastewater was prepared using a 50% HF solution (Daejung, Siheung-si, Republic of Korea), with a target fluoride concentration of 300 mg/L at pH 2.5.

2.3. Batch Adsorption Experiments

All batch adsorption experiments were conducted for 24 h in a thermostatic water-bath shaker at 100 rpm. To evaluate the effect of particle size, EOS and ERS were classified into three ranges (850 μm–1.7 mm, 250–850 μm, and <250 μm), and each sample (1 g) was reacted with the fluoride solution (30 mL, 300 mg/L). The pH-dependent study was carried out by adjusting the solution pH from 2 to 11 using NaOH and H₂SO₄ at a constant temperature of 25 °C. To examine the effect of temperature variation, the experiments were conducted at 25, 35, and 45 °C. Adsorption isotherm experiments were performed using the modified slag under a constant solid-to-liquid ratio (1 g/30 mL), initial fluoride concentrations ranging from 50 to 600 mg/L, and a temperature of 25 °C. After 24 h, the samples were allowed to settle for 2 h, and the supernatant was used directly for analysis without filtration.

2.4. Analytical Methods

Fluoride concentrations were measured using a spectrophotometer (HS-400; Humas, Daejeon, Republic of Korea), and pH was determined using a pH meter (pH 330i; WTW, Weilheim, Germany). The concentrations of heavy metals (Cr, Cu, As, Pb, and Cd) were analyzed using inductively coupled plasma mass spectrometry (ICP-MS; NexION 2000 Series, Perkin Elmer, Waltham, MA, USA). X-ray diffraction (XRD) patterns were recorded using an X’Pert³ diffractometer (Panalytical, Almelo, The Netherlands) with Cu Kα radiation over a 2θ range of 5–80°.

2.5. Adsorption Isotherm Models

To analyze the adsorption of fluoride on the slag, the experimental data were interpreted using the Freundlich and Langmuir isotherm models.

The Freundlich model is an empirical equation that assumes the presence of heterogeneous multilayer adsorption sites. It is particularly suitable for describing nonlinear adsorption processes such as chemisorption or surface precipitation. This model is expressed as follows [20]:

$${\displaystyle \frac{x}{m}}=K_f{C}_e^{{\displaystyle \frac{1}{n}}}$$

(1)

where $x$ (mg) is the amount of adsorbed fluoride, $m$ (g) is the mass of the adsorbent, $C_e$ (mg/L) is the equilibrium concentration of fluoride in the solution, and $K_f$ and $n$ are the Freundlich constants. When $n$ = 1, the model is reduced to a linear isotherm. If $n$ > 1, the adsorption density decreases with increasing concentration, indicating a decrease in the adsorption affinity. A higher value of $K_f$ typically implies a lower adsorption effectiveness.

By contrast, the Langmuir model assumes monolayer adsorption onto a surface with a finite number of identical sites. This is described by the following equation [21]:

$${\displaystyle \frac{x}{m}}={\displaystyle \frac{X_{m}b{C}_e}{1+b{C}_e}}$$

(2)

where $X_m$ (mg/g) is the maximum adsorption capacity and $b$ is a dimensionless constant related to the binding energy of adsorption. As the equilibrium concentration $C_e$ increases, the number of adsorption sites on the surface approaches saturation, and the adsorbed amount $x/m$ approaches $X_m$.

3. Results and Discussion

3.1. Fluoride Removal Characteristics of Modified Slag for Different Particle Sizes

To evaluate the dependence of the fluoride adsorption performance of the slag on the particle size and slag type, synthetic HF wastewater with 300 mg/L fluoride at pH 2.5 was used. Each slag sample (EOS-B, EOS-C1, EOS-C2, ERS-B, ERS-C1, and ERS-C2, 1 g) was added to the fluoride solution (30 mL). The samples were categorized into three particle size ranges: ① 850 μm–1.7 mm, ② 250–850 μm, and ③ <250 μm. The reaction was carried out in a water bath at 25 °C and 100 rpm for 24 h, and the results are shown in Figure 2.

As shown in Figure 2a, slag samples with a particle size of <250 μm exhibited superior fluoride removal performance regardless of modification. The residual fluoride concentrations were 125.0, 29.0, and 18.5 mg/L for EOS-B, EOS-C1, and EOS-C2, respectively, which were significantly lower than those of the groups with larger particle sizes (250–850 μm and 850 μm–1.7 mm). Regarding the CaCl₂ modification effect, the adsorption capacity of EOS-C2 was 3.2 mg/g higher than that of EOS-B for the particle sizes of <250 μm and 3.7 mg/g higher for the particle sizes of 850 μm–1.7 mm, indicating a more pronounced enhancement in the fluoride uptake due to the modification.

A similar trend was observed for the reduced slag, as shown in Figure 2b. For particle sizes of <250 μm, the residual fluoride concentrations were 85.0, 17.0, and 8.5 mg/L for ERS-B, ERS-C1, and ERS-C2, respectively, indicating superior removal at smaller particle sizes. Compared to the oxidized slag under the same particle size conditions, the ERS samples exhibited residual fluoride concentrations that are lower by 10–40 mg/L. The improvement in the adsorption capacity due to CaCl₂ modification was also evident: for the particle sizes of <250 μm, EOS-C2 showed an increase of 2.3 mg/g over EOS-B, and for the particle sizes of 850 μm–1.7 mm, an increase of 3.2 mg/g was observed, demonstrating a stronger modification effect.

Calcium oxide (CaO) present in the slag can react with fluoride ions and remove them via precipitation, as described in the following reaction:

$${\mathrm{Ca}}^{2+}+2{\mathrm{F}}^{-}\to \mathrm{Ca}{\mathrm{F}}_2\downarrow +2\mathrm{O}{\mathrm{H}}^{-}$$

(3)

Islam (2011) [6] showed that thermally treated basic oxygen furnace (BOF) slag can be effectively used for fluoride removal. Similarly, Claveau-Mallet (2013) [22] reported that the application of a steel slag column to mining wastewater containing phosphorus, fluoride, Mn, and Zn resulted in a fluoride removal efficiency of approximately 85.3% for influent concentrations ranging from 9 to 37 mg/L. The removal mechanism was attributed to the formation of fluoroapatite (Ca₅(PO₄)₃F) via precipitation.

According to adsorption theory, reducing the particle size of an adsorbent increases its available surface area, thereby enhancing the adsorption capacity and reducing the time required to reach equilibrium [23]. The positive correlation between smaller particle size and adsorption efficiency has been widely reported in the literature [24,25]. For example, Kwon (1999) [26] reported that in the treatment of acid mine drainage using steelmaking slag with particle sizes in the ranges of <5 mm, 5–20 mm, and >20 mm, smaller particles exhibited superior removal of heavy metals, such as Fe, Al, Mn, Ni, and Zn.

3.2. Fluoride Removal Characteristics of Modified Slag at Different Temperatures

To evaluate the effect of the reaction temperature on fluoride removal, synthetic HF wastewater containing 300 mg/L F⁻ at pH 2.5 was used. Each slag sample (EOS-B, EOS-C1, EOS-C2 and ERS-B, ERS-C1, and ERS-C2, 1 g) was added to the fluoride solution (30 mL) and reacted for 24 h at 25, 35, and 45 °C in a water bath at 100 rpm. The results are shown in Figure 3.

As presented in Figure 3a, for the unmodified oxidized slag (EOS-B), the residual fluoride concentration decreased from 125.0 mg/L at 25 °C to 97.5 mg/L at 35 °C and 80.0 mg/L at 45 °C, indicating that fluoride removal improves with increasing temperature. The adsorption capacity increased from 5.3 mg/g (35 °C) to 6.6 mg/g (45 °C). By contrast, the EOS-C1 and EOS-C2 modified slags showed relatively consistent performance across the examined temperature range, with residual fluoride concentrations between 18.5 and 35.0 mg/L and adsorption capacities between 8.1 and 8.4 mg/g, indicating minimal influence of the temperature.

A similar trend was observed for the reduced slags, as shown in Figure 3b. The residual fluoride concentration of ERS-B decreased from 85.0 mg/L at 25 °C to 52.5 mg/L at 35 and 45 °C, suggesting improved removal at higher temperatures, even though the effect was less pronounced than that in EOS-B. The ERS-C1 and ERS-C2 modified slags also maintained stable performance across all temperatures, with residual fluoride concentrations ranging from 10.0 to 26.0 mg/L.

In previous studies on the effect of temperature on fluoride adsorption, Valdivieso (2006) [27] reported that fluoride adsorption on α-Al₂O₃ decreased with increasing temperature (25→40 °C), indicating an exothermic reaction mechanism, whereas the study of graphene-based adsorbents performed by Li (2011) [28] found that as the temperature increased, the adsorption capacity increased from 12.29 to 19.59 mg/g, suggesting an endothermic process. Mondal (2016) [29]. also reported that temperature affected the fluoride removal performance of CaCO₃, Al₂O₃, and sugarcane charcoal-based activated carbon in the 300–335 K range, with the adsorption mechanism varying between endothermic and exothermic depending on the material.

Kumari (2020) [30] demonstrated that as the temperature increased, EAF slag modified by heat treatment at 600 °C and acid activation showed an improvement in the adsorption capacity from 10.368 at 298 K to 13.436 mg/g at 318 K. These findings are consistent with the results of the present study, where the EAF slag showed increased fluoride adsorption at elevated temperatures, likely due to its composite composition involving Ca, Al, and other active components, and is governed by an endothermic adsorption mechanism.

3.3. Fluoride Removal Characteristics of Modified Slag for Different Initial pH

To evaluate the effect of initial pH on the fluoride removal performance, synthetic HF wastewater containing 300 mg/L F⁻ at pH 2.5 was used. Each slag sample (EOS-B, EOS-C1, EOS-C2 and ERS-B, ERS-C1, and ERS-C2, 1 g) was added to the fluoride solution (30 mL), and the pH was adjusted in a range of 2–11 using 1 M NaOH and H₂SO₄. The reactions were conducted at 25 °C and 100 rpm for 24 h. pH 2.5 represents the original solution without any adjustment. The results are shown in Figure 4.

As shown in Figure 4a, for the unmodified oxidized slag (EOS-B), the residual fluoride concentration was 125.0 mg/L at pH 2.5 and decreased to 82.5 mg/L at pH 2. However, when the initial pH was adjusted to 5–11, the residual fluoride concentration increased significantly to 246.0–262.0 mg/L, and the corresponding adsorption capacities dropped to 1.14–1.62 mg/g, indicating a clear decline in the removal effectiveness compared to acidic conditions.

By contrast, the modified samples (EOS-C1 and EOS-C2) exhibited much smaller variations in the removal efficiency. At pH 2, the residual fluoride concentrations were 21.0 and 17.5 mg/L for EOS-C1 and EOS-C2, respectively, and were 63.0 and 24.8 mg/L at pH 11, respectively. Particularly, EOS-C2 showed highly consistent performance with adsorption capacity ranging from 8.26 to 8.48 mg/g across the entire pH range. For EOS-C1, the capacity decreased from 8.37 mg/g at pH 2 to 7.11 mg/g at pH 11, indicating a somewhat greater sensitivity under alkaline conditions.

Similar trends were observed for the reduced slags (Figure 4b). At pH 2, the adsorption capacities were 7.28, 8.88, and 8.81 mg/g for ERS-B, ERS-C1, and ERS-C2, respectively, which were 0.33–0.75 mg/g higher than those of the corresponding oxidized slags. The residual fluoride concentration of ERS-B increased to 253.0 mg/L at pH 11, and its adsorption capacity declined accordingly. The ERS-C1 and ERS-C2 modified samples exhibited excellent performance at pH 2–2.5, with residual fluoride concentrations between 4.0 and 8.5 mg/L. However, ERS-C1 exhibited a noticeable decrease in performance at pH 11, reaching 89.0 mg/L, showing greater sensitivity to pH than EOS-C1.

According to Ren (2022) [31], for modified carbide slag, fluoride adsorption was shown to be most effective at pH 2–5 and declined significantly at pH > 6. This was attributed to the protonation of the adsorbent surface at lower pH, which enhanced the electrostatic attraction with fluoride ions, whereas deprotonation at higher pH resulted in repulsion and decreased adsorption. Similarly, Oh (2012) [32,33] reported maximum fluoride removal using biochar at pH 5.1–6.2, with removal efficiency decreasing at higher pH values. This behavior was linked to the presence of aluminum and iron oxides on the adsorbent surface, which may form fluoroaluminate complexes (AlF_i^3−I) under acidic conditions, thereby enhancing fluoride uptake.

In another study, Lacson (2021) [34] found that Ca-based coagulants exhibited the highest fluoride removal efficiency at pH 5.0, whereas lower pH (e.g., pH 2) promoted HF formation, hindering CaF₂ precipitation. At pH values above 5, excess OH⁻ also interfered with the precipitation process.

The post-reaction pH values are listed in

Table 3. Changes in solution pH after fluoride adsorption under different initial pH conditions.

Initial pH	EOS-B	EOS-C1	EOSC-2	ERS-B	ERS-C1	ERS-C2
pH 2	7.22	7.12	6.83	7.86	7.76	7.36
pH 2.5 *	8.28	7.74	7.45	8.57	7.97	7.65
pH 5.0	11.1	11.15	10.52	11.62	11.52	10.87
pH 7.0	11.27	11.19	10.77	11.68	11.62	11.16
pH 9.0	11.26	11.15	10.37	11.71	11.6	10.96
pH 11.0	11.41	11.34	10.96	11.76	11.66	11.11

Note: * At pH 2.5, the experiment was performed without pH modification, reflecting the acidity of the solution.

. Samples initially adjusted to pH 2–2.5 showed only moderate increases in final pH (up to 8.57), whereas those initially adjusted to pH 5 or higher exhibited final pH values exceeding 11, indicating a strong alkaline shift. Such pH elevation may inhibit Ca²⁺–F⁻ precipitation due to OH⁻ accumulation.

In the study by Singh (2021) [35], the leaching experiment of EAF slag also showed that the pH exceeded 13 regardless of the initial pH. This is because EAF slag contains CaO, Ca-Mg oxides, and silicates, which react with water to form Ca(OH)₂, thereby increasing the pH of the solution. Similarly, Riley (2015) [36] reported a positive correlation between calcium content and pH in strongly alkaline leachates from steelmaking slag, which is consistent with the findings of this study.

The unmodified slag showed a strong dependence on the pH, with the fluoride removal efficiency significantly reduced at higher pH. By contrast, modified slags maintained relatively stable performance across the entire pH range. This is attributed to the increased availability of the Ca²⁺ ions and enhanced CaF₂ precipitation following modification, which is consistent with the finding of Lacson (2021) [34] that optimal CaF₂ precipitation occurs near pH 5. By contrast, unmodified slag, which contains higher proportions of Fe₂O₃ and Al₂O₃, may rely more strongly on the acid-driven mechanisms, such as surface complexation or fluoroaluminate formation under acidic conditions.

3.4. Isotherm Modeling of Fluoride Adsorption onto EAF Slag

To evaluate the fluoride adsorption characteristics of EAF slag, slag samples (EOS-B, EOS-C1, EOS-C2, ERS-B, ERS-C1, and ERS-C2) with particle sizes less than 250 μm were used. Each sample (1 g) was added to fluoride solutions (30 mL) with the initial concentrations of 50, 150, 300, 450, and 600 mg/L. The mixtures were reacted at 25 °C and 100 rpm for 24 h until equilibrium was reached. After the reaction, the residual fluoride concentration in the solution was measured, and the adsorption data were fitted to the Freundlich and Langmuir isotherm models.

As shown in Figure 5a, an increase in the initial F-concentration led to an increase in the residual F-concentration. Accordingly, the adsorption capacity increases, reaching a maximum of 16.13 mg/g for ERS-C2 at an initial concentration of 600 mg/L (Figure 5b). No maximum adsorption plateau was observed within the concentration range used in this study.

The isotherm fitting results are presented in Figure 6, and the corresponding model parameters are listed in

Table 4. Isotherm model parameters (Freundlich and Langmuir) for fluoride adsorption by EAF slag samples.

Sample	Freundlich			Langmuir
	$\mathit{n}$	${\mathit{K}}_{\mathit{f}}$	${\mathit{R}}^2$	a	b	${\mathit{R}}^2$
EOS-B	1.205	0.148	0.946	52.083	0.002	0.097
EOS-C1	1.133	0.278	0.929	39.216	0.006	0.504
EOS-C2	1.309	0.464	0.962	25.126	0.013	0.913
ERS-B	1.191	0.186	0.983	41.667	0.028	0.364
ERS-C1	1.364	0.59	0.864	23.753	0.016	0.801
ERS-C2	1.212	0.599	0.856	34.843	0.013	0.486

. In the Freundlich model, the adsorption affinity constant $K_f$ was the highest for ERS-C2 (0.599), indicating superior fluoride uptake. By contrast, the Langmuir model predicted the highest theoretical maximum adsorption capacity ($X_m$) for EOS-B (52.083 mg/g), albeit with a low correlation coefficient (R² = 0.097), indicating a poor model fit. Therefore, the overall adsorption behavior observed in this study is more consistent with the Freundlich model, which assumes multilayer adsorption.

In previous studies, the adsorption capacity ($X_m$) of thermally modified EAF slag was reported as 10.368 mg/g [30], whereas other adsorbents, such as waste mud (4.20 mg/g) [37], quick lime (16.67 mg/g) [38], and brick powder (1.84 mg/g) [39], showed relatively lower values. The maximum adsorption capacity ($q_e$) observed in this study (16.13 mg/g) was within the upper range of the reported values. Particularly, the Langmuir model estimated $X_m$ values ranging from 23.753 to 52.083 mg/g, further demonstrating the high fluoride removal potential of the modified slag compared with conventional adsorbents.

3.5. XRD Analysis

To investigate the structural characteristics of the modified slags, X-ray diffraction (XRD) analysis was performed for EOS-B, EOS-C1, EOS-C2, ERS-B, ERS-C1, and ERS-C2, and the results are shown in Figure 7. The slag samples exhibited a complex mixture of amorphous and crystalline phases, including chromite (C), kermanite–gehlenite (A–G), dicalcium silicate (D), and brownmillerite (B), with some overlapping peaks observed among these phases.

The unmodified oxidized slag (EOS-B) exhibited sharp and intense diffraction peaks, indicating high crystallinity. However, after modification, the peaks became weaker and broader, suggesting an increase in the amorphous character. Particularly, a partial collapse of the crystalline structure was observed for EOS-C2, along with a relative increase in the dicalcium silicate (D) and brownmillerite (B) phases in the 2θ range of 30–34°, which is highlighted by a red frame in the figure.

The reduced slag (ERS) also exhibited increased amorphization after modification, although the diffraction peaks corresponding to brownmillerite (B) and dicalcium silicate (D) became more pronounced than those of the unmodified sample. Comparing the oxidized (EOS) and reduced (ERS) slags, stronger peaks for chromite (C) and magnetite (M) were observed in the EOS samples, whereas brownmillerite (B) and dicalcium silicate (D) were more prominent in the ERS samples.

3.6. Post-Adsorption Leaching Behavior of Heavy Metals from EAF Slag

In this study, fluoride adsorption experiments were conducted using electric arc furnace (EAF) slag and synthetic hydrofluoric acid (HF) wastewater at pH 2.5. Owing to the scrap-based origin of EAF slag, concerns remain regarding potential heavy metal leaching, particularly under varying pH and environmental conditions, as reported in previous studies [40,41]. To evaluate the possible leaching, the supernatants from the post-adsorption experiments were filtered through a 0.45 μm membrane and analyzed for Cr, Cu, As, Pb, and Cd using inductively coupled plasma mass spectrometry (ICP-MS).

The leaching concentrations under varying particle sizes, pH values, and temperatures are shown in Figure 8. Particle size was categorized into D1 (850–1700 μm), D2 (250–850 μm), and D3 (<250 μm). The Pb and Cd concentrations were consistently below the detection limits under all conditions, indicating high stability. By contrast, Cr, Cu, and As were detected only under certain conditions.

In the particle size experiments, the maximum concentrations observed were Cr 17.28 μg/L (D1, EOS-B), Cu 7.16 μg/L (D1, EOS-C1), and As 35.31 μg/L (D1, ERS-C2). However, no clear trends were observed based on the particle size and modification status. Cr was found predominantly in the pH 2–7 range, with a peak concentration of 115.51 μg/L at pH 5 (ERS-C1). Cu concentrations increased at pH ≥ 7, with values reaching 45–50 μg/L on average at pH 11. Arsenic showed relatively low concentrations overall, with a mean value of 6.90 μg/L, even though elevated values were observed in some modified samples: 23.72 μg/L (pH 2, EOS-C2) and 22.40 μg/L (pH 2, ERS-C2). These levels remain extremely low compared to the naturally occurring arsenic concentrations in the soil (~15 mg/L) [42].

Under temperature variation, the highest observed concentrations were As 35.31 μg/L (25 °C, ERS-C2), Cu 5.07 μg/L (25 °C, ERS-B), and Cr 8.79 μg/L (35 °C, ERS-B), indicating no significant correlation between temperature and metal leaching. Overall, the leaching of Cr, Cu, and As remained within environmentally acceptable limits, and the absence of detectable Pb and Cd further suggested that the environmental risk associated with heavy metal leaching from EAF slag is relatively low when the slag is used as a fluoride adsorbent.

4. Conclusions

In this study, CaCl₂-modified electric arc furnace (EAF) slag was used as an adsorbent for fluoride removal from synthetic hydrofluoric acid (HF) wastewater. The adsorption performance was evaluated under different particle sizes, temperatures, and initial pH conditions using oxidized (EOS) and reduced (ERS) slag samples in the raw (B) and modified forms (C1 and C2). Isothermal adsorption experiments, XRD analysis, and heavy metal leaching assessments using ICP-MS were conducted to examine the physicochemical properties and environmental safety of the slag.

The fluoride removal performance was evaluated based on the residual fluoride concentrations. Smaller particle sizes (<250 μm) yielded higher removal efficiency, and the CaCl₂-modified slag samples (EOS-C2 and ERS-C2) consistently exhibited excellent performance under all conditions. Although fluoride uptake by the unmodified slag increased with temperature, the modified slag maintained stable adsorption performance for all examined temperatures.

In pH-dependent experiments, the unmodified slags (EOS-B and ERS-B) showed a marked decrease in the removal efficiency at elevated pH, with residual fluoride concentrations exceeding 250 mg/L above pH 5 and adsorption capacities decreasing to 1.14–2.14 mg/g. Conversely, EOS-C2 maintained adsorption capacities above 8.0 mg/g across the entire pH range (2–11), indicating a significant enhancement in the pH tolerance due to the modification.

The XRD analysis revealed increased amorphization and relative intensification of the dicalcium silicate and brownmillerite phases after modification, which contributed to the improved adsorption performance. Isotherm fitting showed a better correlation of the results with the Freundlich model (R² = 0.856–0.983) than with the Langmuir model (R² = 0.097–0.913), indicating that multilayer adsorption was the dominant mechanism. ERS-C2 achieved the highest adsorption capacity of 16.13 mg/g, which exceeds that of the conventional adsorbents reported in previous studies.

Heavy metal leaching tests indicated trace levels of Cr, Cu, and As under specific conditions, whereas Pb and Cd remained below the detection limits in all cases. The overall environmental risk associated with slag reuse was low.

In conclusion, the modified EAF slag demonstrated high fluoride removal efficiency and excellent stability under diverse operating conditions, confirming its potential as a promising and environmentally safe adsorbent for the treatment of HF-containing industrial wastewater.