Surface Hydrophobization of Glass Wool Fibers via Methyltrimethoxysilane Silanization for Polyethylene-Based Microplastic Filtration

Abstract

This study examines the surface modification of glass wool waste for the removal of polyethylene (PE) microplastics from aqueous suspensions. The modification strategy included sequential cleaning, pretreatment with NaOH at various concentrations (0.1, 0.5, and 1 M), and coating with methyltrimethoxysilane (MTMS) at various temperatures (room temperature (RT), 40, and 60 °C). The aim is to modify the surface properties of glass wool and evaluate its potential as a filtration medium. FTIR spectroscopy, scanning electron microscopy, contact angle measurements, and X-ray photoelectron spectroscopy (XPS) showed that alkaline pretreatment increases surface hydroxylation, facilitating effective silanization. The introduction of methyl groups resulted in a transition from hydrophilic to hydrophobic behavior, with contact angles for the modified samples exceeding 130°. XPS results showed an increase in surface carbon content, accompanied by decreases in oxygen and silicon signals, indicating the formation of an organosilane-based layer. Filtration experiments revealed that untreated glass wool exhibited unstable and limited microplastic retention, whereas the modified samples demonstrated significantly improved and more consistent results. Among the conditions tested, the best results were achieved with a combination of 1 M NaOH pretreatment and silanization at 60 °C, yielding low, stable turbidity values. The observed performance is attributed to the combined effect of enhanced hydrophobic interactions, increased surface roughness, and mechanical entrapment within the fibrous structure. Reusability tests showed that the modified material retained its microplastic-removal ability across multiple cycles, although some variability in performance was observed. Overall, this study demonstrates that an MTMS-based surface modification can improve the suitability of waste glass wool as a potential filtration medium for microplastic-containing suspensions.

1. Introduction

Due to their many useful and adaptable applications, plastics, discovered in the early 20th century, have become an essential component of contemporary society. However, a major factor in the global plastic pollution problem is the widespread and often negligent disposal of plastic products. Furthermore, microplastics (MPs) are found in a wide range of commercial products, including detergents and cosmetics, which increases their prevalence, especially in aquatic environments.

Microplastics are particles with a diameter of 1 μm to 5 mm. They can further break down through abrasion processes to become nanoplastics (NPs), which are particles with a diameter of 1 nm to 1 μm. MPs are generally dominated by polystyrene (PS) (20–90 %), poly­(ethylene terephthalate) (PET) (40–75 %), polyethylene (PE) (5–60 %), and polypropylene (PP) (2–40 %) in wastewater discharged into treatment plants. More than 90 % of MPs can be removed from raw wastewater using traditional wastewater treatment techniques such as flocculation and coagulation. Nonetheless, it is estimated that 10¹¹–10¹² MP particles can still be released daily in treated effluent from a single wastewater treatment plant. There are serious ecotoxicological risks associated with MPs, and it has been demonstrated that the majority of fish in areas affected by effluent discharge are MP-contaminated. Aquatic organisms that consume MPs may experience a range of adverse effects, including impaired growth, digestive problems, and intestinal abrasion.

In addition, MPs can act as carriers of harmful metals, pathogenic microbes, and persistent organic pollutants in aquatic food webs by adsorbing these contaminants. As a result, several approaches have been investigated to improve MP removal from wastewater. The most widely used techniques are biological, chemical, and physical techniques. Coagulants or solvents are frequently used in chemical processes to alter the surface characteristics of MPs or promote aggregation. Coagulation-sedimentation raises questions about residual chemical pollution, even though it is economical for large-scale operations. Although solvent-based techniques can be customized for particular MP types, they are typically expensive and intricate. Although biological methods are considered environmentally friendly, they are often slow and not industrially scalable. They use microorganisms or enzymes to break down MPs.

Filtration and density-based separation are the most common physical techniques. , These techniques exploit variations in the size and density of MPs. Removal efficiencies from filtration, including micro-, ultra-, and nanofiltration, have been close to 100 %. Nevertheless, fouling and high operating costs frequently impede the widespread implementation of membrane-based systems, reducing their applicability as a universal solution. Recent studies have demonstrated that the performance of membrane-based filtration systems can be further enhanced through surface engineering approaches. In particular, surface-modified polyamide membranes have shown improved separation efficiency and selective contaminant removal due to tailored surface properties and enhanced interfacial interactions. These findings highlight the importance of surface modification strategies in filtration technologies and demonstrate how controlled surface chemistry can significantly influence separation performance. However, despite their high efficiency, polymeric membrane systems may still suffer from fouling, regeneration challenges, and relatively high operational costs, motivating the search for alternative low-cost filtration materials derived from waste resources. In addition, MPs can be removed by gravity-driven filtration systems, such as multimedia or deep-bed filters that use sand, zeolite, or biochar as filtration media. Compared with sophisticated membrane processes, these systems can reduce unit treatment costs by up to 60 %. Although its effectiveness in removing smaller MPs remains unknown, sand filtration, in particular, offers removal efficiencies of 74–97%. Over 95 % of MPs have been removed by biochar-based filters, although issues of regeneration and reuse remain. In addition, tests have shown that natural zeolites (approximately 3 mm) have an efficiency of less than 50 %. However, when surfaces were modified with hexadecylpyridinium bromide, zeolite efficiency increased significantly (to over 96 %), particularly for particles such as 100 μm PA fibers and 10 μm PE. Despite their low cost, density-based techniques struggle with high-density MPs.

Physical and chemical techniques typically offer high efficiency when compared to these methods; however, the use of reagents in chemical processes may cause environmental concerns. In the meantime, the low cost and environmentally friendly nature of biological methods highlight the need for more study. This study suggests using waste glass wool as a possible filtration medium for MP removal. Glass wool has a naturally fibrous, web-like structure that resembles filter media when unground. Despite being widely used worldwide as insulation, glass wool often ends up in landfill after buildings are decommissioned. Given that buildings typically last 40 years, this kind of waste is set to become a major worldwide problem.

Mineral wool (glass and stone wool) is a low-density, porous, inorganic fibrous material made from spun or drawn molten rock or silica that has been bound with oils or resins. Its exceptional fire resistance, thermal insulation, and sound absorption contributed to its quick industrialization in the early 20th century. Despite these uses, abandoned mineral wool is primarily dumped in landfills and is still categorized as non-recyclable waste. Its lightweight structure necessitates large disposal areas and presents logistical and environmental challenges. Prior uses of mineral wool include ceramics, fiber-based composites, drywall, cement and wood-polypropylene composites, , and briquettes. Newer uses include geopolymers, , and wastewater adsorbents.

This study aims to prepare a filtration medium for MPs without grinding, using glass wool in its natural fibrous state. Glass wool provides a more affordable and environmentally friendly alternative to well-known materials such as zeolites, geopolymers, and activated carbon. The preparation of a superhydrophobic micro-net for MP removal is a significant method. Since MPs are hydrophobic, MP adhesion will be facilitated by hydrophobic–hydrophobic interactions when glass wool surfaces are modified to increase hydrophobicity with silane agents such as methyltrimethoxysilane (MTMS).

MTMS was selected as the silanization agent because it offers a practical balance between surface reactivity, hydrophobic performance, and economic feasibility for the modification of glass-based materials. Trialkoxysilanes readily undergo hydrolysis and condensation reactions with hydroxyl groups present on silica-rich surfaces, leading to the formation of stable Si–O–Si covalent bonds while exposing organic functional groups that reduce the surface free energy. Because of these characteristics, alkoxysilanes have been extensively employed for the surface modification of glass, silica, and other inorganic oxide materials. , In comparison with fluorinated silanes, which can provide extremely low surface energies but are generally associated with substantially higher costs and increasing environmental and regulatory concerns related to per- and polyfluoroalkyl substances (PFAS), MTMS offers an effective fluorine-free alternative for producing hydrophobic surfaces. Recent reviews have therefore highlighted non-fluorinated organosilane coatings as promising candidates for sustainable surface engineering. Furthermore, although longer-chain alkyltrialkoxysilanes such as octyltrimethoxysilane or octadecyltrimethoxysilane are also capable of producing hydrophobic surfaces, their bulkier alkyl chains may slow hydrolysis and condensation reactions and often require longer processing times or more carefully controlled deposition conditions to achieve homogeneous coatings. In contrast, the relatively small methyl group of MTMS enables rapid hydrolysis and condensation while still providing sufficient reduction in surface free energy to achieve highly hydrophobic or superhydrophobic surfaces when combined with appropriate surface roughness. In addition, MTMS is commercially available, relatively inexpensive, and has been widely used for water-repellent treatment of silica-containing materials, making it a suitable choice for the scalable surface modification of waste glass wool in the present study. ,

Previous studies have reported the presence of reactive surface hydroxyl groups on mineral wool fibers using complementary spectroscopic techniques, supporting the feasibility of subsequent silane grafting. A new Si–O–Si surface layer is produced by the anticipated chemical modification, which entails silanol groups reacting with silane agents to form Si–C linkages. For MPs ≤100 μm in wastewater, this method can produce efficient, reasonably priced, and environmentally friendly filtration solutions. The main objectives of the study are to (i) modify the glass wool surface using methyltrimethoxysilane, (ii) confirm the modification via FTIR, contact angle analysis, and XPS, and (iii) evaluate the filtration performance of modified and unmodified wool micro-nets in separating polyethylene MPs under varying flow rates.

To the best of our knowledge, this study represents one of the first attempts to utilize hydrophobically modified glass wool waste for microplastic separation. The work introduces a cost-effective, environmentally friendly micro-net approach and systematically examines the role of surface hydrophobization in governing microplastic interactions and separation efficiency.

2. Materials and Methods

2.1. Pre-Treatment and Surface Modification of Glass Wool

In this study, process waste glass wool with binder was used (Figure ). To capture microplastics (MPs), surface-modified glass-wool-based micro-nets were prepared in several steps. Polyethylene (PE), Macklin (≈1000 mesh) (CAS:9002-88-4) microplastic particles were used as model contaminants in the filtration experiments. Sodium hydroxide (NaOH) (Merck, 97 %), hydrogen peroxide (H₂O₂) (Kimya lab 35 %), methyltrimethoxysilane (MTMS, 98 %) (CAS:1185-55-3), sulfuric acid (H₂SO₄, 95 %), isopropyl alcohol (Kimya lab, 99.97 %), and deionized water were used during the cleaning and surface modification steps. All chemicals were of analytical grade and were used as received without further purification.

1.

Process waste glass wool supplied from Türkiye.

The initial cleaning step aimed to remove surface-bound dust and impurities from the as-received glass wool. Deionized water was used to wash the material at a liquid-to-solid ratio of 20, and the material was shaken in an orbital shaker for 30 min at 180 rpm. After that, the washed glass wool was filtered under vacuum and dried in an oven (WiseVen Fuzzy Control System) at 80 °C until it reached a constant weight. Following washing with water and drying, an oxidative cleaning step was performed with H₂O₂. 5 g of the pre-washed glass wool was completely submerged in 1 L of a 3 wt % H₂O₂ solution and allowed to react for 10 min. After that, the samples were vacuum-filtered and thoroughly rinsed with deionized water to remove residual H₂O₂. Subsequently, NaOH solutions at 0.1, 0.5, and 1 M were used to treat the samples using an alkaline method. For each treatment, 5 g of glass wool was placed in NaOH solution for 5 h. Following the alkaline treatment, the samples were filtered under vacuum and then thoroughly washed with deionized water to remove any remaining NaOH from the surface. Washing was continued until the pH of the final washing effluent reached approximately neutral values. The pH measurements were performed using a Milwaukee MW806 MAX pH/EC/TDS meter. The samples were then dried in an oven at 80 °C until constant weight was reached.

To prepare a silane-based surface modifier, 180 mL of isopropyl alcohol and 20 mL of deionized water were combined to yield a total of 200 mL of silanization solution. Methyltrimethoxysilane (MTMS) was dissolved in isopropyl alcohol (IPA), which was selected as the reaction medium since it provides suitable conditions for MTMS hydrolysis and condensation and facilitates stable silanization under the selected experimental conditions. After that, 2 mL of MTMS was added to the mixture, and the solution was stirred at 300 rpm for 10 min using a magnetic stirrer. To keep the pH between 4 and 5, 100 μL of H₂SO₄ was added to the solution. Pre-treated and oven-dried glass wool samples were cut into smaller pieces to ensure homogeneous coating and immersed in the silane solution for 30 min at temperatures of room temperature (RT), 40 and 60 °C. During the silanization process, the solution was continuously stirred at 100 rpm to prevent gelation. After the silanization treatment, the silane solution was completely removed by vacuum filtration. The coated glass wool samples were dried in an oven at 80 °C until constant weight was reached and then collected and considered ready for subsequent microplastic removal experiments.

The samples were weighed after the silanization step using an analytical balance. These measurements were performed to provide a simple gravimetric estimation of the subsequent mass increase associated with MTMS deposition.

2.2. Microplastic Filtration Experiments

A series of filtration experiments was conducted to assess the effectiveness of glass-wool-based filters for removing MPs from water. The study focused on PE, which is widely used in food packaging and plastic bottles. A stock solution of PE microplastics was prepared in deionized water at a concentration of 1000 mg L^–1. The stock solution was subsequently diluted to a working concentration of 5 mg L^–1, which was used consistently throughout all filtration experiments to ensure comparability of results.

Both surface-modified and unmodified glass wool samples were tested as filtration media (Figure ). For each experiment, 2.5 g of glass wool was packed into a 50 mL column to form a fixed filter bed. The microplastic-containing solution was then passed through the column using a peristaltic pump at controlled flow rates. During filtration, the PE microplastic suspension was continuously stirred using a magnetic stirrer (Lachoi, LCH-MSB-HD) at 400 rpm to maintain a homogeneous particle distribution throughout the experimental period. The filtrate was collected at the outlet in a separate container, and effluent samples were collected every 30 min over a total filtration period of 3 h. The filtration performance of the glass wool filters was evaluated by comparing removal efficiencies before and after surface modification. In addition, the effect of flow rate on microplastic removal was investigated by conducting experiments at 8, 12, and 20 rpm, corresponding to flow rates of 5, 7, and 10 mL min^–1. The empty bed contact time (EBCT) for flow rates of 5, 7, and 10 mL min^–1 was about 6.4, 4.6, and 3.2 min, respectively. The EBCT values were determined by dividing the packed-bed volume by the volumetric flow rate used. All experiments were performed in quadruplicate.

2.

Experimental setup for polyethylene microplastic separation from aqueous suspension.

Effluent samples were analyzed using a turbidimeter (PhotoFlex, WTW, Germany) as a rapid indicator of the filter medium’s performance in capturing microplastic particles. , Turbidity measurements were used as a rapid, indirect method to evaluate microplastic removal efficiency. Before the filtration experiments, a calibration curve was established by correlating known concentrations of PE microplastics with their corresponding turbidity values under identical experimental conditions (Table S1 and Figure S1), and the results were interpreted based on this relationship. The calibration curve showed a generally increasing trend with concentration, though minor deviations from linearity were observed, likely due to particle aggregation and light-scattering effects. For comparative evaluation, turbidity-based removal efficiency was calculated as (T ₀–T_t )/T ₀ × 100, where T ₀ represents the initial turbidity of the feed solution, and T_t is the turbidity measured at a given filtration time. All turbidity measurements are reported in nephelometric turbidity units (NTU).

Researchers have widely used turbidity in studies of drinking water and wastewater treatment to determine the extent to which suspended particles, such as microplastics, are removed during coagulation, filtration, and membrane-based processes. ,,,, Accordingly, decreases in turbidity were interpreted as indicative of reduced suspended microplastic particles, and the results are presented as comparative indicators of filtration performance rather than absolute quantification. It should be noted that turbidity measurements do not selectively distinguish microplastic particles from other suspended materials such as sediments, algae, colloidal matter, or dissolved organic compounds. However, since the present study was conducted using controlled PE microplastic suspensions prepared in deionized water, interference from other suspended species was minimized. Therefore, turbidity was considered a suitable comparative indicator of filtration performance under the experimental conditions employed.

2.3. Repeated Use of Surface-Treated Filter

The reusability of the prepared filters was evaluated through repeated filtration cycles. Following pretreatment and surface modification under different conditions, the sample exhibiting the best filtration performance and its corresponding flow rate were selected for reuse experiments. The selected filter was subjected to consecutive filtration runs under identical operating conditions to assess performance stability. After each 3 h filtration cycle, the filters were dried at 50 °C and reused without additional treatment. This approach was designed to evaluate short-term operational stability rather than full regeneration behavior.

2.4. Superhydrophobic Glass Wool Characterization

A Bruker VERTEX 70v spectrometer was used to analyze functional groups at a resolution of 2 cm^–1 over the range 450–4000 cm^–1. A comparison was made between the modified and unmodified samples.

A KSV Attension ThetaLite TL 101 optical tensiometer was used to measure the water contact angle. Water droplets were applied to fiber surfaces using a Hamilton microsyringe, and the apparent contact angles were noted. Surfaces with contact angles ≥130° were considered superhydrophobic. Since the measurements were performed on a porous fibrous material rather than an ideal flat surface, the measured contact angles should be interpreted as comparative indicators of wettability for the treated samples rather than intrinsic Young contact angles.

A LEO 1430 VP scanning electron microscope operating at 5 kV was used to analyze the effectiveness of surface modification treatment and surface morphology.

XPS measurements were carried out using a SPECS instrument equipped with a PHOIBOS hemispherical analyzer and a FOCUS 500 monochromatic X-ray source (SPECS Surface Nano Analysis GmbH, Berlin, Germany). The elemental composition and surface chemical states were determined from the obtained XPS spectra. Hydrophobization success was evaluated by comparing the C 1s peak intensity of silanized fibers with that of untreated glass wool.

3. Results and Discussion

3.1. FTIR Results

FTIR spectroscopy was used to better understand the temperature-dependent and surface-chemical modifications resulting from NaOH pretreatment and subsequent MTMS silanization of glass wool fibers (Figure a–c). The FTIR spectra suggest that NaOH concentration and silanization temperature influence the surface chemistry of the glass wool and are consistent with changes associated with hydroxyl-containing species and MTMS-derived structures. The wide absorption band that peaks at about 3400 cm^–1 is produced by O–H stretching vibrations from surface silanol (Si–OH) groups and water that is physically adsorbed. − As the concentration of NaOH increases, the intensity and width of this band gradually increase. These changes are consistent with the formation of additional hydroxyl-containing surface species following alkaline treatment. The observation is consistent with an increased availability of surface hydroxyl groups that may serve as reaction sites during subsequent silanization. The bands in the 2300–2500 cm^–1 range become more apparent, especially as the NaOH concentration for surface treatment and temperature increase. These characteristics are linked to carbonate compounds generated through the interaction of alkaline-treated surfaces with atmospheric CO₂. The presence of these bands suggests that residual alkaline species may interact with atmospheric CO₂ during sample handling and drying. The absorption band at about 1650 cm^–1 corresponds to H–O–H bending vibrations of water molecules that have been adsorbed onto a surface. , The differences in band intensity across samples suggest that surface polarity and water affinity may vary due to the combined alkaline treatment and silanization.

3.

FTIR spectra of glass wool samples subjected to NaOH pretreatment at varying concentrations (0.1, 0.5, and 1 M), followed by MTMS silanization at (a) room temperature, (b) 40 °C, and (c) 60 °C.

The main band in the 1100–1000 cm^–1 range is due to asymmetric stretching vibrations of Si–O–Si bonds, which are typical of the silicate network. After silanization, this band exhibits broadening and intensity changes that are consistent with the presence of additional siloxane-containing species. Such changes are consistent with the hydrolysis and condensation reactions commonly reported during MTMS-based surface modification.

As the literature on MTMS hydrolysis suggests, a band around 930 cm^–1 can be assigned to Si–OH vibrations associated with hydrolyzed silane species. The presence and increased intensity of this band, particularly in samples treated to higher NaOH concentrations, are consistent with changes in surface chemistry associated with alkaline pretreatment before silanization. This result is consistent with previous studies that identified Si–OH bands around ∼922 cm^–1 during alkoxysilane hydrolysis.

The appearance of methyl-related absorption bands provides additional evidence for the presence of MTMS-derived species on the treated glass wool surface. The peaks observed in the 2950–2850 cm^–1 range are due to C–H stretching vibrations of methyl (−CH₃) groups. The deformation bands in the 1450–1250 cm^–1 and ∼846 cm^–1 ranges are also due to terminal methyl groups from MTMS. The presence of these bands is consistent with the incorporation of organosilane-derived structures following the silanization treatment. Methyl-containing surface species are known to contribute to lower surface free energy and enhanced hydrophobic behavior, which is consistent with the contact-angle measurements obtained in this study. The relatively stable intensity of the methyl-related bands across different treatment conditions suggests that MTMS-derived species were present on the surface of all silanized samples, although the extent and uniformity of surface coverage may vary depending on the treatment conditions.

The C–H deformation bands in the 1450–1250 cm^–1 range remain relatively stable across all the silanized samples. They are caused by methyl groups from MTMS. The persistence of these bands provides additional support for the presence of organosilane-derived species on the surface of the glass wool. On the other hand, untreated samples do not exhibit these characteristics, further supporting the assignment of these bands to MTMS-derived species introduced during the silanization treatment.

The band observed near ∼580 cm^–1 at lower wavenumbers is linked to Si–O–Al vibrations in the glass wool’s aluminosilicate framework. Variations in the intensity of this band may reflect changes in the local aluminosilicate environment following the treatment steps. Temperature-dependent differences were observed among the silanized samples. In general, samples treated at elevated temperatures exhibited more pronounced methyl-related and siloxane-associated spectral features than those prepared at room temperature. These spectral changes are consistent with a greater extent of MTMS-derived surface modification at elevated temperature. At higher alkali concentrations, carbonate-related bands become more apparent, suggesting increased interaction between residual alkaline species and atmospheric CO₂. Overall, the FTIR results suggest that both NaOH concentration and silanization temperature influence the surface chemistry of the glass wool. The observed spectral variations are consistent with differences in hydroxyl-containing species, MTMS-derived groups, and siloxane-related structures among the treatment conditions.

3.2. Water Repellency

Figure compares the wettability of untreated and MTMS-modified glass wool. The untreated glass wool exhibited complete wetting, with the water droplet rapidly spreading and penetrating the fibrous structure, whereas the MTMS-modified sample retained nearly spherical water droplets, indicating a substantial increase in surface hydrophobicity following silanization.

4.

Wettability of untreated and MTMS-modified glass wool (a) untreated glass wool exhibiting complete wetting, (b) water droplets on the MTMS-modified glass wool surface, and (c) apparent water contact angle measurement of the modified sample.

Unlike the glass wool materials after surface modification, which demonstrated hydrophobic properties, the as-received glass wool showed strongly hydrophilic characteristics, with the water droplet spreading rapidly and penetrating the surface; hence, no contact angle could be determined. This is due to the hydrophilic nature of glass wool, which is caused by its hydroxyl groups (Figure a). Contact angle measurements were performed to evaluate the water repellency characteristics of surface-modified glass wool samples with varying concentrations of NaOH (0.1, 0.5, and 1 M) and then silanized with MTMS at different temperatures (Figure ). However, after surface modification by alkaline activation and silanization with MTMS, a significant increase in contact angle was observed, indicating a shift from hydrophilic to hydrophobic surface characteristics. As shown in Figure , the concentration of the NaOH solution and the temperature of the silanization process are found to greatly affect the wettability of the glass wool surface, though in a nonlinear manner. With increased NaOH concentration and temperature during MTMS silanization, the contact angle increased significantly, indicating more effective surface modification and increased hydrophobicity. The hydrophobic nature of glass wool materials after surface modification can be attributed to the presence of methyl (−CH₃) groups covalently bonded to the surface, as confirmed by FTIR spectroscopy. Similar contact angle values (>130°) have also been reported in previous studies. ,,

5.

Water contact angle values for the samples treated with different concentrations of NaOH (0.1, 0.5, and 1 M) and then silanized with MTMS at different temperatures. Each value is an average of several measurements carried out at various points on the surface of each glass wool sample. Note: The untreated sample shows complete wetting and is not included in the graph.

3.3. SEM Analysis

The untreated glass wool fibers have a typically smooth, inactive surface morphology with low roughness (Figure ). After washing and reacting with an alkaline solution, combining it with a methyltrimethoxysilane (MTMS) treatment alters the surface structure and coating behavior in multiple ways, depending on the NaOH concentration and the reaction temperature (Figures , , and ). Silanization at room temperature (RT) yields relatively thin and partially heterogeneous coatings. In the 0.1 M NaOH-treated sample, surface deposits appear to be localized to specific regions of the fiber surface. The observed morphology indicates a relatively limited degree of surface coverage under these treatment conditions. Increasing the NaOH concentration to 0.5 M results in a greater number of visible surface deposits and roughness features. The observed increase in surface deposits may be associated with enhanced surface activation after alkaline pretreatment, as indicated by the appearance of uniformly distributed nanoscale deposits on the fiber surface. At room temperature 1 M NaOH, the silane layer becomes more apparent but also less uniform (Figure ). The observed morphology is characterized by localized accumulation and heterogeneous surface coverage.

6.

SEM micrographs of untreated glass wool fibers at different magnifications (20,000×; 5000×; and 500×).

7.

SEM micrographs of MTMS-modified glass wool fibers at room temperature following NaOH pretreatment with varying concentrations (0.1, 0.5, and 1 M) and varying magnifications (20,000×; 5000×; and 500×).

8.

SEM micrographs of MTMS-functionalized glass wool fibers at 40 °C after NaOH activation at different concentrations (0.1, 0.5, and 1 M) and varying magnifications (20,000×; 5000×; and 500×).

9.

SEM micrographs of MTMS-functionalized glass wool fibers at 60 °C after NaOH activation at different concentrations (0.1, 0.5, and 1 M) and varying magnifications (20,000×; 5000×; and 500×)

When the silanization process is carried out at 40 °C, the surface morphology becomes more homogeneous than that observed for the room-temperature samples, with the surface deposits appearing more uniformly distributed across the fiber surface than in the room-temperature samples. The fibers in the 0.5 M condition exhibited hierarchical roughness, with nodular and granular features. Similar surface morphologies have been reported following silane-based surface treatments. At 1 M NaOH and 40 °C, the silane layer becomes denser and less even, with evidence of localized thickening (Figure ). These observations indicate increased accumulation of surface deposits under these treatment conditions.

At 60 °C, more pronounced surface deposits and roughness features are observed (Figure ). Taking a close look at the 0.1 M sample, clear spherical and semi-spherical features on the fiber surfaces can be noticed. These features are consistent with the presence of surface deposits formed during the treatment process. The appearance of spherical and semi-spherical deposits indicates noticeable morphological changes compared with the room-temperature samples. In the 0.5 M condition, the surface is covered with many micro- and nanoscale particles, forming a complex hierarchical structure. A similar morphology has been reported for silane-treated mineral surfaces. , The SEM observations are consistent with the presence of surface deposits and morphological changes following treatment. Chemical interpretation of the surface modification is supported separately by FTIR and XPS analyses. For the 1 M sample at 60 °C, extensive surface deposits and aggregation partially obscure the original fibrous structure. The observed morphology indicates substantial accumulation of deposited material and non-uniform surface coverage under these conditions.

To further investigate the nature of the particulate deposits observed on the fiber surfaces after silanization, spot EDS analysis was performed directly on representative surface deposits (Figure ). The EDS spectrum showed that the analyzed region was dominated by silicon and oxygen, together with the characteristic elements of the glass wool substrate. Carbon was also detected, which is consistent with the presence of organosilane-derived species introduced during MTMS treatment. Since the EDS interaction volume also includes the underlying glass fiber, the measured composition represents contributions from both the surface deposit and the substrate. Nevertheless, the results suggest that the observed particles are associated with the MTMS silanization process.

10.

SEM micrograph of a representative particulate deposit formed on the MTMS-modified glass wool fiber and the corresponding spot EDS spectrum acquired from the marked region.

3.4. XPS Analysis

Since SEM observations alone cannot verify the chemical nature of the deposited layer, XPS analysis was performed to confirm the incorporation of MTMS-derived species onto the glass wool surface. The survey XPS spectra were recorded over a wide binding energy range (0–1100 eV) to determine the surface elemental composition of the samples (Figure a). The surface compositions of the raw and modified samples are summarized in Table . The untreated glass wool surface is mainly composed of oxygen (O) and silicon (Si), which are associated with the underlying silicate structure. Carbon (C) is also detected on the surface and is attributed mainly to adventitious carbon commonly observed in XPS analyses of air-exposed materials. After MTMS modification, a clear change in surface composition is observed (Figure a). Following MTMS treatment, the sample mass increased by approximately 0.10–0.15 g per 5 g of glass wool, corresponding to about 2–3 wt % of deposited material. Although this gravimetric approach does not distinguish chemically bonded silane from physically retained species, the observed mass increase provides additional quantitative evidence for MTMS deposition on the glass wool surface. In particular, the C content increases significantly, while the relative atomic percentages of O and Si decrease. In addition, sodium (Na) shows a slight reduction after modification, indicating that the NaOH treatment primarily serves as a surface-activation step and does not remain on the surface after washing.

11.

. XPS survey and high-resolution spectra of raw and 1 M NaOH–60 °C samples: (a) survey spectra; O 1s spectra of (b) raw and (c) modified samples; Si 2p spectra of (d) raw and (e) modified samples; and C 1s spectra of (f) raw and (g) modified samples.

1. Surface Elemental Composition of Raw and 1 M NaOH-60 °C Sample Obtained from XPS Survey Spectra.

Element	Raw (at. %)	Modified (at. %)
O 1s	45.51	28.85
C 1s	26.45	52.32
Si 2p	16.43	10.88
Na 1s	5.21	3.87
S 2p	5.26	4.59
N 1s	1.12	0.57

The increase in C content indicates the successful incorporation of organosilane species onto the glass wool surface. At the same time, the decrease in O and Si suggests that the original silicate surface is partially covered by an organic layer. As shown in Table , the C/O ratio increases markedly after modification, whereas the O/Si ratio shows only a minor change, indicating that the bulk silicate framework remains largely unaffected and that the modification is mainly confined to the outermost surface. , It is important to note that the decrease in O content does not imply the absence of O functionalities. Instead, O undergoes a redistribution from lattice O in the silicate network to more chemically diverse environments associated with surface hydroxyls and silane-related O species. These hydroxyl groups are initially generated during NaOH pretreatment and subsequently serve as reactive sites for silane attachment.

2. Elemental Ratios Derived from XPS Analysis of Raw and 1 M NaOH-60 °C Sample.

Elemental ratio	Raw	Modified
C/O	0.58	1.81
O/Si	2.77	2.65

To gain further insight into the surface chemical states and bonding environments, the high-resolution O 1s, Si 2p, and C 1s spectra were deconvoluted using component peak fitting (Figure b–g). The minimum number of chemically meaningful components required to reproduce the experimental spectra was employed in order to avoid overfitting.

The O 1s spectra of both the raw and MTMS-modified glass wool were resolved into three contributions assigned to Si–OH, Si–O-M, and Si–O–Si environments. Here, M denotes network-modifying metal cations (e.g., Na, Ca, and Mg) present in the glass wool composition; consequently, Si–O-M corresponds to non-bridging oxygen (NBO) species within the silicate network, whereas Si–O–Si represents bridging oxygen (BO) atoms linking adjacent SiO₄ tetrahedra. The Si–OH contribution is attributed to hydroxylated surface sites. Following NaOH activation and MTMS treatment, the relative distribution of these oxygen environments changed noticeably. Although the Si–O–Si contribution remained dominant, indicating preservation of the silicate framework, the modified sample exhibited a redistribution of oxygen among Si–OH, Si–O-M, and siloxane-related environments. These changes are consistent with modification of the surface oxygen environment during alkali activation and subsequent condensation of hydrolyzed MTMS species. Since the binding-energy ranges of substrate-derived Si–O–Si and MTMS-derived siloxane species strongly overlap, complete discrimination of all oxygen environments from the O 1s spectrum alone is not feasible. Therefore, the O 1s spectra were interpreted in conjunction with the C 1s, Si 2p, and survey XPS results.

The Si 2p spectra further support the surface modification. The untreated glass wool was adequately described by two contributions assigned to Si–O-M and Si–O–Si environments, corresponding to less polymerized and more polymerized silicon species within the glass network, respectively. Following MTMS modification, two principal components were likewise sufficient to reproduce the Si 2p envelope. Besides the Si–O–Si contribution originating predominantly from the silicate substrate, the lower-binding-energy component in the modified sample may include MTMS-derived organosilicon species (R-SiO _x ) in addition to substrate-related less-polymerized silicon environments. The change in the relative contribution of the lower-binding-energy component is consistent with the formation of a silane-derived surface layer. Because of the substantial overlap between silicon species originating from the substrate and those associated with the condensed siloxane network, further subdivision of the Si 2p spectrum was not considered chemically justified and could lead to overinterpretation.

The C 1s spectra exhibited the most pronounced changes following MTMS modification. The untreated sample was dominated by a C–C/C–H contribution arising primarily from adventitious carbon commonly present on air-exposed surfaces, together with a minor contribution assigned to oxygen-containing carbon species (C–O/CO), which are generally attributed to surface oxidation and atmospheric carbonates. In contrast, the modified sample required an additional low-binding-energy component assigned to Si–C/Si–CH₃, which is characteristic of the methyl functionality of MTMS and provides supporting evidence for the incorporation of organosilane species onto the glass wool surface. , The modified spectrum also contained C–C/C–H and C–O contributions corresponding to hydrocarbon and oxygen-containing carbon environments. In agreement with the survey spectra, the marked increase in surface carbon concentration together with the appearance of the Si–C/Si–CH₃ supports successful surface modification with MTMS-derived species.

Overall, the combined survey and high-resolution XPS analyses demonstrate that NaOH activation followed by MTMS treatment alters both the surface elemental composition and the chemical bonding environments of the outermost glass wool surface. The increased surface carbon concentration, the appearance of Si–C/Si–CH₃ species in the C 1s spectrum, the redistribution of O 1s components, and the changes in the Si 2p region consistent with MTMS-derived organosilicon environments collectively provide supporting evidence for the successful silanization of the glass wool surface. These findings are consistent with the FTIR results and the substantial increase in water contact angle, confirming the formation of a hydrophobic organosilane layer. , The gravimetric mass increase observed after silanization is also consistent with the XPS results, which show a substantial increase in surface carbon concentration after MTMS treatment.

3.5. Effect of NaOH Pretreatment, Silanization Temperature, and Flow Rate on Microplastic Removal Performance

The untreated glass wool showed limited and unstable microplastic retention behavior (Figure a). There was some microplastic retention, but the turbidity profiles changed significantly during the experiment, indicating inconsistent performance (Figure a and Table S2). Such behavior can be attributed to weak surface-particle interactions, as it is well established that insufficient surface modification results in reversible adsorption and easier detachment of microplastics under hydrodynamic conditions. ,

12.

Time-dependent turbidity profiles obtained during microplastic filtration using untreated and surface-modified glass wool samples. (a) Untreated glass wool tested at flow rates corresponding to 8, 12, and 20 rpm. For the surface-modified samples, all glass wool samples were pretreated with 0.1, 0.5, or 1 M NaOH and subsequently silanized at 60 °C, 40 °C, or room temperature (RT). Filtration experiments were conducted at 20 rpm (b–d), 12 rpm (e–g), and 8 rpm (h–j). The rotational speeds of 8, 12, and 20 rpm correspond to flow rates of approximately 5, 7, and 10 mL min^–1, resulting in empty bed contact times (EBCTs) of approximately 6.4, 4.6, and 3.2 min, respectively. All experiments were performed in quadruplicate. Data points represent mean values, and error bars indicate one standard deviation (n = 4). Complete numerical data are provided in the Supporting Information (Tables S2–S5).

For the modified samples, the turbidity profiles did not show a strictly monotonic trend with filtration time or NaOH concentration (Figure b–d). Temporary fluctuations in turbidity were observed under several operating conditions and may be attributed to the combined effects of particle aggregation, local detachment and reattachment of PE particles, redistribution of flow paths within the fibrous bed, and the depth-filtration nature of the glass wool network. ,

The effect of NaOH concentration became more evident when the overall performance of the modified samples was compared. The 0.1 M series exhibited more scattered performance, with only a few conditions showing low turbidity, while several others resulted in significantly higher values (Tables S3, S4, and S5). This variability may indicate less uniform surface modification and a more heterogeneous distribution of interaction sites. In contrast, the 0.5 M series demonstrated intermediate performance, whereas the 1 M series consistently produced lower and more stable turbidity values under most operating conditions.

Silanization temperature also influenced filtration performance. In general, 60 °C yielded the most favorable results, particularly in the 1 M series. The 1 M-20 rpm-60 °C sample maintained turbidity values between approximately 0.10 and 0.16 NTU throughout the filtration period, representing the most stable condition in the dataset. SEM images of this sample revealed a highly textured surface morphology with pronounced surface deposits and micro- and nanoscale roughness (Figure ). The presence of silane-derived surface species was confirmed independently by FTIR (Section ) and XPS (Section ) analyses, while contact-angle measurements confirmed substantial changes in surface wettability following treatment. Similar findings have been reported for silane-modified materials, where the introduction of hydrophobic functional groups significantly enhances the affinity toward microplastics and improves capture efficiency through combined adsorption and physical trapping mechanisms. , The improved filtration performance is therefore attributed to the combined effects of surface morphological and chemical modifications rather than to any single parameter. ,,,

At lower temperatures, surface modification was less effective. The 40 °C condition produced a more homogeneous surface morphology than the room-temperature samples, whereas the room-temperature treatments generally exhibited more heterogeneous surface coverage. Notably, the benefits of higher silanization temperatures were most pronounced when combined with sufficient alkaline pretreatment. At low molarity (0.1 M), higher temperatures did not consistently improve performance, likely due to localized silane accumulation rather than uniform surface coverage. Flow rate did not exhibit a single dominant trend but rather depended on the interaction between surface chemistry and hydrodynamic conditions. In the 1 M series, the best performance was observed at 20 rpm, indicating that well-developed surfaces can maintain stability even at higher flow rates. Differences in surface morphology may contribute to variations in particle retention under hydrodynamic conditions; however, filtration performance is likely governed by the combined effects of surface morphology, surface chemistry, wettability, and flow conditions. , In moderately structured systems (e.g., 0.5 M), lower flow rates improved performance by increasing contact time, partially compensating for weaker adhesion.

Overall, the results suggest that untreated glass wool exhibits less stable microplastic retention behavior than the modified samples investigated in this study. Instead, the improved performance observed in treated samples arises from surface modifications induced by alkaline activation and silanization, which enhance particle-surface interactions. The observed removal performance is consistent with the combined contribution of hydrophobic interactions, van der Waals forces, and mechanical trapping within the fibrous structure, all of which are expected to be enhanced after surface modification. ,

Based on its consistently low turbidity values and stable filtration behavior throughout the experimental period, the 1 M-20 rpm-60 °C sample was selected for the repeated-use experiments. The 1 M-20 rpm-60 °C sample consistently exhibited low turbidity values, suggesting that the combined effects of surface chemistry and hierarchical morphology are critical for achieving reliable microplastic removal. Based on turbidity reduction, removal efficiencies of 87–98 % were achieved under optimal conditions (Tables S6, S7, and S8). These values are reported as turbidity-based comparative filtration efficiencies and should not be interpreted as absolute microplastic removal values. Using the processed suspension volume together with the turbidity-derived removal efficiencies, the amount of PE retained during a single 3 h filtration cycle under the optimum operating conditions was estimated to be approximately 7.8–8.8 mg. This corresponds to an apparent PE loading of 3.1–3.5 mg g^–1 of glass wool. Since the filter was not operated until breakthrough or saturation, these values should be interpreted as operating-condition-dependent loadings rather than maximum filtration capacities. In contrast, untreated glass wool exhibited unstable and fluctuating retention behavior under different flow conditions, indicating insufficient particle-surface interactions and poor filtration stability in the absence of surface modification (Table S9).

The gravimetric measurements further support the occurrence of MTMS deposition under the selected modification conditions. Consistent with the XPS results, which reveal changes in the surface chemical composition after alkaline activation and MTMS treatment, the improved filtration performance of the modified samples suggests that the deposited organosilane layer remained sufficiently stable throughout the filtration experiments.

The absence of a visible cake layer on the filter surface (Figure a) indicates that microplastic retention occurs primarily through distributed capture within the fibrous network rather than conventional surface cake formation. , Unlike membrane-based filtration systems where particle accumulation often results in a dense surface cake layer, the open and highly porous structure of glass wool allows particles to penetrate the filter bulk, promoting depth filtration instead of surface filtration. , Consequently, regeneration based on surface cleaning is limited, as retained particles are not confined to the external surface but are instead embedded within the fiber network, consistent with the behavior predicted by Hermia’s model. In addition, the relatively low particle loading (5 mg L^–1) used in this study may not be sufficient to induce cake layer formation, as the particle flux is likely below the threshold required for surface accumulation and compression. This behavior is consistent with the observed stable yet slightly variable turbidity profiles during repeated use, indicating that particle retention occurs through distributed trapping rather than surface blocking. Overall, the results show that the material can still remove microplastics after several cycles. However, changes in turbidity show that performance can vary with repeated use.

13.

(a) Visual appearance of the filter after operation, showing no observable cake layer or removable accumulation on the surface. (b) The performance of the 1 M-60 °C sample during repeated use cycles for PE microplastics removal (1 M NaOH pretreatment, MTMS silanization at 60 °C, 20 rpm, 5 mg L^–1 PE suspension, 3 h filtration cycle). Data points represent mean values, and error bars indicate one standard deviation (n = 4).

The performance of the filter during repeated use cycles, following drying after each 3 h operation period, is presented in Figure b. The first cycle had the lowest and most stable turbidity values. There was a slight increase in turbidity during the second cycle, but the overall removal performance remained similar to that of the first cycle. The third cycle showed more pronounced fluctuations in turbidity, particularly at longer contact times, whereas the fifth cycle showed relatively higher turbidity values at the beginning of the experiment. On the other hand, the fourth cycle showed relatively stable behavior, indicating that the system did not follow a strictly monotonic trend. These differences may be due to the buildup of particles trapped in the fibrous structure, as well as changes in flow paths and local filtration dynamics after drying and reusing the material several times. Nevertheless, the filter remained capable of reducing turbidity throughout all tested cycles, indicating that the modified glass wool retained its filtration functionality after repeated use, although some variability in performance was observed.

It should be noted that the repeated-use experiments were not designed to determine the maximum microplastic loading capacity of the filter medium. Instead, the experiments were performed under a relatively low and constant influent concentration (5 mg L^–1) to evaluate short-term operational stability. Consequently, although the filter maintained its performance over five consecutive cycles, the present results do not establish the saturation capacity or complete regeneration capability of the material. Rather, they demonstrate that, under the investigated operating conditions, the accumulated microplastic loading was insufficient to cause substantial deterioration in filtration performance. Determination of the maximum loading capacity and regeneration efficiency will require dedicated breakthrough and regeneration experiments conducted at progressively higher particle loadings and is beyond the scope of the present study.

4. Conclusion

In this study, glass wool waste was successfully surface-modified and evaluated as a potential filtration medium for microplastic-containing suspensions through a combination of NaOH pretreatment and MTMS-based silanization. The results indicate that surface modification influences the filtration performance of glass wool under the investigated conditions. Alkaline pretreatment was associated with spectral changes consistent with an increased abundance of hydroxyl-containing surface species. FTIR and XPS analyses showed the formation of an organosilane-based surface layer, while SEM observations revealed hierarchical surface structures that depended on the processing conditions. Contact angle measurements further demonstrated a significant transition from hydrophilic to hydrophobic behavior after modification. Filtration experiments revealed that untreated glass wool exhibited variable turbidity profiles, whereas modified samples generally produced lower turbidity values under comparable operating conditions. Among the tested samples, the material pretreated with 1 M NaOH and silanized at 60 °C consistently exhibited the lowest turbidity values during filtration experiments. The observed performance is likely related to changes in surface chemistry and morphology, including increased hydrophobicity, surface roughness, and particle retention within the fibrous network. Although the material retained its functionality over multiple uses, some variability in performance was observed, indicating the need for further research into long-term stability and regeneration strategies. Overall, this study demonstrates that MTMS-based surface modification can successfully alter the surface properties of waste glass wool and may enhance its suitability as a filtration medium for microplastic-containing suspensions while also promoting sustainable waste management and the implementation of circular economy principles.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c05772.

• PE microplastic calibration data and calibration curve; turbidity data for untreated and surface-modified glass wool under different pretreatment, silanization, and filtration conditions; turbidity-based removal efficiencies (Tables S1–S9 and Figure S1) (PDF)

The authors declare no competing financial interest.