Fe₃O₄/SiO₂ Nanocomposite Derived from Coal Fly Ash and Acid Mine Drainage for the Adsorptive Removal of Diclofenac in Wastewater

Abstract

The ubiquity of diclofenac (DCF) in the environment has raised significant concerns. Diclofenac is a non-steroidal anti-inflammatory drug that has been found in various environmental matrices at minimum concentrations that are harmful to aquatic and terrestrial organisms. Traditional wastewater treatment plants (WWTPs) are not fully equipped to remove a range of pharmaceuticals, and that explains the continued ubiquity of DCF in surface waters. In this study, an Fe₃O₄/SiO₂ nanocomposite prepared from acid mine drainage and coal fly ash was applied for the removal of DCF from wastewater. Major functional groups (Si–O–Si and Fe–O) were discovered from FTIR. TEM revealed uniform SiO₂ nanoparticle rod-like structures with embedded dark spherical nanoparticles. SEM-EDS analysis discovered a sponge-like structure fused with Fe₃O₄ nanoparticles that had significant Si, O, and Fe content. XRD demonstrated the crystalline nature of the nanocomposite. The surface properties of the nanocomposite were evaluated using BET and were 67.8 m²/g, 0.39 cm³/g, and 23.2 nm for surface area, pore volume, and pore size, respectively. Parameters that were suspected to be affecting the removal process were evaluated, including pH, nanocomposite dosage, and sample volume. The detection of DCF was conducted on high-performance liquid chromatography with diode-array detection (HPLC-DAD). Under optimum conditions, the adsorption process was monolayer, and physisorption was described using the Langmuir and Dubinin-Radushkevich (D-R) isotherm models. The kinetic data best fitted the pseudo-first order kinetic model, indicating a physisorption adsorption process. The thermodynamic experimental data confirmed that the adsorption process was spontaneous. The results obtained from real water samples showed 95.28% and 97.44% removal efficiencies from influent and effluent: 94.83% and 88.61% from raw sewage and final sewage, respectively. Overall, this work demonstrated that an Fe₃O₄/SiO₂ nanocomposite could be successfully prepared from coal fly ash and acid mine drainage and could be used to remove DCF in wastewater.

1. Introduction

Deterioration of water quality has been attributed to the presence of residues of pesticides, industrial solvents, personal care products, and pharmaceuticals, among other commonly detected emerging contaminants [1]. Among the pharmaceutical residues, diclofenac (DCF) is present—a common non-steroidal anti-inflammatory drug for mild to severe pain, including migraine and toothache. In addition, DCF is used as an analgesic and anti-arthritic compound. Diclofenac has been found in amounts between ng/L and µg/L in environmental water and may pose a danger to aquatic life and human health [2,3]. This drug mostly enters the aquatic environment and groundwater through direct human excretion, leakage of sewage pipelines, and partially treated wastewater [4,5]. Therefore, researchers need to ensure that DFC is removed from water points since it is toxic to aquatic organisms and a threat to human health.

Recently, several techniques including reverse osmosis [6], biodegradation [7], filtration [8], photodegradation [9], photocatalysis [10], oxidation [11], and bank filtration [12] were used for the removal of DCF from water. Nevertheless, most of these techniques operate at a high cost and are ineffective. Consequently, it is necessary to enhance treatment processes by removing DCF. The adsorption method can be used to reduce these types of contaminants [13]. Adsorption is preferred since it is easy to operate, environmentally friendly, and inexpensive [14].

A wide range of adsorbent materials, such as rice husk biochar [15], mesoporous silica nanospheres [16], organoclay [17], and molecularly imprinted polymers [18] have been studied for their potential to remove DCF from wastewater. The adsorption of DCF onto activated carbon (AC) derived from the husk of einkorn (Triticum monococcum L.) was investigated, and the maximum adsorption capacity was found to be 147.06 mg/g [19]. Guava seed AC-loaded calcium alginate aerogel was applied for the removal of DCF, which had 489.97 mg/g adsorption capacity [20]. Moreover, a nanoporous carbon adsorbent produced from red pepper (Capsicum annuum L.) showed 303.0 mg/g adsorption capacity on the adsorption of DCF from water [21].

Silica nanoparticles (SiO₂ NPs) are known to be stable with a high surface area, which makes SiO₂ NPs a good adsorbent [22]. However, SiO₂ NP adsorption capacity for DCF is generally limited since there is a lack of necessary adsorptive sites for effective interaction with pollutant molecules [23,24]. On the other hand, Fe₃O₄ exhibits excellent magnetic properties, facilitating easy separation of the adsorbent from the aqueous solutions. Nonetheless, Fe₃O₄ alone may agglomerate and lower the surface area [25], consequently reducing the effectiveness of Fe₃O₄ as an adsorbent. Integrating SiO₂ NPs with Fe₃O₄ forms a composite that results in a high surface area and with magnetic properties [26]. This composite can provide more adsorptive sites for adsorption. It is crucial to explore various waste materials and convert them into valuable products. High surface area and magnetic properties allow the Fe₃O₄/SiO₂ composites to effectively remove organic contaminants such as dyes, heavy metals, and pharmaceuticals, which could work well for DCF removal from water [27,28]. Furthermore, Fe₃O₄/SiO₂ nanocomposites have been broadly explored for drug delivery [29].

In the present work, an Fe₃O₄/SiO₂ nanocomposite was successfully prepared from coal fly ash (CFA) and acid mine drainage (AMD). The Fe₃O₄/SiO₂ nanocomposite exhibited functional groups, as reflected on the FTIR, that interact with DCF molecules during the removal process. Consequently, numerous optimization parameters, isotherms, kinetics, thermodynamics, and adsorption mechanisms, together with desorption studies and regeneration cycles, were analyzed to understand the potential, performance, recovery, and possibility for further reuse of the Fe₃O₄/SiO₂ nanocomposite.

2. Results and Discussion

2.1. Characterization of SiO₂ NPs and Fe₃O₄/SiO₂

2.1.1. Fourier Transform Infrared Spectroscopy

The FTIR spectra of SiO₂ NPs, Fe₃O₄ NPs, and Fe₃O₄/SiO₂ nanocomposites are shown in Figure 1, which reveal major functional groups. Strong and broad bands at 3500 cm⁻¹ in Figure 1a,b were ascribed to the silanol, Si–OH [30]. The shoulder peak appearing around 3100–3200 cm⁻¹ in Figure 1b corresponds to the OH group. This is because surface hydroxyl groups can form hydrogen bonds, which can cause their stretching frequencies to be slightly shifted and result in shoulder peaks in the 3100–3200 cm⁻¹ region [31,32]. Nevertheless, the SiO₂ NPs prepared and calcined for 30 min presented in Figure 1a, did not show the shoulder peak that signifies the OH group. The peaks at 1400 and 1640 cm⁻¹ in Figure 1a,b were allotted to the physisorbed water molecules, H–O–H [33]. The peak exhibited at 1100 cm⁻¹ in both materials was attributed to the Si–O–Si asymmetric stretching vibration. Furthermore, Si–OH stretching appeared at 495 cm⁻¹ in Figure 1a,b, respectively. Pure silica nanoparticles have a similar peak of around 470–495 cm⁻¹ [34]. The FTIR spectrum of Fe₃O₄ NPs (Figure 1c) revealed a broad peak and a shoulder around 3200–3500 cm⁻¹, which are due to O–H vibrational stretching, suggesting the presence of hydroxyl groups. Absorption bands at 1612, 1400, and 1106 cm⁻¹ peaks are associated with the bending vibrations of physisorbed water molecules, H–O–H [35]. Moreover, a strong peak around 619 cm⁻¹, suggests Fe−O stretching vibration, confirming the successful synthesis of Fe₃O₄ NPs. In Figure 1d,e, the broad bands at 3500 cm⁻¹ were ascribed to the silanol, Si–OH [36], and the shoulder peak appearing around 3100–3200 cm⁻¹ corresponds to the OH group. The 1400 and 1640 cm⁻¹ peaks in Figure 1d,e were assigned to the adsorbed water molecules, H–O–H [37]. The peak exhibited at 1100 cm⁻¹ was ascribed to the Si–O–Si asymmetric stretching vibration. The peaks observed at 646 cm⁻¹ were assigned to the Fe–O stretching vibrations [38], indicating that the nanocomposites were successfully prepared.

2.1.2. X-Ray Diffraction

X-ray diffraction (XRD) analysis was executed to study the crystalline structure of SiO₂ NPs, Fe₃O₄ NPs, and Fe₃O₄/SiO₂ nanocomposites, and the X-ray patterns are presented in Figure 2. For SiO₂ NPs (Figure 2a,b) and Fe₃O₄/SiO₂ (Figure 2d,e), a sharp peak roughly at 2θ = 26.5° is perceived and it corresponds to (101) crystal planes, which is attributed to the SiO₂ NPs [39,40]. This crystal plane is mostly observed in numerous XRD patterns of silica and its composites [41]. Peaks appeared at 2θ = 35.6 and 53.1° for the nanocomposites with (113) and (224) crystal planes, demonstrating that the magnetic nanoparticles were present [42]. The existence of (101) crystal plane in the composite indicates that the magnetite phase has maintained its crystalline structure within the composite [43]. The peaks labelled with asterisks come from the impurity components. Furthermore, the XRD pattern of the magnetite (Figure 2c) exhibited the characteristic peaks of Fe₃O₄ NPs at 2θ = 32.7°, 35.4°, 43.0°, 54.2°, 56.1°, and 62.6°, which corresponds to (220), (311), (440), (422), (511), and (440) crystal planes of pure Fe₃O₄ NPs [44,45]. The absence of broad peaks shows that there is no amorphous structure in all the materials, indicating the crystalline nature of the SiO₂ NPs, Fe₃O₄ NPs, and Fe₃O₄/SiO₂ nanocomposites. The atoms and molecules of crystalline material are organized in a definite lattice pattern. The SiO₂ NPs produced in 120 min of calcination time are likely to be more uniform and well-defined compared to those prepared in 30 min, leading to slight differences in performance [46]. Combining SiO₂ NPs and Fe₃O₄ NPs results in changes in the XRD patterns, such as decreased peak intensity and shifted peak positions. SiO₂ NPs can disrupt the crystalline order of Fe₃O₄ NPs, leading to increased disorder and a reduction in crystallinity [47,48]. This results in less intense peaks. Additionally, SiO₂ can absorb or scatter part of the incident X-rays, reducing the intensity of the diffracted peaks from Fe₃O₄ [49]. This phenomenon was observed in the nanocomposites in Figure 2d,e. Furthermore, chemical interaction between SiO₂ and Fe₃O₄ can change the electronic environment and affect the lattice spacing [50], which causes shifts in the peak positions. The average particle sizes of the SiO₂ NPs, Fe₃O₄ NPs, and Fe₃O₄/SiO₂ nanocomposites were calculated using the Debye–Scherrer Equation (1) as shown below:

d = Kλ/βcosθ

(1)

where variable d indicates particle size (nm), k indicates lattice constant (0.9), λ represents the wavelength (0.154 nm), β is the full-width half maximum (FWHM) of the maximum intensity (radians), and θ is the Bragg peak angle (degrees). Thus, the SiO₂—30 min, SiO₂—2 h, Fe₃O₄ NPs, Fe₃O₄/SiO₂—30 min, and Fe₃O₄/SiO₂—2 h particle sizes were calculated as 26, 21.5, 9.0, 15, and 13 nm, respectively. SiO₂ NPs showed more intense peaks; and the Fe₃O₄/SiO₂ nanocomposites exhibited intermediate sizes with peak intensities between SiO₂ NPs and Fe₃O₄ NPs.

2.1.3. Scanning Electron Microscopy

SEM analysis was carried out to investigate the surface morphology of the synthesized materials as presented in Figure 3. Figure 3a,b demonstrate sponge-like structures with particles dispersed in Figure 3a. In contrast, the particles are packed together in Figure 3b. Similar results were reported by Meng [51]. Relatively small spherical shapes, together with irregular shapes of Fe₃O₄ NPs, were observed in Figure 3c. Furthermore, the Fe₃O₄ NPs were agglomerated. The Fe₃O₄ NPs were embedded in the pores of SiO₂ NPs in the SEM images displayed in Figure 3d,e, which further corroborates the successful synthesis of Fe₃O₄/SiO₂ nanocomposites.

2.1.4. Energy-Dispersive X-Ray Spectroscopy

The EDS spectra of SiO₂ NPs synthesized from CFA, Fe₃O₄ NPs, and Fe₃O₄/SiO₂ nanocomposites are shown in Figure 4. The presence of Si and O in spectra Figure 4a (O = 40.6% and Si = 12.3%) and Figure 4b (O = 42.4% and Si = 13.7%) verified successful formation of SiO₂ NPs. A very high content of Fe (Fe = 65.4%) with O (O = 23.0%) proved that Fe₃O₄ NPs were synthesized successfully as seen in Figure 4c. Conversely, the EDS spectra of Fe₃O₄/SiO₂ nanocomposites shown in Figure 4d, with (O = 37.9%, Si = 12.1%, and Fe = 4.1%), and Figure 4e (O = 42.0%, Si = 18.4%, and Fe = 17.6%) corroborated the formation of Fe₃O₄/SiO₂ nanocomposites. Furthermore, the existence of Al, C, Ti, Mg, Na, P, Ca, K, Cl, and Cu is ascribable to impurities linked with inappropriate washing. These impurities (Ti, Mg, Na, P, Ca, K, Cl, and Cu) would not compromise the performance of the SiO₂ NPs and Fe₃O₄/SiO₂ nanocomposites during the adsorption process because they exist at a very low amount, less than 3%. C is perceived as due to the CFA carbon [52]. The significant amount of Al observed in all EDS spectra is attributed to the contact of the specimen aluminum stubs with the sample CFA (contains various oxides with alumina (Al₂O₃) included) [53,54] and AMD (which often contains dissolved [55,56] metals such as aluminum) [57].

2.1.5. Transmission Electron Microscopy

The particle size distribution and shape of the SiO₂ NPs, Fe₃O₄ NPs, and Fe₃O₄/SiO₂ nanocomposites were evaluated using TEM, as presented in Figure 5a–e. Figure 5a,b reveal uniform rod-like structures of SiO₂ NPs, which are aggregated. The TEM image in Figure 5c displays spherical Fe₃O₄ NPs, which are distributed evenly and slightly agglomerated at other spots. The dark spherical nanoparticles embedded in thin rod-like structures are observed in Figure 5d,e. These spherically shaped nanoparticles in Figure 5d,e are associated with Fe₃O₄ NPs, while thin rod-like structures are amalgamated to SiO₂ NPs, respectively. However, the embedded spherical nanoparticles are clustered in Figure 5d, whereas in Figure 5e, they are well-dispersed, implying that the number of active sites for adsorption would be high [58]. BET results emphasized this point as the highest surface areas of the SiO₂ NPs and Fe₃O₄/SiO₂ nanocomposites were found to be 10.5 and 67.7 m²/g (see

Table 1. Surface area, pore volume, and pore size of SiO₂ NPs and magnetic nanocomposites.

Sample	Surface Area (m2/g)	Pore Volume (cm3/g)	Pore Size (nm)
SiO2—30 min	3.7	0.21	229
SiO2—2 h	10.6	0.23	87.3
Fe3O4	90.8	0.27	12.1
Fe3O4/SiO2—30 min	37.7	0.19	20.1
Fe3O4/SiO2—2 h	67.8	0.39	23.2

). This is evident in that the surface area increased after the incorporation of Fe₃O₄ NPs into SiO₂ NPs.

2.1.6. Nitrogen Adsorption–Desorption Studies

The N₂ adsorption–desorption isotherms of the prepared materials were studied to acquire the surface area, pore volume, and pore size, as illustrated in Figure 6. The SiO₂ NPs (Figure 6a,b) depict type IV isotherms with an H₃-type hysteresis loop in the range 0.6–1.0 (P/Po), indicating the presence of mesopores in the materials [59]. The Fe₃O₄ NPs (Figure 6c) depict type IV isotherm. In addition, Fe₃O₄/SiO₂ nanocomposite (Figure 6d,e) exhibited type IV isotherm with H₃-type hysteresis loop in a 0.4 to 1.0 (P/Po) range, and this confirms meso-porosity of the nanocomposite [60,61,62]. Table 1 depicts the surface properties of the prepared materials. More adsorptive sites that are available for adsorption are indicated by a higher surface area [63]. The BET surface area of SiO₂ NPs (SiO₂—30 min and SiO₂—2 h) increased from (3.7 and 10.6 m²/g) to (37.7 and 67.8 m²/g) after the incorporation of Fe₃O₄ NPs. The BET surface area of Fe₃O₄ NPs was discovered as 90.8 m²/g. After incorporation, surface area, pore size, and pore volume increased significantly, indicating improved performance of the nanocomposite. Nonetheless, the variations in pore size between SiO₂—30 min and SiO₂—2 h can be ascribed to synthesis conditions, structural changes, and the nature of the precursor. A shorter treatment time may result in a more open structure with less density [64], enabling the formation of bigger pores (228 nm). This could be the result of incomplete network formation [65]. On the other hand, an extended treatment duration promotes complete condensation and polymerization of the silica network [66], leading to a denser structure with smaller pores (87 nm). Moreover, the pore size is reduced by the growth of silica particles and increased cross-linking [40,67]. Ultimately, the addition of the Fe₃O₄ NPs reduces the pore size by taking up space and changing the structure of the silica network. Thus, Fe₃O₄/SiO₂—30 min exhibits smaller pores than Fe₃O₄/SiO₂—2 h, mainly due to differences in the silica shell condensation and network formation around the Fe₃O₄ core [68].

2.1.7. Point of Zero Charge (pH_PZC)

The point of zero charge is a crucial property that represents the pH at which the material’s surface has a net zero charge. Figure S1 reveals a pH_PZC of SiO₂—30 min, SiO₂—2 h, Fe₃O₄/SiO₂—30 min and Fe₃O₄/SiO₂—2 h, which were found at pH = 3.4, pH = 2.8, pH = 3.5 and pH = 5.2, respectively. The pH_PZC postulates that when the pH of the solution is below and above the pH_PZC, the adsorbent’s surface becomes positively and negatively charged, respectively. The pKa value of DCF is 3.15 [69]. It has a greater ability to donate protons in aqueous media/solutions, and that results in the adsorbent surface strongly favoring the removal of negatively charged species (anions). When the pH of the solution is greater than the pH_PZC, the adsorbent’s surface becomes negatively charged and adsorbs positively charged species (cations) [70].

2.1.8. Vibrating Sample Magnetometer

A vibrating sample magnetometer (VSM) was used to investigate the magnetic properties of Fe₃O₄ NPs and Fe₃O₄/SiO₂ nanocomposite, with the results presented in Figure 7. Based on the results in Figure 7, it was revealed that the saturation magnetisation values for Fe₃O₄ NPs and Fe₃O₄/SiO₂ nanocomposite were calculated as 14 and 6.10 emu/g, indicating that Fe₃O₄ NPs were more magnetic than Fe₃O₄/SiO₂ nanocomposite. The prepared nanocomposite’s saturation magnetisation value expectedly dropped due to the incorporation of SiO₂, which reduced the magnetic moment, making it less magnetic than Fe₃O₄ NPs. This is also indicative that Fe₃O₄ was successfully incorporated with SiO₂, which expectedly lowered the magnetic field. However, it increased the surface area and was expected to have a higher removal efficiency than Fe₃O₄ NPs.

2.2. Optimisation of Adsorbents (Selection of Best Adsorbent Material)

The selection of a better adsorbent depends on the analyte and solution pH. The performance of SiO₂ NPs, Fe₃O₄, and Fe₃O₄/SiO₂ nanocomposites at different pH levels is shown in Figure 8. Notably, Fe₃O₄/SiO₂ nanocomposite exhibited remarkable removal efficiencies compared to SiO₂ NPs and Fe₃O₄ at acidic pH (pH = 3). This may be due to the synergistic effects and additional adsorption sites, which facilitated the adsorption of more DCF. This also agrees with the SEM and TEM results in Figure 3 and Figure 4, which prove that the SiO₂ NPs and Fe₃O₄ NPs were properly integrated. Moreover, the adsorbent’s surface charge varies with pH, thus affecting the removal efficiencies. This point was thoroughly stressed in the above section when PZC was discussed (Figure S1). Consequently, Fe₃O₄/SiO₂—2 h nanocomposite was selected based on the performance to conduct additional research.

2.3. Optimization Scheme

Response surface methodology (RSM), which is made possible by CCD, was used to investigate the impact of significant parameters, including the MA, SV, and sample pH, in order to remove DCF from the aqueous solution. The Pareto chart in Figure 9 depicts the outcomes. Information about the importance of the variables and how they interact is provided by the horizontal bars. If the horizontal bar passes the red line (95 % confidence level), it indicates that there is statistical significance for the factor or interactions [71]. Based on the Pareto chart (Figure 9), it is evident that the pH demonstrated statistical significance for the process of adsorptive removal. Additionally, the negative pH value (−4.820) suggests that sorption is highly favored at lower pH levels.

Three-dimensional (3-D) response surface plots in Figure 10 were used to evaluate the impact of variables interacting when DCF is being removed. Based on the RSM plots, it was observed that there were notable interactions between the sample pH and adsorbent mass. It was evident from RSM plots that MA and %RE were proportional to one another, i.e., when MA increased, RE% also increased. This was because of the increase in the surface area and availability of adsorptive sites on the adsorbent’s surface. Additionally, the effect of pH was conducted, and the highest recoveries were attained at a pH below 8. The pKa value of DCF is 3.15 [72]. When the pH ≥ pKa, DCF exists as an anion and when pH ≤ pKa, it exists in its neutral form. The PZC of the adsorbent was discovered to be 5.2. Thus, when the pH of the solution is lower (below 3.15) than the pH_PZC, the surface of the adsorbent becomes positively charged. Therefore, at pH values below 5.2, the %RE increases drastically. This was due to the electrostatic attraction between the nanocomposite and DCF. As seen in 3-D plots, at pH values above 5.2, the %RE significantly decreased. This was due to the repulsive forces which existed between the nanocomposite and DCF. Moreover, there were no significant interactions between SV and MA. Therefore, it should be clear from the 3-D plots that there was no significant %RE increase as both the SV and MA were increasing.

The desirability function was used to optimize all the evaluated parameters simultaneously. After the method of the desired function shown in Figure 11 was employed, optimum conditions were discovered to be as follows: sample pH value 2.4, SV 221.5 mL, and MA at 111.5 mg with total desirability of 1.00. The chosen SV was 200 mL since it is on the same desirability value of 1.00. The predicted optimum mass of adsorbent was also high; however, different masses of adsorbent ranging from 10−111.49 mg lie on the desirability value of 1.00; any MA in this range is suitable for further application. Therefore, 20 mg was chosen for the rest of the study. The experiments were conducted in triplicate with the optimal conditions, which produced %RE of 82.8%. Table S1 shows a list of runs attained using CCD and their responses.

2.4. Isotherm Studies

Adsorption isotherms models were critical in understanding the interaction behavior of DCF and Fe₃O₄/SiO₂ nanocomposite [73]. In this study, nonlinear Langmuir, Freundlich, Temkin, Sips, Redlich–Peterson (R−P), and Dubinin–Radushkevich (D−R) adsorption isotherms were employed to investigate the measured data shown in

Table 2. Adsorption isotherms of DCF onto Fe₃O₄/SiO₂.

Isotherms	Parameters	Values
Langmuir	qmax (mg/g)	55.7
	KL (L/mg)	0.38
	R2	0.9766
Freundlich	n	2.77
	KF (mg/g)	18.9
	R2	0.9236
Temkin	AT (L/g)	11.2
	BT (J/mol)	4.87
	R2	0.9574
Sips	n	1.34
	qS (mg/g)	54.4
	KS (L/mg)n	0.58
	R2	0.9777
Redlich–Peterson (R−P)	β	0.09
	KRP (L/g) αRP (L/mg)	13.6 1.38
	R2	0.9882
Dubinin–Radushkevich (D−R)	E (kJ/mol)	4.55
	qRP (mg/g)	45.2
	R2	0.9673

.

The Langmuir isotherm model (Figure S2a) is an important tool for analyzing adsorption processes, particularly for systems where monolayer adsorption on homogeneous surfaces is predominant. A measure of how well the model describes the experimental data is provided by the coefficient of determination (R²), and the parameters q_max and K_L provide important details about the adsorption capacity and affinity [74]. As depicted in Table 2, the correlation coefficient (R² = 0.9766) substantiates the adsorption of DCF on a monolayer surface of Fe₃O₄/SiO₂ nanocomposite with no lateral interactions between DCF molecules. The K_L value of 0.38 obtained indicates a high affinity of the Fe₃O₄/SiO₂ nanocomposite towards DCF. Consequently, the experimental data were best described by the Langmuir isotherm.

The Freundlich isotherm model (Figure S2a) plays a pivotal role in understanding the adsorption process on heterogeneous surfaces with different adsorption energies [75]. Freundlich’s adsorption strength is measured by the parameter K_F, while n indicates the adsorption’s feasibility and intensity [76]. These parameters were calculated to be 18.9 mg/g and 2.77 (Table 2), which stipulates that the adsorption process was favorable since values of n indicate a good adsorption process between 1 and 10, and less than 1 indicates inadequate adsorption [77]. Furthermore, the results obtained showed that the Freundlich model has a lower coefficient of determination (R² = 0.9236) than the Langmuir model, denoting that the equilibrium data were best explained by the Langmuir isotherm.

The Temkin isotherm model (Figure S2b) comprises a parameter that accounts for adsorbent–adsorbate interactions [78]. According to the model, all of the molecules in the layer would experience a linear drop in heat of adsorption when the coverage increases, by neglecting very small and very large values of concentrations [79,80]. Based on the results presented in Table 2, the binding energy (B_T) and binding constant (A_T) were discovered to be 11.2 J/mol and 4.87 L/g. The coefficient of determination of this model is R² = 0.9574, which evidently shows that it fits the equilibrium data better than the Freundlich isotherm. However, compared to the Langmuir isotherm, with a coefficient of determination of 0.9766, the Temkin isotherm model does not adequately fit the equilibrium data.

The Sips isotherm model was also used to fit the measured data (Figure S2b), which is the derivative of Freundlich and Langmuir isotherms, to affirm the homogeneity or heterogeneity of the surface of the adsorbent [81]. As the parameters illustrated in Table 2, if n lies between 0 and 1, it represents heterogeneity. However, if n = 1 or close to 1, it suggests a homogenous adsorption process [82]. Based on the findings displayed in Table 2, the n value was discovered to be 1.34, slightly above 1, stipulating a homogenous adsorption process. These results align with the conclusions made from the Langmuir isotherm, which assumed homogenous adsorption.

Figure S2c illustrates Redlich–Petersen (R–P) isotherm model, which is a hybridized adsorption model that combines features of the Freundlich and Langmuir isotherm models. The Redlich–Petersen isotherm has parameters such as K_RP, α_RP, and β, which were calculated as 13.6 L/g, 1.38 L/mg, and 0.09 as shown in Table 2, respectively. The β value of the R–P isotherm model ranges from 0 and 1. If β approaches or equals to 1, the R–P isotherm suggests a monolayer, and if β approaches 0, the R−P isotherm indicates a multilayer feature [83]. As seen in Table 2, the value of β was 0.09, with a coefficient of determination of 0.9882. The equilibrium data described the monolayer characteristic, which was described by the Langmuir isotherm.

Figure S2c illustrates the Dubini–Radushkevich (D–R) isotherm model, which is normally used to predict whether adsorption is based on a chemical or physical process [84]. The prediction was based on the obtained mean free energy (E, kJ/mol), which suggests that mean free energy above and below 8 kJ/mol suggests chemisorption and physisorption, respectively [85]. From the results obtained in Table 2, the value of E was below 8 kJ/mol (4.6 kJ/mol), which showed that physisorption was the primary mechanism by which DCF was adsorbed onto Fe₃O₄/SiO₂.

2.5. Kinetics Studies

To examine the mechanism for DCF onto Fe₃O₄/SiO₂, nonlinear kinetic models (Figure S3) such as pseudo-first order (PFO), pseudo-second order (PSO), and intraparticle diffusion were plotted with the calculated parameters shown in

Table 3. Adsorption kinetic parameters for the adsorption of DCF onto Fe₃O₄/SiO₂ nanocomposite.

Isotherms	Parameters	Values
Pseudo-first order	K1 (1/min)	0.10
	qe (mg/g)	44.0
	R2	0.9819
Pseudo-second order	K2 (g/mg.min)	0.002
	qe (mg/g)	50.2
	R2	0.9569
Intra-particle diffusion	C1 (mg/g)	4.23
	Kid1 (mg/g) min1/2	9.11
	R12	0.8757
	C2 (mg/g)	41.1
	Kid2 (mg/g) min1/2	0.28
	R22	0.7052

. The R² values of PFO and PSO are used to assess the appropriateness of the kinetic adsorption. Therefore, the highest R² value, which is closest to 1, indicates more suitable kinetic adsorption [86,87]. Based on the plots (Figure S3), the PFO plot shows an R² value of 0.9819, which is close to 1. On the other hand, the PSO plot reveals an R² value of 0.9569. The adsorption of DCF was, therefore, driven by a PFO mechanism and was not amenable to the PSO model rendition. According to the PFO model, physisorption is the primary mechanism driving DCF adsorption [88], which is in agreement with the D−R isotherm model, as discussed. On the PFO model, it can be noted that the qt values increased gradually and eventually became nearly constant.

In contrast to the PSO model, with an increase in contact time, the qt values kept increasing. To better understand the adsorption behavior, the experimental data were fitted into the intraparticle diffusion model. From the obtained plot, as presented in Figure S4, it is evident that there are two phases; the initial phase (Figure S4a) entails DCF molecules from the liquid phase migrating to the outer surface of the nanocomposite; the second phase (Figure S4b) involves the diffusion of the DCF molecules within the adsorbent’s pores and finally bonding to the internal structures [89]. As seen in Table 3, the boundary layer diffusion parameters (C₁ and C₂) have values greater than zero, indicating that intraparticle diffusion does not appear as the rate-limiting step. Consequently, intraparticle diffusion mostly controlled the adsorption mechanism.

2.6. Thermodynamic Studies

The study of thermodynamics was carried out to evaluate the impact of temperature on the adsorptive removal of DCF. The plot of 1/T vs. lnK at different temperatures is demonstrated in Figure S5, and temperatures ranged from 25–60 °C. The calculated Gibbs free energies described the thermodynamic behavior of DCF onto Fe₃O₄/SiO₂. Gibbs free energy was calculated using Equation (2). According to the data in

Table 4. Thermodynamic parameters for the adsorption of DCF onto Fe₃O₄/SiO₂ nanocomposite.

Analyte	T (K)	qe (mg/g)	ΔG° (kJ/mol)	ΔS° (J/mol K)	ΔH°(kJ/mol)
DCF	298	44.0	−3.89	51.81	11.38
	303	50.87	−4.52
	313	52.71	−4.82
	333	61.20	−5.84

, the ΔG° values ranged from −5.84 to −3.89 kJ/mol, demonstrating that the process of adsorption was spontaneous. As such, the decrease in ΔG° as the temperature increases proves that the adsorption process that occurred is inversely proportional to the temperature [90]:

$${\mathsf{\Delta }G}^{°}={\mathsf{\Delta }H}^{°}-T\mathsf{\Delta }{S}^{°}$$

(2)

The positive ΔH° value (11.38 kJ/mol) indicated that the adsorption process was endothermic. Furthermore, the ΔS° value was calculated to be 51.81 J/mol·K, demonstrating that the adsorption process has a high degree of disorder [91].

2.7. Possible Adsorption Mechanism

Diclofenac adsorption onto the Fe₃O₄/SiO₂ nanocomposite surface was investigated to provide understanding into potential interactions. According to the BET results, the porous nature of the Fe₃O₄/SiO₂ nanocomposite facilitates the DCF molecules’ penetration into the interior pores. The PZC of Fe₃O₄/SiO₂ nanocomposite was identified (as exhibited in Figure S1), which was obtained at 5.2. At this point, the Fe₃O₄/SiO₂ nanocomposite surface is neutral. Therefore, at the pH below pH_PZC = 5.2, the Fe₃O₄/SiO₂ nanocomposite surface is protonated, and the nanocomposite surface is negatively charged at pH values higher than pH_PZC = 5.2 [92]. The optimum pH is 2.4, and DCF is in an anionic form in the solution. Electrostatic attraction occurs between positively charged Fe₃O₄/SiO₂ and negatively charged DCF molecules, thereby promoting high adsorption efficiency. FTIR spectra and the adsorption mechanism presented in Figure 12 and Figure 13 further elucidated how the hydrogen bond formed between O atoms from the Fe₃O₄/SiO₂ and H atoms on the aromatic ring and the secondary amine group from DCF. Moreover, Figure 13 illustrates the metallic bond formed between the pool of electrons in the aromatic ring of the DCF and iron (Fe) from the Fe₃O₄/SiO₂. According to Figure 12b (after adsorption), the shape, broadness, and intensity of the hydroxy group peaks and asymmetrical peaks from the Si–O–Si changed. Furthermore, the adsorption isotherms demonstrated that the results could be described by the Langmuir isotherm model, which postulates that the adsorbate molecule can only occupy an adsorption site at the Fe₃O₄/SiO₂ nanocomposite once [93,94]. The results from kinetic studies demonstrated that the experimental data best fitted the PFO model, which presumed that physisorption was a dominant mechanism for the removal of DCF, which agrees with the results of the D−R isotherm. The chromatograms of before and after adsorption are presented in Figure S6.

2.8. Desorption Mechanism

Desorption of DCF was conducted to evaluate the behavior of the adsorbent in the process. HNO₃ was used as it created an acidic environment and protonated the functional groups of Fe₃O₄/SiO₂ nanocomposite. Consequently, the interaction forces (hydrogen and metallic bonds) between DCF and Fe₃O₄/SiO₂ nanocomposite were weakened. Therefore, DCF’s H atoms detached from the O atoms of the adsorbent, and the adsorbent was regenerated. The schematic representation of the desorption mechanism is presented in Scheme 1.

2.9. Analysis of Wastewater and Sewage Samples

The spiked real water samples (influent, effluent, raw sewage, and final sewage) were analyzed using the adsorptive removal process described in this study to assess the performance of Fe₃O₄/SiO₂ in the presence of sample matrices. After optimum conditions were employed, high removal efficiencies were obtained, as presented in

Table 5. Adsorptive removal of DCF in real water samples (wastewater and sewage samples).

Sample	Prior Adsorption (mg/L)	Post Adsorption (mg/L)	%RE
Influent	7.311	0.345	95.28
Effluent	6.948	0.178	97.44
Raw sewage	8.376	0.433	94.83
Final sewage	8.629	0.983	88.61

. In the influent and effluent wastewater, the removal efficiencies were determined as 95.28% and 97.44%, whereas 94.83% and 88.61% were obtained for raw sewage and final sewage samples. These findings demonstrated that the complex matrices had no detrimental effects on the removal of DCF.

The physicochemical parameters, such as total dissolved solids (TDS), electrical conductivity (EC), turbidity, and pH (

Table 6. Physico–chemical analysis of wastewater and sewage samples.

Parameter	Influent		Effluent		Raw Sewage		Final Sewage
	Before	After	Before	After	Before	After	Before	After
pH	8.09	7.46	9.86	7.37	8.21	7.42	9.51	7.03
Turbidity (NTU)	29.0	11.7	1.31	0.95	15.8	1.33	23.0	11.1
EC (mS)	0.19	0.12	0.37	0.24	1.48	1.02	1.25	0.95
TDS (mg/L)	130	65	192	96	776	411	595	299

) were assessed before and after applying Fe₃O₄/SiO₂ nanocomposite in sewage samples and wastewater to adsorb DCF. The pH of influent, effluent, raw sewage, and final sewage before the adsorption process was high, ranging from 8.21 to 10.1. However, the optimum conditions were used, and the adsorption process was executed. After adsorption, the pH was perceived to have dropped (7.03–7.46). The turbidity values before adsorption were extraordinarily high in the influent wastewater, raw sewage, and final sewage (29.0, 15.8, and 23.0 NTU); but were low in the effluent (1.31 NTU), which was within the permissible level of 0.4–2.5 NTU [95]. Nevertheless, after adsorption, the turbidity of the effluent and raw sewage was reduced significantly to 0.95 and 1.33 NTU. Additionally, the turbidity of the influent and final sewage dropped to 11.7 and 11.1 NTU, respectively. Nevertheless, untreated wastewater and sewage samples often exhibit high conductivity because they contain dissolved salts and other ionic species. The EC before adsorption was measured to be 0.19, 0.37, 1.48, and 1.25 mS for the influent, effluent, raw sewage, and final sewage, respectively; whereas, after adsorption, the EC decreased to 0.12, 0.24, 1.02, and 0.95 S/m. The measured TDS of the influent, effluent, raw sewage, and final sewage before the adsorption process could be carried out were 130, 192, 776, and 595 mg/L. Then, after adsorption, the TDS decreased to 65, 96, 411, and 299 mg/L. A TDS level of less than 300 mg/L is advised for drinking water by the World Health Organization [96]. Therefore, the TDS levels detected after adsorption were less than 300 mg/L except for the raw sewage sample (411 mg/L). The investigation of these physicochemical parameters demonstrates the efficacy of Fe₃O₄/SiO₂ in adsorbing DCF from wastewater and sewage samples and improving water quality. The adsorption process reduced the TDS, EC, and turbidity significantly. This technique proved to be a promising solution for treating contaminated wastewater and sewage.

2.10. Comparison Study

Table 7. Comparison of other adsorbents employed in removing DCF.

Adsorbent	pH	Adsorption Capacity (mg/g)	MA (g)	Contact Time (min)	References
RM-PPy	5	195	0.20	50	[97]
GO@CoFe2O4 nanocomposite	4	32.4	0.74	60	[98]
Olive stones-AC	4.2	11.0	0.30	30	[99]
MWCNTs-Fe3O4 nanocomposite	5.6	52.1	0.15	82	[100]
Cocoa pod husk-AC	7	5.53	0.25	15	[101]
Fe3O4/SiO2 Nanocomposite	2.4	55.7	0.02	40	This work

displays the maximum capacities of different adsorbents employed in removing DCF. Fe₃O₄/SiO₂ nanocomposite synthesized in this work had 55.7 mg/g maximum adsorption capacity compared to adsorbent materials discussed in the literature for the removal of DCF. However, it is noteworthy that other adsorbents might have relatively high adsorption capacities towards the removal of DCF from aqueous solution or wastewater. For instance, polypyrrole (RM-PPy) showed 195 mg/g adsorption capacity (Table 7).

2.11. Regenerability and Reusability Studies

Regenerability and reusability studies of Fe₃O₄/SiO₂ nanocomposite were evaluated to determine the capacity to remove DCF from wastewater. Successive adsorption–desorption experiments were performed based on the methodology reported in the previous studies [102]. Figure 14 demonstrates the removal efficiencies of the spent Fe₃O₄/SiO₂ nanocomposite after seven adsorption–desorption cycles. As seen in Figure 14, the removal efficiency dropped significantly from the first cycle to the second, then started reducing with a slight margin until the end of the seventh cycle of adsorption–desorption with a removal efficiency of 78.8%. Therefore, it can be concluded that the Fe₃O₄/SiO₂ nanocomposite showed excellent regeneration and reusability properties, with a slight decrease (from the second cycle) in removal efficiency over seven cycles. The results also demonstrated that the Fe₃O₄/SiO₂ nanocomposite was structurally stable, indicating that it could be used repeatedly to remove DCF. Nonetheless, the prepared nanocomposite demonstrated exceptional reusability and a sustainable solution to remove DCF in wastewater.

3. Materials and Methods

3.1. Reagents and Materials

Only analytical grade reagents were used. Diclofenac sodium salt (≥99%), iron (II) chloride tetrahydrate (FeCl₂·4H₂O), hydrochloric acid (HCl), sodium hydroxide (NaOH), methanol (CH₃OH), ammonia hydroxide (NH₄OH), acetonitrile (HPLC grade), and formic acid (HCOOH) were all procured from Sigma-Aldrich (St. Louis, MO, USA). Acid mine drainage waste was collected from the central basin acid mine drainage treatment plant in Germiston, Gauteng province, South Africa. Coal fly ash was obtained from Lethabo Power Station, situated in Free State province, South Africa. Ultra-pure water was used throughout the experiments.

3.2. Instrumentation

The masses of chemical salts and adsorbent materials were recorded using a Radwag AS220/C/2 analytical balance. Transmission electron microscopy (TEM) studies were performed at an acceleration voltage of 200 kV by using a Jeol JEM-2100F transmission electron microscope instrument (Akishima, Tokyo, Japan) equipped with a LaB6 source. The morphological and structural properties of the synthesized materials were studied using a scanning electron microscopy (SEM, Tescan Vega 3 LMH) instrument (Tescan Company, Brno, Czech Republic) coupled with an energy dispersive X-ray spectroscopy (EDS) and an Empyrean X-ray diffractometer (PANalytical, Almelo, The Netherlands). Analysis of functional moieties for the prepared materials was conducted on a MIRacle-10 Fourier-transform infrared spectrometer (FT-IR) (Shimadzu, Japan). A benchtop pH/mV Meter (HI5221, Hanna Instruments, Woonsocket, RI, USA) was used for pH adjustments. The point of zero charge (pH_PZC) was evaluated by using a zetasizer (Malvern PANalytical, Worcestershire, UK). The solids were separated from the supernatant through the Eppendorf 5702 centrifuge (Eppendorf, Hamburg, Germany). The filtrate was analyzed on an Agilent HPLC 1200 Infinity series coupled to a DAD (AgilentTechnologies, Waldbronn, Germany). An Agilent Zorbax Eclipse Plus C18 column (3.5 mm × 150 mm × 4.6 mm) (Agilent, Newport, CA, USA) was operated at an oven temperature of 25 °C.

3.3. Synthesis of SiO₂ Nanoparticles from Coal Fly Ash

To synthesize SiO₂ NPs, a sodium silicate solution was first prepared by following the previous method with a few small adjustments [27]. Coal fly ash was obtained from the Power Station. Sodium silicate solution was prepared according to the following recipe: 10 g of CFA was mixed with 10 M NaOH solution. The mixture was subjected to heat in a closed high-density polyethylene (HDPE) bottle at 97 °C inside a silicone oil bath, and the mixture was constantly agitated at 300 rpm for 60 min. The concentration of NaOH and reaction time were chosen based on the promotion of the highest silica extraction. Silica nanoparticles were synthesized from the sodium silicate solution obtained through the precipitation process. The silicate solution was stirred vigorously for 60 min. During the precipitation process, nitric acid was added dropwise into the sodium silicate solution until the pH of the solution was reduced from approximately 11 to 8. After that, ethanol (20 mL) was added to the mixture. The mixture was stirred continuously for 45 min and then centrifugated at 4000 rpm for a precipitate to form. The formed precipitate was washed several times with deionized water and further placed in a muffle furnace for 30 min at 600 °C to achieve SiO₂ NPs [103]; they were labelled SiO₂—30 min. The SiO₂—2 h was synthesized by following a similar procedure of heating for 120 min in the last step.

3.4. Synthesis of Fe₃O₄/SiO₂ Nanocomposite

Fe₃O₄/SiO₂ nanocomposite was synthesized following the method reported previously [104]. An amount of 1000 mL of AMD was taken for the preparation of Fe₃O₄ NPs. The pH of the AMD was set to 4.5 by adding 1 M HCl or 1 M NaOH to the mixture, and an Fe³⁺ precipitate was formed and settled at the bottom of the beaker. The Fe³⁺ precipitate was separated from the supernatant through centrifugation. Fe₃O₄/SiO₂ nanocomposite was synthesized by following the method reported by Munonde [105] with a few modifications. Briefly, 10.4 g precipitate and 5.2 g FeCl₂·4H₂O were mixed in 200 mL deionized water and rapidly agitated at 85 °C. This process took place under a nitrogen atmosphere. After quickly adding NH₄OH solution (50 mL) and 7.8 g of SiO₂ NPs, the mixture abruptly turned black. The contents were further agitated for 15 min under the same conditions and cooled to 25 °C. The formed precipitate was segregated through a magnetic decantation process. The precipitate was washed several times with deionized water to pH 10. Finally, the achieved Fe₃O₄/SiO₂ nanocomposite was dried for 10 h at 60 °C in an oven and crushed into small grains. The nanocomposite was labelled Fe₃O₄/SiO₂—30 min. The Fe₃O₄/SiO₂—2 h was synthesized by following a similar procedure with an addition of SiO₂—2 h into the mixture. The schematic representation of the nanocomposite synthesis is presented in Scheme 2.

3.5. Batch Adsorption Experiments

Erlenmeyer flasks were used for the adsorption experiments, which were conducted at room temperature. Central composite design (CCD) was utilized to optimize the variables that influenced sorption. The adsorption of DCF was carried out following the method reported by Pereira Gómez and co-workers [106]. Briefly, distinct masses of Fe₃O₄/SiO₂—2 h (10 to 100 mg) were added into 200 mL of 1 mg/L DCF solution. The pH (2–10) of the solutions was set by adding 1.0 M HCl or 1.0 M NaOH dropwise. After sonication for 60 min, the solids were separated from the supernatant by centrifugation at 4000 rpm for 10 min. The filtrates were filtered through a 0.22 μm acrodisc membrane filter and then analyzed with HPLC-DAD to determine the concentration of DCF after adsorption. The percentage removal efficiency (%RE) of DCF was determined using Equation (3):

$$\mathrm{\%}RE={\displaystyle \frac{\left(C_o-C_e\right)}{C_o}}\times 100$$

(3)

where: C₀ (mg/L) is the initial concentration, and C_e (mg/L) is the equilibrium concentration of DCF.

3.6. Isotherms Studies

In order to study the isotherms, various concentrations of DCF (2–20 mg/L) were prepared in 200 mL, and then 1.0 M HCl and 1.0 M NaOH were used to bring the solution’s pH down to 2.4. Twenty mg of Fe₃O₄/SiO₂ nanocomposite was introduced into the solutions, and the contents were sonicated for 60 min at 25 °C. The nanocomposite was isolated from the supernatant through centrifugation (4000 rpm; 10 min), and the filtrates were analyzed by HPLC-DAD. The adsorption capacities (q_e (mg/g)) of DCF were obtained using Equation (4):

$$q_e={\displaystyle \frac{\left(C_o-C_e\right)V}{W}}$$

(4)

where q_e (mg/g) is the amount of DCF adsorbed at equilibrium, W (mg) and V (mL) are mass of adsorbent (MA) and sample volume (SV), respectively.

3.7. Kinetics Studies

Kinetics studies were conducted at varying contact times (5–60 min) with 20 mg Fe₃O₄/SiO₂ nanocomposite, pH = 2.4, and 200 mL SV. The experiments were carried out using an initial concentration of 14.1 mg/L. The concentrations after adsorption were determined by the use of HPLC-DAD, and the adsorption capacities (q_t (mg/g)) of DCF were calculated using Equation (5):

$$q_t={\displaystyle \frac{\left(C_o-C_t\right)V}{W}}$$

(5)

where q_t (mg/g) represents the amounts of DCF adsorbed at equilibrium, W (mg) and V (mL) are the MA and SV.

3.8. Thermodynamics Studies

Temperatures between 25 and 60 °C were used to investigate how temperature affected the adsorption process. The MA, pH, contact time, SV, and initial concentration were kept constant.

3.9. Real Water Sample Analysis

Wastewater samples (effluent, influent, and raw sewage) were collected from Steelpoort and Marble Hall (Steelpoort and Marble Hall, Limpopo, South Africa). Bottles of 800 mL were used to collect these samples, which were then refrigerated at 8 °C. The samples were filtered via 0.22 μm syringe filters.

3.10. Point of Zero Charge

The pH_PZC of the synthesized SiO₂—30 min, SiO₂—2 h, Fe₃O₄/SiO₂—30 min, and Fe₃O₄/SiO₂—2 h was evaluated through the zetasizer technique. During the preparation of 50 mL solutions, the pH was adjusted with 0.10 mol/L NaOH or 0.1 mol/L HCl solutions, and 120 mg of adsorbent materials were introduced into the solutions. The mixtures were sonicated and analyzed to obtain pH_PZC.

4. Conclusions

A Fe₃O₄/SiO₂ nanocomposite was synthesized from CFA and AMD, and various analytical characterization techniques confirmed its successful synthesis. Then, Fe₃O₄/SiO₂ nanocomposite was applied to remove DCF from wastewater, and excellent removal was achieved with a 97.4% maximum removal efficiency. The equilibrium data were fitted into nonlinear Langmuir, Freundlich, Temkin, Sips, Redlich–Peterson (R−P) and Dubinin–Radushkevich (D−R) adsorption isotherms models, and according to the Langmuir isotherm model, the adsorption was monolayer. The PFO kinetic model was best fitted by kinetic data, which suggested that the controlling mechanism for the uptake of DCF was physisorption, and the D–R isotherm model also confirmed this phenomenon. The thermodynamic data established that the system was endothermic and spontaneous. Fe₃O₄/SiO₂ nanocomposite demonstrated high removal efficiencies of DCF from wastewater and sewage. Therefore, it is noticeable that Fe₃O₄/SiO₂ nanocomposite can be applied to reduce the pollution caused by pharmaceuticals (namely diclofenac) in aquatic bodies. By incorporating Fe₃O₄ to enhance the physicochemical properties of the adsorbent, the Fe₃O₄/SiO₂ nanocomposite reduced the turbidity, EC, and TDS of wastewater and sewage while keeping the pH almost constant. The results obtained demonstrated the potential applicability of Fe₃O₄/SiO₂ nanocomposite as an adsorbent to remove DCF in complex matrices of real water samples. Furthermore, the Fe₃O₄/SiO₂ nanocomposite could be regenerated and reused seven times, maintaining high performance.