Mercury Removal and Antibacterial Performance of A TiO₂–APTES Kaolin Composite

Abstract

Mercury (Hg²⁺) contamination in water systems poses a severe environmental and health hazard due to its high toxicity and bioaccumulation potential. In this study, a novel adsorbent was developed by sequentially modifying kaolin via acid–base treatment, titanium dioxide (TiO₂) incorporation, and 3-aminopropyltriethoxysilane (APTES) grafting. Batch adsorption experiments revealed that the fully modified kaolin (TiO₂-loaded and APTES grafted) exhibited the highest adsorption capacity (25.6 mg/g) compared to the acid–base-treated (5.8 mg/g) and TiO₂-loaded (17.7 mg/g) kaolin. Under optimal conditions (75 mg adsorbent dosage; 70 mg/L Hg²⁺; pH 5), the fully modified kaolin maintained its performance even in the presence of varying ionic strengths, natural organic matter, and competing metal ions. Adsorption kinetics followed a pseudo-second-order model, and the equilibrium data were well fitted by the Langmuir isotherm. Antibacterial activity assay revealed that the TiO₂-loaded kaolin effectively inhibited S. aureus (minimum inhibitory concentration = 2.5 mg/mL) and showed moderate activity against E. coli (BL21) (minimum inhibitory concentration = 5 mg/mL). However, antibacterial activity decreased after amine functionalization, indicating a compromise between enhancing adsorption capacity and preserving antibacterial functionality. This study presents a promising cost-efficient approach for the simultaneous removal of Hg²⁺ ions from water matrices and inhibiting bacterial growth, aligning with SDG 6 (Clean Water and Sanitation).

1. Introduction

Rapid population growth and industrialization have intensified the discharge of pollutants into aquatic environments [1]. Industrial effluents often contain hazardous metal ions such as mercury, lead, and cadmium, which are nonbiodegradable and tend to bioaccumulate in ecosystems [2,3,4]. Mercury, in particular, is a priority pollutant due to its extreme toxicity and persistence. Once released into water bodies, inorganic mercury can be transformed through microbes into organic mercury (e.g., methylmercury), a form that bioaccumulates in aquatic food chains [5,6]. This leads to high mercury concentrations in predatory fish and poses serious health risks to wildlife and humans upon consumption [7]. Even at very low concentrations, mercury can damage the nervous, immune, and renal systems and is particularly harmful to developing fetuses [8,9]. The World Health Organization (WHO) considers mercury one of the top ten chemicals of major public health concern. Accordingly, regulatory agencies have set extremely low limits for mercury in drinking water (e.g., the WHO guideline of ~0.001 mg/L) [10], necessitating the development of effective removal technologies for mercury-contaminated waters.

To remediate mercury contamination in water systems, a variety of treatment methods such as chemical precipitation [11], membrane filtration [12], ion exchange [13], and adsorption [14] have been developed over the years. Most of these methods have been found to achieve high removal efficiencies, often exceeding 90% and even reaching 100% in ideal cases, owing to their high selectivity, operational simplicity, and scalability [15]. However, these technologies face practical limitations such as high operational costs and sludge generation. Adsorption is a promising alternative due to its simplicity, cost-effectiveness, and versatility [16,17]. Activated carbon is the most common adsorbent; however, its rising cost necessitates the search for cheaper alternatives [18]. Therefore, low-cost porous solids such as natural clays [19,20], zeolites [21], and agricultural wastes [22] have been extensively studied as adsorbents for mercury remediation. Among these, clay minerals are particularly attractive due to their abundance, affordability, and ability to be chemically modified to enhance performance [23,24].

Kaolinite, which is the primary component of kaolin, is a widely available 1:1 layered aluminosilicate that has been explored as an adsorbent for mercury removal from aqueous solutions [25,26,27]. It typically has a moderate cation exchange capacity and surface area, and its adsorption capability for metal ions is often lower than that of 2:1 clay, such as montmorillonite [23]. However, the large natural abundance and chemical stability of kaolin make it an appealing base material for water treatment applications [25,28]. Studies have found that various surface modifications can substantially improve the metal uptake capacity of kaolin. For example, acid or alkaline activation can increase surface area and create new active sites [28,29]. Inorganic oxide incorporation, such as titanium dioxide (TiO₂) loading, can also enhance the adsorption properties of kaolin by increasing its surface area, porosity, and introducing new functional sites [28]. Surface modification with organic compounds such as organosilanes is another effective strategy [30,31]. For instance, 3-aminopropyltriethoxysilane (APTES), an organosilane that grafts amine (–NH₂) groups onto the surface of materials, endows adsorbents with chelating groups that strongly bind heavy metal ions via coordination or electrostatic attraction [32,33]. Previous studies with APTES-modified adsorbents have demonstrated marked improvements in capacity. Abdellaoui et al., (2024) found that APTES-grafted zeolite and silica materials exhibited lead and cadmium adsorption capacities in the range of 200–400 mg/g, exceeding those for unmodified materials [34]. In the case of kaolin, APTES treatment has been found not only to enhance metal uptake but also to increase the stability and dispersion of kaolin in water [35]. By combining TiO₂ with APTES grafting, the resulting composite is expected to synergistically benefit from both inorganic and organic modification: TiO₂ provides a high-surface-area scaffold and additional binding sites, while APTES supplies targeted functional groups for mercury capture [36,37].

Water contaminated with heavy metals often harbors pathogenic microorganisms [38,39], so an ideal treatment medium would remove toxic metals while simultaneously inhibiting microbial growth. TiO₂ is a widely studied antimicrobial agent. Under ultraviolet illumination, it generates reactive oxygen species (ROS) such as •OH radicals that can destroy bacteria and viruses [40], and the incorporation of APTES can further improve antibacterial performance. Recent studies on APTES-modified TiO₂ nanoparticles (NPs) showed that APTES not only helps disperse NPs but also introduces a positively charged surface (due to protonated –NH₃⁺ groups) that promotes adhesion to negatively charged bacterial cell walls [41]. TiO₂–APTES-modified kaolin is anticipated to inhibit bacterial growth via a combination of contact killing and photocatalytic disinfection mechanisms. Thus, this material can address both chemical pollution (Hg²⁺ ions removal) and biological contamination (bacterial growth inhibition), making it a versatile candidate for water purification and contributing directly to Sustainable Development Goal (SDG) 6 (Clean Water and Sanitation).

In this study, TiO₂–APTES-modified kaolin was developed to evaluate its performance in mercury adsorption and antibacterial activity. Key objectives include: (1) determining the mercury adsorption capacity of the modified kaolin and comparing it to unmodified kaolin, (2) studying the adsorption isotherms and kinetics to understand the uptake mechanism, and (3) assessing the antibacterial activity of synthesized materials against representative Gram-positive (S. aureus) and Gram-negative (E. coli) bacteria. The multi-step surface engineering approach employed in this study is novel and is expected to yield a dual-functional adsorbent with both mercury-binding capacity and inherent antibacterial activity.

2. Materials and Methods

2.1. Materials and Chemicals

Raw kaolin (RK) was purchased from Sigma–Aldrich and used without purification. Hydrochloric acid (HCl; 37%), sodium hydroxide (NaOH; >97%), titanium tetraisopropoxide (TTIP; ≥99%), APTES (≥98%), ethanol (96%), ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA 2Na·2H₂O; ≥99%), potassium iodide (KI; 99%), copper(II) nitrate trihydrate (98–103%), nickel(II) sulfate hexahydrate (≥98%), zinc sulfate heptahydrate (≥99%), cadmium nitrate tetrahydrate ≥99%), and mercury(II) chloride (≥99.5%) were all purchased from Sigma–Aldrich. All reagents were of analytical grade and were used without further purification. Deionized (DI) water (18.2 MΩ·cm) was used for all preparations and washing steps.

2.2. Material Preparation

2.2.1. Acid–Base Treatment of Kaolin

RK was activated through a sequential acid and base treatment, adapted from methods reported by Abdulsalam et al., (2025) [42]. First, RK was thermally activated by heating to 600 °C (ramp rate of 10 °C/min) and holding for 5 h in a muffle furnace. This calcination step converts kaolin to a more reactive metakaolin phase. After cooling, the calcined kaolin was subjected to acid leaching in 2.5 M of HCl at 80 °C for 7 h to remove a portion of Al³⁺ from the lattice and increase surface porosity. This treatment generated additional silanol groups, increasing the density of reactive sites. The acid-treated kaolin was filtered and washed repeatedly with DI water until pH 7 to remove excess acid and dried overnight at 90 °C. Next, the material was treated with 0.5 M NaOH solution at 80 °C for 7 h to induce mild alkaline activation. This step partially etched the silica-rich surface, increasing surface roughness and porosity without causing complete dissolution of the aluminosilicate framework. The alkaline treatment also led to partial modification of the surface silicate network, further enhancing the reactivity of the kaolin surface for subsequent functionalization. The resulting acid–base-treated kaolin was thoroughly washed with DI water until neutral pH and then oven-dried at 90 °C overnight. This product is hereafter referred to as ABK.

Nano-TiO₂ was synthesized via a sol–gel process adopted from Mustapha et al., (2021) to ensure uniform particle size, high purity, and a well-defined anatase phase [37]. Here, 20 mL of TTIP was added to 200 mL of DI water under vigorous stirring. The hydrolysis of the titanium precursor was initiated by adjusting the mixture to pH ≈ 8 with dropwise addition of 0.1 M NaOH, resulting in the formation of a white titania sol. The suspension was continuously stirred and allowed to age for 24 h at room temperature (25 °C) to ensure complete hydrolysis and polycondensation. The precipitated TiO₂ was then collected and washed several times with DI water to remove any impurities. The washed solid was dried at 100 °C for 3 h, and subsequently calcined at 450 °C for 3 h to obtain crystalline TiO₂ NPs.

2.2.2. TiO₂ Loading and APTES Grafting on Kaolin

To prepare a TiO₂–APTES-modified kaolin composite, ABK was first incorporated with the synthesized TiO₂ NPs and then functionalized with –NH₂ groups via APTES grafting. First, ABK and the prepared TiO₂ NPs were mixed at a mass ratio of 20:80 (TiO₂:ABK) in DI water, and the slurry was stirred for 24 h at room temperature. This process allowed TiO₂ NPs to distribute and attach to the kaolin surface. After mixing, the solid composite was recovered by filtration and washed with DI water to remove unattached particles. The product was dried at 90 °C overnight, yielding TiO₂-loaded kaolin, denoted as ABK–TiO₂. Subsequently, ABK–TiO₂ was functionalized with –NH₂ groups via APTES grafting. Here, 5 mL of APTES was added to a mixture of ethanol and water (95 mL of ethanol:5 mL of DI water) and stirred for 30 min. Then, 5 g of ABK–TiO₂ was introduced. The suspension was stirred at room temperature for 7 h, allowing the silane to hydrolyze and subsequently form a covalent bond with hydroxyl groups on the TiO₂–kaolin surface. After grafting, the sample was centrifuged at 4000 rpm and washed thoroughly with ethanol, followed by DI water, to remove any unreacted APTES or by-products. The APTES-modified composite was then dried overnight at 90 °C. The final functionalized material is referred to as ABK–TiO₂–APTES. Figure 1 shows a schematic illustration of the preparation process of AB, ABK–TiO₂, and ABK–TiO₂–APTES.

2.3. Characterization

Raw and modified kaolin were characterized by various analytical techniques to confirm their composition, structure, and morphology. X-ray fluorescence (XRF) analysis was conducted by preparing fused beads of the samples using the Scientific xrFuse 6 instrument (XRF Scientific, Avenue de Roodebeek, Brussels, Belgium) and then analyzing them for chemical composition using an Axios Max wavelength dispersive XRF spectrometer (Malvern Panalytical, Almelo, The Netherlands). X-ray diffraction (XRD) patterns were recorded with a Rigaku diffractometer (Rigaku Corporation, Tokyo, Japan). The instrument was operated at 40 kV and 50 mA with Cu Kα radiation (λ = 1.5406 Å). Scans were collected over a 2θ range of 5–70° at a step size of 0.01°, with a scanning speed of 3° min⁻¹. Fourier transform infrared spectroscopy (FTIR) spectra were obtained using an ALPHA II FTIR spectrometer (Bruker Optik GmbH, Ettlingen, Germany) equipped with a platinum attenuated total reflectance module over the range 4000–400 cm⁻¹ at a resolution of 4 cm⁻¹ with 32 scans per sample. Surface morphologies were examined through scanning electron microscopy (SEM; Carl Zeiss AG, Oberkochen, Germany), while a JEOL JEM-1400 Plus transmission electron microscope (TEM; JEOL Ltd., Tokyo, Japan) operated at an accelerating voltage of 120 kV was used to observe the fine structure of the TiO₂ NPs and their attachment on the kaolin surface. Nitrogen adsorption–desorption isotherms were measured at 77 K using a 3Flex gas adsorption analyzer (Micromeritics Instrument Corporation, Norcross, GA, USA), and the specific surface area of the samples was calculated from the adsorption data using the Brunauer–Emmett–Teller (BET) model. Zeta potential over the range of pH 3 to 11 was measured using a ZetaView PMX-230 TWIN nanoparticle analyzer (Particle Metrix, Ammersee, Germany). Surface elemental composition of the modified kaolin was investigated before and after adsorption through X-ray photoelectron spectroscopy (XPS) using a NEXSA XPS system (Thermo Scientific, Waltham, MA, USA).

2.4. Adsorption Studies

Batch adsorption experiments were conducted to evaluate the performance of the materials in removing Hg²⁺ ions from aqueous solution and to study the effects of various operational parameters. In each experiment, a known mass of adsorbent was added to 50 mL of Hg²⁺ solution in a paraffin-sealed conical flask and agitated at a constant speed (150 rpm) using an orbital shaker. A preliminary adsorption study was conducted to determine the equilibrium time using an initial Hg²⁺ concentration of 50 mg/L and a dosage of 25 mg. Subsequently, the effect of dosage (12.5–100 mg), initial Hg²⁺ concentration (10–90 mg/L), and solution pH (3–11) was investigated. The effect of natural organic matter (NOM) was determined using humic acid (HA) and glucose at concentrations ranging from 1 to 50 mg/L. Finally, to examine the effect of ionic strength and the influence of coexisting heavy metals, adsorption experiments were performed in the presence of NaNO₃ (0.005–0.1 M) and selected heavy metals, including Cu²⁺, Cd²⁺, Zn²⁺, and Ni²⁺. Two sets of adsorption experiments were conducted by spiking the optimal concentration of Hg²⁺ solution with 5 and 10 mg/L of each coexisting metal, resulting in total coexisting metal concentrations of 20 and 40 mg/L, respectively. Mercury concentrations before and after adsorption were measured using a Lumex RA-915M mercury analyzer (Lumex Instruments, St. Petersburg, Russia), while Cu²⁺, Cd²⁺, Zn²⁺, and Ni²⁺ concentrations were determined using an inductively coupled plasma optical emission spectroscopy system (iCAP 6300; Thermo Scientific, Renfrew, United Kingdom). For each set of conditions, the experiments were performed in triplicate to ensure reproducibility. All reported results are the mean values, and error bars on graphs represent the percentage error (typically within ±5% of the value). The percentage removal and the amount of Hg²⁺ adsorbed at equilibrium (qe) were calculated using Equations (1) and (2), respectively.

$$\mathrm{R}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}={\displaystyle \frac{{\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{e}}}{{\mathrm{C}}_0}}\times 100\%$$

(1)

$${\mathrm{q}}_{\mathrm{e}}={\displaystyle \frac{{\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{e}}}{\mathrm{m}}}\times \mathrm{V}$$

(2)

where C₀ and C_e are the initial and equilibrium Hg²⁺ concentrations (mg/L), respectively, m is the mass (g) of the adsorbent, and V is the volume (L) of Hg²⁺ solution. For kinetic studies, the adsorption capacity (qt) of Hg²⁺ adsorbed at time t (h) was calculated using Equation (3) as follows:

$${\mathrm{q}}_{\mathrm{t}}={\displaystyle \frac{{\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{t}}}{\mathrm{m}}}\times \mathrm{V}$$

(3)

where C_t is the concentration (mg/L) of Hg²⁺ at time t (h).

2.5. Adsorption Modeling

Adsorption kinetics and isotherms were analyzed using standard models to describe the mechanism of Hg²⁺ uptake. Two kinetic models, the pseudo-first-order (PFO) and pseudo-second-order (PSO) models, were applied to the time-dependent adsorption data. The nonlinear forms of these models are given in Equations (4) and (5), respectively:

$${\mathrm{q}}_{\mathrm{t}}={\mathrm{q}}_{\mathrm{m}}(1-{\mathrm{e}}^{-{\mathrm{K}}_1\mathrm{t}})$$

(4)

$${\mathrm{q}}_{\mathrm{t}}={\displaystyle \frac{{\mathrm{q}}_{\mathrm{m}}^2{\mathrm{K}}_2\mathrm{t}}{1+{\mathrm{q}}_{\mathrm{m}}{\mathrm{K}}_2\mathrm{t}}}$$

(5)

where qt (mg/g) is the amount of Hg²⁺ adsorbed at time t (h), qm (mg/g) is the maximum adsorption capacity, K₁ (h⁻¹) is the rate constant for the PFO model, and K₂ (g mg⁻¹ h⁻¹) is the rate constant for the PSO model.

Similarly, the equilibrium adsorption data were fitted with the Langmuir and Freundlich isotherm models. The Langmuir isotherm [43] assumes monolayer adsorption onto a homogeneous surface with finite identical sites, while the Freundlich isotherm [44] is an empirical equation assuming a heterogeneous surface with sites of varied affinities. These models are mathematically expressed by Equations (6) and (7), respectively:

$${\mathrm{q}}_{\mathrm{e}}={\displaystyle \frac{{\mathrm{q}}_{\mathrm{m}}{\mathrm{K}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{e}}}{1+{\mathrm{K}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{e}}}}$$

(6)

$${\mathrm{q}}_{\mathrm{e}}={\mathrm{K}}_{\mathrm{F}}{\mathrm{C}}_{\mathrm{e}}^{{\displaystyle \frac{1}{\mathrm{n}}}}$$

(7)

where qe (mg/g) is the amount of Hg²⁺ adsorbed at equilibrium, qm (mg/g) is the maximum adsorption capacity, K_L (L/mg) is the Langmuir constant related to adsorption affinity, C_e (mg/L) is the equilibrium concentration of Hg²⁺, and K_F ((mg/g)/(mg/L)ⁿ) and n are Freundlich constants representing adsorption capacity and intensity, respectively.

2.6. Desorption and Reuse Studies

In each desorption cycle, 75 mg of Hg²⁺-loaded ABK–TiO₂–APTES was placed into a 100 mL Erlenmeyer flask containing 50 mL of desorbing solution (0.1-M EDTA, KI, and HCl). The flask was sealed with paraffin and agitated at 150 rpm for 2 h. After shaking, the solid and liquid phases were separated through centrifugation at 4000 rpm, rinsed with DI water, and oven-dried at 70 °C overnight. The dried adsorbents were reused directly for each cycle. The percentage of Hg²⁺ desorbed in each cycle was calculated using Equation (8) as follows:

$$\mathrm{D}\mathrm{e}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{b}\mathrm{e}\mathrm{d}={\displaystyle \frac{{\mathrm{M}}_{\mathrm{d}}}{{\mathrm{M}}_{\mathrm{a}}}}\hspace{0.17em}\times \hspace{0.17em}100\%$$

(8)

where M_d (mg/g) is the amount of desorbed Hg²⁺, and M_a (mg/g) is the amount of adsorbed Hg²⁺.

2.7. Antibacterial Activity Assay

Antibacterial activity test was conducted following the method outlined by Mergenbayeva et al., (2024) [45]. Here, a micro-dilution technique was used to test the antibacterial activity of RK, ABK, ABK–TiO₂, and ABK–TiO₂–APTES through the determination of their minimum inhibitory concentration (MIC) against E. coli (BL21) and S. aureus. To prepare the bacteria for inoculation, a single colony of each bacterium was chosen from an agar plate stock and cultivated in 10 mL of nutritional broth for 16 h at 37 °C while being continuously shaken at 250 rpm. E. coli (BL21) was cultured in Luria–Bertani broth, and S. aureus in Tryptic Soy Broth. Fifty microliters of the diluted bacterial suspension and 50 μL of culture broth with twice the quantity of microparticles needed were then mixed in each well of a sterile 96-well microplate to form a final volume of 100 μL. VICTOR Nivo S multiplate reader was used to monitor the bacterial growth in each well through absorbance reading while the plate was incubated at 37 °C and shaken at 600 rpm. The optical density at 600 nm was recorded every 15 min over a period of 24 h. The results were subsequently standardized and plotted from 1 h after the initial recording to ensure that maximum mixing and uniformity in absorbance had been achieved.

3. Results and Discussion

3.1. Characterization

3.1.1. Chemical Composition

XRF results (

Table 1. Chemical composition of raw and modified adsorbent.

Sample	Content (wt%)
	SiO2	Al2O3	Fe2O3	TiO2	Na2O	K2O	CaO	SiO2/Al2O3
RK	49.56	37.58	0.10	0.70	-	1.90	0.12	1.32
ABK	56.06	27.78	0.81	0.84	3.44	1.82	0.08	2.02
ABK–TiO2	47.78	23.05	0.86	23.92	1.51	1.75	0.06	2.07
ABK–TiO2–APTES	49.81	22.43	0.74	22.88	1.60	1.55	-	2.22

) revealed clear compositional trends corresponding to each modification step. RK is rich in alumina and silica, but after acid–base treatment, the Al₂O₃ content dropped from ~37.58 wt% to 27.78 wt%, whereas the SiO₂ content increased from 49.56 wt% to 56.06 wt%. This change reflects the leaching of Al³⁺ from the octahedral layers of RK under harsh acid conditions [46], increasing the SiO₂/Al₂O₃ ratio from 1.32 to 2.02, thereby leaving a silica-enriched structure [47]. The presence of K₂O in all samples (1.55–1.90 wt%) can be attributed to the occurrence of minor accessory minerals such as illite, while the appearance of Na₂O after acid–base can be attributed to residual Na⁺ from the alkaline activation step. Upon TiO₂ incorporation to form ABK–TiO₂, the content of TiO₂ increased significantly from 0.84 to 23.92 wt%, directly confirming its successful introduction into the kaolinite matrix [37]. For ABK–TiO₂–APTES, TiO₂ remained high at 22.88 wt%, indicating that the TiO₂ phase is retained after silane functionalization. Al₂O₃ dropped further to 22.43 wt%, while SiO₂ increased to 49.81 wt%. The slight increase in SiO₂ from ABK–TiO₂ to ABK–TiO₂–APTES likely reflects additional silicon introduced by the APTES molecules grafted onto the kaolin/TiO₂ surface.

3.1.2. Structural Characterization

FTIR spectra (Figure 2a) revealed that RK exhibits the characteristic O–H stretching bands of inner-surface and inner hydroxyl groups at ∼3695 and 3620 cm⁻¹, respectively, and the bending mode of molecular water at ∼1638 cm⁻¹ [42]. In the low wavenumber region, the peaks at 1110 and 991 cm⁻¹ are attributed to Si–O stretching in quartz, while the peak at 913 cm⁻¹ corresponds to inner Al–OH bending [47,48]. Additionally, the peak at 773 cm⁻¹ is associated with the bending vibration of Al–O–Si bonds, whereas the peak at 540 cm⁻¹ is attributed to Si–O/Al–O deformation. In ABK, the intensity of the OH bands (~3600–3700 cm⁻¹) is greatly weakened, indicating a loss of structural –OH groups due to thermal treatment and leaching of Al–OH during acid–base treatment [46]. ABK–TiO₂ and ABK–TiO₂–APTES samples showed an even further reduction in the –OH band intensity, suggesting that the thermal and chemical conditions of TiO₂ incorporation (e.g., hydrolysis of Ti precursors and calcination) consumed or replaced additional –OH groups [49]. Meanwhile, the primary Si–O–Si framework vibrations of kaolin (stretching around ∼1000 cm⁻¹) remain present in all samples, indicating that the silicate lattice largely persists through the treatments [42,50].

XRD patterns (Figure 2b) showed that RK exhibited distinct crystalline peaks corresponding to its mineral constituents. These peaks include strong reflections from kaolinite at 2θ ≈ 12.3° (001) and 25° (111). After acid–base treatment, these peaks drastically reduced in intensity [29]. This decrease in diffraction peaks signals that ABK has an amorphous structure [47]. For nano-synthesized TiO₂, a notable peak appeared at 2θ ≈ 25.3°, corresponding to the (101) reflection of anatase TiO₂ [49]. Additionally, smaller anatase peaks can be identified at 2θ values of ∼37.8°, 48.0°, 53.9°, and 62.5°. These matches standard JCPDS data for anatase TiO₂ (JCPDS 00-021-1272). With TiO₂ incorporation (ABK–TiO₂), new crystalline features emerged in the XRD patterns, with the anatase TiO₂ peaks overlapping partly with the broad background. For ABK–TiO₂–APTES, the XRD pattern remained largely similar to ABK–TiO₂ in terms of phase composition. The anatase TiO₂ peaks (e.g., at ~25°) are still observable, showing that the TiO₂ phase persists through the silanization step. Thus, XRD aligns with the other analyses: ABK is largely amorphous dealuminated silica, while ABK–TiO₂ and ABK–TiO₂–APTES contain new anatase TiO₂ crystallites embedded in the amorphous matrix.

3.1.3. Morphological Characterization

SEM images (Figure 3a) show that RK consists of micrometer-sized, plate-like particles with relatively smooth faces and tightly stacked layers. After acid–base treatment, surface etching occurred, creating meso- to macro-porosity and rougher textures in ABK (Figure 3b). The overall architecture is a loosely layered structure, as acid attack partially delaminated the stacks and base washing removed residues, thereby yielding a more open microstructure [51]. Upon TiO₂ incorporation to form ABK–TiO₂ (Figure 3c), TiO₂ NPs dispersed as granules are observed on the kaolinite surface. ABK–TiO₂–APTES (Figure 3d) exhibits more edge-to-face contacts and random orientations of plates, as APTES grafting disrupts the close face-to-face packing of the plates. TEM image for ABK (Figure 3e) further reveals a partially exfoliated, sheet-like structure with irregular edges and increased porosity. Nano-sized clusters around the kaolinite plate were observed in ABK–TiO₂ (Figure 3f), appearing as a dark, scattered region due to the higher atomic number of Ti. In ABK–TiO₂–APTES (Figure 3g), the APTES layer was not directly visible, but its presence could be inferred from the perfect anchoring of TiO₂ NPs on the kaolinite surface.

3.1.4. Textural Properties

RK exhibited a relatively low BET surface area and pore volume (17 m²/g and 0.063 cm³/g, respectively) due to its platy morphology and limited internal porosity. However, the BET surface area and pore volume increased significantly (94 m²/g and 0.238 cm³/g, respectively) for ABK. This is likely due to the formation of new mesopores created by Al³⁺ leaching and structural collapse. TiO₂ incorporation further modified the textural properties of ABK–TiO₂ by slightly increasing the overall surface area (95 m²/g) and pore volume (0.243 cm³/g). After APTES grafting, the BET surface area and pore volume of ABK–TiO₂–APTES were reduced to 68 m²/g and 0.203 cm³/g, respectively. This reduction is due to –NH₂ groups blocking the pores of the material. The porosity type for RK, ABK, ABK–TiO₂, and ABK–TiO₂–APTES exhibits a Type IV nitrogen isotherm with an H3 hysteresis loop (Figure 4a), indicating mesopore formation [37].

Table 2. BET surface characteristics of raw and modified adsorbent.

	RK	ABK	ABK–TiO2	ABK–TiO2–APTES
Surface area (m2/g)	17	94	95	68
Pore volume (cm3/g)	0.063	0.238	0.243	0.203

summarizes the BET surface characteristics.

Figure 4b shows zeta potential measurements of ABK, ABK–TiO₂, and ABK–TiO₂–APTES as a function of pH. The surface charge of ABK (−14.54 mV to −43.52 mV) and ABK–TiO₂ (−12.34 mV to −33.01 mV) were all negative across pH 3 to 11. This is due to the leaching of Al–OH groups in these materials, resulting in good colloidal stability that is favorable for binding Hg²⁺ via Coulombic attraction. For ABK–TiO₂–APTES, positive surface charges were observed at pH 3 (20.6 mV) and pH 5 (15.38 mV). This is because the –NH₂ groups on ABK–TiO₂–APTES are readily protonated to –NH₃⁺, imparting a positive surface charge [47]. However, at pH 7, 9, and 11, surface charges were between −21.56 mV and −39.58 mV. Note that while a negatively charged surface generally enhances Hg²⁺ adsorption through electrostatic attraction, mercury uptake is not solely governed by surface charge. Even under conditions where the surface is positively charged, Hg²⁺ ions can still interact strongly with functional groups such as –NH₂ and –OH through coordination or complexation mechanisms [52].

3.2. Adsorption Studies

3.2.1. Contact Time

Figure 5a presents the effect of contact time on Hg²⁺ uptake for ABK, ABK–TiO₂, and ABK–TiO₂–APTES. ABK–TiO₂–APTES showed the highest removal, reaching ~39% Hg²⁺ removal within 60 min and ~40% by 120 min, whereas ABK–TiO₂ achieved 26.3% in 60 min and plateaued at ~29% thereafter. By contrast, ABK attained only 7.3% by 60 min and dropped to 6.1% at 120 min. This two-stage adsorption (initial rapid surface binding followed by slower intraparticle diffusion) is typical for metal sorption on porous solids [53,54]. The much higher Hg²⁺ uptake by ABK–TiO₂–APTES highlights the effectiveness of surface functionalization. By comparison, Hg²⁺ uptake on ABK likely occurs only at silanol and aluminol sites via cation exchange, which is slower and limited [25]. ABK–TiO₂ intermediate performance suggests TiO₂ NPs add adsorptive hydroxyl sites and possibly increase surface area, thus improving performance over ABK [37]. All adsorbents achieved near-equilibrium by 90–120 min, so a 120 min contact time was chosen for subsequent experiments to ensure maximum Hg²⁺ removal.

3.2.2. Effect of Adsorbent Dosage

Hg²⁺ removal efficiency increased with adsorbent dosage for all three materials (Figure 5b), due to the greater availability of binding sites at higher loading [37]. At a low dose of 12.5 mg, ABK removed only ~1% of Hg²⁺, while ABK–TiO₂ and ABK–TiO₂–APTES achieved 27.2% and ~32% removal, respectively. Increasing the dose to 50 mg raised Hg²⁺ removal to 8.9% for ABK, 45.3% for ABK–TiO₂, and 51.2% for ABK–TiO₂–APTES. An optimal dosage was achieved at 75 mg, where ABK–TiO₂–APTES resulted in 70.4% removal, compared to 57.3% for ABK–TiO₂ and 11.6% for ABK. Beyond this point, the removal percentage declined for all materials. This is likely due to possible adsorbent agglomeration, thereby reducing the effective surface area [55]. The incorporation of TiO₂ and –NH₂ functional groups onto the ABK surface increases the number of active sites and enhances the binding affinity toward Hg²⁺ ions, thereby reducing the amount of adsorbent required to achieve effective removal. These results agree with previous reports that functionalizing clays with inorganic oxides or organic ligands greatly improves their adsorptive efficiency [35,37].

3.2.3. Effect of Initial Mercury Concentration

Equilibrium uptake rose with increasing initial Hg²⁺ concentrations (Figure 5c) for all adsorbents, approaching saturation at higher concentrations. For example, at 5 mg/L of Hg²⁺, the adsorption capacity of ABK was 1.3 mg/g, whereas ABK–TiO₂ reached 3.1 mg/g and ABK–TiO₂–APTES 3.6 mg/g. At 50 mg/L Hg²⁺, adsorption capacities were significantly higher: ABK 5.3 mg/g, ABK–TiO₂ 14.1 mg/g, and ABK–TiO₂–APTES 19.2 mg/g. Finally, at the optimal concentration (70 mg/L), ABK–TiO₂–APTES achieved an adsorption capacity of 25.6 mg/g, compared to 17.7 mg/g for ABK–TiO₂ and 5.8 mg/g for ABK. These trends indicate that a greater driving force (Hg²⁺ concentration) increases the amount of uptake. ABK–TiO₂–APTES clearly exhibits superior performance with a capacity roughly five times that of ABK. This aligns with the literature, which shows that –NH₂-functionalized adsorbents have substantially higher Hg²⁺ capacities due to specific metal–ligand complexation [54,56]. ABK–TiO₂ also achieved a capacity thrice that of ABK, likely due to the presence of additional hydroxyl binding sites and increased surface area.

3.2.4. Effect of Solution pH

Solution pH substantially affects Hg²⁺ adsorption by influencing both metal speciation and adsorbent surface charge. Figure 5d shows that ABK had 2.8% removal at pH 3 but increased to 13.7% at pH 5 (optimum). This is due to the deprotonation of kaolinite sites and reduced competition from H⁺ [54,56]. At pH 7, the performance of ABK dropped to 10.3% and remained low at ~9% for pH 9 and 11. ABK–TiO₂ exhibited an optimal pH of around 7, with removal increasing from 14.3% at pH 3 to ~35% at pH 5 and 40.5% at pH 7. Subsequently, removal decreased to 37.5% at pH 9 and 30.1% at pH 11. The TiO₂-loaded adsorbent thus favors neutral pH, consistent with the point of zero charge of TiO₂, which is between 6 and 7. ABK–TiO₂–APTES exhibited optimal Hg²⁺ removal at pH 5, achieving a removal efficiency of 54.9%, despite its positively charged surface as indicated by zeta potential analysis. This performance is likely attributed to the coordination between surface –NH₂ groups and Hg²⁺ ions, facilitating inner-sphere complexation that overcomes electrostatic repulsion [57]. At pH 3, Hg²⁺ removal was 33.3%, reaching 54% at pH 7, and then decreasing to ~37% at pH 9 and 11. At pH 3, the –NH₂ groups are protonated to –NH₃⁺, while kaolinite surface sites are neutralized by H⁺, thereby suppressing electrostatic attraction [47]. By contrast, at pH 7 to 11, the formation of hydroxo-complexes such as Hg(OH)₂, Hg(OH)₃⁻, Hg₂(OH)₂²⁺ limits the available Hg²⁺ in solution [56]. These species exhibit very low solubility in aqueous media, and as the hydroxide ion concentration increases, they readily condense or aggregate to form colloidal and precipitated mercury hydroxide phases. This process effectively decreases the concentration of free Hg²⁺ ions available in solution for adsorption, thereby reducing the apparent adsorption capacity of the adsorbent. Overall, the pH-dependent trends agree with literature for Hg²⁺ on clays: minimal uptake at pH <4 due to proton competition, a maximum in mildly acidic to neutral conditions, and interference from metal hydroxide precipitation at high pH [54]. Therefore, subsequent experiments were conducted using ABK–TiO₂–APTES at pH 5, with an optimized dosage of 75 mg and an initial Hg²⁺ concentration of 70 mg/L to ensure maximum adsorption performance under representative conditions.

3.3. Effect of Water Matrix Components on Adsorption Efficiency

3.3.1. Effect of Natural Organic Matter

Figure 6a shows the effect of HA and glucose during Hg²⁺ uptake on ABK–TiO₂–APTES. HA caused a slight overall reduction in Hg²⁺ removal, whereas glucose had no marked effect. In the presence of 1 mg/L of HA, 52.3% of Hg²⁺ removal was observed. With 5 mg/L of HA, Hg²⁺ removal was 55.1%, but dropped to 43.9% at 10 mg/L. Higher HA concentrations of 25 and 50 mg/L resulted in Hg²⁺ removal of 47.8% and 49.1%, respectively. By contrast, with 10–50 mg/L glucose, Hg²⁺ removal remained between 51% and 53% with no notable decline. These results indicate that humic substances can moderately interfere with Hg²⁺ uptake, whereas an inert organic carbon source such as glucose does not. The transient ~10% decrease in removal at 10 mg/L of HA suggests competitive binding. This can be due to the abundance of functional groups (carboxyl, carbonyl, hydroxyl, thiol, and amino) present in HA, which form strong complexes with Hg²⁺ in solution, thereby preventing it from adsorbing onto ABK–TiO₂–APTES. These functional groups (oxygen- and sulfur-containing moieties) readily form stable Hg–O and Hg–S complexes, thereby reducing the availability of free Hg²⁺ ions for adsorption onto ABK–TiO₂–APTES. In addition, the adsorption of HA onto the adsorbent surface may partially block active sites, further hindering mercury uptake [56]. This observation highlights the importance of considering NOM competition in practical water treatment systems when using functionalized clay-based adsorbents such as ABK–TiO₂–APTES.

3.3.2. Effect of Ionic Strength

Figure 6b shows that a significant decrease in Hg²⁺ removal was observed, from 51.7% at 0.005 M NaNO₃ to 28.7% at 0.1 M. This trend indicates that higher ionic strength decreases Hg²⁺ uptake. Increased ionic strength generally screens the electrostatic attractions between charged adsorbent sites and metal ions, and cations like Na⁺ can compete for exchange sites on the kaolinite surface [56]. In the case of ABK–TiO₂–APTES, the decline in performance suggests that a portion of Hg²⁺ binding involves electrostatic outer-sphere interactions that are weakened by a compressed electric double layer [58]. However, a substantial fraction of Hg²⁺ was still adsorbed, even at an ionic strength of 0.1 M. This resilience is due to the chemical binding of Hg²⁺ on ABK–TiO₂–APTES. Previous studies on Hg²⁺–clay systems have indeed found that ionic strength decreases mercury uptake. For instance, Huang et al., (2019) found that NaNO₃, with a concentration ranging from 0 to 0.1 M, decreased Hg²⁺ uptake from 88.7% to 79.3% and CH₃Hg⁺ uptake from 98.8% to 86.0% [56]. This finding highlights the sensitivity of mercury species to competitive ionic environments, where increased background electrolyte concentrations can hinder adsorption efficiency.

3.3.3. Effect of Coexisting Heavy Metals

The presence of Cu²⁺, Zn²⁺, Ni²⁺, and Cd²⁺ suppressed Hg²⁺ uptake (Figure 6c). In the control (no co-ions), Hg²⁺ removal was ~55%. With 5 mg/L of Cu²⁺, Zn²⁺, Ni²⁺, and Cd²⁺ each (20 mg/L in total), Hg²⁺ removal fell to ~52%, whereas at 10 mg/L each, it fell to ~40%. Meanwhile, Cu²⁺ removal was ~46% (dropping to ~31% at higher loading), Zn²⁺ removal was low (~9% rising to ~22%), and Ni²⁺/Cd²⁺ uptake remained negligible (<5%). These results reflect competitive adsorption, where additional cations occupy binding sites and reduce Hg²⁺ binding, yet Hg²⁺ remains the predominant adsorbed ion. The fact that Hg²⁺ retention remains relatively robust indicates that ABK–TiO₂–APTES binding sites preferentially chelate soft, low-hydration metals. Huang et al., (2019) demonstrated that the thiol-functionalized graphene oxide/Fe–Mn composite exhibited similar trends where metals with strong water coordination (e.g., Ni²⁺ and Zn²⁺) showed minimal adsorption, while softer metals dominated uptake [56].

3.4. Adsorption Kinetics and Isotherms

Figure 7a compares the PFO and PSO kinetic models for Hg²⁺ adsorption onto ABK, ABK–TiO₂, and ABK–TiO₂–APTES, with their corresponding fitting parameters presented in

Table 3. Adsorption kinetics and isotherm fitting parameters.

Model	Parameter	ABK	ABK–TiO2	ABK–TiO2–APTES
Pseudo-first order	K1	0.031	0.079	0.054
	qm	6.2	15.1	22.2
	R2	0.981	0.987	0.985
	SSE	0.762	2.803	7.768
Pseudo-second order	K2	0.003	0.006	0.002
	qm	8.1	17.1	26.2
	R2	0.978	0.999	0.996
	SSE	0.858	0.315	1.923
Langmuir	qm	8.4	18.0	28.4
	R2	0.974	0.968	0.961
	KL	0.032	0.134	0.105
	SSE	0.560	5.261	15.395
Freundlich	n	1.902	2.803	2.400
	R2	0.960	0.895	0.920
	KF	0.632	3.942	4.953
	SSE	0.856	17.681	32.395

. For ABK–TiO₂ and ABK–TiO₂–APTES, the PSO model yielded an excellent fit (R² ≈ 0.996–0.999) and much lower sum of squared residuals (SSR) than the PFO model. A similar study on Hg²⁺ uptake using functionalized kaolin reported R² values of 0.966 and 0.983 for PFO and PSO models, respectively [26]. Similarly, Moussout et al., (2018) observed that PSO offers superior correlation (R² ≈ 1), but recommended using both R² and error metrics to judge kinetic fits [59]. The PSO fit for both ABK–TiO₂ and ABK–TiO₂–APTES yields not only a higher R² but also a significantly lower SSE (0.315 and 1.923, respectively), indicating a more accurate representation of the uptake rate. The superior fit to the PSO model suggests that chemisorption is the governing mechanism, as this model assumes the rate-limiting step involves valence forces between the adsorbate and the adsorbent [42]. By contrast, ABK exhibited much lower capacity and a more gradual approach to equilibrium. Here, PFO and PSO fits are more similar, with both models yielding an R² of 0.981 and 0.979, respectively. This suggests that the physisorption process dominates Hg²⁺ uptake, indicating that ABK likely relies on weaker interactions.

Figure 7b shows the adsorption isotherms for ABK, ABK–TiO₂, and ABK–TiO₂–APTES, with the corresponding fitting parameters detailed in Table 3. The Langmuir model provided excellent fits for ABK, ABK–TiO₂, and ABK–TiO₂–APTES with R² values of 0.974, 0.968, and 0.961, respectively. This behavior aligns with findings for other functionalized adsorbents (e.g., biochar) where Langmuir isotherms dominate, indicating monolayer adsorption on homogeneous sites [60]. The progression of Langmuir monolayer capacity values from 8.4 mg/g for ABK to 18 mg/g for ABK–TiO₂, and then to 28.4 mg/g for ABK–TiO₂–APTES, indicates a progressive enhancement of adsorption capacity with each successive surface modification.

3.5. Adsorption Mechanism

The performance of ABK–TiO₂ and ABK–TiO₂–APTES in Hg²⁺ removal can be attributed to multiple complementary mechanisms at the surface. Electrostatic attraction is one contributing factor: at an initial solution pH of around 5 to 7, where ABK–TiO₂ and ABK–TiO₂–APTES surfaces acquire a negative charge, Hg²⁺ cations are drawn to the surface by Coulombic forces [57]. This initial attraction can increase local Hg²⁺ concentration near the interface, facilitating further binding. However, pure electrostatic adsorption (outer-sphere) is relatively weak and not solely responsible for the uptake observed. More importantly, ligand exchange and surface complexation play a dominant role in the adsorption process. TiO₂ and Si–OH groups on ABK–TiO₂ binds Hg²⁺ through an inner-sphere mechanism: Hg²⁺ coordinates directly to oxygen atoms by displacing a proton or H₂O ligand, essentially exchanging ligands and forming a surface–Hg bond. This ligand exchange is a classic route for heavy-metal adsorption on oxides and creates a surface complex anchored to the kaolinite/TiO₂ framework [61]. In ABK–TiO₂–APTES, the presence of grafted –NH₂ groups introduces additional strong binding sites. These –NH₂ coordinates to Hg²⁺ by donating the lone pair of electrons from nitrogen, forming Hg–N [57]. Such chemical bonding with N- and O-ligands firmly immobilizes the Hg²⁺ on the surface.

XPS analysis (Figure 8a–d) was conducted before and after Hg²⁺ adsorption for ABK–TiO₂ and ABK–TiO₂–APTES to provide direct evidence of inner-sphere interactions. After Hg²⁺ adsorption, new peaks appeared in the high-resolution XPS corresponding to Hg 4f. Specifically, Hg 4f_7/2 and 4f_5/2 doublet peaks were observed at binding energies of 100.2 and 104.3 eV for ABK–TiO₂ and 100.9 and 104.8 eV for ABK–TiO₂–APTES with a spin–orbit separation of ∼4.0 eV [54]. This spectral signature is characteristic of Hg²⁺ species chemisorbed onto ligating atoms (such as O or N) on a surface [62]. This rules out the presence of metallic Hg⁰ (which would appear at a lower binding energy, ~99 eV) and confirms that Hg remains in the oxidized state, likely bound as surface complexes (–O–Hg²⁺ and –N–Hg²⁺). Notably, the Si 2p region of XPS in ABK–TiO₂ and ABK–TiO₂–APTES showed a shift toward lower binding energy upon Hg²⁺ adsorption. This is due to charge redistribution within the Si–O–Si/Ti–O–Si network, indicating that the Si–O bonds in the kaolin matrix remained chemically stable and were not directly involved in Hg²⁺ coordination. These observations support a mechanism where Hg²⁺ is chemisorbed via specific coordination to surface functional groups.

3.6. Desorption and Reuse Studies

Figure 9a,b show the desorption and reuse studies, respectively, for ABK–TiO₂–APTES using different eluents (HCl, KI, and EDTA). Desorption efficiencies declined across three cycles for all eluents, with HCl yielding the highest (~95%) in the first cycle, followed by KI (~90%) and EDTA (~88%). By Cycle 3, desorption levels dropped to ~60% (HCl), 67% (KI), and 54% (EDTA). Similarly, the adsorption capacity of the desorbed ABK–TiO₂–APTES after reuse decreased (~21 mg/g to 13 mg/g for HCl, ~22 mg/g to 16 mg/g for KI, and ~20 mg/g to 11 mg/g for EDTA). The stronger regeneration effect of HCl is consistent with other Hg²⁺ adsorbents, where 0.1-M HCl recovered most adsorbed Hg²⁺ [47]. By Cycle 3, the performance of KI is comparable to HCl, reflecting the high affinity of iodide for Hg²⁺, whereas the lower efficiency of EDTA implies incomplete chelation. Overall, only ~50%–70% of the initial capacity was retained after three cycles, indicating that a significant fraction of Hg²⁺ is bound irreversibly. This behavior is consistent with a chemisorptive mechanism via Hg–N and Hg–O coordination [63].

Table 4. Summary of studies on the use of kaolin for Hg²⁺ adsorption.

Adsorbent	Size	Initial Concentration (mg/L)	Dosage (g/L)	Adsorption Capacity (mg/g)	Removal (%)	Study
Natural Kaolin	<100 µm	300	0.5	10.1–10.9	-	[25]
3-mercaptopropyltrimethoxysilane modified kaolin	≤125 µm	30.83	0.1	30.1	98.0	[64]
Polypyrrole-functionalized Fe3O4/kaolin	-	50	0.05	317.7	-	[26]
Metakaolin-based geopolymer	150 µm	50	0.05	38.0	65.1	[65]
L-cysteine and polypyrrole-functionalized magnetic kaolin	-	40	0.05	482.7	-	[27]
TiO2-loaded APTES grafted kaolin	≤125 µm	70	1.5	25.6	54.5	This study

compares this study with previous ones on the removal of Hg²⁺ using kaolin and modified kaolin adsorbents.

3.7. Antibacterial Activity

Figure 10a shows that RK exhibited antibacterial activity against E. coli (BL21) with an MIC of 10 mg/mL. Both ABK (Figure 10b) and ABK–TiO₂ (Figure 10c) exhibited enhanced antibacterial activity with an MIC of 5 mg/mL, indicating that the antibacterial ability of RK was improved by acid–base treatment. Despite the known antibacterial properties of TiO₂, ABK–TiO₂–APTES (Figure 10d) did not exhibit enhanced antibacterial activity against E. coli (BL21), as evidenced by the unchanged MIC compared to RK.

For S. aureus, RK (Figure 11a) and ABK (Figure 11b) exhibited antibacterial activity with MIC = 5 mg/mL, whereas ABK–TiO₂ (Figure 11c) showed the greatest antibacterial efficacy with an MIC of 2.5 mg/mL. Notably, ABK–TiO₂–APTES (Figure 11d) again exhibited the weakest antibacterial activity (MIC = 10 mg/mL). The greater susceptibility of S. aureus to ABK–TiO₂ may be due to stronger interactions between the positively charged TiO₂ surface and the Gram-positive bacterial membrane, which may be mediated by increased ROS generation.

Previous studies have shown that ultraviolet-activated TiO₂ surfaces inactivate S. aureus more rapidly than E. coli (BL21), which aligns with the observed pattern [66]. This is most likely due to the absence of an outer membrane in S. aureus, which typically restricts ROS penetration in Gram-negative organisms [67,68]. According to Tang et al., (2019), TiO₂ NPs generate photocatalytic ROS that damage intracellular proteins, DNA, and bacterial membranes, resulting in antibacterial actions [69]. While NH₂ functional groups can confer antibacterial properties to APTES, the reduced efficacy of ABK–TiO₂–APTES suggests a more complex relationship. APTES is commonly used as a surface modifier or linker rather than an active biocidal agent. It primarily alters the zeta potential, bandgap characteristics, and surface chemistry of TiO₂ NPs through the formation of Ti–O–Si bonds [70]. Functionalization may result in a multilayer surface coating that covers active sites and thus reduces the efficiency of ROS-mediated bacterial killing, even as it increases susceptibility to visible light. Rokicka-Konieczna et al., (2023) found that increasing the APTES concentration from 300 to 400 mM decreased antibacterial activity due to a masking effect and a reduction in surface area and pore volume, despite a higher surface charge [41].

Table 5. Minimum inhibitory concentration for RK, ABK, ABK–TiO₂, and ABK–TiO₂–APTES.

Bacteria	RK (mg/mL)	ABK (mg/mL)	ABK–TiO2 (mg/mL)	ABK–TiO2–APTES (mg/mL)
E. coli (BL21)	10	5	5	10
S. aureus	5	5	2.5	10

summarizes the MIC values for RK, ABK, ABK–TiO₂, and ABK–TiO₂–APTES.

4. Conclusions

In this study, sequential acid–base treatment, TiO₂ NP incorporation, and APTES grafting of kaolin successfully yielded a dual-functional adsorbent. The modified kaolin (ABK–TiO₂ and ABK–TiO₂–APTES) demonstrated a Hg²⁺ adsorption capacity (17.7 and 25.6 mg/g, respectively) approximately three and five times higher than that of acid–base-treated kaolin (ABK; 5.8 mg/g). Among the tested materials, ABK–TiO₂–APTES was the most effective and was therefore evaluated under challenging water matrices. ABK–TiO₂–APTES maintained its Hg²⁺ removal efficiency across a wide range of conditions, including variations in ionic strength, NOM, and the presence of competing heavy metal ions (Cu²⁺, Zn²⁺, Cd²⁺, and Ni²⁺). Isotherm modeling and XPS analysis confirmed that Hg²⁺ uptake occurs primarily through chemisorption via inner-sphere complexation to the TiO₂ hydroxyls and grafted –NH₂ groups. ABK–TiO₂–APTES maintained stable performance across multiple regeneration cycles, demonstrating its reusability. For antibacterial activity, ABK–TiO₂ inhibited the growth of S. aureus (MIC of 2.5 mg/mL), while showing a moderate effect against E. coli (BL21). However, ABK–TiO₂–APTES exhibited a slight reduction in bacterial inhibition, suggesting a trade-off between maximizing adsorption capacity and preserving antimicrobial sites. This multifunctional adsorbent presents a promising solution for the simultaneous removal of toxic metals and microbial pathogens, aligning with SDG 6.