Fe3O4@C Matrix with Tailorable Adsorption Capacities for Paracetamol and Acetylsalicylic Acid: Synthesis, Characterization, and Kinetic Modeling

In this study Fe3O4@C matrix was obtained by combustion method and used hereafter as adsorbent for paracetamol and acetylsalicylic acid removal from aqueous solutions. The Fe3O4@C matrix was characterized by electronic microscopy, X-ray diffraction, thermal analysis, Fourier-transform infrared spectroscopy, and magnetic measurements. Two kinetic models of pseudo first-order and pseudo-second-order for both paracetamol and acetylsalicylic acid were studied. The experimental data were investigated by Langmuir, Freundlich, and Redlich–Peterson adsorption isotherm models. The adsorption followed the Redlich–Peterson and pseudo-second-order models with correlation coefficients R2 = 0.98593 and R2 = 0.99996, respectively, for the adsorption of paracetamol; for the acetylsalicylic acid, the adsorption followed the Freundlich and pseudo-second-order model, with correlation coefficients R2 = 0.99421 and R2 = 0.99977, respectively. The equilibrium was quickly reached after approximately 1h for the paracetamol adsorption and approximately 2h for acetylsalicylic acid adsorption. According to the Langmuir isotherm, the maximum adsorption capacity of the magnetic matrix was 142.01 mg·g−1 for the retention of paracetamol and 234.01 mg·g−1 for the retention of acetylsalicylic acid. The benefits of using the Fe3O4@C matrix are the low cost of synthesis and its easy and fast separation from solution by using an NdBFe magnet.


Introduction
Population development over the years had led to industrial progress [1,2] accompanied by undesired consequences, such as the negative polluting effects. These changes could not go for long untraced by the environment and humans. The pollution-related problems emerged years ago and continue to persist due to the contamination found in soil [3], air [4], and water [5]; therefore, solving these problems became an urgent matter.
The aim of this study was the synthesis of the Fe3O4@C matrix by the solution combustion method and the investigation of its adsorption capacity for the removal of paracetamol and acetylsalicylic acid from aqueous solution. The use of the Fe3O4@C matrix combines the advantages of magnetic separation with the benefits of a porous material and allows the achievement of a highly efficient adsorbent material. To the best of our knowledge, this is the first study focusing on the use of a magnetic matrix in the removal of both drugs from aqueous solution.

XRD Characterization of Fe3O4@C Matrix
The XRD pattern ( Figure 1) revealed that the analyzed sample contains only magnetite (Fe3O4) as single-phase according to the DB card 1011084; also, the active carbon is present in an amorphous phase which is not evidenced on the XRD pattern.  Figure 2 shows the magnetic hysteresis loop of the Fe3O4@C matrix. The superparamagnetic behavior involves low values of remnant magnetization (Mr) and coercive field (Hc); the synthesized magnetic matrix is near to superparamagnetic behavior. Due to the fact that the Fe3O4@C matrix does not reach saturation over the applied magnetic field (H), saturation magnetization, Ms, was calculated by using the linear representation of magnetization versus 1/H and the extrapolation of 1/H→0.

Magnetic Properties of Fe3O4@C Matrix
The presence of carbon in the magnetic matrix led to the decrease of the saturation magnetization (Ms = 22.9 emu·g −1 ) when compared to the saturation magnetization of bulk magnetite (Ms = 84 emu·g −1 ) [40]. It is well known that a high magnetization value is correlated with larger particles; this observation is congenial with our results regarding the saturation magnetization and the matrix size evaluated by transmission electron microscopy Furthermore, our results are in agreement with previously reported literature data [41,42].  Figure 2 shows the magnetic hysteresis loop of the Fe 3 O 4 @C matrix. The superparamagnetic behavior involves low values of remnant magnetization (Mr) and coercive field (Hc); the synthesized magnetic matrix is near to superparamagnetic behavior. Due to the fact that the Fe 3 O 4 @C matrix does not reach saturation over the applied magnetic field (H), saturation magnetization, Ms, was calculated by using the linear representation of magnetization versus 1/H and the extrapolation of 1/H→0.

Magnetic Properties of Fe 3 O 4 @C Matrix
The presence of carbon in the magnetic matrix led to the decrease of the saturation magnetization (Ms = 22.9 emu·g −1 ) when compared to the saturation magnetization of bulk magnetite (Ms = 84 emu·g −1 ) [40]. It is well known that a high magnetization value is correlated with larger particles; this observation is congenial with our results regarding the saturation magnetization and the matrix size evaluated by transmission electron microscopy Furthermore, our results are in agreement with previously reported literature data [41,42].

Thermal Analysis of Fe 3 O 4 @C Matrix
The thermogravimetric (TG) and differential scanning calorimetry (DSC) curves of the Fe 3 O 4 @C matrix prepared by combustion method are depicted in Figure 3.
Thermal analysis revealed a total mass loss of 63.34% in two stages. In the first stage, the sample loses 8.17% of the total mass, and an endothermic process can be noticed with a maximum at 80.6 • C, attributed to the water removal. The second stage is wider, located between 300 • C and 650 • C, with a maximum at 620.1 • C; within this interval the sample loses 55.17% of the total mass and shows a strong exothermic effect assigned to the oxidation of the carbon content present in the sample.

Thermal Analysis of Fe3O4@C Matrix
The thermogravimetric (TG) and differential scanning calorimetry (DSC) curves of the Fe3O4@C matrix prepared by combustion method are depicted in Figure 3. Thermal analysis revealed a total mass loss of 63.34% in two stages. In the first stage, the sample loses 8.17% of the total mass, and an endothermic process can be noticed with a maximum at 80.6 °C, attributed to the water removal. The second stage is wider, located between 300 °C and 650 °C, with a maximum at 620.1 °C; within this interval the sample loses 55.17% of the total mass and shows a strong exothermic effect assigned to the oxidation of the carbon content present in the sample.
The thermal analysis of our sample revealed neither the exothermic effect at ~212 °C characteristic to the magnetite oxidation to maghemite accompanied with mass growth on TG curve, nor the exothermic effect at ~500 °C without mass loss attributed to the maghemite transition to hematite. This behavior can be explained by the fact that the magnetite is embedded in a high content of carbon which protects it against oxidation to maghemite.

FTIR Investigations of Fe3O4@C Matrix
The analysis of the FTIR spectrum ( Figure 4) reveals the presence of bands located at 575 cm −1 and 442 cm −1 corresponding to the stretching vibration of Fe-O from the Fe3O4@C matrix [43][44][45]. This result is in agreement with the XRD analysis that proved magnetite to be a single phase. In

Thermal Analysis of Fe3O4@C Matrix
The thermogravimetric (TG) and differential scanning calorimetry (DSC) curves of the Fe3O4@C matrix prepared by combustion method are depicted in Figure 3. Thermal analysis revealed a total mass loss of 63.34% in two stages. In the first stage, the sample loses 8.17% of the total mass, and an endothermic process can be noticed with a maximum at 80.6 °C, attributed to the water removal. The second stage is wider, located between 300 °C and 650 °C, with a maximum at 620.1 °C; within this interval the sample loses 55.17% of the total mass and shows a strong exothermic effect assigned to the oxidation of the carbon content present in the sample.
The thermal analysis of our sample revealed neither the exothermic effect at ~212 °C characteristic to the magnetite oxidation to maghemite accompanied with mass growth on TG curve, nor the exothermic effect at ~500 °C without mass loss attributed to the maghemite transition to hematite. This behavior can be explained by the fact that the magnetite is embedded in a high content of carbon which protects it against oxidation to maghemite.

FTIR Investigations of Fe3O4@C Matrix
The analysis of the FTIR spectrum ( Figure 4) reveals the presence of bands located at 575 cm −1 and 442 cm −1 corresponding to the stretching vibration of Fe-O from the Fe3O4@C matrix [43][44][45]. This result is in agreement with the XRD analysis that proved magnetite to be a single phase. In The thermal analysis of our sample revealed neither the exothermic effect at~212 • C characteristic to the magnetite oxidation to maghemite accompanied with mass growth on TG curve, nor the exothermic effect at~500 • C without mass loss attributed to the maghemite transition to hematite. This behavior can be explained by the fact that the magnetite is embedded in a high content of carbon which protects it against oxidation to maghemite.

FTIR Investigations of Fe 3 O 4 @C Matrix
The analysis of the FTIR spectrum ( Figure 4) reveals the presence of bands located at 575 cm −1 and 442 cm −1 corresponding to the stretching vibration of Fe-O from the Fe 3 O 4 @C matrix [43][44][45]. This result is in agreement with the XRD analysis that proved magnetite to be a single phase. In addition, the FTIR spectrum shows the presence of amorphous phases that cannot be identified by X-ray diffraction. The band located at 1650 cm −1 can be attributed to the C=O stretching vibration, while the band located at 3347 cm −1 can be assigned to the O-H stretching vibration. These bands characterize the organic residue present in the sample after the raw materials combustion, in particular, the combustion fuel.
addition, the FTIR spectrum shows the presence of amorphous phases that cannot be identified by X-ray diffraction. The band located at 1650 cm −1 can be attributed to the C=O stretching vibration, while the band located at 3347 cm −1 can be assigned to the O-H stretching vibration. These bands characterize the organic residue present in the sample after the raw materials combustion, in particular, the combustion fuel.

Electron Microscopy Characterization of Fe3O4@C Matrix
The details regarding the morphology and ultrastructure of the samples were established by scanning electron microscopy -SEM ( Figure 5) and transmission electron microscopy ( Figure 6). The SEM general overview micrograph ( Figure 5A) shows the Fe3O4@C matrix; one can notice that the magnetic matrix is characterized by differently sized spherical nanoparticles forming aggregates. Figure 5B shows the topology of the nanoparticles and their aggregates. The energy dispersive X-ray analysis (EDAX) was employed to determine the elemental composition of the matrix ( Table 1). The EDAX analysis showed that only C, O, and Fe were present in the Fe3O4@C matrix.

Electron Microscopy Characterization of Fe 3 O 4 @C Matrix
The details regarding the morphology and ultrastructure of the samples were established by scanning electron microscopy -SEM ( Figure 5) and transmission electron microscopy ( Figure 6). The SEM general overview micrograph ( Figure 5A) shows the Fe 3 O 4 @C matrix; one can notice that the magnetic matrix is characterized by differently sized spherical nanoparticles forming aggregates. Figure 5B shows the topology of the nanoparticles and their aggregates.
addition, the FTIR spectrum shows the presence of amorphous phases that cannot be identified by X-ray diffraction. The band located at 1650 cm −1 can be attributed to the C=O stretching vibration, while the band located at 3347 cm −1 can be assigned to the O-H stretching vibration. These bands characterize the organic residue present in the sample after the raw materials combustion, in particular, the combustion fuel.

Electron Microscopy Characterization of Fe3O4@C Matrix
The details regarding the morphology and ultrastructure of the samples were established by scanning electron microscopy -SEM ( Figure 5) and transmission electron microscopy ( Figure 6). The SEM general overview micrograph ( Figure 5A) shows the Fe3O4@C matrix; one can notice that the magnetic matrix is characterized by differently sized spherical nanoparticles forming aggregates. Figure 5B shows the topology of the nanoparticles and their aggregates. The energy dispersive X-ray analysis (EDAX) was employed to determine the elemental composition of the matrix ( Table 1). The EDAX analysis showed that only C, O, and Fe were present in the Fe3O4@C matrix. The energy dispersive X-ray analysis (EDAX) was employed to determine the elemental composition of the matrix ( Table 1). The EDAX analysis showed that only C, O, and Fe were present in the Fe 3 O 4 @C matrix.
As shown in the TEM micrograph ( Figure 6), the Fe 3 O 4 @C matrix consists of spherical nanoparticles with diameters of~30 nm. The cluster size distribution was determined at higher magnification (220 Kx) in order to visualize the particles' topology, and their organization as aggregates in the near-native solvent state was revealed ( Figure 6A). At 300 Kx magnification (20-nm scale), the spherical form of the magnetic nanoparticles can be more accurately observed; they display multiple faceted surfaces with activated carbon deposits ( Figure 6B).
As shown in the TEM micrograph ( Figure 6), the Fe3O4@C matrix consists of spherical nanoparticles with diameters of ~30 nm. The cluster size distribution was determined at higher magnification (220 Kx) in order to visualize the particles' topology, and their organization as aggregates in the near-native solvent state was revealed ( Figure 6A). At 300 Kx magnification (20-nm scale), the spherical form of the magnetic nanoparticles can be more accurately observed; they display multiple faceted surfaces with activated carbon deposits ( Figure 6B).

Effect of the Initial Concentration on the Removal Efficiency
The effect of the initial concentration of paracetamol and acetylsalicylic acid on the adsorption efficiency is shown in Figure 7. Using a Fe3O4@C matrix mass of 50 mg, the removal efficiency of paracetamol increases from 89% (at 10 mg/L) to 97% (at 100 mg/L). In the case of acetylsalicylic acid, the removal efficiency increases from 66% (at 10 mg/L) to 86% (at 100 mg/L). As one can notice, the adsorption efficiency is much higher for paracetamol than for acetylsalicylic acid; also, the adsorption efficiency is directly proportional to the drug concentration.   The effect of the initial concentration of paracetamol and acetylsalicylic acid on the adsorption efficiency is shown in Figure 7. Using a Fe 3 O 4 @C matrix mass of 50 mg, the removal efficiency of paracetamol increases from 89% (at 10 mg/L) to 97% (at 100 mg/L). In the case of acetylsalicylic acid, the removal efficiency increases from 66% (at 10 mg/L) to 86% (at 100 mg/L). As one can notice, the adsorption efficiency is much higher for paracetamol than for acetylsalicylic acid; also, the adsorption efficiency is directly proportional to the drug concentration. As shown in the TEM micrograph ( Figure 6), the Fe3O4@C matrix consists of spherical nanoparticles with diameters of ~30 nm. The cluster size distribution was determined at higher magnification (220 Kx) in order to visualize the particles' topology, and their organization as aggregates in the near-native solvent state was revealed ( Figure 6A). At 300 Kx magnification (20-nm scale), the spherical form of the magnetic nanoparticles can be more accurately observed; they display multiple faceted surfaces with activated carbon deposits ( Figure 6B).

Effect of the Initial Concentration on the Removal Efficiency
The effect of the initial concentration of paracetamol and acetylsalicylic acid on the adsorption efficiency is shown in Figure 7. Using a Fe3O4@C matrix mass of 50 mg, the removal efficiency of paracetamol increases from 89% (at 10 mg/L) to 97% (at 100 mg/L). In the case of acetylsalicylic acid, the removal efficiency increases from 66% (at 10 mg/L) to 86% (at 100 mg/L). As one can notice, the adsorption efficiency is much higher for paracetamol than for acetylsalicylic acid; also, the adsorption efficiency is directly proportional to the drug concentration.  The removal efficiency of paracetamol from aqueous solution was higher than the removal of acetylsalicylic acid, probably due to the amino and phenol groups, which show affinity for the carbon shell present on the Fe 3 O 4 surface. Our results are in perfect correlation with the data previously presented in the literature [46]. Ianoş et al. prepared magnetite nanoparticles embedded within a matrix of activated carbon using the combustion synthesis technique. The authors prepared magnetite/carbon nanocomposites with varied carbon ratio in order to remove phenol and p-chlorophenol from wastewater. They demonstrated that the removal efficiency of these organic pollutants (with very similar structure to the paracetamol-N-acetyl-para-aminophenol) from aqueous solution increased as the magnetite/carbon ratio decreased. They also tested the adsorption capacity of activated carbon in the removal of phenol and p-chlorophenol. Using a mass of 2 g/L activated carbon, the removal efficiency was 99%, while using a mass of 2 g/L nanocomposite (magnetite/carbon = 1:2), the removal efficiency was 94% [46].
Our synthesized compound (Fe 3 O 4 @C matrix) has an adsorption capacity of 97%, almost identical to the adsorption capacity of pure activated carbon, when 2 g/L of magnetic matrix (with mass ratio Fe 3 O 4 /C = 1:2 was used). Based on this result, we can state that the Fe 3 O 4 @C matrix can be successfully employed in the paracetamol and acetylsalicylic acid adsorption from wastewater.
The benefits of using the Fe 3 O 4 @C matrix instead of pure activated charcoal include the fast and easy magnetic separation instead of the centrifugation and filtration of activated charcoal and the possibility of its use in several desorption-adsorption cycles; pure active carbon can only be used once.
2.6.2. Effect of Contact Time Figure 8 shows the effect of the contact time on the adsorption of paracetamol and acetylsalicylic acid onto the Fe 3 O 4 @C matrix. One can notice that in the case of acetylsalicylic acid almost the entire drug amount is slowly adsorbed within 100 min, after which the system reaches equilibrium. For paracetamol, the adsorption process is very fast in the first 25 min, and then equilibrium is achieved. This behavior is due to the Fe 3 O 4 @C matrix surface, more precisely its carbon outer layer, which has numerous vacant sites that can be gradually occupied over time as a result of the adsorption process. It is also of note that the amount of drug adsorbed in the case of acetylsalicylic acid is higher than the adsorbed amount of paracetamol. This result is consistent and in perfect agreement with the adsorption efficiency (Figure 7), given the drug concentrations used in the experiment (C = 100 mg/mL acetylsalicylic acid; C = 50 mg/mL paracetamol).
The removal efficiency of paracetamol from aqueous solution was higher than the removal of acetylsalicylic acid, probably due to the amino and phenol groups, which show affinity for the carbon shell present on the Fe3O4 surface. Our results are in perfect correlation with the data previously presented in the literature [46]. Ianoş et al. prepared magnetite nanoparticles embedded within a matrix of activated carbon using the combustion synthesis technique. The authors prepared magnetite/carbon nanocomposites with varied carbon ratio in order to remove phenol and p-chlorophenol from wastewater. They demonstrated that the removal efficiency of these organic pollutants (with very similar structure to the paracetamol-N-acetyl-para-aminophenol) from aqueous solution increased as the magnetite/carbon ratio decreased. They also tested the adsorption capacity of activated carbon in the removal of phenol and p-chlorophenol. Using a mass of 2 g/L activated carbon, the removal efficiency was 99%, while using a mass of 2 g/L nanocomposite (magnetite/carbon = 1:2), the removal efficiency was 94% [46].
Our synthesized compound (Fe3O4@C matrix) has an adsorption capacity of 97%, almost identical to the adsorption capacity of pure activated carbon, when 2 g/L of magnetic matrix (with mass ratio Fe3O4/C = 1:2 was used). Based on this result, we can state that the Fe3O4@C matrix can be successfully employed in the paracetamol and acetylsalicylic acid adsorption from wastewater.
The benefits of using the Fe3O4@C matrix instead of pure activated charcoal include the fast and easy magnetic separation instead of the centrifugation and filtration of activated charcoal and the possibility of its use in several desorption-adsorption cycles; pure active carbon can only be used once.
2.6.2. Effect of Contact Time Figure 8 shows the effect of the contact time on the adsorption of paracetamol and acetylsalicylic acid onto the Fe3O4@C matrix. One can notice that in the case of acetylsalicylic acid almost the entire drug amount is slowly adsorbed within 100 min, after which the system reaches equilibrium. For paracetamol, the adsorption process is very fast in the first 25 min, and then equilibrium is achieved. This behavior is due to the Fe3O4@C matrix surface, more precisely its carbon outer layer, which has numerous vacant sites that can be gradually occupied over time as a result of the adsorption process. It is also of note that the amount of drug adsorbed in the case of acetylsalicylic acid is higher than the adsorbed amount of paracetamol. This result is consistent and in perfect agreement with the adsorption efficiency (Figure 7), given the drug concentrations used in the experiment (C = 100 mg/mL acetylsalicylic acid; C = 50 mg/mL paracetamol).

Kinetic Studies
The adsorption kinetics of paracetamol and acetylsalicylic acid onto the Fe 3 O 4 @C matrix were investigated using the pseudo-first-order (Equation (4)) and the pseudo-second-order kinetic models (Equation (6)). Figure 9 displays the pseudo-first-order model ( Figure 9A) and the pseudo-second-order model ( Figure 9B) of the paracetamol adsorption onto the Fe 3 O 4 @C matrix. Figure 10 shows the pseudo-first-order model ( Figure 10A) and the pseudo-second-order model ( Figure 10B) in the case of acetylsalicylic acid adsorption onto the Fe 3 O 4 @C matrix. The calculated values of k 1 , k 2 , q e , and R 2 for paracetamol and acetylsalicylic acid adsorption are shown in Table 2.
The value of the correlation coefficient (Table 2) obtained for the pseudo-first-order model is lower than the one calculated for the pseudo-second-order model in both cases, which suggests that the paracetamol and acetylsalicylic acid kinetic adsorptions onto the Fe 3 O 4 @C matrix are more accurately described by the pseudo-second-order kinetic model.

Kinetic Studies
The adsorption kinetics of paracetamol and acetylsalicylic acid onto the Fe3O4@C matrix were investigated using the pseudo-first-order (Equation (4)) and the pseudo-second-order kinetic models (Equation (6)). Figure 9 displays the pseudo-first-order model ( Figure 9A) and the pseudo-second-order model ( Figure 9B) of the paracetamol adsorption onto the Fe3O4@C matrix. Figure 10 shows the pseudo-first-order model ( Figure 10A) and the pseudo-second-order model ( Figure 10B) in the case of acetylsalicylic acid adsorption onto the Fe3O4@C matrix. The calculated values of k1, k2, qe, and R 2 for paracetamol and acetylsalicylic acid adsorption are shown in Table 2.
The value of the correlation coefficient (Table 2) obtained for the pseudo-first-order model is lower than the one calculated for the pseudo-second-order model in both cases, which suggests that the paracetamol and acetylsalicylic acid kinetic adsorptions onto the Fe3O4@C matrix are more accurately described by the pseudo-second-order kinetic model.

Kinetic Studies
The adsorption kinetics of paracetamol and acetylsalicylic acid onto the Fe3O4@C matrix were investigated using the pseudo-first-order (Equation (4)) and the pseudo-second-order kinetic models (Equation (6)). Figure 9 displays the pseudo-first-order model ( Figure 9A) and the pseudo-second-order model ( Figure 9B) of the paracetamol adsorption onto the Fe3O4@C matrix. Figure 10 shows the pseudo-first-order model ( Figure 10A) and the pseudo-second-order model ( Figure 10B) in the case of acetylsalicylic acid adsorption onto the Fe3O4@C matrix. The calculated values of k1, k2, qe, and R 2 for paracetamol and acetylsalicylic acid adsorption are shown in Table 2.
The value of the correlation coefficient (Table 2) obtained for the pseudo-first-order model is lower than the one calculated for the pseudo-second-order model in both cases, which suggests that the paracetamol and acetylsalicylic acid kinetic adsorptions onto the Fe3O4@C matrix are more accurately described by the pseudo-second-order kinetic model.

Adsorption Isotherms
The experimental data were investigated with Langmuir (Equation (7)), Freundlich (Equation (8)), and Redlich-Peterson (Equation (11)) adsorption isotherms ( Figure 11). These isotherms provide significant information concerning the surface properties of the carbon matrix and the affinities between the Fe 3 O 4 @C matrix and the adsorbed drugs (paracetamol and acetylsalicylic acid, respectively), parameters of paramount importance for the adsorption process, and its efficiency. The experimental data were investigated with Langmuir (Equation (7)), Freundlich (Equation (8)), and Redlich-Peterson (Equation (11)) adsorption isotherms ( Figure 11). These isotherms provide significant information concerning the surface properties of the carbon matrix and the affinities between the Fe3O4@C matrix and the adsorbed drugs (paracetamol and acetylsalicylic acid, respectively), parameters of paramount importance for the adsorption process, and its efficiency. The main parameters obtained from the fittings of the three equations and the corresponding correlation coefficients (R 2 ) are listed in Table 3. By comparing the R 2 values of the analyzed isotherms, one can notice that the Redlich-Peterson model is the most suitable to describe the paracetamol adsorption onto the Fe3O4@C matrix, while the Freundlich model is the most appropriate to describe the acetylsalicylic acid adsorption onto the Fe3O4@C matrix. From the literature [47] it is well known that the Redlich-Peterson is an empiric isotherm that combines the characteristics of both Langmuir and Freundlich models. This isotherm contains three parameters, one of which (β) is an exponent ranging between 0 and 1. When β = 1, the Redlich-Peterson model (Equation (11)) becomes the Langmuir isotherm (Equation (7)). Our results reveal that the adsorption of paracetamol onto the Fe3O4@C matrix follows a combination of the Langmuir and Freundlich isotherms, closer to the Langmuir model taking into account the value of β = 0.71299, which is near to β = 1.
Regarding the Freundlich isotherm, which describes the acetylsalicylic acid adsorption onto the Fe3O4@C matrix, the value of the exponent n also offers information about the adsorption process. A good adsorption process is characterized by values of exponent n between 2 and 10, according Hamdaoui and Naffrechoux, 2007 [48]. In our case, the value n = 1.91832 (Table 3) indicates that the matrix based on Fe3O4@C is an appropriate adsorbent for acetylsalicylic acid.
Also, the maximum adsorption capacity of Fe3O4@C matrix resulting from the Langmuir isotherm is 234.01 mg g −1 , which is higher than the other data previously reported in the literature [49][50][51]. The main parameters obtained from the fittings of the three equations and the corresponding correlation coefficients (R 2 ) are listed in Table 3. By comparing the R 2 values of the analyzed isotherms, one can notice that the Redlich-Peterson model is the most suitable to describe the paracetamol adsorption onto the Fe 3 O 4 @C matrix, while the Freundlich model is the most appropriate to describe the acetylsalicylic acid adsorption onto the Fe 3 O 4 @C matrix. From the literature [47] it is well known that the Redlich-Peterson is an empiric isotherm that combines the characteristics of both Langmuir and Freundlich models. This isotherm contains three parameters, one of which (β) is an exponent ranging between 0 and 1. When β = 1, the Redlich-Peterson model (Equation (11)) becomes the Langmuir isotherm (Equation (7)). Our results reveal that the adsorption of paracetamol onto the Fe 3 O 4 @C matrix follows a combination of the Langmuir and Freundlich isotherms, closer to the Langmuir model taking into account the value of β = 0.71299, which is near to β = 1.
Regarding the Freundlich isotherm, which describes the acetylsalicylic acid adsorption onto the Fe 3 O 4 @C matrix, the value of the exponent n also offers information about the adsorption process. A good adsorption process is characterized by values of exponent n between 2 and 10, according Hamdaoui and Naffrechoux, 2007 [48]. In our case, the value n = 1.91832 (Table 3) indicates that the matrix based on Fe 3 O 4 @C is an appropriate adsorbent for acetylsalicylic acid.
Also, the maximum adsorption capacity of Fe 3 O 4 @C matrix resulting from the Langmuir isotherm is 234.01 mg g −1 , which is higher than the other data previously reported in the literature [49][50][51]. Tables 4 and 5 show data from the literature regarding the maximum adsorption capacity as well as the adsorption parameters concerning paracetamol and acetylsalicylic acid removal. One can notice that our synthesized magnetic matrix (Fe 3 O 4 @C) exhibits a very good adsorption capacity for the removal of paracetamol and acetylsalicylic acid.
The Fe 3 O 4 @C matrix conjoins both the enhanced adsorption capacity and the high separation ability due to the magnetic properties of F 3 O 4 nanoparticles. The magnetic separation is an unconventional technology, economic, and one-step solution because it does not involve extra materials and supplementary steps such as centrifugation and filtration. The magnetic separation took place instantaneously using an NdBFe magnet, thus reducing the manipulation time of the experiment. In addition, the as synthesized magnetic matrix comes with the advantage of a short adsorption time until reaching equilibrium.

Preparation of Aqueous Solution Containing Drugs
Paracetamol (C 8 H 9 NO 2 -acetaminophen, N-acetyl-para-aminophenol (APAP)) is a member of the class of analgesic and antipyretic drugs widely used for the relief of severe/minor pain and as fever reducer. Acetylsalicylic acid (C 9 H 8 O 4 -2-acetoxybenzoic acid, aspirin) is considered an NSAID (nonsteroidal anti-inflammatory drug) due to its anti-inflammatory activity as well as analgesic and antipyretic properties. Paracetamol was acquired from Sigma-Aldrich and it was prepared an aqueous stock solution of 900 mg/L, pH = 3. Acetylsalicylic acid was achieved from Acros Organics and the aqueous stock solution prepared was of 800 mg/L, pH = 3. From each stock solution an additional six solutions were prepared, with concentrations between 10 and 900 mg/L for paracetamol and 10 and 800 mg/L for acetylsalicylic acid, in order to realize the adsorption isotherms. For the kinetic study and for the determination of the equilibrium time, it only one concentration for each drug solution was used: 50 mg/L for paracetamol and 100 mg/L for acetylsalicylic acid.
In order to establish the removal efficiency, in case of both drugs, another six solutions were prepared with the following concentrations; 10, 20, 40, 60, 80, and 100 mg/L.

Fe 3 O 4 @C Matrix Preparation
A magnetic matrix consisting of iron oxide magnetic nanoparticles (magnetite-Fe 3 O 4 ) and activated carbon [46] was prepared for the adsorption of the two analgesic drugs from aqueous solutions. Briefly, 36.4 g Fe(NO 3 ) 3 ·9H 2 O (Roth, Karlsruhe, Germany, 96%) and 20.7 g tartaric acid (C 4 H 6 O 6 ) (Merck, Darmstadt, Germany, 99.5%) were dissolved in 50 mL distilled water and the resulting solution was then transferred into a round-bottom flask containing 13.9 g activated carbon.
The mixture was kept for 24 h to impregnate and was afterwards placed inside a heating mantle at 400 • C so that the combustion reaction could take place ( Figure 12).
The combustion reaction took place in the absence of air just to prevent the formation of γ-Fe 2 O 3 instead of Fe 3 O 4 . The gases released after the combustion process were evacuated in a distilled water dish according to the method described by Ianoş and coworkers [39]. After gaseous emission stopped, the obtained magnetic black powder was washed several times with warm distilled water (60-80 • C) and then dried in an oven at 70 • C.

Preparation of Aqueous Solution Containing Drugs
Paracetamol (C8H9NO2-acetaminophen, N-acetyl-para-aminophenol (APAP)) is a member of the class of analgesic and antipyretic drugs widely used for the relief of severe/minor pain and as fever reducer. Acetylsalicylic acid (C9H8O4-2-acetoxybenzoic acid, aspirin) is considered an NSAID (nonsteroidal anti-inflammatory drug) due to its anti-inflammatory activity as well as analgesic and antipyretic properties. Paracetamol was acquired from Sigma-Aldrich and it was prepared an aqueous stock solution of 900 mg/L, pH = 3. Acetylsalicylic acid was achieved from Acros Organics and the aqueous stock solution prepared was of 800 mg/L, pH = 3. From each stock solution an additional six solutions were prepared, with concentrations between 10 and 900 mg/L for paracetamol and 10 and 800 mg/L for acetylsalicylic acid, in order to realize the adsorption isotherms. For the kinetic study and for the determination of the equilibrium time, it only one concentration for each drug solution was used: 50 mg/L for paracetamol and 100 mg/L for acetylsalicylic acid.
In order to establish the removal efficiency, in case of both drugs, another six solutions were prepared with the following concentrations; 10, 20, 40, 60, 80, and 100 mg/L.

Fe3O4@C Matrix Preparation
A magnetic matrix consisting of iron oxide magnetic nanoparticles (magnetite-Fe3O4) and activated carbon [46] was prepared for the adsorption of the two analgesic drugs from aqueous solutions. Briefly, 36.4 g Fe(NO3)3·9H2O (Roth, Karlsruhe, Germany, 96%) and 20.7 g tartaric acid (C4H6O6) (Merck, Darmstadt, Germany, 99.5%) were dissolved in 50 mL distilled water and the resulting solution was then transferred into a round-bottom flask containing 13.9 g activated carbon.
The mixture was kept for 24 h to impregnate and was afterwards placed inside a heating mantle at 400 °C so that the combustion reaction could take place ( Figure 12).
The combustion reaction took place in the absence of air just to prevent the formation of γ-Fe2O3 instead of Fe3O4. The gases released after the combustion process were evacuated in a distilled water dish according to the method described by Ianoş and coworkers [39]. After gaseous emission stopped, the obtained magnetic black powder was washed several times with warm distilled water (60-80 °C) and then dried in an oven at 70 °C.

Fe 3 O 4 @C Matrix Characterization
The phase composition of the sample was investigated by X-ray diffraction (XRD) using a Rigaku Ultima IV instrument (Tokyo, Japan) operating at 40 kV and 40 mA. The XRD pattern was recorded using CuK α radiation. The magnetic property of the sample was measured at room temperature by vibrating sample magnetometry using a VSM 880 ADE/DMS instrument (Massachusetts, USA). The heating behavior of the sample was studied within the range of 25 to 1000 • C by thermal analysis using a Netzsch STA 449 C instrument (Selb, Germany). The thermogravimetric (TG) and differential scanning calorimetry (DSC) curves were recorded using aluminum crucibles under artificial air flow of 20 mL/min at a heating rate of 10 • C/min.
The FTIR spectrum of the sample was carried out using a Shimadzu Prestige-21 spectrometer (Duisburg, Germany) within the range of 400 to 4000 cm −1 , using KBr pellets and a resolution of 4 cm −1 .
SEM-EDAX analysis was carried out with EDX detection on scanning electron microscope (SEM), using an EDAX detector (ZAF Quantification-Standardless, Element Normalized) with FEI Quanta 250 microscope (Eindhoven, Holland). SEM analysis parameters were HV mode, 30 kV, ETD (Everhart-Thornley detector for secondary electrons), with two magnification orders, one for a general overview of the image/measurements and another for higher surface topography for regions of interest. The identified chemical species were expressed in weight percent (Wt %) or atomic percent (At %).

Paracetamol and Acetylsalicylic Acid Adsorption Studies
The adsorption experiments for the removal of the drugs from aqueous solution were performed at 25 • C, using 2 g/L adsorbent mass, and various initial concentrations: 10-900 mg/L for paracetamol and 10-800 mg/L for acetylsalicylic acid solutions. In order to study the adsorption kinetics, 25 mL of both paracetamol and acetylsalicylic acid solutions (initial concentration-50 mg/L for paracetamol solution and 100 mg/L for acetylsalicylic acid solution) were mixed with 50 mg of Fe 3 O 4 @C matrix in an Erlenmeyer flask and continuously stirred at 250 rpm and controlled temperature of 25 • C in an Environmental Shaker Incubator ES-20/60 BioSan (Riga, Latvia).
The recording started when the stirring began and six samples were collected between 15 min and 4 h. All adsorption assays were made in duplicate. After the magnetic separation, the concentrations of both paracetamol and acetylsalicylic acid solutions remaining in the Erlenmeyer flask were assessed with a UV-Vis spectrophotometer T70 UV/Vis Spectrometer, PG Instruments (Leicestershire, UK) at 202 nm for paracetamol and 226 nm for acetylsalicylic acid. The adsorbed amount (q t [mg/g]) and the removal efficiency (R [%]) as a result of the adsorption process were determined according to Equations (1) and (2): where C 0 -the initial concentrations of the paracetamol and acetylsalicylic acid solutions, respectively The time needed to reach equilibrium was chosen based on the kinetic studies of Fe 3 O 4 @C matrix-3 h for both drugs. The kinetics of paracetamol and acetylsalicylic acid adsorption, respectively, onto the Fe 3 O 4 @C matrix was evaluated by means of two different models: the Lagergren pseudo-first-order [55] and pseudo-second-order, respectively [56].
The rate expression for the Lagergren pseudo-first-order is represented by the equation where, k 1 -the adsorption rate constant of the pseudo-first-order [min −1 ]; q e -amount of solute adsorbed at equilibrium [mg/g]; and q t -amount of solute adsorbed at time t [mg/g]. The integration of Equation (3) between the limits t = 0 and t = t, with the values of q t = 0 and q t = t, respectively, gives the linear form (Equation (4)) of the Lagergren equation: where, q e -amount of solute adsorbed at equilibrium [mg/g]; q t -amount of solute adsorbed at time t [mg/g]. By graphically representing the mathematical function a straight line will be obtained, from which one can calculate the adsorption rate constant of the pseudo-first-order (k 1 ) by using the slope of the straight line.
The rate expression for the pseudo-second-order is given by the equation [57] dq t dt = k 2 × (q e − q t ) 2 where, k 2 -adsorption rate constant of the pseudo-second-order [g/mg·min]. By integrating the Equation (5) between the limits t = 0 and t = t and by using the values q t = 0 at q t = t, respectively, the linear expression (Equation (6)) of the pseudo-second-order kinetics model will result, as follows The graphical representation of the mathematical function will consist of a straight line from which the adsorption rate constant of the pseudo-second-order (k 2 ) can be calculated by using the slope of the straight line.
The experimental adsorption data in the case of paracetamol and acetylsalicylic acid were fitted to the linear form of the kinetic models. The linear regression analysis was employed in order to validate the kinetic models.
The experimental equilibrium data were fitted by the consecrated isotherm models: Langmuir (Equation (7)), Freundlich (Equation (8)), and Redlich-Peterson (Equation (11)). The adsorption isotherms provide important information regarding the adsorption mechanism, the surface properties of the adsorbent, the affinity between the adsorbent and the adsorbate as well as the quantitative evaluation of the adsorption process performance [58][59][60][61].
The Langmuir isotherm is described by the equation where, q e -the amount of solute adsorbed at equilibrium [mg/g]; q m -the maximum monolayer adsorption capacity at the formation of the molecular layer [mg/g]; C e -the equilibrium concentration of the solute [mg/L]; and K L -the Langmuir adsorption constant [L/mg]. The Freundlich isotherm is expressed as follows where, K F -the Freundlich constant [mg 1−1/n L 1/n g −1 ], indicating the adsorption capacity of the adsorbent; n-a constant, indicating the intensity of adsorption (dimensionless). The Langmuir and Freundlich equations allow the assessment of the adsorption capacity of the carbon surface magnetic matrix. The q m and K L constants can be obtained by linearization of Equation (7): C e q e = 1 q m × K L + 1 q m × C e By graphically representing the function a straight line will emerge from which the values of q m and K L constants can be calculated by using the slope and the intercept of the straight line. The Freundlich equation can also be expressed as a linear function: ln q e = ln K F + 1 n × ln C e The graphical representation of the function will produce a straight line from which the n parameters and K F constant can be calculated by using the slope and the intercept of the straight line, respectively.
The Redlich-Peterson isotherm is an empiric isotherm that contains three parameters; this isotherm combines the characteristics of both Langmuir and Freundlich isotherms. The Redlich-Peterson isotherm is described below. q e = K RP × C e where, K RP -the Redlich-Peterson constant [L/g]; α RP -constant [L/mg] β ; and β-an exponent ranging between 0 and 1. Equation (11) reduces to Henry's Law when β = 0 and becomes the Langmuir isotherm when β = 1 [47].
The correlation coefficient (R 2 ) was obtained by nonlinear regression in order to choose the most suitable isotherm model for the experimental equilibrium data.

Conclusions
The potential of using a Fe 3 O 4 @C matrix obtained by combustion method as adsorption material for the removal of paracetamol and acetylsalicylic acid was investigated. Results show that the as synthesized magnetic matrix is suitable for both paracetamol and acetylsalicylic acid removal; this being the first such study reported in the literature.
Taking into account that at a specific dose both drugs become toxic, their removal from the biological/aqueous medium is imperative. It is well known that magnetite (Fe 3 O 4 ) shows superparamagnetic properties and is biocompatible in specific conditions. In addition, activated charcoal is the most commonly used adsorbent in almost all cases of drug intoxication.
Altogether, our synthesized matrix, based on both magnetite and activated charcoal, can be successfully used in the adsorption of various drugs from aqueous solution, wastewater and even biological medium.
Besides the novelty offered by this study (two drugs were adsorbed onto a magnetite-core active charcoal-shell matrix), the magnetic properties of Fe 3 O 4 @C matrix are of great significance. Such properties allow these type of nanoparticles to be magnetically controlled efficiently, and they can also be reused in multiple adsorption-desorption cycles. These aspects represent important findings that can be used as various starting points for future studies.