Removal of Metal Ions in Spin-on Hardmask Using Functionalized Porous Silica Adsorbents

Abstract

The ongoing miniaturization of semiconductor devices necessitates continuous advancements in lithographic processes, which are critical for high-precision circuit formation. To prevent substrate damage during the etching step, a spin-on hardmask (SOH) layer is often introduced between the photoresist (PR) and the substrate. However, residual metal ions in SOH solutions can adversely affect integrated circuit performance, underscoring the need for efficient and chemically compatible removal strategies. This study investigates the adsorption of metal ions (Al³⁺, Cr³⁺, Cu²⁺, Fe³⁺, Ni²⁺, and Ti⁴⁺) from SOH solutions using mesoporous silica materials—MCM-41 and SBA-15—functionalized with various groups (–OH, –NH₂, –SH, and –CH₃). Adsorption performance was evaluated under solvent-only, monomer-containing, and polymer-containing conditions. Among the tested materials, amine-functionalized mesoporous silica exhibited the highest adsorption efficiency, with SBA-15-NH₂ showing relatively effective and uniform performance in polymer-containing systems. Isotherm analysis supported a monolayer chemical adsorption mechanism, suggesting the significance of surface functional groups in the adsorption process. These findings demonstrate the potential of functionalized mesoporous silica as a promising candidate for trace metal ion removal in semiconductor manufacturing, offering enhanced yield and improved process reliability.

1. Introduction

As the size of semiconductor devices continues to shrink, the demand for miniaturized electronic circuits has increased significantly. Lithography processes, which play a crucial role in circuit patterning, are among the most sensitive stages in semiconductor manufacturing. To fabricate high-resolution circuits, it is essential to form a relatively thin photoresist (PR) layer [1]. However, reducing the thickness of the PR layer may increase the risk of substrate damage during the etching process. To overcome these challenges, various strategies have been proposed to enhance or complement conventional lithography techniques [2]. One such approach involves the use of an auxiliary material known as a spin-on hardmask (SOH), which is applied between the PR and the substrate to reinforce etch resistance [3]. SOH have gained widespread attention for their high plasma etch resistance and compatibility with advanced lithographic processes, especially in the sub-100 nm regime [4,5]. The presence of metal ions in the SOH solution is a critical concern, as these impurities can negatively impact the performance and reliability of integrated circuits. Therefore, the removal of these metal ions is essential to improve product yield. A common method for metal ion removal involves the use of porous membrane filters with ion-exchange capability. However, such filters are often costly to purchase and maintain. Moreover, since ion exchange membranes typically employ strongly acidic functional groups, they may cause unintended chemical reactions particularly when used with materials that are sensitive to acidic environments. Therefore, there is a need for an alternative approach that can effectively and economically remove metal ions without adversely affecting the SOH solution.

Various methods have been proposed for removal of the metal ions including chemical precipitation, reverse osmosis, and electrocoagulation. However, these techniques present drawbacks such as high operational costs, the generation of hazardous sludge, incomplete removal efficiency, and excessive consumption of energy and materials. Adsorption, in contrast, is a well-established method for metal ion removal, widely recognized for its simplicity, cost-effectiveness, and adaptability to various treatment environments. This technique typically utilizes porous materials with high surface area as adsorbents, such as activated carbon and clay minerals. However, the irregular pore structure of activated carbon and the high cost associated with its activation process limit its applicability. In addition, the weak adsorption capacity of untreated clay minerals for metal ions results in insufficient ion-exchange performance for large-scale applications [6,7,8,9].

Mesoporous silica materials have attracted considerable interest as efficient adsorbents for metal ion removal. Characterized by pore sizes in the range of 2 to 50 nm, they offer a high surface area, uniform and tunable pore structures, and favorable adsorption characteristics. Furthermore, they exhibit excellent thermal and mechanical stability, low toxicity, and ease of surface functionalization, enabling the adsorption of specific metal ions. These properties make mesoporous silica particularly suitable for removing trace metal contaminants from chemically sensitive systems such as SOH solutions [10,11]. Previous studies have demonstrated the potential of amine-functionalized mesoporous silica for the selective removal of hexavalent chromium (Cr(VI)) from aqueous environments. In particular, Lee et al. reported that mesoporous silica functionalized with (3-aminopropyl) trimethoxysilane (APTMS) exhibited strong electrostatic interactions between protonated amine groups (–NH₃⁺) and anionic Cr(VI) species, leading to efficient adsorption even under acidic conditions. Their work highlighted that while excessive APTMS loading reduced the surface area due to pore narrowing, the increase in functional group density significantly improved adsorption performance [12]. This suggests that the chemical nature and protonation state of surface functionalities play a more critical role than the pore structure itself.

In this study, two types of mesoporous silica, MCM-41 and SBA-15, were investigated as adsorbents for the removal of metal ions from SOH solutions. These materials were selected due to their well-defined pore structures, pore uniformity, large surface areas, and suitability for surface functionalization, which make them ideal platforms for adsorption-based applications. In particular, MCM-41 and SBA-15 represent widely used mesoporous silica frameworks with similar structural order but different pore sizes, allowing for a comparative evaluation of pore size effects on adsorption performance [10]. To tailor their surface properties, each material was functionalized with different groups, including hydroxyl (–OH), amine (–NH₂), thiol (–SH), and methyl (–CH₃). The metal ion removal efficiency of the functionalized adsorbents was evaluated for various metal ions (Al³⁺, Cr³⁺, Cu²⁺, Fe³⁺, Ni²⁺, and Ti⁴⁺) that are considered critical ionic contaminants in SOH and PR processes. These residual metal ions, even in trace amounts, can have a significant negative impact on semiconductor performance. For instance, Wang et al. showed that transition metal contaminants degrade device performance by introducing recombination centers or by affecting etch selectivity. Therefore, efficient removal of such trace metal ions is critical to ensuring high process reliability and product quality in semiconductor fabrication [13]. The adsorption performance was tested under three different solution conditions: a solvent-only solution, a monomer-dispersed solution, and a polymer-dispersed solution that represents the actual PR formulation condition. In addition, adsorption isotherm modeling was employed to analyze the adsorption behavior and underlying mechanisms. The study also systematically investigated how adsorption characteristics varied depending on the type of metal ion and the surface functional group, thereby evaluating the applicability of functionalized mesoporous silica for effective ion removal in SOH-related semiconductor processes. These findings are expected to contribute to the development of a more effective and chemically compatible ion-removal strategy for advanced lithographic applications. While previous studies have explored mesoporous silica materials for environmental applications such as wastewater treatment, few have focused on their use in removing residual metal ions from SOH solutions—a source of contamination in semiconductor manufacturing. This study uniquely investigates the performance of functionalized mesoporous silica (MCM-41 and SBA-15) in the presence of actual process-relevant solvents, monomers, and polymers, offering insights into real-world applicability. The novelty of this work lies in its systematic comparison of multiple functional groups (–NH₂, –SH, –CH₃, and –OH) under realistic industrial conditions, which has not been reported in the previous literature.

2. Experimental

2.1. Chemicals

Tetraethyl orthosilicate (TEOS, 95.0%), triblock copolymer Pluronic P-123 (Poly (ethylene glycol)-block-poly(propylene glycol)-block-poly(ethylene glycol)), 3-mercaptopropyltrimethoxysilane (MPTMS, 95%), 3-aminopropyltriethoxysilane (APTES, 99%), and methyltrimethoxysilane (MTES, ≥98%) were purchased from Sigma-Aldrich (St. Louis, Mo, USA). Ammonia solution (NH₄OH, 28.0–30.0%), Cetyltrimethylammonium bromide (CTAB, ≥99.0%), toluene (99.5%), ethyl alcohol (99.5%) were obtained from Samchun Chemicals (Seoul, Republic of Korea). Hydrochloric acid (HCl, 35%) was sourced from Daejung Chmicals & Metals (Siheung, Republic of Korea), and high-purity water was purchased from Duksan Chemicals (Seoul, Republic of Korea).

For adsorption studies, chromium(Ⅲ) nitrate nonahydrate (Cr(NO₃)₃·9H₂O, 98%), nickel(Ⅱ) nitrate hexahydrate (Ni(NO₃)₂·6H₂O, 98%), and iron(Ⅲ) nitrate nonahydrate (Fe(NO₃)₃·9H₂O, 98.5%), were purchased from Sigma-Aldrich. Aluminium nitrate nonahydrate (Al(NO₃)₃·9H₂O, 98%), titanium(Ⅳ) chloride (TiCl₄) solution (1.0 M in methylene chloride, 85%), methyl alcohol (99.5%), formaldehyde solution (35.0%), phenol (99.0%), and cyclohexanone (≥99.5%) were purchased from Samchun Chemicals. Copper(Ⅱ) nitrate trihydrate (Cu(NO₃)₂·3H₂O, 99%) was obtained from Junsei (Tokyo, Japan). Novolac resin (phenol novolac epoxy, YDPN-631) was provided by Kukdo Chemical (Seoul, Republic of Korea), and propylene glycol monomethyl ether acetate (PGMEA, 99%) was purchased from Daejung Chemicals & Metals.

2.2. Preparation of the Adsorbents

MCM-41 was synthesized by scaling up a previously reported method [14]. CTAB (3.06 g, 8.40 mmol) was dissolved in a 1 L beaker containing ammonia solution (171.84 mL, 28%) and distilled water (630 mL). The mixture was stirred for 30 min at 50 °C. Subsequently, TEOS (14.1 g, 66 mmol) was then added dropwise and stirred at 400 rpm for 3 h at 75 °C. The resulting precipitate was separated by vacuum filtration, thoroughly washed with ethanol, and dried at room temperature for 24 h.

For the synthesis of SBA-15, P-123 (12 g) was dissolved in a mixture of hydrochloric acid (62.5 mL, 35%) and distilled water (297.5 mL) in a 500 mL beaker. After complete dissolution, TEOS (25.7 mL, 24 g) was added, and the solution was stirred at 400 rpm at 40 °C for 24 h. The resulting mixture was transferred to a polypropylene (PP) bottle and maintained at 100 °C for 48 h without agitation. After cooling to room temperature, the solid product was separated by filtration, thoroughly washed with ethanol, and air-dried overnight [15].

To remove the surfactant and template, solvent extraction was performed on the as-synthesized MCM-41 and SBA-15 samples [16]. The pre-synthesized materials (8 g) were added to a three-neck flask containing 180 mL of ethanol, followed by the addition of HCl (8 mL, 35%). The mixture was stirred and heated to 75 °C using a reflux condenser at a heating rate of 3 °C min⁻¹. The temperature was maintained at 75 °C for 3 h for MCM-41 and 6 h for SBA-15. After cooling to room temperature, the materials were filtered, thoroughly washed with ethanol, and dried overnight.

For surface functionalization, 1.5 g of mesoporous silica was dispersed in 150 mL of dry toluene in a 250 mL three-neck flask. Then, 6 mL of MPTMS was added for thiol functionalization. The flask was equipped with a reflux condenser and heated to 110 °C at a rate of 3 °C min⁻¹ under stirring for 24 h. The resulting material was filtered, washed with ethanol, and air-dried overnight [17]. The same procedure was applied for amine and methyl functionalization using APTES and MTES, respectively.

2.3. Material Characterization

To investigate the morphology, particle size, and pore structure of synthesized MCM-41 and SBA-15, transmission electron microscopy (TEM, JEM-2100, JEOL, Tokyo, Japan) was employed. Small-angle X-ray diffraction (SAXRD, Ultima3, Rigaku, Tokyo Japan, 0.5–10°, 40 kV/40 mA) was used to confirm the structural ordering of the mesoporous frameworks. Nitrogen adsorption–desorption isotherms were obtained to analyze porous characteristics. Specific surface area, average pore size, and total pore volume of the adsorbents were determined using the Brunauer–Emmett–Teller (BET) method, and the pore diameter distribution was calculated using the Barrett–Joyner–Halenda (BJH) method. Zeta potential measurements were conducted using a zetasizer (ZETASIZER ADVANCE Pro, Malvern Panalytical, Malvern, UK). Metal ion concentrations were analyzed using inductively coupled plasma optical emission spectrometry (ICP-OES, Agilent 5900, Santa Clara, CA, USA).

2.4. Preparation of Metal Ion Solutions

To prepare the metal ion solutions used in the adsorption experiments, three different solution systems were prepared: a solvent-only solution, a monomer-dispersed solution, and a polymer-dispersed solution. In all cases, the final total metal ion concentration was adjusted to 100 ppm.

For the solvent-only system, a stock solution was prepared by dissolving Al(NO₃)₃·9H₂O (183.53 mg), Cr(NO₃)₃·9H₂O (101.57 mg), Cu(NO₃)₂·3H₂O (50.18 mg), Fe(NO₃)₃·9H₂O (95.48 mg), Ni(NO₃)₂·6H₂O (65.40 mg), and TiCl₄ solution (275.75 µL, 1 M) in 10 mL of methanol. This stock solution, corresponding to a 10 wt% metal ion concentration, was diluted 100-fold with methanol to yield a 100 ppm working solution.

For the monomer-dispersed system, phenol (186.92 mL) and formaldehyde (181.92 mL) were mixed with methanol (749.47 mL) at a mass ratio of 1:1:3. The same amount of metal salts used in the solvent-only system was dissolved in 10 mL of this monomer mixture to prepare a stock solution containing 10 wt% total metal ions. This stock solution was then diluted 100-fold with the same monomer mixture to obtain the final solution at 100 ppm.

For the polymer-dispersed system, novolac resin (112.47 mL) and PGMEA (528.50 mL) were mixed with cyclohexanone (359.03 mL) to form a solution with a mass ratio of 15:51:34. Due to the poor metal salt solubility in the polymer solution, a stock solution with a metal ion concentration 20 times lower than that of the other stock solutions was prepared and subsequently diluted 5-fold with the same polymer mixture to yield a final total metal ion concentration of 100 ppm.

2.5. Adsorption Experiment

To evaluate the metal ion adsorption performance, 30 mL of each prepared metal ion solution was placed in a 50 mL vial. Each prepared adsorbent was then dispersed in the solution at a ratio of 10 mg mL⁻¹ (adsorbent/solution), and the mixture was stirred at room temperature for 24 h. Afterward, the adsorbent was separated from the solution using a syringe filter. The initial and final concentrations of metal ions were measured by ICP-OES. The metal ion removal efficiency (%) of the adsorbents was calculated using the following equation:

$$\%R={\displaystyle \frac{C_0-C_e}{C_0}}\times 100\%$$

(1)

where C₀ and C_e represent the initial and equilibrium concentrations of metal ions, respectively.

2.6. Adsorption Isotherm Modeling

To plot the adsorption isotherms, solutions were prepared individually for each metal ion [18]. Single-metal ion solutions with initial concentrations of 0.02, 0.05, 0.1, 0.2, 0.4, 0.6, 0.7, 0.8, 0.9, and 1.0 M for each metal ion in methanol were prepared. For each concentration, 30 mL of the solution was placed in a 50 mL vial. MCM-41-NH₂ adsorbent was then dispersed in each solution at a ratio of 2 mg mL⁻¹ (adsorbent/solution). The mixtures were stirred at room temperature for 24 h, and the initial and equilibrium concentrations of metal ions were measured using ICP-OES.

3. Result and Discussion

3.1. Synthesis and Characterization of Mesoporous Adsorbents

Figure 1 shows a schematic illustration of the synthesis of functionalized mesoporous silica adsorbents (MCM-41 and SBA-15) and the metal ion removal process from SOH solution. MCM-41 and SBA-15 were synthesized using CTAB and P123, respectively, as structure-directing surfactants. The surfactants were removed under acidic conditions through an extraction process, followed by surface functionalization to introduce specific functional groups onto the mesoporous silica framework. The resulting functionalized adsorbents were subsequently applied in metal ion removal experiments under SOH solution environments.

TEM and small-angle XRD analyses were employed to confirm the structures of MCM-41 and SBA-15, and the results are presented in Figure 2. MCM-41 (Figure 2a–c) exhibited spherical particles with sizes ranging from 400 to 500 nm. SBA-15 (Figure 2d–f) appeared as chain-like rod-shaped particles with sizes of approximately 700–800 nm. Both samples displayed well-ordered mesoporous structures with uniform pore arrangements. Small-angle XRD analysis revealed that MCM-41 exhibited peaks corresponding to the (100), (110), (200), and (210) at 2θ values of 2.15°, 3.72°, 4.3°, and 5.69°, respectively, confirming the hexagonal lattice structure of the synthesized MCM-41. These peaks correspond to the characteristic diffraction planes, indicating the well-ordered mesoporous structure [14]. In contrast, SBA-15 showed a strong peak corresponding to the (100) plane at 2θ = 1.035°, along with smaller peaks at 1.675° and 1.915° that correspond to the diffraction of (110) and (200), suggesting a well-ordered structure with a hexagonal lattice symmetry of p6mm [19,20].

To confirm the pore characteristics of functionalized MCM-41 and SBA-15, nitrogen adsorption–desorption analysis based on the Brunauer–Emmett–Teller (BET) method was conducted, and the pore size distribution was illustrated in Figure 3. Additionally, the pore characteristics of the surface-functionalized MCM-41 and SBA-15 materials were summarized in

Table 1. Textural parameters of pristine and functionalized mesoporous silica materials.

Plot Data	Specific Surface Area (m2 g−1)	Pore Size (nm)	Pore Volume (cm3 g−1)
MCM-41-OH	1120.4	3.5605	0.9973
MCM-41-CH3	927.21	3.4795	0.8066
MCM-41-NH2	656.81	3.1381	0.5153
MCM-41-SH	785.3	3.2415	0.6364
SBA-15-OH	593.75	7.737	1.1455
SBA-15-CH3	462.63	7.3737	0.8528
SBA-15-NH2	376.28	7.1297	0.6707
SBA-15-SH	500.89	7.6897	0.9629

. All of the functionalized MCM-41 exhibited a Type-Ⅳ adsorption–desorption isotherm with an H1 hysteresis loop attributed to capillary condensation of mesopores at a relative pressure (p/p₀) range of 0.3–0.4. Functionalized SBA-15 samples also displayed a similar Type-Ⅳ isotherm with an H1 hysteresis loop at a p/p₀ range of 0.65–0.8. This indicates that both synthesized MCM-41 and SBA-15 possess well-developed cylindrical mesoporous structures [19,21].

The specific surface area, average pore size, and total pore volume of MCM-41 were 1120 m² g⁻¹, 3.6 nm, and 1.0 cm³ g⁻¹, respectively. In contrast, SBA-15 exhibited values of 593.8 m² g⁻¹, 7.7 nm, and 1.1 cm³ g⁻¹ for specific surface area, average pore size, and total pore volume, respectively. Compared to SBA-15, MCM-41 exhibited a smaller pore diameter and more than twice the surface area. In the case of surface-modified materials, it was observed that the specific surface area, pore size, and pore volume all decreased compared to the pristine materials. This observation indicates the formation of functional groups not only on the adsorbent’s surface but also within the pores. Among the functionalized samples, the amine-modified (-NH₂) materials exhibited the smallest pore sizes in both MCM-41 and SBA-15. This can be attributed to the nucleophilic or alkaline nature of the amine, which is presumed to facilitate surface modification by promoting additional layer formation within the pores. Consequently, both the specific surface area and pore volume were reduced [22].

To confirm the successful functionalization of mesoporous silica, zeta potential measurements were conducted to evaluate the surface charge variations resulting from the introduction of different functional groups. As shown in Figure 4, MCM-41-OH and SBA-15-OH exhibited a negative zeta potential of approximately −43.6 mV and −21.9 mV at pH 7, which is attributed to the presence of deprotonated silanol (–Si–OH) groups [23]. Upon methyl functionalization (–CH₃), the zeta potential values increased to around −19 mV for both materials and thiol-functionalized silica (–SH) also showed a moderate increase (−28.7 mV for MCM-41 and −22.3 mV for SBA-15), consistent with the partial ionizability of thiol groups.

Notably, amine-functionalized mesoporous silica exhibited a significant shift to positive zeta potentials, reaching +11.3 mV (MCM-41) and +14.3 mV (SBA-15), due to the protonation of surface amine groups at neutral pH. This confirms the successful grafting of positively charged -NH₃⁺ groups and highlights the utility of zeta potential analysis for distinguishing surface functionalization. The magnitude and direction of potential shifts across different functional groups demonstrate the varying surface charge environments, thereby validating the presence and identity of introduced functional groups [24].

3.2. Metal Ion Adsorption

To evaluate the adsorption performance of the functionalized mesoporous silica materials, metal ion removal experiments were initially conducted under solvent-only conditions to isolate the intrinsic adsorption capacity of each surface functional group without interference from other organic components. This approach allows for a direct comparison of the affinity of each functional group (–OH, –CH₃, –SH, and –NH₂) toward the target metal ions (Al³⁺, Cr³⁺, Cu²⁺, Fe³⁺, Ni²⁺, and Ti⁴⁺), providing a clear understanding of how surface chemistry influences ion removal efficiency. The removal efficiency (%) was calculated based on the difference between the initial and equilibrium metal ion concentrations using ICP-OES analysis, as described in Section 2.5.

A consistent trend in ion removal efficiency was observed across different adsorbents, primarily influenced by the type of functional group introduced on the surface rather than by pore size or surface area. This highlights the dominant role of surface chemistry in adsorption behavior.

As illustrated in Figure 5a, unmodified MCM-41-OH and SBA-15-OH both demonstrated a common feature of selectively removing Al³⁺ and Ti⁴⁺ ions. Notably, MCM-41-OH also demonstrated moderate adsorption of Ni²⁺, which was not observed for SBA-15-OH.

Figure 5b shows that methyl-functionalized samples (MCM-41-CH₃, SBA-15-CH₃) exhibited the lowest overall adsorption performance. Only limited removal of Al³⁺ and Ti⁴⁺ was observed, indicating the hydrophobic nature of the –CH₃ group may hinder effective interaction with solvated metal ions.

In Figure 5c, MCM-41-SH, and SBA-15-SH modified with thiol groups (–SH) on the surface, selectively removed Al³⁺, Cu²⁺, and Ti⁴⁺ ions. In particular, the selective adsorption of Cu²⁺ ions, which is not achieved by other functionalized adsorbents, is especially noteworthy.

As observed in Figure 5d, MCM-41-NH₂ and SBA-15-NH₂ modified with amine groups (-NH₂) on the surface, demonstrated exceptional ion removal efficiency compared to other functionalized materials. They exhibited consistently superior ion removal efficiency for all tested ions, especially with MCM-41-NH₂ achieving over 99% removal efficiency across the board. This outstanding performance is attributed to the strong chelation ability and high affinity of amine groups for metal cations. Chelation occurs as the lone pair of electrons on nitrogen atoms in the amine groups coordinate with metal cations to form stable complexes. In addition, amine groups enable multiple interaction mechanisms, including electrostatic attraction, hydrogen bonding, and covalent coordination, all of which contribute to enhanced adsorption efficiency [25]. These results suggest that amine-functionalized mesoporous silica is a promising adsorbent for the removal of trace metal contaminants in semiconductor applications, offering high efficiency with low environmental and economic impact.

When comparing the overall adsorption performance across different functional groups, the trend –NH₂ > –SH > –OH > –CH₃ was observed, showcasing superior adsorption performance in that order. Based on these results, MCM-41-NH₂ and SBA-15-NH₂ were selected for further experiments using metal ion solutions containing monomers and polymers. For the adsorption isotherm experiments, only MCM-41-NH₂ was chosen due to its outstanding performance. Moreover, the adsorption affinity of functional groups was closely related to metal ion properties such as valence, size, and softness. Thiol-functionalized mesoporous silica (–SH) showed strong affinity for soft divalent cations like Cu²⁺, Pb²⁺, and Cd²⁺, due to covalent bonding based on Pearson’s hard-soft acid-base (HSAB) theory. Its selectivity over hard trivalent ions (e.g., Cr³⁺ and Al³⁺) is attributed to the high charge density and strong hydration shells of the latter [26]. Methyl-functionalized mesoporous silica (–CH₃) exhibited negligible uptake due to its hydrophobicity and lack of donor atoms. However, co-functionalization with thiol groups can improve steric accessibility and prevent oxidation [26]. Unmodified mesoporous silica with hydroxyl groups (–OH) selectively adsorbed hard trivalent cations such as Al³⁺, Ti⁴⁺, and Fe³⁺ through inner-sphere complexation with deprotonated silanols at a favorable pH, although its capacity was modest [27,28].

In the subsequent experiments using monomer-containing metal ion solutions, the ion removal efficiencies of the adsorbents were evaluated, as presented in Figure 6. Compared to the results obtained in the solvent-only system (Figure 5), both MCM-41-NH₂ and SBA-15-NH₂ exhibited significantly reduced adsorption performance. Moreover, the previously observed trend of uniform ion adsorption across different metal ions was no longer maintained. The effective molecular size of phenol is approximately 0.75 nm, and that of formaldehyde is approximately 0.25 nm [29,30]. According to Table 1, the pore sizes of MCM-41-NH₂ and SBA-15-NH₂ are 3.1381 nm and 7.1297 nm, respectively. Therefore, it can be interpreted that the dissolved monomers partially infiltrate the mesopores of the adsorbents, thereby hindering access of metal ions to the active adsorption sites on the internal surfaces.

In addition to this steric hindrance, potential chemical interactions between phenol and metal ions may further contribute to the observed reduction in adsorption efficiency. Recent studies have demonstrated that phenol can undergo oxidation via proton-coupled electron transfer (PCET) or stepwise proton transfer/electron transfer (PTET) mechanisms in the presence of Cu²⁺ complexes, leading to the generation of phenoxy radicals and the reduction of Cu²⁺ to Cu⁺ [31]. Such reactivity suggests that phenol may chemically bind or react with certain metal ions in solution, altering their speciation and reducing their availability for adsorption. Furthermore, phenol is also known to form hydrogen bonds with amine groups on the adsorbent surface, which could competitively occupy active sites or modify the surface polarity. While these effects were not the primary focus of this study, they may partially explain the reduced and non-uniform adsorption performance in monomer-containing systems.

Adsorption experiments were also carried out in polymer-containing metal ion solutions, and the resulting ion removal efficiencies are presented in Figure 7. Compared to the solvent-only condition (Figure 5), a decrease in adsorption performance was observed, although the values were notably higher than those obtained in the monomer-containing solution (Figure 6). This trend can be interpreted as follows: due to the relatively larger molecular size of the polymer compared to monomers, it cannot easily diffuse into the mesopores of the adsorbents, thereby limiting its interference with metal ion adsorption. In the case of MCM-41-NH₂, adsorption efficiency in the polymer-containing solution was improved compared to that in the monomer-containing condition. However, it did not exhibit uniform adsorption behavior, showing high removal efficiency only for certain metal ions. In contrast, SBA-15-NH₂ demonstrated a significant enhancement in adsorption performance relative to its behavior in the monomer-containing solution. Although its efficiency remained lower than that observed under solvent-only conditions, it effectively removed all target ions with relatively uniform performance.

3.3. Adsorption Isotherm

To better understand the ion adsorption mechanism involved, adsorption isotherms for ions were obtained by plotting the equilibrium concentration (C_e, mmol L⁻¹) versus the amount of metal ion removed at equilibrium state (q_e, mmol g⁻¹), as shown in Figure 8. The amount of removed metal ion was calculated using the following equation:

$$q_e={\displaystyle \frac{C_i-C_e}{C_m}}\times V$$

(2)

where C_i and C_e represent the initial and equilibrium concentration, measured before and after the adsorption process, respectively, and expressed in mmol L⁻¹. V is the volume of the solution (L); and m is the mass of the adsorbent (g).

The obtained adsorption isotherms were fitted to three common models: Langmuir, Freundlich, and Dubinin–Radushkevich (D-R) models. The Langmuir adsorption isotherm model, which assumes monolayer adsorption onto a homogeneous surface, is expressed as [32,33]:

$$q_e={\displaystyle \frac{\left(q_{m}b\right)}{1+{ bC}_m}}{ C}_e$$

(3)

$${\displaystyle \frac{C_e}{q_e}}={\displaystyle \frac{1}{q_{e}b}}+{\displaystyle \frac{1}{ q_m}}{ C}_e$$

(4)

where $q_m$ represents the maximum adsorption capacity (mmol g⁻¹), and b is the Langmuir constant, a parameter associated with the adsorption energy. The maximum adsorption capacities (q_m) for Al³⁺, Cu²⁺, Fe³⁺, and Ni²⁺ were found to be 0.6206, 0.8672, 0.5179, and 0.3653 mmol g⁻¹, respectively, indicating the order of Cu²⁺ > Al³⁺ > Fe³⁺ > Ni²⁺ as shown in

Table 2. Langmuir isotherm parameters including maximum adsorption capacity (q_m), Langmuir constant (b), and Gibbs free energy (ΔG) for the adsorption of metal ions onto amine-functionalized MCM-41 at 298 K.

	Al3+	Cu2+	Fe3+	Ni2+
qm (mmol g−1)	0.6206	0.8672	0.5179	0.3653
b	12.515	1691.2	736.92	221.56
ΔG (J mol−1)	−6260.67	−18,416.3	−16,358.1	−13,379.5

. Furthermore, the negative values of Gibbs free energy (ΔG) for all metal ions indicate that the adsorption processes onto amine-functionalized MCM-41 are spontaneous in nature.

The Freundlich isotherm, which assumes multilayer adsorption on heterogeneous surfaces, is given as [32]:

$$\mathit{log}\left(q_e\right)=log{K}_F+{\displaystyle \frac{1}{n}}\mathit{log}\left(C_e\right)$$

(5)

where $K_F$ is the adsorption capacity, and $n$ is indicating adsorption intensity or surface heterogeneity.

The Dubinin–Radushkevich Isotherm (D-R) is expressed as follows [32].

$$ln{q}_e=ln{q}_m-\beta {ϵ}^2$$

(6)

$$ϵ=RTln(1+{\displaystyle \frac{1}{C_e}})$$

(7)

$$E={\displaystyle \frac{1}{\sqrt{2\beta }}}$$

(8)

where $ϵ$ is Polanyi potential, $\mathsf{\beta }$ is Dubinin–Radushkevich constant, $R$ is gas constant (8.31 J mol⁻¹∙k⁻¹), $T$ is absolute temperature, and $E$ is mean adsorption energy.

The experimental isotherm data for four metal ions (Al³⁺, Cu²⁺, Fe³⁺, and Ni²⁺) were obtained and fitted to the previously described isotherm models. Among these, the Langmuir isotherm consistently provided the best fit across all ions, indicating that the adsorption process was characterized by monolayer adsorption on the surface of the adsorbent with uniform energy distribution. This agreement with the Langmuir model suggests that the adsorption mechanism was primarily governed by chemical interactions between the metal ions and the surface of the functionalized mesoporous silica, rather than a physical one. In particular, the nature of the surface functional groups appears to play a decisive role in determining the overall adsorption efficiency. These results underscore the importance of functional group selection and surface chemistry in designing mesoporous adsorbents for effective metal ion removal in complex chemical environments [30,33,34,35,36].

4. Conclusions

In this study, two types of mesoporous silica materials with different pore sizes—MCM-41 and SBA-15—were functionalized with various surface groups and evaluated as adsorbents for metal ion removal from SOH solutions under different chemical environments. Adsorption experiments were conducted using solvent-only, monomer-containing, and polymer-containing solutions to reflect relevant conditions in semiconductor manufacturing processes. Among the tested functional groups, amine-modified mesoporous silica (-NH₂) exhibited the highest metal ion removal efficiency. MCM-41-NH₂, in particular, achieved over 99% removal in solvent-only conditions.

Amine-modified mesoporous silica showed better adsorption performance in polymer-containing solutions than in monomer-containing solutions. In particular, SBA-15-NH₂ successfully removed all target metal ions in polymer-containing solutions with relatively uniform and high efficiency. Adsorption isotherm analysis showed that all tested ions fit the Langmuir model most closely, indicating monolayer adsorption behavior governed primarily by chemical interactions between the metal ions and the functionalized surface. These results underline the critical influence of surface functional groups in determining the adsorption efficiency of mesoporous materials.

This work provides meaningful insight for the development of next-generation adsorbents tailored for advanced lithography processes requiring precise contamination control and highlights the promise of functionalized mesoporous silica as a viable solution for trace metal removal in semiconductor manufacturing.