Recoverable Fe₃O₄@SiO₂@TMC magnetic sorbent for efficient Congo red removal: performance and adsorption mechanism

Abstract

The increasing discharge of dye-laden wastewater necessitates the development of efficient and easily recoverable adsorbents. In this study, a trimesoyl chloride (TMC) functionalized Fe₃O₄@SiO₂ magnetic sorbent was synthesized via a sol–gel route for the removal of Congo red (CR) from aqueous media. Dynamic light scattering indicated a uniform particle size distribution (∼130–140 nm). FT-IR analysis confirmed the presence of –COOH, –OH, and aromatic moieties, enabling multi-modal interactions including hydrogen bonding, electrostatic attraction, π–π stacking, and coordination with CR molecules. Structural and surface properties were further validated by XRD, FE-SEM, and zeta potential analyses. Operational parameters (initial concentration, pH, adsorbent dosage, and contact time) were systematically optimized, with the maximum adsorption capacity of 4614.62 (pH = 5, dosage = 2 mg) for CR dye. Kinetic modeling revealed an excellent fit to the Bangham model (R² = 0.999), which yielded a calculated equilibrium adsorption capacity of 231.95 mg g⁻¹, closely matching the experimental value of 233.85 mg g⁻¹ and indicating a substantial contribution of pore diffusion. Isotherm analysis suggested multistage adsorption behavior. The Jovanović model provided a reasonable fit (R² = 0.960) and a q_m value of 4100 mg g⁻¹, which was relatively close to the experimental adsorption capacity of 4614.61 mg g⁻¹. The Koble–Corrigan model provided the best fit to the equilibrium data (R² = 0.995), suggesting the contribution of heterogeneous adsorption sites. Reusability tests demonstrated that the magnetic adsorbent retained ∼80% of its initial removal efficiency after 4 consecutive cycles, confirming its practical applicability and facile separation. Overall, the Fe₃O₄@SiO₂@TMC system presents a robust and magnetically recoverable platform for efficient dye remediation, offering a promising approach for future studies and wastewater treatment.

Trimesoyl chloride (TMC) functionalised Fe₃O₄@SiO₂ magnetic adsorbent, exhibits as a promising candidate for Congo Red dye decontamination from wastewater.

1. Introduction

As a major contributor to water pollution, industrial effluents pose a significant threat to our ecosystems. These effluents, if left untreated, will have devastating effects on ecosystems, marine life, and human health. As a result, the primary waste streams are those from the food, printing, textile, paint, and cosmetics domains. Heavy metals, dyes, pesticides, organic acids, pharmaceuticals, microbes and pigments are among the pollutants that these discharges might carry.¹ Around 10 000 tonnes of synthetic dyes are utilised by the textile industry each year, out of a global output of roughly 7 × 10⁷ tons.² Dyeing sector activities create significant water contamination by discharging enormous volumes of effluents into water bodies. Diverse dye compounds exist based on the raw materials utilized in the manufacturing process. Some dye effluents include abundant suspended particles and hazardous compounds.³ Congo red dye is a red crystalline powder characterized by a heterocyclic aromatic chemical composition. It is an anionic azo dye predominantly utilized in the textile sector.⁴

The treatment of textile effluent contaminated with dyes currently involves a variety of techniques, including adsorption, biological, membrane filtration, UV degradation, reverse osmosis, ultrafiltration, oxidation, ion exchange, coagulation, flocculation, sedimentation, electrochemical oxidation, etc.^5,6 Among them, adsorption is the most appealing and appropriate method for the removal of organic pollutants. This is due to the fact that adsorption is economical, straightforward, adaptable, eco-friendly, ecologically benign, efficient, and does not produce sludge.⁷ To achieve these objectives, a diverse array of organic and inorganic materials has been developed, including zeolites,⁸ clays,⁹ metal oxide nanoparticles,¹⁰ agricultural waste,¹¹ fly ash,¹² nanofibers,¹³ metal–organic frameworks (MOFs),¹⁴ bionanomaterials,¹⁵ and both synthetic and natural polymers (such as cellulose, chitosan, PAA, PAM, PAN, etc.)^16–21 polymeric composites^22,23 and biochar.^24–27

As a result of its unique characteristics, including nanoparticle size, mixing ability, and solution mobility, metal oxide-based nano sorbents have been used in wastewater treatment. Metal oxide materials has potential uses in chemistry and engineering, including catalysis and environmental remediation.²⁸ The properties of magnetic NPs (Fe₃O₄ and γ-Fe₃O₄) including, such as superparamagnetism, low toxicity, chemical stability, dispersity in liquid media, and uniform size, have garnered attention in recent decades. These properties have led to their potential applications in various fields. Several methods have been used to synthesize and characterize high-quality magnetic iron oxide NPs.

A variety of coatings, such as inorganic coatings [(silica, SiO₂), noble metal shells (Au, Ag), metal oxide layers (ZrO₂, CrO_x)], polymers (PEG, chitosan), or functional molecules (DMSA, APTES, folic acid), can be applied to the surface of magnetic nanoparticles (MNPs) to enhance biocompatibility, adsorption, target cells, or facilitate drug delivery.^29–33 Because of their high adsorption capacity, wide surface-area-to-volume ratio, and ease of magnetic separation, core–shell magnetic NPs are employed as extremely effective adsorbents. Surface functionalization using materials such as silica, polymers, or ligands can prevent aggregation, while enhancing interaction with contaminants.^34–36

Although Fe₃O₄-based magnetic adsorbents have been investigated for the removal of dyes and other aqueous pollutants, the application of bare Fe₃O₄ nanoparticles can be limited by particle aggregation, relatively limited surface functionality, and reduced colloidal stability. Surface modification and core–shell engineering provide effective strategies to overcome these limitations by improving the stability, dispersibility, and chemical functionality of magnetic nanoparticles. In the present study, Fe₃O₄ nanoparticles were coated with an amorphous SiO₂ layer through a sol–gel process. The silica layer can suppress aggregation of the iron oxide nanoparticles and improve their chemical stability, while the abundant surface silanol (Si–OH) groups provide additional reactive sites for subsequent surface functionalization. These silanol groups can also participate in condensation reactions with suitable organosilane functionalities, facilitating the construction of chemically tailored silica interfaces.³⁷ The SiO₂ surface was subsequently functionalized with trimesoyl chloride (TMC), introducing aromatic and oxygen-containing functionalities that can provide additional interaction sites for Congo red (CR) molecules. The resulting Fe₃O₄@ SiO₂@TMC architecture integrates the magnetic responsiveness of Fe₃O₄ with the protective and functional role of SiO₂ and the pollutant-interacting characteristics of the TMC-functionalized surface within a single adsorbent platform. In particular, the combination of aromatic domains and oxygen-containing functionalities (–OH, –COOH, and –COO⁻) can promote multiple interactions with CR, including electrostatic attraction, hydrogen bonding, and π–π interactions, while the pH-responsive surface charge of the adsorbent is favorable for the uptake of anionic CR dye molecules.

Therefore, the distinctive feature of the present material is not merely the incorporation of Fe₃O₄, but the integration of magnetic recoverability, structural stability, and a chemically tailored adsorption interface within a single sorbent. This design provides a potentially advantageous platform for efficient CR removal while enabling facile magnetic separation and recovery of the adsorbent. Accordingly, the present study focuses on the synthesis and characterization of the TMC-functionalized Fe₃O₄@SiO₂ magnetic sorbent and systematically evaluates the effects of solution pH, contact time, adsorbent dosage, and initial CR concentration on adsorption performance. Equilibrium isotherm and kinetic models were further employed to investigate the adsorption behavior and provide insights into the governing adsorption process.

2. Experimental

2.1. Materials

Each material, including iron(ii) chloride tetrahydrate (FeCl₂·4H₂O, ≥99%), iron(iii) chloride hexahydrate (FeCl₃·6H₂O, ≥98%), trimesic acid (benzene-1,3,5-tricarboxylic acid) (C₉H₆O₆, 95%), thionyl chloride(SOCl₂ ≥99%), triethyl orthosilicate (TEOS, Si(OC₂H₅)₄, 98%), (3-aminopropyl) trimethoxysilane (APTMS, C₆H₁₇NO₃Si, 97%), 3,3′-[(1,1′-Biphenyl)-4, 4′-diylbis(azo)] bis (4-amino-1-naphthalenesulfonic acid) disodium salt (Congo red, C₃₂H₂₂N₆Na₂O₆S₂, 85%), purchased from the Sigma-Aldrich corporation. Additionally, ethanol, dry toluene (C₇H₈, 99.8%, anhydrous), ammonia solution (28%), hydrochloric acid (HCl, 37%), sodium hydroxide (NaOH, 99%) were received from Loba Chemie Co. (India) and used without further purification.

2.2. Synthesis of Fe₃O₄

A mixture comprising 5 g of iron(ii) chloride tetrahydrate and 10 g of iron(iii) chloride hexahydrate was dissolved in 50 ml of water under a nitrogen-rich environment. The resultant solution was subjected to sonication in an ultrasonic bath for 10 minutes. Subsequently, a 28% ammonia solution was added dropwise at a rate of 1 ml per minute while maintaining magnetic stirring at 35 °C. Upon achieving a pH of 12, stirring was continued for an additional 40 minutes. The resulting particles were extracted using an external magnet, then washed 3 to 4 times with a 20% ethanol and water solution, and finally dried at 70 °C in a vacuum oven to avert iron oxidation.

2.3. Synthesis of Fe₃O₄@SiO₂

Initially, 1 g of synthesized Fe₃O₄ was dispersed in 100 mL of an 80% (v/v) ethanol/water mixture and sonicated for 15 min to ensure uniform dispersion. Under a nitrogen atmosphere, a solution of 2 g tetraethyl orthosilicate in 50 mL ethanol was introduced, followed by the dropwise addition of 6 mL of 28% ammonia over 10 min to initiate the solgel process. The reaction mixture was then vigorously stirred at 35 °C for 24 h to allow complete silica shell formation. The resulting Fe₃O₄@SiO₂ particles were magnetically separated, washed three times with 20% ethanol solution to remove residual reagents, and subsequently dried in a vacuum oven at 70 °C for 24 h.

2.4. Synthesis of TMC functionalized- Fe₃O₄@SiO₂

1 g of Fe₃O₄@SiO₂ was dispersed in 50 mL of dry toluene under a N₂ atmosphere with vigorous mechanical stirring. In parallel, trimesoyl chloride was synthesized via the chlorination of trimesic acid using thionyl chloride at 70 °C with a molar ratio of 1 : 3 (trimesic acid:thionyl chloride). Subsequently, 5 mL of the prepared TMC was added to the suspension, and the reaction mixture was stirred under N₂ for 8 h to enable surface functionalization. The resulting TMC-functionalized Fe₃O₄@SiO₂ particles (Fe₃O₄@SiO₂@TMC) were then collected, washed with dichloromethane and vacuum-dried at 60 °C for 12 h. A schematic representation of the synthesis route is presented in Fig. 1.

Fig. 1. Synthesis of Fe₃O₄@SiO₂@TMC adsorbent patricles.

2.5. Characterization

The physicochemical characteristics of Fe₃O₄, Fe₃O₄@SiO₂, and Fe₃O₄@SiO₂@TMC were investigated using Fourier-transform infrared spectroscopy (FTIR), UV-visible spectroscopy, powder X-ray diffraction (XRD), thermogravimetric analysis (TGA), and scanning electron microscopy (SEM). The zeta potential, specific surface area and pore characteristics, and magnetic properties of the final Fe₃O₄@SiO₂@TMC adsorbent were determined using a zeta potential analyser, Brunauer–Emmett–Teller (BET) surface area analyser, and vibrating sample magnetometer (VSM), respectively.

The functional groups of the prepared samples were qualitatively identified using an FTIR spectrometer (Shimadzu IRSpirit-X) over the range of 4000–400 cm⁻¹ using the KBr pellet technique. A UV-vis spectrometer (Shimadzu UV-3600, Japan) equipped with a quartz cuvette was used to determine the initial and residual dye concentrations.

Thermogravimetric analysis (TGA) was performed using a thermogravimetric analyser (Mettler-Toledo, Switzerland) at a heating rate of 10 °C min⁻¹ from 25 to 700 °C under an N₂ atmosphere to evaluate the thermal stability and thermal behavior of the prepared materials.

The surface morphology and elemental composition of the starting materials and final adsorbent were examined using a scanning electron microscope (Thermo Scientific Phenom Pharos G2) equipped with energy-dispersive X-ray spectroscopy (EDX). The specimens were mounted on carbon tape and gold-coated prior to analysis at different magnifications.

The XRD patterns of the intermediates and final product were recorded over a 2θ range of 10–70° with a step size of 0.02°, a fast scanning mode, and an integration time of 0.5 s using a D2 PHASER X-ray diffractometer equipped with Cu K_α radiation (λ = 1.5406 Å). The measurements were conducted at an operating voltage of 40 kV and a current of 30 mA.

The specific surface area and pore characteristics of the final adsorbent were determined using a Brunauer–Emmett–Teller (BET) surface area analyser (Quantachrome TouchWin v1.2.4). The magnetic properties of the samples were evaluated at room temperature using a Lakeshore Model 7404 vibrating sample magnetometer (VSM-7304, USA).

3. Results & discussion

3.1. Fourier transform infrared spectroscopy

The infrared spectra of Fe₃O₄, Fe₃O₄@SiO₂, and Fe₃O₄@SiO₂@TMC are illustrated in Fig. 2a. Bonds around 635 cm⁻¹ and 570 cm⁻¹ are in the fingerprint area and confirm the presence of a Fe–O bond (Fe–O bending vibration),³⁸ while additional peaks at 1642 cm⁻¹ and 1398 cm⁻¹ contribute for the bending vibration of adsorbed water molecules in Fe₃O₄. The asymmetric stretching bonds of Si–O–Si at 1092 cm⁻¹ are linked to a large high-intensity band in Fig. 2a. While the Si–O–Si or O–Si–O bending modes are associated to the intense band at 468 cm⁻¹, the Si–O–Si symmetric stretch is associated to the band at 798 cm⁻¹.³⁹ The band at 570 cm⁻¹ indicates the presence of a Fe–O–Si bond. A minor band at 2930 cm⁻¹ suggests C–H Vibration in TMC modified Fe₃O₄@SiO₂. Furthermore, along with 1642 cm⁻¹ signifying C C stretching vibration in TMC-functionalized SiO₂-decorated Fe₃O₄ magnetic paticles (Fe₃O₄@SiO₂@TMC, organic material). A minor indication around 1723 cm⁻¹ shows a C O carbonyl group associated with carboxyl groups. Two peaks at 1542 cm⁻¹ and 1453 cm⁻¹ correspond with C C stretching vibrations in the aromatic ring. A band for C–O stretching vibrations observed at 923 cm⁻¹.⁴⁰ The 3440 cm⁻¹ wide band represents the stretching vibration mode of O–H functional groups in water. All of this characteristic peak indicates the successful formation of the functional group within the synthesized Fe₃O₄@SiO₂@TMC adsorbent.

Fig. 2. FTIR spectra of Fe₃O₄, Fe₃O₄@SiO₂, Fe₃O₄@SiO₂@TMC (a), XRD difractograms (b) of Fe₃O₄ (i), Fe₃O₄@SiO₂ (ii), Fe₃O₄@SiO₂@TMC adsorbent (iii).

3.2. X-ray diffraction analysis

Fig. 2b represents the XRD difractogram of Fe₃O₄, Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂@TMC. Fig. 2b(i) shows diffraction peaks with 2θ at 30.2°, 35.6°, 43.2°, 53.31°, 57.2°, and 62.8°, corresponding to the crystallographic planes (220), (311), (400), (422), (511) and (440) showing the reference crystalline characteristics of Fe₃O₄ (JCPDS 19-0629).⁴¹Fig. 2b(ii) shows that the XRD pattern of Fe₃O₄@SiO₂ particles is comparable to that of Fe₃O₄ nanoparticles, keeps the distinctive peaks of Fe₃O₄, demonstrating that the silica covering preserves the crystal structure of the core and establishes phase coexistence. Furthermore, in addition to the characteristic diffraction peaks of the constituent materials, a broad diffraction hump centered at 2θ ≈ 22.8° was observed for the Fe₃O₄@SiO₂@TMC adsorbent. This broad feature is indicative of the amorphous nature of the surface coating and supports the successful functionalization of Fe₃O₄@SiO₂ with TMC (Fig. 2b(iii)).⁴²

3.3. Thermogravimetric analysis

Thermogravimetric analysis (TGA) and derivative thermogravimetric (DTG) analysis were performed to compare the thermal behavior of Fe₃O₄, Fe₃O₄@SiO₂, and Fe₃O₄@SiO₂@TMC and to evaluate the changes associated with sequential surface modification (Fig. 3a–c). The progressive changes in the mass–loss profiles provide evidence for the incorporation of the SiO₂ shell and subsequent organic functionalization of the magnetic core.

Fig. 3. TGA and DTG graphs of Fe₃O₄ (a), Fe₃O₄@SiO₂ (b), Fe₃O₄@SiO₂@TMC (c), zeta potential plot of Fe₃O₄@SiO₂@TMC adsorbent (d).

The TGA/DTG profile of pristine Fe₃O₄ (Fig. 3a) shows an overall mass loss of approximately 10.29 wt% up to 800 °C, occurring mainly in two stages. An initial mass loss of 5.84 wt% is observed at lower temperatures and can primarily be attributed to the removal of physically adsorbed water and residual volatile species. A further gradual mass loss of 4.45 wt% occurs at higher temperatures, which may be associated with surface dehydroxylation and the decomposition. The relatively limited mass loss at temperatures above approximately 450 °C indicates the high thermal stability of the Fe₃O₄ framework.

Following silica coating, the Fe₃O₄@SiO₂ sample (Fig. 3b) exhibits a total mass loss of approximately 12.93 wt% up to 800 °C, which is slightly higher than that of pristine Fe₃O₄. The initial mass loss of 9.32 wt% is mainly associated with the removal of physically adsorbed water and volatile species. This relatively larger low-temperature loss compared with Fe₃O₄ is consistent with the hydrophilic nature of the SiO₂ surface and the presence of surface silanol (Si–OH) groups capable of retaining moisture.⁴³ The subsequent mass losses of 2.11 wt% and 1.50 wt% at higher temperatures associated to progressive condensation/dehydroxylation of surface silanol groups and further thermal changes within the silica-coated structure. The relatively small additional mass loss at higher temperatures indicates that the Fe₃O₄@SiO₂ architecture retains good thermal stability. The increase in total weight loss from 10.29 wt% for Fe₃O₄ to 12.93 wt% for Fe₃O₄@SiO₂, together with the altered thermal profile, is consistent with the incorporation of the hydrophilic SiO₂ shell.

In comparison, the Fe₃O₄@SiO₂@TMC adsorbent (Fig. 3c) exhibits a substantially different thermal degradation profile, with a total mass loss of approximately 32.67 wt% between 25 and 700 °C. Three major stages can be distinguished. The first stage, corresponding to a mass loss of 6.86 wt% between approximately 30 and 150 °C, with a DTG maximum at 65.18 °C, is primarily attributed to the removal of physically adsorbed moisture and residual volatile species associated with the silica-containing surface and hydrophilic organic functionalities. The second stage occurs approximately between 170 and 390 °C, with a total mass loss of 9.12 wt% and two DTG maxima at 264.05 and 349.91 °C. These thermal events can be associated with the decomposition of surface-bound organic functionalities and thermally labile oxygen-containing groups (–OH, and –COOH) introduced during surface functionalization,⁴⁴ together with possible condensation/dehydroxylation processes involving surface Si–OH groups. The third and largest mass loss of 16.69 wt%, occurring approximately between 390 and 600 °C with a DTG maximum at 476.69 °C, can be attributed primarily to the progressive thermal decomposition of the more thermally stable organic functionalities associated with the TMC-functionalized surface, including degradation of aromatic and oxygen-containing organic structures.

A clear comparison among the three samples demonstrates the effect of sequential modification. The total mass loss increases from 10.29 wt% for Fe₃O₄ to 12.93 wt% for Fe₃O₄@SiO₂ and subsequently to 32.67 wt% for Fe₃O₄@SiO₂@TMC. The relatively small increase after silica coating is consistent with the incorporation of the inorganic SiO₂ shell and its associated surface hydroxyl groups, whereas the pronounced increase following TMC functionalization indicates a substantially higher contribution from thermally degradable surface-bound organic functionalities. The appearance of distinct DTG events at 264.05, 349.91, and 476.69 °C in the final product, which are not observed with the same intensity in the Fe₃O₄ and Fe₃O₄@SiO₂ profiles, further supports the successful introduction of additional organic functionalities onto the silica-coated magnetic core. Overall, the comparative TGA results demonstrate the progressive modification of Fe₃O₄ through SiO₂ coating and subsequent TMC functionalization while maintaining the thermally stable inorganic Fe₃O₄@SiO₂ framework.

3.4. Zeta potential

The surface charge measurements were performed in the range of pH 3–11, and the findings are displayed against pH in Fig. 3d. The adsorbent exhibits a positive surface charge value (ζ = +30.77 mV) at extremely acidic atmospheres at pH = 3. This might be attributed to the protonation of acyl chloride groups on the surface of the magnetic particles. As pH increases, the particles exhibited a declining pattern in surface charge. The surface charge reduced as pH increases, and the pH_PZC was found at pH ∼ 9.2. Under simple circumstances, the negatively charged surface is created by the deprotonation of the –COCl group, and it rises to a nearly constant value. Ultimately, the surface charge attains its greatest value at pH 11 (ζ = −34.54 mV). The pH experiments are validated by this result. Overall, the magnetic sorbent exhibits a positive zeta potential in acidic environments, which is favorable for the adsorption of negatively charged species and ultimately determines the stability of the adsorbent in solution.

3.5. Vibrating sample magnetometer (VSM) analysis

Fig. 4a exhibits the VSM hysteresis loop of the developed magnetic adsorbent. The magnetic characterization indicares a saturation magnetization (M_s) of approximately 14.75 emu g⁻¹, a relatively low remanent magnetization (M_r) of approximately 0.67 emu g⁻¹, and a low coercivity (H_c) of approximately 16.2 Oe. The narrow hysteresis loop and low coercivity indicate that the synthesized adsorbent exhibits soft magnetic nature with low magnetic remanence. The M_s value confirms the presence and retention of a magnetically responsive phase within the adsorbent and suggests that the adsorbent can be effectively separated from aqueous media using an external magnetic field. The lower M_s compared with bulk or pristine Fe₃O₄ ⁴⁵ can be attributed to the incorporation of organic and inorganic nonmagnetic components and surface functional groups, which mask the relative fraction of the magnetic phase within the composite. The relatively low remanent magnetization minimizes residual magnetization after removal of the external magnetic field, which is favorable for magnetic separation and subsequent regeneration. Overall, the observed magnetic results demonstrate that the synthesized adsorbent possesses favorable magnetic responsiveness for recovery from aqueous media and support its potential application as a magnetically recoverable adsorbent.

Fig. 4. Hysteresis loop of Fe₃O₄@SiO₂@TMC adsorbent (a) BET isotherm diagram and pore size distribution of Fe₃O₄@SiO₂@TMC sorbent (b).

3.6. BET (Brunauer–Emmett–Teller) analysis

The N₂ adsorption–desorption isotherm of the Fe₃O₄@SiO₂@TMC adsorbent is presented in Fig. 4b. The BET adsorption–desorption isotherm of Fe₃O₄@SiO₂@TMC exhibits a classic type IV isotherm, typical of mesoporous substances, showing a rise in N₂ absorption at elevated relative pressures, signifying capillary condensation in the mesopores. The BET-specific surface area of the produced adsorbent was found to be 9.91 m² g⁻¹, whereas the BJH analysis indicated a pore volume of 0.0313 cm³ g⁻¹ and a typical pore diameter of approximately 3.42 nm. The detected hysteresis across the adsorption and desorption phases corresponds with capillary condensation in mesopores and indicates the existence of heterogeneous pore architectures inside the Fe₃O₄@SiO₂@TMC structure. The comparatively wide distribution of pore sizes can enhance the accessibility of the functionalized surface and encourage the diffusion of pollutant molecules to the accessible adsorption sites. The BET findings indicate that Fe₃O₄@SiO₂ exhibits a mesoporous architecture. Despite the composite's comparatively modest specific surface area, its available mesoporous structure, along with the surface characteristics imparted by TMC, can offer favorable active sites for interaction with target pollutant molecules.

3.7. Particle size analysis of adsorbent

The particle size distributions of Fe₃O₄, Fe₃O₄@SiO₂, and Fe₃O₄@SiO₂@TMC are presented in Fig. 5a–c. The average size distribution of Fe₃O₄, Fe₃O₄@SiO₂, and Fe₃O₄@SiO₂@TMC are about 38, 102, and 135 nm, respectively. These findings confirm the successful stepwise formation of the SiO₂ shell and subsequent TMC surface functionalization, in good agreement with the morphological features observed in SEM images.

Fig. 5. Size distribution of (a) Fe₃O₄ NPs, (b) Fe₃O₄@SiO₂ and (c) Fe₃O₄@SiO₂@TMC particles.

3.8. Elemental analysis

Fig. 6 presents the EDX elemental mapping and quantitative composition of the synthesized adsorbent. The results confirm a homogeneous distribution of C (28.3%), O (40.74%), N (2.7%), Si (21.51%), Fe (5.69%), and Cl (1.16%) throughout the material matrix. The relatively high oxygen content is attributed to the presence of Fe–O, Si–O, and carbonyl (C O) functional groups in the constituent components. The detected carbon signal corresponds to C–C and C O bonding environments, while the chlorine content is associated with C–Cl bonds originating from the trimesoyl chloride structure. Furthermore, the presence of Fe and Si indicates the formation of their respective oxide phases, as evidenced by Fe–O and Si–O linkages within the adsorbent.

Fig. 6. Elemental mapping and atomic percentage of Fe₃O₄@SiO₂@TMC adsorbent.

3.9. Morphological analysis

The SEM micrographs show that the adsorbent particles has a well-constructed spherical shape with a range of size increase from 38 to 135 nm. In contrast to previous studies on Fe₃O₄ and Fe₃O₄@SiO₂ (Fig. 7A–D), the spherical formations are homogeneous in size but appear coagulated due to their small shape.⁴⁶ The dispersed morphology and spherical form of the particles suggested that they were porous and amorphous. The changed spherical form has been reported by several studies. Due to the presence of hydroxyl, carbonyl and carboxyl functional groups in the TMC, Fe₃O₄ @SiO₂@TMC adsorbent particles (Fig. 7E and F), the surface of particle appears to have risen after modification. The DLS data shows that size expands after adjustments, and this modified Fe₃O₄ is compared to previously reported Fe₃O₄ results.^47,48

Fig. 7. SEM micrographs of Fe₃O₄ nanoparticles (A and B), Fe₃O₄@SiO₂ particles (C and D), Fe₃O₄@SiO₂@TMC adsorbent (E and F).

3.10. Adsorption studies of Congo red dye

The adsorption performance of Fe₃O₄@SiO₂@TMC toward CR was systematically evaluated through batch experiments, with key operational parameters optimized to maximize removal efficiency. To this end, batch adsorption experiments were conducted to evaluate the performance of the synthesized metal oxide by systematically varying pH, initial dye concentration (C_i), contact time (t), and adsorbent dosage (Fig. 9). The experiment conducted with 10 mL of a 50 mg L⁻¹ CR dye solution, with pH variations ranging from 3 to 11, an adsorbent amount of 2 to 20 mg, and a time interval of 15 to 180 minutes, while the concentration of the CR dye solution varied from 20 to 1200 mg L⁻¹ at room temperature. The adsorbent collected after each test and the residual dye concentrations were quantified using a UV-visible spectrophotometer (Shimadzu UV3600, Japan) at λ_max (497 nm). The following equations were applied for calculating the percentage of removal efficiency (% RE) as well as the adsorption capacity of adsorbent at equilibrium (q_e):²⁶where, C_e (mg L⁻¹) denotes dye concentration at equilibrium, C_i (mg L⁻¹) for initial dye concentration, q_e is amount adsorbed at equilibrium(mg g⁻¹), V for solution volume (L) and m (mg) for adsorbent dosage.

Fig. 9. Effect of pH on q_e (a), %RE (b), effect of adsorbent dose on q_e (c), %RE (d).

The adsorption isotherm and kinetic parameters were determined by nonlinear fitting of the experimental data to the corresponding models listed in Table 1 and 2. In particular, the maximum adsorption capacity (q_m) was obtained as the fitted parameter of the corresponding isotherm model, rather than being directly calculated from a single experimental data point.

Table 1. Isotherm parameters obtained from nonlinear fitting of the experimental CR adsorption data using selected models.

Model	Equation	Parameter	CR	References
Langmuir		q e (mg g−1)	4614.61	49
		q m (mg g−1)	5885.98
		K L (L mg−1)	0.015
		R L (Ci = 100 mg L−1)	0.666
		R 2	0.994
Freundlich		K f (mg g−1) (L mg−1)1/n	526.56	50
		1/n	0.410
		R 2	0.962
Temkin		B T (kJ mol−1)	2.60	51
		A T (L g−1)	0.337
		R 2	0.903
Koble–Corrigan		A KC (Ln mg1−n g−1)	256.250	52
		B KC (L mg−1) n	0.048
		n KC	0.756
		R 2	0.995
Jovanović	q e = qm(1 − ekJCe)	q m (mg g−1)	4100	53
		k J (L g−1)	0.020
		R 2	0.960

Table 2. Kinetic parameters of CR adsorption onto adsorbent.

Model	Equation	Parameters	CR	References
Pseudo first order	q t = qe(1 − exp−k1t)	q t (mg g−1)		54
		q cal (mg g−1)	231.761
		k 1 (min−1)	0.040
		R 2	0.880
Pseudo second order		q cal (mg g−1)	263.92	55
		k 2 (g mg−1 min−1)	1.153
		R 2	0.961
Intra particle diffusion	q t = kidt0.5 + C	k id (mg g−1 min0.5)	28.757	56
		C (Mg g−1)	8.777
		R 2	0.990
		k id (mg g−1 min0.5)	6.828
		C (Mg g−1)	159.277
		R 2	0.918
Bangham	q t = qm[1 − exp(−kbtn)]	q m (mg g−1)	231.949	57
		k b	0.054
		n	0.932
		R 2	0.999
Elovich kinetic model		α El (mg g− 1 min− 1)	0.022	58
		β EI (g mg−1)	66.046
		R 2	0.973

3.10.1. Effect of pH on CR adsorption

The CR dye indicates a pH-dependent colour change; in acidic conditions, the dye molecule becomes protonated, resulting in a blue hue (below pH 5.0). With increasing pH, the dye molecule undergoes deprotonation, yielding a red hue (above pH 5.0). Additionally, the adsorption of Congo red dye onto surfaces is influenced by the pH of the substance. Both the surface charge of the adsorbent and the structure of CR are influenced by pH, which leads to the pH dependent adsorption process. In very low acidic environment (below pH = 5) Congo red is positively charged (–SO₃H, NH₂⁺) but with increasing pH, CR converted to negatively charged (SO₃⁻, NH₂) form and can be adsorbed onto materials that are positively charged when the environment is acidic.^17,59 On the other hand, when the environment is alkaline, both Congo red and the adsorbent can be negatively charged, which results in electrostatic repulsion and decreased adsorption. Strong effects on the adsorption process, adsorption active sites, and physiochemical interactions between the adsorbent and dye molecule are caused by the pH of the target dye solution and the adsorbents surface charge.⁶⁰

Zeta potential measurements ranging across the pH range of 3–11 were used to examine the surface charge properties of the Fe₃O₄@SiO₂@TMC adsorbent (Fig. 3d). The zeta potential steadily dropped from +30.8 mV at pH 3 to +26.7 mV at pH 5, +13.6 mV at pH 7, and +2.1 mV at pH 9 as the pH of the solution rose. At pH 11, it became significantly negative (−34.8 mV). The shift from positive to negative surface charge shows that the adsorbent's point of zero charge (pH_PZC) is 9.2.

To examine the adsorption of Congo red dye, a pH range of 3 to 11 was adjusted. 2 mg of adsorbents were added to 10 ml of a 50 mg L⁻¹ dye solution, which was then stirred for 180 minutes. The HCl and NaOH solutions (0.1 mol L⁻¹), were used in order to get an adequate pH. Fig. 9a and b illustrate that, raising the pH results in an increase in the % RE and the q_e values for CR up to a pH of 5 having %RE is 93.74%.

Under these conditions, strong electrostatic attraction occurs between the positively charged Fe₃O₄@SiO₂@TMC surface and the negatively charged CR molecules, owing to the sulfonate (–SO₃⁻) groups present in the anionic azo dye. The enhanced adsorption observed at pH ≈ 5 can be attributed to the combined effects of electrostatic attraction, hydrogen bonding, and π–π interactions between the aromatic rings of CR and the aromatic functionalities of the TMC-modified surface (Fig. 8).

Fig. 8. Effect of pH on the dye structure and interactions.

3.10.2. Effect of adsorbent dosage on CR dye adsorption

Adsorbent dosage is a critical parameter for evaluating adsorption performance at a fixed initial dye concentration. Batch experiments were conducted under optimized pH conditions using 10 mL of Congo red solution (50 mg L⁻¹) over a contact time of 180 min to assess the effect of adsorbent dose on removal efficiency and adsorption capacity. Fig. 9c and d illustrate the effect of adsorbent dosage on the q_e and RE% of CR. Increasing the adsorbent dosage from 2 to 20 mg resulted in a marginal enhancement in RE%, rising from 93.33% to 98.78% within 180 min. This trend is attributed to the increased availability of vacant active sites at higher dosages. However, a quicker attaining of equilibrium results from fewer interactions with the dye molecules at low doses. The acquiring of equilibrium is slowed down as the dose is increased because there are more effective interactions with the dye molecule (Fig. 9). Accordingly, 2 mg was selected as the optimal adsorbent dosage for subsequent experiments.⁶¹

3.10.3. Effect of concentration and isothermal study

Concentration-dependent adsorption experiments were conducted to evaluate the effect of the initial CR concentration on adsorption performance over the range of 20–1200 mg L⁻¹ at pH 5, with an equilibrium time of 180 min, and an adsorbent dosage of 2 mg. The corresponding adsorption capacities are presented in Fig. 10a. As the initial CR concentration increased, the q_e increased markedly from 93.85 to 4614.62 mg g⁻¹, which is attributed to the increased concentration gradient and the resulting stronger driving force for mass transfer between the bulk solution and the adsorbent surface. At lower CR concentrations, a large proportion of the available adsorption sites remained unoccupied, allowing the dye molecules to interact readily with the accessible active sites favours the high RE%. With increasing CR concentration, the number of dye molecules available for adsorption increased, resulting in progressively greater uptake per unit mass of adsorbent. In contrast, the RE% gradually decreased from 93.85% at 20 mg L⁻¹ to 76.91% at 1200 mg L⁻¹. This decrease can be attributed to the progressive occupation of available adsorption sites and the increasing competition among CR molecules for the remaining accessible sites as the adsorbent approached saturation. Notably, even at 1000 mg L⁻¹, the adsorbent exhibited a high removal efficiency of 86.07%, corresponding to a q_e of 4303.59 mg g⁻¹, demonstrating its substantial adsorption capacity toward CR. At the highest investigated concentration of 1200 mg L⁻¹, the q_e reached 4614.62 mg g⁻¹, while the RE% declined to 76.91%, indicating that the available adsorption sites were progressively approaching saturation.⁶²

Fig. 10. Effect of C_i on adsorption capacity of CR onto adsorbent (a) (m = 2 mg, V = 10 mL, C_i = 20–1200 mg L⁻¹, t = 180 min, optimal pH); isothermal fitting plot of Langmuir and Freundlich (b); Temkin, K–C, and Jovanovic isotherm models (c); effect of time on adsorbrnt performance (d) (m = 2 mg, V = 10 mL, C_i = 50 mg L⁻¹, t = 15–180 min, optimal pH); kinetics fitting plot for PFO, PSO, and Elovich (e); Banhgam and IPD, models (f).

To analyze the adsorption mechanism, the concentration-dependent data were fitted using the nonlinear Langmuir, Freundlich, Temkin, Koble–Corrigan, and Jovanović models. The estimated parameters and fitting curves are presented in Table 1 and Fig. 10.

The Langmuir model assumes that adsorption occurs through monolayer coverage of energetically equivalent and homogeneous adsorption sites, with no lateral interactions between adsorbed molecules.⁴⁹ The Langmuir adsorption isotherm equation has been represented in Table 1 (ref. 49) where, C_e is the equilibrium concentration of dye in solution (mg L⁻¹), q_m is the maximal adsorption capacity (mg g⁻¹), q_e is the quantity of adsorbed dye at equilibrium (mg g⁻¹), and K_L is a Langmuir coefficient. The Langmuir model provided an excellent description of the CR adsorption equilibrium data, yielding a high coefficient of determination (R² = 0.994). The calculated maximum adsorption capacity was 5885.98 mg g⁻¹, which is higher than the experimentally obtained equilibrium adsorption capacity of 4614.61 mg g⁻¹ at 1200 mg L⁻¹. The higher fitted (q_m) indicates that complete saturation of the available adsorption sites was not reached within the investigated concentration range. Furthermore, the Langmuir equilibrium constant (K_L) was 0.015 L mg⁻¹. The R_L value of 0.666 indicates that the adsorption of CR onto the adsorbent is favorable according to the Langmuir criterion (Fig. 10b). The high R² value and favorable R_L range indicate that the Langmuir model adequately represents the equilibrium adsorption behavior of CR, although the model fit alone should not be considered definitive evidence of a perfectly homogeneous surface or an exclusively monolayer adsorption mechanism (Fig. 10b).

The Freundlich isotherm is an empirical model commonly used to describe adsorption on heterogeneous surfaces with a non-uniform distribution of adsorption energies and adsorption sites. Table 1 displays the Freundlich adsorption isotherm equation.⁵⁰ where, C_e is the concentration of dye in solution at equilibrium (mg L⁻¹), q_e is the amount of dye that has been adsorbed at equilibrium (mg g⁻¹), and K_f is a Freundlich constant. The Freundlich model yielded a 1/n value of 0.410, which is within the range of 0–1 and indicates favorable adsorption with decreasing adsorption affinity as the surface becomes progressively occupied. The K_F value was 526.56 (mg g⁻¹)(L mg⁻¹)^1/n, while the model provided an R² value of 0.962. The 1/n < 1 value suggests that the CR molecules interact with adsorption sites possessing different affinities, which is consistent with the heterogeneous nature of the adsorbent surface. Thus, the Freundlich model indicates favorable adsorption involving a distribution of adsorption energies and non-equivalent active sites (Fig. 10b).

According to the Temkin isotherm, the heat of adsorption decreases linearly with increasing surface coverage due to progressive adsorbent–adsorbate interactions. The Temkin adsorption isotherm equation is shown in Table 1.⁵¹ Here, R and T are universal gas constant (8.314 J Kmol⁻¹) and temperature, q_e is the uptake capacity of adsorbent (mg g⁻¹), B_T and A_T are binding energy constant (J mol⁻¹) and Temkin isotherm constant (L g⁻¹). The Temkin model yielded an R² value of 0.903 for CR, indicating a relatively weak correlation with the experimental equilibrium data. The calculated B_T value was 2.60 kJ mol⁻¹, indicating a positive heat-related Temkin parameter and suggesting that the adsorption process is energetically favorable. However, the relatively low R² value indicates that the Temkin model does not satisfactorily define the equilibrium adsorption behavior of CR (Fig. 10c).

The Koble–Corrigan isotherm combines the characteristics of the Langmuir and Freundlich models and is commonly applied to describe adsorption systems involving heterogeneous surfaces and non-ideal adsorption behavior. The K–C model provided an excellent fit to the experimental CR adsorption data (Table 1), with an R² value of 0.995. Here, A_KC and B_KC, and referes Koble–Carrigan's isotherm constant and n_KC is the exponent of Koble–Carrigan's isotherm.⁵² The model yielded (A_KC), (B_KC) and (n_KC) values of 256.250 (Lⁿ mg¹⁻ⁿ g⁻¹), 0.048 (L mg⁻¹)ⁿ, and 0.756, respectively. The (n_KC) value below unity indicates deviation from ideal Langmuir-type behavior and suggests that the adsorption process is influenced by surface heterogeneity. The high R² value demonstrates that the Koble–Corrigan model can satisfactorily represent the equilibrium adsorption behavior of CR over the investigated concentration range (Fig. 10c).

The Jovanović isotherm is a modified Langmuir-type model that describes adsorption on a finite number of adsorption sites while considering interactions between the adsorbate and adsorbent surface. Table 1 displays the Jovanović adsorption isotherm equation. Where, q_e is amount of adsorbate in the adsorbent at equilibrium (mg g⁻¹), q_m is maximum uptake of adsorbate, and K_j is Jovanovic constant.⁵³ The Jovanović model yielded a q_m of 4100 mg g⁻¹ and an equilibrium constant (k_J) of 0.020 L g⁻¹, with an R² value of 0.960. Notably, the estimated q_m of 4100 mg g⁻¹ showed a relatively close correspondence with the experimentally obtained maximum adsorption capacity of 4614.61 mg g⁻¹, with a difference of approximately 11.2%. This agreement suggests that the Jovanović model provides a reasonable estimation of the adsorption capacity within the investigated concentration range. However, the moderate R² value indicates that the model does not describe the overall equilibrium data as closely as the better-fitting isotherm models. Thus, while the Jovanović model provides a comparatively close estimation of the q_max, its overall fit proposes that adsorption behavior is not fully captured by this model (Fig. 10c).

3.10.4. Effect of contact time and kinetic study

The effect of contact time on the adsorption performance of Congo red (CR) onto Fe₃O₄@SiO₂@TMC was evaluated at a fixed initial CR concentration of 50 mg L⁻¹, an adsorbent dosage of 2 mg, and pH 5. The adsorption process was rapid during the initial stage, with the q_t increasing from 108.46 mg g⁻¹ at 15 min to 212.74 mg g⁻¹ at 60 min, accompanied by an increase in RE% from 43.38% to 85.10% (Fig. 10d). This rapid initial uptake is attributed to the abundance of readily accessible vacant adsorption sites and the strong concentration gradient between CR molecules in the solution and the adsorbent surface. With further extension of the contact time, the adsorption rate gradually decreased, as reflected by the smaller increase in q_t from 212.74 mg g⁻¹ at 60 min to 224.87, 227.95, 231.28, and 233.85 mg g⁻¹ at 90, 120, 150, and 180 min, respectively (Fig. 10d). The corresponding RE% increased from 85.10% at 60 min to 93.54% at 180 min (Fig. 10d). The progressive reduction in adsorption rate is credited to the gradual occupation of readily available adsorption sites and the decreasing concentration gradient as equilibrium was approached. The continued increase in adsorption capacity after 60 min suggests that CR molecules progressively migrated toward less accessible adsorption sites and pores within the adsorbent. The adsorption process therefore involved a rapid initial uptake followed by a slower stage associated with diffusion into the internal structure and progressive occupation of the remaining adsorption sites, with equilibrium approached at approximately 180 min.⁶³

When describing the rate at which an adsorbate is adsorbed onto an adsorbent, kinetic research is the method of choice. This method focuses on the rate and functioning of the process, which is key to analysing and optimizing adsorption.

The pseudo-first-order (PFO) kinetic model simplifies reaction rates by assuming that a single step controls adsorption and follows first-order kinetics. In cases where one reactant has a much greater concentration than the others, this model is employed to maintain its concentration.⁵⁴

Table 2 contains PFO kinetic model equation.⁵⁴ Where, q_e represents the amount of adsorbate adsorbed at equilibrium (mg g⁻¹), q_t represents the amount of dye adsorbed at any time t (mg g⁻¹), k₁ represents the PFO rate constant, t Represents time (min). The pseudo-first-order (PFO) kinetic model yielded a calculated equilibrium adsorption capacity (q_cal) of 231.76 mg g⁻¹ and a k₁ of 0.040 min⁻¹, with an R² value of 0.880. The relatively low correlation coefficient indicates that the PFO model does not adequately fits the kinetic behavior of CR adsorption. Moreover, the calculated q_cal was slightly lower than the experimentally obtained equilibrium capacity of 233.85 mg g⁻¹, with a difference of approximately 0.9%. Although the close agreement between the calculated and experimental capacities suggests that the PFO model reasonably estimates the equilibrium uptake, its relatively lower R² indicates that it does not satisfactorily explains the overall adsorption kinetics (Fig. 10e).

The pseudo-second-order (PSO) kinetic model is commonly employed to describe adsorption processes governed by chemisorption, where the rate-limiting step involves valence forces and typically requires longer contact times for effective adsorption. The adsorption rate is presumed to be affected by the interaction between adsorption sites on the surface of the adsorbent and the adsorbate during the adsorption process.

The pseudo-second-order (PSO) kinetic model equation stands in Table 2.⁵⁵ Here, q_e represents the quantity of adsorbate adsorbed at equilibrium (mg g⁻¹), q_t represents the amount of dye adsorbed at a time t (mg g⁻¹), k₂ is the rate constant for pseudo-second-order model, t is time (min). PSO model resulted in q_cal of 263.92 mg g⁻¹ and a k₂ of 1.153 g mg⁻¹ min⁻¹, with an R² value of 0.961. The relatively high correlation coefficient indicates that the PSO model provides a favorable explanation for CR adsorption kinetics. However, the calculated equilibrium capacity is higher than the experimentally obtained value of 233.85 mg g⁻¹, with a difference of approximately 12.9%. Although the relatively good PSO fit indicates that the model describes the kinetic data reasonably well, the PSO fit alone does not establish a chemisorption-dominated mechanism. Such a mechanistic assignment requires complementary kinetic models (Fig. 10e).

It is possible to estimate the amount of mass and surface transfer, as well as both the activation and deactivation energy of a system using the Elovich model. It also helps to assess whether or not chemisorption is taking place on the heterogeneous surface of the adsorbent, resulting in another perk.⁵⁸

Table 2 represents the equation for Elovich kinetic model.⁵⁸ Where, where, q_t represents the amount of dye adsorbed at a time t (mg g⁻¹), the constant α is associated with the initial rate of adsorption, the constant β is associated with the desorption process, t is time (min). The Elovich kinetic model provided a good description of the CR adsorption kinetics, with an R² value of 0.973. The calculated α_E and β_E were 0.022 mg g⁻¹ min⁻¹ and 66.046 g mg⁻¹, respectively. The relatively high β_E value indicates a progressive decrease in the adsorption rate as the available adsorption sites became increasingly occupied, consistent with the heterogeneous nature of the adsorbent surface. Although the good Elovich fit suggests that surface heterogeneity and changes in adsorption energy may contribute to the adsorption process, the model alone cannot be considered definitive evidence of chemisorption (Fig. 10e).⁶⁴

The intra-particle diffusion model is useful for understanding and quantifying the function that diffusion plays within pores throughout the adsorption process, as well as for interpreting the rate regulating step when it occurs.

The intraparticle diffusion analysis revealed a nonlinear, multistage adsorption process with two distinguishable diffusion regions (Table 2).⁵⁶ Where, q_t represents the quantity of dye adsorbed at a time t (mg g⁻¹), k_id is the intra-particle diffusion rate constant, t^0.5 is the square root of time, C is the constant related to initial adsorption rate. The first stage (15–45 min) exhibited a higher k_id value of 28.757 mg g⁻¹ min^−0.5 and a stronger correlation (R² = 0.990) than the second stage (45–90 min), where k_id decreased to 6.828 mg g⁻¹ min^−0.5 (R² = 0.918). The higher k_id during the initial stage indicates faster mass transfer, whereas the subsequent decrease reflects progressively slower diffusion as the readily accessible adsorption sites became occupied. The non-zero intercepts of both regions indicate that intraparticle diffusion was not the sole rate-controlling process, suggesting the contribution of other mass-transfer and surface-adsorption processes during CR uptake (Fig. 10f).

The Bangham kinetic model exhibited the best correlation with the experimental CR adsorption data. The Bangham kinetic model equation (Table 2) yielding an R² value of 0.999. Where, q_t represents the amount of dye adsorbed at a time t (mg g⁻¹), q_m maximum adsorption capacity, the constant α is associated with the initial rate of adsorption, k_b is the Bangham constant, t is time (min).⁵⁷ The q_m was 231.95 mg g⁻¹, which is in very close proximity with the experimentally obtained equilibrium capacity of 233.85 mg g⁻¹, with a difference of only approximately 0.8%. The Bangham constant (k_b) and n were 0.054 and 0.932, respectively. The excellent agreement between the calculated and experimental adsorption capacities, together with the very high R², indicates that the Bangham model satisfactorily defines the adsorption mechanism of CR. The results further suggest that pore diffusion and the progressive penetration of CR molecules into the porous structure contribute substantially to the overall adsorption process.

Overall, the kinetic analysis indicates that CR adsorption onto Fe₃O₄@SiO₂@TMC involves a multistage adsorption process rather than a single rate-controlling mechanism. The Bangham model provided the best fit to the experimental data (R² = 0.999) and yielded a calculated equilibrium capacity (231.95 mg g⁻¹) very close to the experimental value (233.85 mg g⁻¹), indicating a substantial contribution of pore diffusion to the overall adsorption process. The intraparticle diffusion analysis further supported a multistage process, with faster diffusion during the initial stage followed by a slower diffusion stage as the readily accessible adsorption sites became progressively occupied. The PSO model also provided a decent fit (R² = 0.961), while the PFO model showed a weaker correlation (R² = 0.880). The good fit of the Elovich model (R² = 0.973) additionally suggests the contribution of heterogeneous adsorption sites and progressively changing adsorption rates. Collectively, these results indicate that pore diffusion, surface adsorption, and boundary-layer/mass-transfer effects collectively contribute to CR uptake, rather than a single mechanism governing the overall adsorption kinetics.

3.10.5. Effect of salt concentration on adsorption performance

Fig. 11a illustrates the effect of ionic strength on the adsorption performance of the adsorbent. NaCl, CaCl₂, and Na₂SO₄·2H₂O solutions were selected to assess the impact of ionic strength on the adsorption of Congo red by Fe₃O₄@SiO₂@TMC magnetic particles. The removal efficiency reduced as the salt concentration rose from 5 to 40 g L⁻¹, demonstrating that adsorption is sensitive to ionic strength. Compression of the electrical double layer and shielding of the electrostatic attraction between the anionic Congo red molecules and the positively charged Fe₃O₄@SiO₂@TMC surface are responsible for the decrease in adsorption. Due to the comparatively poor screening effect of monovalent Na⁺ and Cl⁻ ions, NaCl showed the least impact among the salts under investigation, with the removal efficiency only declining from 88.41 to 79.18%. Na₂SO₄·2H₂O, on the other hand, induced a more noticeable decrease (87.69–69.23%) because the divalent sulphate ions enhanced the ionic strength of the solution and more successfully competed with Congo red anions for positively charged adsorption sites. CaCl₂ showed the more pronounce reduction (86.26–59.08%), which is explained by the substantial electrostatic shielding effect of Ca²⁺ ions due to their higher charge density and valency. Additionally, Ca²⁺ may interact with Congo red's sulfonate groups, decreasing the dye's affinity for the adsorbent surface. Significant adsorption capacity was maintained even at the highest salt concentration despite the inhibitory effect of electrolytes. This suggests that, in addition to electrostatic attraction, hydrogen bonding, π–π interactions between the aromatic trimesoyl chloride moieties and Congo red molecules, and van der Waals interactions also play a major role in the adsorption mechanism.

Fig. 11. (a) Effect of salt concentration on %RE of Congo red (2 mg of adsorbent, 10 mL of 50 mg L⁻¹ dye solution at optimum pH 5 and 60 min), (b) reusability study of Fe₃O₄@SiO₂@TMC adsorbent.

3.10.6. Regeneration and reusability

Regeneration capacity is a key factor determining the practical applicability of adsorption-based water treatment systems. To evaluate reusability, the adsorption–desorption performance of Fe₃O₄@SiO₂@TMC was investigated over six consecutive cycles. In each cycle, 5 mg of the adsorbent was introduced into the 10 mL of Congo red solution (100 mg L⁻¹) under continuous stirring for 2 h. The residual dye concentration in the filtrate was then determined using UV-vis spectroscopy. The recovered Fe₃O₄@SiO₂@TMC was regenerated by sequential washing with a water/methanol mixture, followed by brief agitation to facilitate the desorption of retained CR. The regenerated material was subsequently rinsed with distilled water, dried overnight at 60 °C, and reused in the next adsorption–desorption cycle.

As shown in Fig. 11b, the removal efficiency (RE%) decreased from 93.74% in the first cycle to 88.39%, 80.86%, and 75.09% in the second, third, and fourth cycles, respectively. Relative to the first cycle, the adsorbent retained 86.26% and 80.10% of its initial removal performance after the third and fourth cycles, respectively. The gradual decrease in adsorption performance may be associated with incomplete desorption of CR from the available adsorption sites and/or minor loss or alteration of accessible surface functionalities during repeated regeneration. However, these factors were not directly investigated in the present study and should therefore be considered possible rather than definitive explanations.

Although the adsorbent retained appreciable CR removal performance after four cycles and could be conveniently recovered using an external magnetic field, the observed regeneration stability was acceptable (80.10% of initial performance), highlighting the need for further optimization of the regeneration strategy to improve long-term reusability. It should also be noted that the regeneration performance was evaluated using a relatively high initial CR concentration of 100 mg L⁻¹ and an adsorbent dosage of 0.5 g L⁻¹ (5 mg in 10 mL). Therefore, meaningful comparison of the reusability of Fe₃O₄@SiO₂@TMC with previously reported adsorbents should consider the initial pollutant concentration, adsorbent dosage, contact time, regeneration/desorption protocol, and number of cycles, as these experimental factors can substantially influence the apparent retention of adsorption performance. Further optimization of the regeneration conditions is required to improve the long-term reusability of the developed sorbent.

3.10.7. Comparison of synthesized Fe₃O₄@SiO₂@TMC adsorbent with previous reports

The q_max values of several reported adsorbents and developed Fe₃O₄@SiO₂@TMC adsorbent compared in Table 3. This comparison shows that the adsorbent exhibits superior adsorption capacity than the other documented adsorbent and can be considered as a potential adsorbent to ensure effective remediation in the CR dye removal process. The enhanced adsorption capacity can be credited to the abundance of surface functional groups, particularly –OH, –COOH and –C O functional groups as well as benzene rings, which render the adsorbent capable of engaging in multiple interaction mechanisms, including hydrogen bonding, electrostatic attraction, and hydrophobic interactions.

Table 3. Synthesised Fe₃O₄@SiO₂@TMC adsorbent comparison with earlier studies.

Adsorbent	Adsorbate	q max (mg g−1)	References
FexCo3−xO4 NPs	Congo red	128.6	65
Graphene oxide-CuFe2O4 nanohybrid	Congo red	114.21	66
Fly-ash@Fe3O4	Congo red	153	67
Carbon derived from neem bark	Congo red	25.84	68
Heterophased iron oxides/carbon composite	Congo red	945.84	69
Magnetic iron oxide Kaolinite nanocomposite	Congo red	45.59	70
Mn-doped ZnO NPs	Congo red	232.5	71
Fe2O3-MgO nanorods	Congo red	200	72
GO/MIL-100 (Fe)	Congo red	119	73
Iron oxide NPs (from Piper chaba steam extract)	Congo red	88	74
CC(cotton calyx)-Fe3O4	Congo red	20.66	75
MnFe2O4@MOF-5	Congo red	100	76
NiFe2O4 nanofibers	Congo red	65	77
Activated carbon (from Juniperus Communis seeds)	Congo red	157.13	78
CoMnOx NPs	Congo red	218.54	79
ZnO-bentonite nanocomposite	Congo red	71.43	80
Fe–Sn–Ni ternary metal oxide nanocomposite	Congo red	76.87	81
Fe3O4@SiO2@TMC	Congo red	4614.62	This work

4. Conclusion

This work presents a TMC-functionalized SiO₂-decorated Fe₃O₄ magnetic adsorbent. Several characterisation techniques, including FTIR, XRD, DLS, TGA, SEM, elemental analysis, and zeta potential analysis, were used to investigate the physiochemical characteristics of the adsorbent. The batch adsorption experiment's results showed that, under ideal conditions, adsorbent reached a maximum capacity of 4614.62 mg g⁻¹ at pH 5 using 2 mg adsorbent dose. It was discovered that the adsorption of CR driven by hydrogen bonds, electrostatic interactions, and π–π interactions. The Langmuir isotherm model shows a high correlation coefficient (R² = 0.994) but devites experimental adsorption capacity (5885.98 mg g⁻¹). Whereas, The Koble–Corrigan isotherm exhibit an excellent fit of the adsorption data with an R² value of 0.995, describing the non-ideal Langmuir model with heterogeneous surface. On the otherhand Jovanović isotherm generated an almost perfect match of q_max (4100 mg g⁻¹) with a decent R² value (0.960). Thus, it may infer that a multimechanistic strategy is suggested by isotherm models. In case of kinetic studies, the Bangham model produced a calculated equilibrium capacity (231.95 mg g⁻¹) that was highly comparable to the experimental value (233.85 mg g⁻¹) and offered the best fit to the experimental data (R² = 0.999), supports the pore diffusion. The analysis of intraparticle diffusion provided additional evidence for a multistage process, with faster diffusion in the first stage. Pseudo second order(PSO) and Elovich models also provided a good fit with R² = 0.961 and 0.973 respectively.The developed Fe₃O₄@SiO₂@TMC retained appreciable adsorption performance over four regeneration cycles, with the retaind removal efficiency of 80.10% after 4 cycle. These results demonstrate the effectiveness of the proposed approach in developing an efficient and recyclable adsorbent. Our current study provides a potential magnetic porous adsorbent with superior adsorption capacity, high removal efficiency, and reusability, highlighting its potential applicability in practical water treatment. Future research will be concentrated on improving regeneration procedures and adsorption efficiency across a broader range of pollutants.

Author contributions

DP: methodology, experimental, data analysis, investigation, validation, writing – original draft, review & editing, EM: data validation, data analysis, writing – original draft, review & editing, SC: investigation, data analysis, AB: conceptualization, methodology, data validation, writing – review & editing, funding acquisition, supervision, instrumental facilities.

Conflicts of interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Dr Atanu Banerjee reports financial support was provided by Council for Scientific and Industrial Research. Dr Atanu Banerjee reports equipment was provided by India Ministry of Science & Technology Department of Science and Technology. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data can be made available on request