Innovative utilization of quartz: creating nanosilica adsorbents for dye removal and norfloxacin sensing applications

Abstract

This article describes the synthesis of silica nanoparticles from a quartz-based mineral, which was sequentially subjected to thermal and acid treatments to remove impurities, followed by alkaline treatment to obtain a sodium silicate solution. As a result of precipitation of sodium silicate with nitric acid, followed by washing and drying, silica nanoparticles were obtained. These nanoparticles exhibited spherical morphology, amorphous nature, and chemical purity, as confirmed by SEM, XRD, and chemical-elemental analysis. Through the modification of nanosized silica using N-containing silanes, a series of monofunctional materials containing amino-, diamino-, and secondary amino groups, as well as a bifunctional material with amino and phenyl groups, were synthesised. It was established that the synthesised particles exhibit nanoscale dimensions and surface amino groups, indicating their potential for efficient removal of pollutants from aqueous solutions. The adsorption properties of the obtained functional materials were investigated concerning the fluorescent textile dye disperse yellow 82, which may be present in wastewater from textile industries. It was demonstrated that the dye-loaded materials can serve as potential sensors for the fluorescent detection of norfloxacin in aqueous solutions.

---

Introduction

The removal of dyes from textile wastewater remains an urgent environmental problem year after year.^1,2 This is primarily driven by the global increase in clothing and textile production, stimulated by the spread of the “fast fashion” phenomenon. In recent years, tens of thousands of tons of inexpensive polyester-based garments have been manufactured worldwide,³ typically dyed using disperse dyes. This class of water-insoluble dyes is applied to polyester either from organic solvents or under high-temperature conditions.⁴ However, following the dyeing process, disperse dyes are not fully exhausted in the dye baths, and a significant portion is released into wastewater. This requires mandatory treatment, since disperse dyes – together with formaldehyde, heavy metals, benzene, and phthalates – are classified as substances toxic to reproductive functions (CMR substances).⁵

Dyes can generally be removed from aqueous environments using two main approaches: photocatalytic degradation and adsorption. While photocatalysis is effective, it often leads to the formation of intermediate or final products with unknown composition and potential toxicity, raising concerns about secondary pollution.^6,7 Adsorption, on the other hand, is a cleaner and more controllable method, particularly when both the adsorbent and the target dye are well characterized.

The sorption of a fluorescent dye by synthesised silica-based adsorbents containing amino groups is of interest not only for water purification, but also for the subsequent use of the dye-loaded materials without regeneration, particularly as sensors for the detection of pharmaceutical compounds in aqueous solutions. Adsorbed dyes immobilized on silica materials can act as efficient sensors for various molecules due to their direct photoluminescence, the intensity of which increases in the presence of certain analytes.^8–10 From a practical perspective, the development of silica-based materials with adsorbed fluorescent dye is both promising and relevant, as such materials not only retain fluorescence properties but are also representative of pollutants found in textile wastewater. For detection, norfloxacin – a broad-spectrum antibiotic widely used in medicine and veterinary practice – was selected. However, its occurrence in aquatic environments, food products, and pharmaceuticals may have significant consequences, making its detection highly relevant for multiple fields.¹¹

In our previous works, we investigated the design of nanosized silica-based adsorbents functionalised with N-containing silanes for the removal of heavy metal ions,¹² rare-earth elements,¹³ and synthetic organic dyes^14,15 from aqueous solutions. In those studies, silica was synthesised from tetraethyl orthosilicate or bridged silanes via the sol–gel method. However, traditional synthesis routes often lead to environmental degradation and are economically costly. Therefore, integrating the principles of green chemistry into silica production is essential for improving sustainability and reducing environmental impact.

Agricultural residues, such as rice husk and bagasse, have attracted considerable attention as alternative sources of silica. According to Haider et al.,¹⁶ silica extracted from rice husk ash via alkaline treatment can contain up to 72.4% SiO₂. Additionally, other agricultural by-products, such as sugarcane bagasse ash, can also serve as silica sources.¹⁷ This approach not only enables waste valorization but also offers a sustainable alternative to conventional synthesis routes. Biotechnology has introduced innovative approaches, as certain algae and diatoms naturally biosynthesise silica, providing a renewable resource. For instance, Sidorowicz et al.¹⁸ discussed the potential of microalgae for silica production, demonstrating the feasibility of using such biogenic processes as environmentally friendly alternatives to traditional sources.

Despite the increasing popularity of innovative routes to obtain silica from agricultural waste and biogenic sources,¹⁹ mineral-based extraction methods remain dominant,^20,21 as they are the most widespread and economically viable. Quartz – the most common crystalline form of silica – continues to play a key role in industrial applications owing to its high purity and ordered crystalline structure, which are critical for glass, ceramics, and electronics production.²² Recent studies have shown that optimizing quartz extraction via improved acid leaching and thermal treatment can significantly enhance yield and purity while lowering energy costs and environmental impacts.²³ Combining conventional quartz-oriented methods with modern “green” technologies not only supports the economic viability of silica production but also paves the way toward more sustainable industrial practices.²⁴ Moreover, shale – a naturally abundant aluminosilicate – has been demonstrated to be convertible into high-purity amorphous silica through efficient Al₂O₃ removal via hydrothermal processing,²⁵ making it a promising alternative source.²⁶

In this work, we report the synthesis of silica nanoparticles from a quartz-containing mineral, their modification with N-containing silanes, and the investigation of their adsorption behavior toward the fluorescent dye disperse yellow 82 (Fig. S1). Furthermore, the obtained dye-loaded materials were evaluated as potential sensors for norfloxacin detection. The promising results highlight the significant potential of naturally derived silica for the development of highly functionalised materials and open new pathways for designing efficient adsorption-based sensors for environmental and pharmaceutical applications.

Results and discussion

Characterisation of the silica-containing mineral

The silica-containing mineral was obtained from a deposit in the form of approximately rectangular fragments with lengths ranging from 15 to 24 cm, widths of 8 to 12 cm, and a thickness of about 3 cm (Fig. 1a).

Fig. 1 The external appearance of the silica-containing mineral fragment (a) and its detail at 30× magnification (b).

The mineral is yellowish in color and exhibits a pencil- or rod-like structure resulting from regional metamorphism. Ochre stains are also visible on the surface of the sample.²⁷ Since the mineral was sourced from a deposit near Olevsk, located within the Sushchany-Perga fault zone in the western part of the Ukrainian Shield, it is most likely weathered pencil orthogneiss (or quartzite, quartz dyke). Feldspars are almost completely decomposed, leaving mainly quartz and kaolinite.

Examination of the mineral fragments revealed that hard, rod-like structures were interspersed with regions exhibiting a softer, clay-like texture (Fig. 1b). XRD analysis was performed to gain a deeper understanding of the structure and mineral composition of the silica-containing mineral. The XRD patterns of both the hard and soft regions indicate that they are characterised by a crystalline structure and have a similar compositions (Fig. 2). Quartz (Qz) is the dominant phase in both, with 2θ values of 26.632° for the hard part and 26.671° for the soft one. Quartz is accompanied by kaolinite (Kln), identified at 12.461° and 12.500° for the hard and soft region, respectively, with the soft region containing slightly more kaolinite. Given that the hard component predominates, whole mineral fragments with mechanically cleaned surfaces were used for subsequent crushing and leaching experiments.

Fig. 2 XRD patterns of the hard (a) and soft parts (b) of the silica-containing mineral.

Synthesis and characterisation of silica nanoparticles from silica-containing mineral

The silica-containing mineral, after being ground to a particle size of 100 μm and calcined, was washed with 2.5 M HCl to remove aluminium, iron, calcium, and other metal compounds (Fig. 3).

Fig. 3 SEM image (a) EDXS analysis (b) EDS layered image including all elements and elemental maps of O, Si, Al and Ti (c) of the silica-containing mineral.

The silica-rich material was obtained after filtration and washing to a neutral pH (Fig. 4). At this stage, the mass loss was ≈4.3%. A chemical-elemental analysis was performed for the silica-rich material at this stage. The results of the ICP-MS analysis showed that the SiO₂ content increased relative to the raw mineral, reaching 97.8%, due to a reduction in impurity content from 9.69% to 2.2% (Table 1).

Fig. 4 Scheme for the production of silica nanoparticles from the silica-containing mineral.

Table 1 Composition data for the raw mineral and the acid-treated sample

Sample	Composition, %
	SiO2	Al2O3	Fe2O3	K2O	MgO	CaO	Zn	Cu	Ni
Silica-containing mineral	90.320 ± 0.020	3.470 ± 0.010	1.960 ± 0.010	1.100 ± 0.010	0.340 ± 0.010	0.340 ± 0.010	0.033 ± 0.001	0.017 ± 0.001	0.016 ± 0.001
Silica-rich material	97.770 ± 0.060	0.280 ± 0.010	0.200 ± 0.010	0.170 ± 0.010	0.120 ± 0.010	0.130 ± 0.000	0.030 ± 0.001	0.010 ± 0.000	0.010 ± 0.000

---

Fig. 5a shows the XRD pattern of the silica-containing mineral after calcination and acid leaching (i.e., the silica-rich material). The main crystalline peak of quartz is clearly observed in the XRD pattern at 26.672°, indicating that the applied temperature of 680 °C and the acid leaching process did not affect its structure. However, the disappearance of kaolinite peaks indicates the removal of alumina and acid-soluble metal oxides, as supported by the chemical-elemental analysis (Table 1).

Fig. 5 XRD patterns of the silica-rich material (a) and silica particles (b).

To produce a sodium silicate solution, the silica-rich material was dissolved in NaOH at 100 °C with continuous stirring for 14 h, taking advantage of the mineral's high SiO₂ content. The silica production method was adapted from the study by Zulfiqar et al.²⁸ with some modifications. In that work, silica was synthesised from bentonite clay using an alkaline treatment method. Therefore, given these properties, the silica-rich material (quartz) was subjected to an alkaline leaching process for 14 h (Fig. 4). After filtration, the resulting sodium silicate solution, which initially had a pH of ≈13, was titrated with 7 M HNO₃ until a pH of ≈7 was reached to initiate gelation. The aged gel was centrifuged after 12 h and washed several times with ethanol and water to remove nitrate salts. After drying, the synthesised silica material was ground in an agate mortar. The yield of silica nanoparticles was 70%.

The X-ray diffraction pattern shown in Fig. 5b confirms the amorphous nature of the synthesised silica. This pattern typically does not exhibit the sharp diffraction peaks typical of crystalline materials but instead shows a single broad halo at 2θ = 22°, which is characteristic of amorphous silica. Moreover, according to the ICP-MS analysis, the SiO₂ content in the sample reaches 99.9 ± 0.0%, corroborating its high chemical purity.

The morphology and structure of the synthesised silica particles were analysed using SEM, SEM-EDXS, and nitrogen adsorption/desorption measurements (Fig. 6). The analysis revealed that the silica appears as aggregated spherical particles. The SEM image (Fig. 6a) shows that the average particle diameter is approximately 50 nm. The particle size distribution curve exhibits a distinct peak, indicating sample uniformity and suggesting the potential for dispersing the aggregated particles in water. EDXS analysis confirmed the absence of impurities, with the sample primarily consisting of Si and O (Fig. 6b).

Fig. 6 SEM image and particle size distribution (a), EDXS analysis (b) and nitrogen adsorption/desorption isotherm (c) of the prepared silica particles.

The surface area of the synthesised silica, calculated by BET method, was found to be 179 m² g⁻¹, with an average pore diameter of approximately 12.6 nm, indicating its mesoporous nature. BET analysis revealed that the silica exhibits a type II isotherm with a well-defined hysteresis loop (Fig. 6c), which is consistent with the IUPAC classification for nonporous or macroporous adsorbents.²⁹ The hysteresis loop corresponds to type H3, which typically suggests the presence of particle aggregates or agglomerates forming slit-shaped pores of heterogeneous sizes.³⁰

Thus, silica particles were successfully synthesised from a silica-containing mineral. The silica-rich material was obtained after thermal treatment and acid leaching and was subsequently treated with sodium hydroxide solution to produce sodium silicate. Silica particles were formed by precipitating silica from the sodium silicate solution using nitric acid (Fig. 4). The resulting silica particles are spherical, amorphous, and chemically pure, as confirmed by SEM, XRD, and chemical-elemental analysis, respectively.

The synthesised silica nanoparticles are of significant interest due to their high specific surface area, chemical inertness, and thermal stability, making them promising for a wide range of applications. While they can be effectively utilized in various fields in their pure form, targeted surface modification is essential for specific applications, such as the adsorption of metal ions and organic dyes³¹ and biomolecules.³² Surface functionalisation of silica significantly enhances its selectivity and increases its adsorption capacity, thereby expanding its potential for use in water purification, chromatography, and other technological processes. Consequently, further research has focused on developing efficient chemical modification methods for synthesised silica particles with N-containing functional groups to create highly effective adsorbents with tailored properties.

Synthesis and characterisation of functionalised silica nanoparticles

The silica particles were prepared from a silica-containing mineral and subsequently functionalised with various functional agents in anhydrous toluene. In toluene, the solubility of functional silanes is enhanced, preventing their premature hydrolysis, while its non-polar nature promotes the uniform grafting of amino-functional silanes onto the silica surface and reduces the likelihood of self-condensation. Additionally, the hydrophobic environment of toluene, compared to aqueous or alcoholic media, minimizes the risk of uncontrolled aggregation of silica nanoparticles, resulting in a material with consistent size and morphology. Moreover, unlike alcohol-based systems, toluene minimizes the formation of volatile by-products that could complicate purification steps. These factors make toluene a superior medium compared to polar solvents such as water or ethanol for the preparation of high-quality amino-functionalised silica nanoparticles.³³

In this work, silica nanoparticles functionalised with a single type of functional group (amino, diamino, or secondary amino) were prepared using APTES, TMPED, and BTMPA. Additionally, a bifunctional material was synthesised by combining APTES and PhTES to introduce additional phenyl groups onto the silica surface, as schematically illustrated in Fig. 7.

Fig. 7 The synthesis scheme of silica adsorbents with functional groups on the surface.

The resulting functionalised materials appear as white, powdery substances that are insoluble in water. The morphology of the synthesised materials was examined using SEM. By comparing the SEM images (Fig. 8), it can be concluded that the synthesised monofunctional samples SiO₂/NH₂, SiO₂/N–N, and SiO₂/NH consist of spherical particles that aggregate into loosely packed globules. Meanwhile, as evidenced by the SEM images and the absence of a distinct peak on the particle size distribution curve (Fig. 8a), the APTES-based SiO₂/NH₂ sample exhibits lower uniformity compared with the other samples.

Fig. 8 SEM images and particle size distribution data of functionalised silica nanoparticles: SiO₂/NH₂ (a); SiO₂/N–N (b); SiO₂/NH (c); SiO₂/NH₂/Ph (d).

The SiO₂/N–N and SiO₂/NH samples exhibit uniformly synthesised particles with an average size of 70–80 nm (Fig. 8b and c). Furthermore, the obtained data indicate that although the bifunctional sample SiO₂/NH₂/Ph has a uniform average particle size of 60–70 nm (Fig. 8d), its particles are characterised by denser packing due to the presence of hydrophobic phenyl fragments on the surface.

The porous structure of the obtained materials was evaluated by measuring nitrogen adsorption–desorption isotherms (Fig. 9). The physical characteristics of the synthesised functionalised materials, such as specific surface area (S_sp), total pore volume (V_tot.), and pore diameter (d_p), were determined using BET analysis (Table 2). According to the obtained data, the isotherms of the monofunctional particles SiO₂/NH₂ and the bifunctional one SiO₂/NH₂/Ph, synthesised using APTES and APTES/PhTES, respectively, correspond to type II according to the IUPAC classification.³⁴ It is known that type II isotherms are characteristic of non-porous or macroporous structures, indicating the occurrence of unrestricted mono-multilayer adsorption on the particle surface and a relatively low surface area of the samples. The isotherms of the SiO₂/N–N and SiO₂/NH samples exhibit an almost identical appearance; however, since monolayer saturation is completely absent, they can be classified as type III. In this case, the adsorbed molecules tend to cluster around the most favorable sites on the surface of the non-porous or macroporous solid.³⁵

Fig. 9 Low-temperature nitrogen adsorption/desorption isotherms of the functionalised silica nanoparticles: (a) SiO₂/NH₂; (b) SiO₂/N–N; (c) SiO₂/NH; (d) SiO₂/NH₂/Ph.

Table 2 Structural and adsorption characteristics of the synthesised samples

Sample	N, mass%	C, mass%	H, mass%	C f.gr., mmol g^−1, elem. analysis	pI	S BET, m^2 g^−1	V tot, cm^3 g^−1	d p, nm
SiO 2	—	—	—	—	3.77	179	0.974	12.6
SiO 2 /NH 2	3.09	8.33	2.64	2.14	9.96	122	0.742	12.5
SiO 2 /N–N	3.95	9.01	3.20	1.35	10.57	54	0.973	32.4
SiO 2 /NH	1.33	7.38	2.40	0.94	10.21	112	1.415	32.4
SiO 2 /NH 2 /Ph	5.29	13.77	3.76	3.58	10.35	63	0.754	12.4

---

The observed hysteresis loops of these isotherms were classified as type H3, indicative of distinct porous structures. Type H3 is typically observed in non-rigid aggregates of particles or their loose assemblies, which form slit-like pores with heterogeneous size and shape that remain partially unfilled by the adsorbate.^36,37

The initial silica exhibited a relatively high specific surface area of 179 m² g⁻¹. After modification, all samples showed a decrease in specific surface area, indicating the successful functionalisation of the materials. Since the types of isotherms clearly indicate that the samples are nonporous, the specific surface area values are even higher for samples with smaller particle sizes.

Among the synthesised materials, the nanospheres of the SiO₂/NH₂ sample based on APTES exhibited a relatively high S_BET of 122 m² g⁻¹ (Table 2), which can be attributed to the small particle sizes that form aggregates of varying density with slit-shaped pores. Conversely, the presence of phenyl fragments in the bifunctional sample SiO₂/NH₂/Ph resulted in a reduction of the specific surface area to 63 m² g⁻¹ without any change in the pore diameter. This result was expected, as the introduction of phenyl groups onto the silica surface adds centers that promote increased particle aggregation. The hydrophobic nature of these groups likely leads to a denser packing of the particles, thereby limiting the accessible surface area for gas penetration. The SiO₂/N–N sample, synthesised using TMPED, exhibited the lowest specific surface area of 54 m² g⁻¹. Furthermore, compared both to the initial silica and to the other synthesised materials, SiO₂/N–N and SiO₂/NH samples showed an increase in pore diameter, presumably due to the formation of larger voids and interparticle gaps. However, the SiO₂/NH sample based on BTMPA exhibited a higher specific surface area of 112 m² g⁻¹, even though SEM data indicate (Fig. 8) that their particle sizes are identical. Thus, the size of the functional group influences the particle size during modification, while the nature of the group affects their agglomeration due to the interactions between the surface groups of the resulting particles.

To confirm the successful modification of the silica nanoparticles, FTIR spectra were recorded (Fig. 10). In all spectra of the synthesised samples, asymmetric Si–O–Si vibrations at 1000–1200 cm⁻¹ and symmetric Si–O–Si stretching vibrations with a moderate intensity absorption band at 795 cm⁻¹ are observed, indicating the formation of a polysiloxane network. However, the presence of uncondensed silanol groups and adsorbed water is evidenced by the absorption bands at 3670 and 1630 cm⁻¹, respectively.

Fig. 10 Infrared spectra of silica nanoparticles and functionalised silica nanoparticles SiO₂/NH₂, SiO₂/N–N, SiO₂/NH, SiO₂/NH₂/Ph.

Additionally, a group of bands associated with surface functional groups was identified, confirming the presence of the introduced groups. The presence of –CH– units from alkyl chains of functional groups is indicated by the bands at 2930 and 2850 cm⁻¹, corresponding to the asymmetric and symmetric stretching vibrations of the methylene groups, respectively, as well as by low-intensity bands at 1417 and 1481 cm⁻¹, which can be attributed to their bending vibrations. The IR spectra of all samples exhibit characteristic absorption bands confirming the existence of amino functions, with bands at 1558 cm⁻¹ for the bending vibrations of NH₂ groups and at 698 cm⁻¹ for the wagging vibrations of primary and secondary amines. Also, for the sample with phenyl groups SiO₂/NH₂/Ph, upon heating, the N–H stretching vibrations at 3360 cm⁻¹ are observed after water removal, which further indicates that the introduction of a hydrophobic group hinders water molecules from reaching the amino group. In the case of the bifunctional sample, the presence of phenyl groups introduced on the surface is clearly signaled by absorption bands associated with various phenyl vibrational modes: at 3046 cm⁻¹ for the C–H stretching vibrations in aromatic compounds, at 1431 cm⁻¹ as a weak absorption band corresponding to C=C stretching vibrations in aromatic compounds, and at 741 cm⁻¹ for the C–H bending vibrations in the benzene ring. Thus, FTIR spectroscopy confirms the functionalisation of the synthesised silica particles with silanes, revealing amino functionalities in all investigated samples, while the bifunctional sample additionally contains phenyl groups.

The presence of nitrogen-containing functional groups was verified by elemental analysis. The analysis results, presented in Table 2, indicate that the samples contain varying amounts of amino groups. The highest concentration of amino groups (3.58 mmol g⁻¹) was found in the bifunctional sample SiO₂/NH₂/Ph. This result for the bifunctional sample was obtained in a one-step synthesis as well, where, likely due to the different condensation rates of the silanes, a greater number of amino groups became attached in these systems.³⁸ While the lowest amino groups concentration (0.94 mmol g⁻¹) was observed in the sample SiO₂/NH, which bears secondary amino groups. The other monofunctional samples, SiO₂/NH₂ and SiO₂/N–N, exhibited slightly higher amino group contents of 2.14 and 1.35 mmol g⁻¹, respectively.

The electrophoretic mobility of diluted adsorbent suspensions was measured in aqueous solutions over a broad pH range and subsequently converted to zeta-potentials to characterise the surface electrical properties of the synthesised materials, evaluate the stability of their suspensions in water, and determine the nature of the surface groups. As can be seen from Fig. S2, the pure sample of silica nanoparticles differs markedly from the modified ones. It contains acidic silanol groups and exhibits a negative surface charge within the pH range of 4 to 11. It is well known that when the zeta-potential values are between −30 mV and +40 mV, the suspensions remain stable and particle agglomeration does not occur. Due to the partial protonation of amino and secondary amino groups, the SiO₂/NH₂, SiO₂/NH, and SiO₂/NH₂/Ph samples exhibit positive zeta-potential values and are stable within the pH range of 2–9, with their isoelectric points located in the alkaline region (pH = 9.96–10.35) (Fig. S2). In contrast, the suspension of the SiO₂/N–N sample remains stable within a narrower pH range of 2–7; however, its isoelectric point is the highest at 10.57, which can be attributed to the presence of diamino groups on its structure.

Thus, by modifying nanoscale silica derived from quartz using APTES, TMPED, BTMPA, and a combination of APTES and PhTES, a series of monofunctional materials containing amino, diamino, and secondary amino groups, as well as a bifunctional material bearing both amino and phenyl groups, were synthesised. The synthesised particles are nanosized and feature amino functionalities on the silica surface, indicating their potential for effective application in the removal of contaminants from aqueous solutions.

Adsorption properties of functionalised silica nanoparticles toward a fluorescent dye

The synthesised functional materials were evaluated as adsorbents for the dye C.I. Disperse Yellow 82 (DY 82), also known as Disperse Fluorescent Yellow 8GFF – a synthetic dye primarily used in the textile industry for colouring synthetic fibres (Fig. S1).

The adsorption kinetics curves of DY 82 for the investigated samples are shown in Fig. 11. The evaluation of the contact time effect on adsorption performance shows that the kinetic curve for the SiO₂/NH₂/Ph sample reaches a plateau the fastest, within just 3 hours, achieving an adsorption capacity of 37.89 mg g⁻¹. For the other samples – SiO₂/N–N, SiO₂/NH₂, and SiO₂/NH – equilibrium was reached later, after 12 hours. Among these, the SiO₂/N–N sample demonstrates the highest dye uptake, reaching 63.50 mg g⁻¹, while the SiO₂/NH₂ and SiO₂/NH samples display comparable adsorption capacities of 36.70 and 37.41 mg g⁻¹, respectively.

Fig. 11 Kinetics of DY 82 adsorption by the synthesised samples.

The kinetic curves of DY 82 adsorption were also analysed using pseudo-first order (PFO) and pseudo-second order (PSO) kinetic models, as well as the intraparticle diffusion (IPD) models (Table 3 and Fig. S3). Both PFO and PSO models yielded high correlation coefficients (R²), indicating that both empirical models describe the data well. However, in the case of the PFO model, the calculated adsorption capacity values (q_t) deviated notably from the experimentally observed ones (q). This discrepancy can be attributed to the fact that the PFO model primarily describes the initial stages of adsorption and does not account for surface saturation at longer contact times. Therefore, the PSO model provides a more accurate description of the interaction between the synthesised adsorbent samples and the dye molecules in suspension.

Table 3 Kinetic parameters of DY 82 adsorption on the synthesised samples

Kinetic models	Parameters	Samples
		SiO2/NH2	SiO2/N–N	SiO2/NH	SiO2/NH2/Ph
—	q, mg g^−1	35.7	63.5	37.41	37.89
Pseudo-I order	q t , mg g^−1	20.34	40.96	20.52	20.43
	k 1, h^−1	0.105	0.15	0.07	0.127
	R ^2	0.991	0.957	0.995	0.966
Pseudo-II order	q t , mg g^−1	38.31	68.97	40.16	40.16
	k 2, g (mg h)^−1	0.011	0.006	0.008	0.013
	R ^2	0.991	0.951	0.974	0.992
Intraparticle diffusion	k i1, mg (g h^0.5)^−1	0.188	0.111	0.231	0.260
	C, mg g^−1	2.192	2.517	3.130	4.517
	R ^2	0.987	1.0	0.995	0.991
	k i2, mg (g h^0.5)^−1	0.944	0.069	1.603	0.319
	C, mg g^−1	28.799	0.505	55.057	7.350
	R ^2	1.0	1.0	1.0	0.929
	k i3, mg (g h^0.5)^−1	—	1.300	—	—
	C, mg g^−1	—	77.676	—	—
	R ^2	—	1.0	—	—

---

Although the PSO model is traditionally associated with chemisorption, in this case, the formation of chemical bonds between the adsorbent and the adsorbate is unlikely due to the dye's non-ionic character and the absence of reactive functional groups. Instead, the adsorption behaviour is governed by specific physicochemical interactions between the dye's polar moieties (–NH, –C=O, aromatic rings) and the amino and silanol groups on the silica surface. These interactions may include hydrogen bonding, π–π stacking, and dipole–dipole interactions, which together impart a high affinity of the dye for the surface. Although covalent bonding does not occur, these intermolecular forces can produce kinetics that resembling chemisorption – particularly in terms of surface-site saturation and adsorption rate.

For all studied samples, the IPD model plots do not pass through the origin (Fig. S3c), indicating that diffusion alone is not the rate-limiting step in DY 82 adsorption.¹⁰ For the SiO₂/NH₂, SiO₂/NH, and SiO₂/NH₂/Ph samples, the q_tversus t^0.5 plots show two distinct linear regions, suggesting a multi-step adsorption mechanism. In the first stage, low intraparticle diffusion constants (k_i1 = 0.188–0.260 mg (g h^0.5)⁻¹) indicate the pronounced boundary-layer effect and significant external (film) diffusion. The second stage features higher diffusion constants (k_i2 = 0.319–1.603 mg (g h^0.5)⁻¹) and large intercepts (C = 7.35–55.057 mg g⁻¹), implying progressive surface saturation and increasing mass-transfer resistance. For the SiO₂/N–N sample, IPD analysis reveals a three-phase mass-transfer mechanism. In the first phase, a diffusion constant of k_i1 = 0.111 mg (g h^0.5)⁻¹ and an intercept C = 2.517 mg g⁻¹ indicate the predominance of external (film) diffusion – that is, the transport of dye molecules to the adsorbent surface. The second phase, characterized by the lowest diffusion constant (k_i2 = 0.069 mg (g h^0.5)⁻¹) and a minimal intercept (C = 0.505 mg g⁻¹), corresponds to intrapore diffusion, which appears to be the rate-limiting step. In the third phase, the diffusion constant rises sharply (k_i3 = 1.300 mg (g h^0.5)⁻¹ alongside a very high intercept (C = 77.676 mg g⁻¹), suggesting completion of adsorption, active-site saturation, or secondary adsorption in macropores. Thus, dye uptake proceeds stepwise – from film diffusion to- intraparticle diffusion and, finally, to surface saturation.

Overall, kinetic modelling indicates that DY 82 adsorption on the synthesised samples occurs in multiple stages. Initially, relatively slow film diffusion dominates due to the presence of a boundary layer. This is followed by internal diffusion and gradual surface saturation. Throughout surface adsorption process, specific interactions such as hydrogen bonding and dipole–dipole forces develop between silanol and amino groups on the silica surface and the dye's nitrogen atoms, carbonyl groups, and aromatic fragments.

The adsorption isotherms were used to interpret the experimental adsorption equilibrium data and to analyse the adsorbent–adsorbate system using the linear equations for the Langmuir and Freundlich isotherm models (Fig. 12). The corresponding adsorption parameters of the dye on the studied materials are presented in Table 4.

Fig. 12 Experimental adsorption isotherms of DY 82 on functionalised silica nanoparticles (a) and their fitting to the Langmuir (b) and Freundlich isotherm models (c).

Table 4 Parameters of DY 82 adsorption in the coordinates of the Langmuir and Freundlich equations

Sample	A eq, mg g^−1 (mmol g^−1)	C N/Aeq	K d, dm^3 g^−1	Langmuir equation			Freundlich equation
				C eq/Aeq = (1/KLAmax) + (Ceq/Amax)			Ln Aeq = ln KF + (1/n)ln Ceq
				A max, mg g^−1	K L, L mg^−1	R ^2	K F, mg g^−1	n	R ^2
A eq is adsorption capacity at equilibrium, mg g^−1 and mmol g^−1; CN/Aeq is correlation between number of N-containing groups and adsorption capacity at equilibrium; Kd is an affinity of the adsorbate to the adsorbent, dm^3 g^−1; Amax is maximal adsorption capacity for complete monolayer covering of the surface, mg g^−1; Ceq is equilibrium concentration of metal ions in the solution, mg L^−1; KL is the Langmuir constant, which characterises the adsorption energy, L mg^−1; KF is a Freundlich constant, mg g^−1; n is an empirical parameter related to the intensity of adsorption.
SiO 2 /NH 2	139 (0.44)	2.14	0.42	0.94	0.56	0.6783	2.68	1.08	0.9693
SiO 2 /N–N	193 (0.58)	1.35	0.64	0.72	1.38	0.8254	2.75	1.98	0.8820
SiO 2 /NH	142 (0.44)	0.94	0.43	0.69	1.15	0.9344	2.74	1.40	0.9558
SiO 2 /NH 2 /Ph	115 (0.37)	3.58	0.39	0.56	1.17	0.9317	3.23	1.29	0.9686

---

Disperse Yellow 82 possesses a coumarin core with a carbonyl group that can form hydrogen bonds with protonated and partially protonated amino groups. In addition, its aromatic rings can interact with amino groups via dipole–dipole and π–π interactions, which together facilitate strong interaction between the dye and amino groups.³⁹ The SiO₂/N–N sample (193 mg g⁻¹), prepared using a silane containing diamine groups, exhibited the highest dye adsorption. In contrast, silica particles functionalised with APTES (SiO₂/NH₂) and BTMPA (SiO₂/NH) showed slightly lower adsorption capacities of 139 and 142 mg g⁻¹, respectively. The bifunctional adsorbent SiO₂/NH₂/Ph had the lowest adsorption capacity (115 mg g⁻¹), likely due to the hydrophobicity and steric hindrance of the phenyl groups.

This suggests that a higher number of functional groups does not necessarily enhance adsorption capacity, but rather their proximity and the availability of protons through which binding can occur. As demonstrated earlier for the bifunctional samples, the amino groups are arranged discretely and are less protonated.⁴⁰ Moreover, the adsorption isotherms for the dye fit well with the Freundlich equation, indicating adsorption on a heterogeneous surface.

Additionally, information about the adsorption mechanism was obtained from FTIR spectra recorded before and after dye adsorption (Fig. 13). Following the adsorption of DY 82, the appearance of new bands, as well as the shifting and disappearance of certain absorption bands, was observed in the FTIR spectra. The absence of bands at ∼1417 cm⁻¹, attributed to the bending vibration of C–H, and the appearance of a new one at 2975 cm⁻¹, assigned to the symmetric stretching vibrations of the methyl group, confirms the presence of the dye on the surface. After the adsorption of DY 82, the bands associated with N–H stretching vibrations at 1558 cm⁻¹ and 1481 cm⁻¹ shifted, while the band at 698 cm⁻¹ disappeared. These spectral changes indicate that DY 82 is adsorbed onto the investigated materials through both physical and chemical interactions.

Fig. 13 Infrared spectra of DY 82 and functionalised silica nanoparticles post-adsorption.

Another confirmation of dye adsorption on the synthesised materials is provided by photographs taken under short-wave UV light at λ = 254 nm (Fig. S4). The initial samples exhibit a blue glow at this wavelength, whereas after dye adsorption, they emit a bright green glow. Notably, the dye itself does not exhibit a strong fluorescence within this spectral range.

The photoluminescence spectra of the synthesised samples with adsorbed dye were investigated. The maximum photoluminescence intensity of the dye dispersion was observed at a wavelength of 498 nm. The dye adsorbed amount on the samples was ∼40 mg g⁻¹, and the suspension of dye-adsorbed samples had a concentration of 0.25 g L⁻¹, corresponding to a dye concentration of 0.01 g L⁻¹. The norfloxacin concentration in the system was 3 mg L⁻¹, with an ionic strength maintained at 0.1 M KCl.

It was established that the synthesised samples with adsorbed dye exhibited photoluminescent properties. Notably, the photoluminescence intensity of all studied samples was substantially higher than that of the dye in an individual dispersion at the same concentration (Fig. S5 and Fig. 14).

Fig. 14 The effect of norfloxacin on the photoluminescence spectra of suspensions of DY 82-loaded functionalised silica samples.

This enhancement is presumably due to the influence of the silica matrix. Additionally, a slight shift in the photoluminescence maximum was detected: for the SiO₂/N–N sample, the maximum shifted from 498 nm to 496 nm, and while for the other samples – SiO₂/NH₂, SiO₂/NH, and SiO₂/NH₂/Ph – it shifted to 495 nm. The SiO₂/NH sample, containing secondary amino groups, exhibited the highest initial photoluminescence intensity, whereas the other samples showed slightly lower intensities. Upon the addition of norfloxacin (NOR) to the dye-adsorbed material suspensions, an increase in photoluminescence intensity was observed for the systems with SiO₂/NH₂, SiO₂/N–N, and SiO₂/NH₂/Ph samples. In contrast, the intensity decreased for the SiO₂/NH sample with secondary amino groups, despite its initially high photoluminescence intensity. This behavior can be attributed to the absence of free amino groups on the surface of that material. Moreover, the fact that adding norfloxacin to the individual dye suspension does not affect its photoluminescence intensity confirms the crucial role of the synthesised silica adsorbents as a matrix that immobilizes dye molecules and prevents fluorescence quenching. The most significant increase in photoluminescence was observed for the bifunctional sample, which can be explained by the involvement of phenyl groups in energy-transfer processes.

To determine the origin of the photoluminescence changes induced by norfloxacin, spectra were recorded for the suspension of the initial SiO₂/NH₂/Ph sample, the functionalised SiO₂/NH₂/Ph sample loaded with DY 82, and a DY 82 solution matching the dye concentration on the functionalised sample. The SiO₂/NH₂/Ph sample, chosen because it showed the greatest photoluminescence enhancement in norfloxacin's presence, was used throughout the study.

It was found (Fig. 15) that norfloxacin alone, the initial SiO₂/NH₂/Ph sample and their combination all exhibited minimal photoluminescence. The DY 82 solution by itself produced an intensity of approximately 186 000 a.u., which increased slightly to about 195 000 a.u. upon addition of the sample suspension and to roughly 211 000 a.u. when both the suspension and norfloxacin were present. Direct addition of norfloxacin to the DY 82 solution resulted in a higher intensity of around 248 000 a.u.

Fig. 15 Effect of norfloxacin on the photoluminescence intensity of- suspension of the initial SiO₂/NH₂/Ph sample, the dye-loaded functionalised sample SiO₂/NH₂/Ph/DY 82, and the DY 82 solution (suspension concentration 0.25 g L⁻¹ in 0.1 M KCl, DY 82 concentration 0.01 g L⁻¹, NOR concentration 3 mg L⁻¹).

By contrast, the dye-loaded functionalised SiO₂/NH₂/Ph/DY 82 sample exhibited a much higher baseline intensity of 570 000 a.u., increasing to a maximum of ∼721 000 a.u. upon norfloxacin addition. This enhancement was accompanied by a peak shift from 495 nm to 498 nm.

These results indicate that the enhanced photoluminescence in the system containing the functionalised SiO₂/NH₂/Ph/DY 82 sample and norfloxacin arises from specific interactions between the fluorescent dye and norfloxacin. Moreover, because the dye is immobilised within the adsorbent matrix, the system exhibits enhanced photoluminescent sensitivity.

The effect of norfloxacin concentration on the photoluminescent response of the SiO₂/NH₂/Ph/DY 82 sensor is presented in Fig. S6, with the corresponding calibration curve shown in Fig. 16. As illustrated, the sensor displays a linear fluorescence response over the norfloxacin concentration range of 0.05–4.0 mg L⁻¹ (R² = 0.999). The limit of detection (LOD) was determined to be 0.081 mg L⁻¹, and the limit of quantification (LOQ) was 0.246 mg L⁻¹, demonstrating that the SiO₂/NH₂/Ph/DY 82 suspension is a highly sensitive and reliable platform for detecting norfloxacin in aqueous solutions.

Fig. 16 Linear relationship between the emission intensity at 498 nm of the SiO₂/NH₂/Ph/DY 82 sensor and the norfloxacin concentration.

The linear correlation between norfloxacin concentration and photoluminescence intensity in the presence of SiO₂/NH₂/Ph/DY 82, together with the observed red shift of the emission maximum from 495 nm to 498 nm after dye adsorption on SiO₂/NH₂/Ph sample (Fig. 8 and Fig. S5), indicates the formation of a complex between SiO₂/NH₂/Ph/DY 82 and norfloxacin. In this complex, the adsorbent matrix shields the dye from quenching by excluding water molecules and stabilising the dye within its structure.

The interaction between the SiO₂/NH₂/Ph/DY 82 sensor and norfloxacin was investigated using the Benesi–Hildebrand equation: log((F − F₀)/F₀)) = log K_a + n·log C.⁴¹ The resulting linear plot (R² = 0.9943, Fig. S7) suggests a 1 : 1 stoichiometry for the complex. The binding constant was determined to be K_a = 0.102 L mg⁻¹ (319 M⁻¹), indicating a specific but relatively weak interaction between the components.

To assess the selectivity of the SiO₂/NH₂/Ph/DY 82 sensor (0.25 g L⁻¹ in 0.05 M KCl), a 0.1 mg L⁻¹ norfloxacin solution was tested alongside various pharmaceuticals, including carbamazepine, ciprofloxacin, diclofenac, ibuprofen, sulfamethoxazole and tetracycline, each at the same concentration 0.1 mg L⁻¹. The photoluminescence response of SiO₂/NH₂/Ph/DY 82 showed high selectivity toward norfloxacin over these potential interferents. As illustrated in Fig. 17, only minor luminescence changes were observed in the presence of the other drugs, whereas norfloxacin induced a pronounced enhancement in emission intensity at 498 nm. Moreover, adding 0.1 mg L⁻¹ norfloxacin to each drug mixture further amplified the signal at 498 nm, confirming the SiO₂/NH₂/Ph/DY 82 sensor's excellent selectivity and specificity toward norfloxacin.

Fig. 17 Emission intensities at 498 nm of SiO₂/NH₂/Ph/DY 82 in the presence of individual potential interfering pharmaceuticals at 0.1 mg L⁻¹ (Ctrl: sensor suspension + interferent; test: Ctrl + 0.1 mg L⁻¹ norfloxacin).

Overall, distinct photoluminescence was observed when the fluorescent dye was adsorbed onto silica samples functionalised with amino groups. Moreover, the presence of norfloxacin led to a pronounced enhancement of the photoluminescent signal in suspensions containing dye-loaded silica nanoparticles with both amino and phenyl groups. This effect can be exploited to develop highly sensitive optical sensors for detecting norfloxacin and similar compounds. The enhanced photoluminescence indicates that the dye molecules are effectively immobilized within the silica matrix, thereby reducing non-radiative quenching and providing a clear, direct photoluminescent response. Such dye-silica composites hold great promise for applications in environmental monitoring and clinical diagnostics, where changes in photoluminescence can serve as a reliable indicator of target analytes.

Conclusions

In conclusion, this study demonstrates the successful extraction of silica from mineral sources and its effective application in the synthesis of adsorbents for dye removal from aqueous solutions. Furthermore, our study underscores the critical importance and cost-effectiveness of extracting silica from abundant mineral sources, particularly quartz, which is rich in silica. The process developed here produces silica nanoparticles with high purity and excellent yield, demonstrating a sustainable, economical, and scalable method for obtaining high-quality nanostructured material. The primary objective was to obtain silica microspheres that serve as a platform for surface functionalisation and subsequent various applications. The synthesised silica-based adsorbents exhibited significant efficiency in dye adsorption, particularly in the case of materials functionalised with amino and diamine groups. These materials offer substantial potential for environmental applications in water purification, particularly for the removal of organic pollutants such as textile dyes. However, the study also highlights the importance of the nature, availability and distribution of surface functional groups on the adsorbent in achieving optimal adsorption capacity.

Furthermore, it was established that the synthesised samples with adsorbed textile dye, which may be present in wastewater, exhibit enhanced photoluminescent properties for norfloxacin detection compared with the dye suspension alone, owing to the stabilising effect of the silica matrix. Thus, beyond their role in wastewater treatment, the obtained materials show great promise as dual-function systems that simultaneously address environmental remediation and the fluorescent sensing of pharmaceuticals in aquatic environments.

Author contributions

Conceptualization, O. S. and I. M.; methodology, O. S. and I. M.; formal analysis, O. S., S. H., M. M., J. B., E. D. and I. M.; investigation, O. S., S. H., M. M., J. B., E. D. and I. M.; resources, O. S. and I. M.; data curation, O. S., S. H., M. M., J. B., E. D. and I. M.; writing – original draft preparation, O. S.; writing – review and editing, I. M.; supervision, I. M.; project administration, I. M. All authors have read and agreed to the published version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). The authors have cited additional references within the SI.^42–47 The Supplementary Information includes data related to the Experimental section of the article (materials, synthesis and functionalization of silica nanoparticles, and methods of investigation); the dependence of the zeta potential of particle suspensions on pH; kinetic models (pseudo-first-order, pseudo-second-order, and intraparticle diffusion) for the adsorption of DY 92 by the synthesized samples; photos under UV light of the initial samples, samples after dye sorption, and dye DY 82; photoluminescence spectra of suspensions of functionalized samples with adsorbed dye DY 82 with the addition of norfloxacin; the effect of norfloxacin concentration on the photoluminescence intensity of sample suspensions; and the dependence of photoluminescence intensity on norfloxacin concentration as described by the Benesi–Hildebrand equation. See DOI: https://doi.org/10.1039/d5tc02934a.

Acknowledgements

O. S. thanks the EU's NextGenerationEU through the Recovery and Resilience Plan for Slovakia (project 09I03-03-V01-0098) and the Slovak Grant Agency VEGA (projects 2/0108/23, 2/0138/24); I. M. thanks the EU's NextGenerationEU through the Recovery and Resilience Plan for Slovakia (project 09I03-03-V04-00708).

References

1. A. Sahu and J. C. Poler, J. Environ. Chem. Eng., 2024, 12(5), 113754,  DOI:10.1016/j.jece.2024.113754.
2. L. A. Castillo-Suárez, A. G. Sierra-Sánchez, I. Linares-Hernández, V. Martínez-Miranda and E. A. Teutli-Sequeira, Int. J. Environ. Sci. Technol., 2023, 20, 10553–10590,  DOI:10.1007/s13762-023-04810-2.
3. N. Olivar Aponte, J. Hernández Gómez, V. Torres Argüelles and E. D. Smith, Sustainability, 2024, 20(1), 2381871,  DOI:10.1080/15487733.2024.2381871.
4. A. Tkaczyk, K. Mitrowska and A. Posyniak, Sci. Total Environ., 2020, 717, 137222,  DOI:10.1016/j.scitotenv.2020.137222.
5. F. Uddin, Cellulose, 2021, 28, 10715–10739,  DOI:10.1007/s10570-021-04228-4.
6. M. Mahajan, S. Kumar, J. Gaur, S. Kaushal, J. Dalal, G. Singh, M. Misra and D. S. Ahlawat, RSC Adv., 2025, 15, 2958–2980,  10.1039/D4RA08632E/.
7. K. Bano, P. P. Singh, S. Kumar, S. M. Saeed, S. Aggarwal, R. Kumare and S. Kaushal, Environ. Sci.: Water Res. Technol., 2024, 10, 1714–1725,  10.1039/D4EW00270A.
8. H. Yankovych, E. Dutková, V. Kyshkarova, M. Vaclavikova and I. Melnyk, Chemosensors, 2023, 11(6), 332,  DOI:10.3390/chemosensors11060332.
9. I. V. Melnyk, V. Tomina, H. Yankovych, H. Kolev, E. Dutkova, T. C. Breijaertd, V. G. Kessler and G. A. Seisenbaeva, Nanoscale Adv., 2024, 6, 3146–3157,  10.1039/d3na00909b.
10. J. Chen, Y. Sheng, Y. Song, M. Chang, X. Zhang, L. Cui, D. Meng, H. Zhu, Z. Shi and H. Zou, ACS Sustainable Chem. Eng., 2018, 6(3), 3533–3545,  DOI:10.1021/acssuschemeng.7b03849.
11. N. H. A. Hussen, S. H. Qadir, H. S. Rahman, Y. Y. Hamalaw, P. S. S. Kareem and B. A. Hamza, Drug Chem. Toxicol., 2023, 47(5), 795–806,  DOI:10.1080/01480545.2023.2240036.
12. F. Kriese, S. Lassen and H. Horn, ChemistryOpen, 2022, 11, e202200034,  DOI:10.1002/open.202200034.
13. O. Artiushenko, W. S. Rojano, M. Nazarkovsky, M. F. M. F. Azevedo, T. D. Saint'Pierre, J. Kai and V. Zaitsev, Sep. Purif. Technol., 2024, 330, 125525,  DOI:10.1016/j.seppur.2023.125525.
14. I. Melnyk, V. Tomina, N. Stolyarchuk, D. Marcin Behunova, M. Václavíková and O. Semeshko, Appl. Nanosci., 2023, 13(11), 7349–7363,  DOI:10.1007/s13204-023-02919-x.
15. V. Tomina, N. Stolyarchuk, O. Semeshko, M. Barczak and I. Melnyk, Molecules, 2023, 28, 430,  DOI:10.3390/molecules28010430.
16. J. B. Haider, Md. I. Haque, H. Mozammel, Md. M. Hossen, M. Mottakin, Md. A. Khaleque, M. A. H. Johir, J. L. Zhou, M. B. Ahmed and M. Zargar, J. Cleaner Prod., 2022, 380(2), 135121,  DOI:10.1016/j.jclepro.2022.135121.
17. L. A. September, N. Kheswa, N. S. Seroka and L. Khotseng, RSC Adv., 2023, 13, 1370–1380,  10.1039/D2RA07490G.
18. A. Sidorowicz, G. Fais, M. Casula, M. Borselli, G. Giannaccare, A. M. Locci, N. Lai, R. Orrù, G. Cao and A. Concas, Mar. Drugs, 2023, 21(6), 3522023,  DOI:10.3390/md21060352.
19. E. Errington, M. Guo and J. Y. Y. Heng, Green Chem., 2023, 25, 4244–4259,  10.1039/d2gc01433e.
20. C. V. Hoang, D. N. Thoai, N. T. D. Cam, T. T. T. Phuong, N. T. Lieu, T. T. T. Hien, D. N. Nhiem, T.-D. Pham, M. H. T. Tung, N. T. T. Tran, A. Mechler and Q. V. Vo, ACS Omega, 2022, 7(45), 41687–41695,  DOI:10.1021/acsomega.2c05760.
21. N. Raza, W. Raza, S. Madeddu, H. Agbe, R. V. Kumar and K.-H. Kim, RSC Adv., 2018, 8, 32651–32658,  10.1039/C8RA06257A.
22. M. Liu, G. Wang, F. Zhao, W. Li, G. Zhu, G. Liang, W. Jian, L. Liao and G. Lv, Prog. Nat. Sci.:Mater. Int., 2025, 35(1), 51–64,  DOI:10.1016/j.pnsc.2024.11.008.
23. H. Long, D. Zhu, J. Pan, S. Li, Z. Guo and X. Xu, J. Mater. Res. Technol., 2024, 31, 2094–2102,  DOI:10.1016/j.jmrt.2024.06.212.
24. A. Jennings, A. Senior, K. Guerin, P. Main and J. Walsh, Aust. J. Earth Sci., 2024, 71(8), 1085–1097,  DOI:10.1080/08120099.2024.2362296.
25. S. Arunmetha, A. Karthik, S. R. Srither, M. Vinoth, R. Suriyaprabha, P. Manivasakana and V. Rajendran, RSC Adv., 2015, 5, 47390–47397,  10.1039/C5RA07074K.
26. U. Ghani, S. Hussain, A. Ali, V. Tirth, A. Algahtani, A. Zaman, M. Mushtaq, K. Althubeiti and M. Aljohani, ACS Omega, 2022, 7(7), 6113–6120,  DOI:10.1021/acsomega.1c06553.
27. S. Mychak and L. Farfuliak, Geophys. J., 2021, 43(1), 142–159,  DOI:10.24028/gzh.0203-3100.v43i1.2021.225496.
28. U. Zulfiqar, T. Subhani and S. W. Husain, J. Asian Ceram. Soc., 2016, 4(1), 91–96,  DOI:10.1016/j.jascer.2015.12.001.
29. M. Sarikaya, T. Depci, R. Aydogmus, A. Yucel and N. Kizilkaya, IOP Conf. Ser.: Earth Environ. Sci., 2016, 44 DOI:10.1088/1755-1315/44/5/052004.
30. W. Zhang, T. Zhang, L. Xi, H. Gu, Y. Hu, H. Gu and Y. Zhang, Energy Procedia, 2012, 17(B), 1627–1634,  DOI:10.1016/j.egypro.2012.02.290.
31. E. Rampazzo, E. Brasola, S. Marcuz, F. Mancin, P. Tecilla and U. Tonellato, J. Mater. Chem., 2005, 15(27–28), 2687,  10.1039/B502052B.
32. L. Sennoun, C.-C. Lee, Y. Fretel, M. Clavié, G. Subra, Y. Ladner, A. Napoli, A. Galarneau, P. Hesemann and A. Mehdi, New J. Chem., 2024, 48, 7052–7069,  10.1039/D4NJ00715H.
33. L. C. Mugica, B. Rodríguez-Molina, S. Ramos and A. Kozina, Colloids Surf., A, 2016, 500, 79–87,  DOI:10.1016/j.colsurfa.2016.04.002.
34. M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso, J. Rouquerol and K. S. W. Sing, Pure Appl. Chem., 2015, 87(9–10), 1051–1069,  DOI:10.1515/pac-2014-1117.
35. G. Leofanti, M. Padovan, G. Tozzola and B. Venturelli, Catal. Today, 1998, 41(1–3), 207–219,  DOI:10.1016/S0920-5861(98)00050-9.
36. O. Semeshko, M. Fizer, V. Sliesarenko, J. Briancin, O. Bondarchuk, A. Lobnik and I. Melnyk, Adv. Mater. Interfaces, 2025, 12, 2400761,  DOI:10.1002/admi.202400761.
37. A. Idris, Z. Man, A. S. Maulud, M. A. Bustam, H. A. Mannan and I. Ahmed, Appl. Surf. Sci., 2020, 506, 144978,  DOI:10.1016/j.apsusc.2019.144978.
38. S. S. Kotsyuda, V. V. Tomina, Y. L. Zub, I. M. Furtat, A. P. Lebed, M. Vaclavikova and I. V. Melnyk, Appl. Surf. Sci., 2017, 420, 782–791,  DOI:10.1016/j.apsusc.2017.05.150.
39. M. S. Zakerhamidi, A. Ghanadzadeh, H. Tajalli, M. Moghadam, M. Jassas and R. Hosseininia, Spectrochim. Acta, Part A, 2010, 77, 337–341,  DOI:10.1016/j.saa.2009.12.060.
40. V. V. Tomina, I. M. Furtat, A. P. Lebed, S. S. Kotsyuda, H. Kolev, M. Kanuchova, D. Marcin Behunova, M. Vaclavikova and I. V. Melnyk, ACS Omega, 2020, 5(25), 15290–15300,  DOI:10.1021/acsomega.0c01335.
41. D. Prabha, D. Singh and R. Gupta, J. Chem. Sci., 2021, 133(88) DOI:10.1007/s12039-021-01956-9.
42. M. Triwikantoro, M. Zainuri and D. Darminto, Mater. Sci.-Pol., 2015, 33, 47–55,  DOI:10.1515/msp-2015-0008.
43. M. Kalbarczyk, S. Skupiński and M. Kosmulski, J. Mol. Liq., 2024, 397(1), 124195,  DOI:10.1016/j.molliq.2024.124195.
44. J. W. Anthony, R. A. Bideaux, K. W. Bladh and M. C. Nichol, Handbook of Mineralogy, Mineralogical Society of America, Chantilly, 2004–2020, http://www.handbookofmineralogy.org/, (accessed April 2025).
45. Mineralogy Database, https://www.webmineral.com/, (accessed April 2025).
46. R. T. Downs and M. Hall-Wallace, Am. Mineral., 2003, 88, 247–250,  DOI:10.5860/choice.41sup-0262.
47. L. N. Warr, Mineral. Mag., 2021, 85, 291–320,  DOI:10.1180/mgm.2021.43.

---