2.1. Materials

All used chemicals are analytical grade (99.99%). The synthesis process of AC/Fe₃O₄ composite is presented in Scheme 1. RHs were collected in Phu Thuong ward, Hue city, Thuan Hoa district, Vietnam. After being washed with water, RHs were dried and calcined at 400°C for 3 h to obtain rice husk ash (RHA). The amount of silicon was removed by soaking RHA in NaOH (China, PA) solution before synthesizing AC. RHA was soaked in NaOH solution with a NaOH/SiO₂ molar ratio of 2.5 at 200°C for 30 min and then at normal temperature for 24 h. The solid is filtered, dried, and activated at 700°C for 3 h. Then, it is stirred in an activating agent of HCl (Merck) 0.05 M solution for 12 h. The solid is filtered, dried at 100°C for 24 h, and activated again at 700°C for 3 h to obtain AC.

The resulting AC was added to a 250-mL flask containing 50 mL of Fe₂(SO₄)₃ (China, PA) solution with continuous N₂ aeration. In this condition, (NH₄)₂Fe(SO₄)₂ was added to the above mixture with a defined weight ratio of Fe₃O₄:AC (w/w). Then, the mixture of NH₃:ethanol with a volume ratio of 1:1 (v/v) was added to form Fe₃O₄ nanocrystals on the surface of AC. The composite was separated from the mixture after that had been under ultrasonic conditions at 80°C for 60 min before and dried at 80°C.

The bonding between functional groups (COO⁻) on the surface of AC and Fe³⁺/Fe²⁺ is formed during the synthesis, resulting in AC/Fe₃O₄ composite.

2.2. Methods

2.2.2. Simultaneous Determination of CIP and TET

CIP and TET concentrations were simultaneously determined on the Cary 60 UV–Vis (Agilent) equipment by UV–Vis molecular absorption spectrometric at nonoverlapping maximum absorption wavelengths which is 274 nm for CIP and 363 nm for TET (Figure 1). The comparison of the individual spectra of each antibiotic with the spectrum of binary mixture of them confirmed their influence on each other. The quantification of CIP and TET was carried out by the standard addition method. The substrate was the solution with a pH of 7.4 collected after soaking AC/Fe₃O₄ composite in distilled water.

3.1. Characterization of AC/Fe₃O₄ Composite

The phase composition of AC/Fe₃O₄ composite containing different weight percentages of Fe₃O₄ (15%, 30%, 45%, 60%, 75%, and 90%) is presented in Figure 2(a). All XRD patterns of materials expose the characteristic diffraction peaks at 2θ of 30.4, 35.7, 43.7, 53.2, and 57.2°, corresponding to (220), (311), (400), (422), and (511) lattice faces which demonstrate the spinel structure of the phase of Fe₃O₄ (JCPDS No 19-0629) [55, 56]. The crystal size of Fe₃O₄ nanoparticles calculated from the XRD diagram (black line) is around 17 nm which is quite similar to this determined by Juturu et al. [57]. The presence of AC reduces the magnetically induced agglomeration of Fe₃O₄ nanoparticles, so the smaller crystal size of nanoparticles in the composite is observed around 11 nm for AC/90Fe₃O₄, 13 nm for AC/75Fe₃O₄, and 16 nm for AC/60Fe₃O₄. The lattice constant obtained is 0.8322 nm which meets the standard of magnetic nanoparticles [57]. The intensity of the peaks gradually decreases as the weight percentage of Fe₃O₄ decreases. For the AC sample, the peak band at 2θ of 24° and the peak band at 2θ of 44.3° corresponding to (002) and (100) lattice faces are characteristic of the amorphous phase of carbon. A sharp and very low peak appears at 2θ of 33.7, corresponding to a small amount of crystalline phase of SiO₂ in the RH that has not been completely removed during the soaking of RHA in NaOH [58].

Raman spectra of materials (Figure 2(b)) are recorded to determine the nature of the iron oxide core. Panta and Bergmann pointed out that low laser powers produced weak Raman scattering for nano-Fe₃O₄ because of the laser-induced phase transition [59]. In the present study, a 785 nm laser (power of 1.2 mW) is used to create the Raman spectrum of the sample. The results in Figure 2(b) show that the main structures of the spectrum are 297, 386, 495, 667, and 1300 cm⁻¹, in which the peak at a wavenumber of 667 cm⁻¹ is characteristic of the Fe₃O₄ phase, and the remaining peaks are predicted to be characteristic of the γ-Fe₂O₃ phase because the magnetic iron oxide sample has been oxidized by a high-power laser. This allowed confirmation of the formation of Fe₃O₄ [59, 60].

For AC materials, the Raman spectrum shows that peak D (D = disorder) at 1337 cm⁻¹ characterizes the presence of amorphous carbon and structural defects. The structure of graphite is shown in peak G (G = graphite) at 1589 cm⁻¹. The level of defects in the material structure is evaluated using the ratio of peak intensity D and G (I_D/I_G). The larger this ratio is, the higher the level of defects is [61, 62].

With the presence of AC in the composite and the increase in weight percentage of AC (from AC/90Fe₃O₄ sample to AC/15Fe₃O₄ sample), the characteristic peaks of AC appear more clearly and the intensity of the characteristic peaks of Fe₃O₄ gradually decreases. It is found that the I_D/I_G ratios of AC/Fe₃O₄ samples are higher than that of AC sample (as shown in Table 1), which confirms that the penetration of Fe₃O₄ nanoparticles into the pores of AC results in the higher level of defects of AC.

Figure 3(a) shows that Fe₃O₄ and AC/Fe₃O₄ materials display superparamagnetic behavior due to the fact that the superparamagnetic curve and hysteresis loop of each material coincide and pass through the origin of the axes [63, 64]. In other words, Fe₃O₄ and AC/Fe₃O₄ materials exhibit negligible coercivity and remanence when removing the applied magnetic field [65]. The saturation magnetization of the Fe₃O₄ material is about 52 emu which is quite higher than that of some previous studies [66–69]. The increase in Fe₃O₄ weight percentage of AC/Fe₃O₄ composite from 15% to 90% corresponding to the reduction of the amount of diamagnetic AC resulted in the gradual enhancement of the saturation magnetization from 6.83 to 41.52 emu (Table 1). Figure 4 displays a strong magnetic interaction of AC/75Fe₃O₄ with a magnet, which means that the separation of the composite from a solution will be very fast and effective. On the contrary, the lowering of the BET surface area of AC/Fe₃O₄ composites (Table 1) from 407.7 to 131.6 m²·g is observed, which may be due to the filling of AC pores with Fe₃O₄ nanoparticles. This could underline the formation of the composite structure. The mesoporous structure of the composites is confirmed based on the pore diameter ranging from 2.2 to 6.6 nm [70]. The N₂ adsorption and desorption isotherm of AC express that the material exhibits a noncapillary structure. For Fe₃O₄ and AC/75Fe₃O₄ composites, hysteresis loops appear which is characteristic of Type IV according to IUPAC classification, indicating the existence of mesoporous material (Figure 3(b)).

The CIP and TET simultaneous adsorption efficiency of AC/Fe₃O₄ composites were checked with an initial concentration of each adsorbate of 20 mg/L and adsorbent dosage of 0.4 g/L. The results as shown in Figure 5 present the lowest efficiency of Fe₃O₄ material (42.5% for CIP and 46.5% for TET). The higher efficiency is obtained for AC/Fe₃O₄ composites and AC material. By partially replacing Fe₃O₄ with AC and gradually increasing the AC content, the adsorption efficiency is greatly improved. The rather strong enhancement of efficiency (from 42.5% to 74.3% for CIP and from 46.5% to 84.2% for TET) is observed when 25% (w/w) of Fe₃O₄ is replaced with AC, and after that, the increase is lighter. More than 83% of CIP and 89% of TET are removed from aqueous solution when replacing 85% (w/w) of Fe₃O₄ with AC. This confirms that the combination of AC and Fe₃O₄ increases the adsorption capacity of AC/Fe₃O₄ composite compared to Fe₃O₄ material. Although AC material exhibits higher adsorption efficiency than bared-Fe₃O₄ material, the efficiency is still lower than that of AC/Fe₃O₄ composite. Besides, compared to AC, the easier separation of the composite after treatment with an external magnetic field is recognized. However, with the aim of overcoming the difficulty in separating AC from solution and the weak adsorption capacity of Fe₃O₄, the composite of AC/75Fe₃O₄ exhibits the most significance because the adsorption efficiency is quite high (more than 74% for CIP and 84% for TET) and the solution becomes transparent after being settled in an external magnetic field for only 15 min.

The elemental composition of AC/75Fe₃O₄ material is presented in Figure 6(a). The results of the weight percent of the main elements which are C, Fe, and O in 3 random material samples (n = 3) are unremarkable different (the standard deviation value of the average value is very small). Their existence suggests that there are no significant contaminants in the obtained composite. The mass percentage of Fe₃O₄ calculated from the EDX analysis is 76.73% which is in accordance with this value theoretically calculated according to the raw material (75%). This partly confirms the fair distribution between Fe₃O₄ and AC. Although it is very small, an amount of Si remains in the composite because the Si removal with NaOH missed it in RHA.

XPS spectrum as shown in Figure 6(b) indicates the elemental and chemical state information of AC/75Fe₃O₄ composite. Two peaks at 724.78 and 711.28 eV corresponding to the Fe2p^1/2 and Fe2p^3/2 spin–orbit peaks which are assigned to the presence of Fe₃O₄ composed of Fe²⁺ octahedron, Fe³⁺ octahedron, and Fe³⁺ tetrahedron. Each peak of Fe2p (Figure 6(c)) contains two peaks corresponding to Fe(II) (711.08 and 725.08 eV) and Fe(III) (712.08 and 728.08 eV). The lattice oxygen at 530.08 eV is attributed to oxygen atoms in the octahedron and tetrahedron lattice. AC is visible in the composite due to the C1s peak at 284.48 eV corresponding to C-C and C=C bonds. All the peaks support the successful anchoring of Fe₃O₄ nanoparticles onto AC [70]. These results are in an agreement with some reports [71–73]. The atomic percentages of Fe2p, C1s, and O1s found from the XPS analysis are 24.03%, 35.55%, and 40.42%, respectively. These results confirmed the weight percent of Fe₃O₄ in the composite is 76.82%, which is an agreement with EDX analysis (76.73%).

Functional group characteristics on the surface of materials are found on the IR spectra as shown in Figure 7(a). For all materials, a wide peak band associated with OH vibration is observed at the wavenumber around 3400 cm⁻¹ [29, 36]. The appearance of the peaks centered at the wavenumbers of 1641, 1408, and 864 cm⁻¹ indicates C=O, C-C, and C-H vibrations, respectively, on the surface of AC [74, 75] and AC/Fe₃O₄ composite [35]. The peak at the wavenumber around 550 cm⁻¹ assigned to Fe-O stretching vibration confirms the existence of Fe₃O₄ in the composite [29, 35, 36].

The PZC is one of the important characteristics of the adsorbent. The PZC is the pH value of the solution when the charge of the adsorbent surface is zero. Figure 7(b) shows the data for determining the PZC of AC/75Fe₃O₄ material. The curves representing the dependence of pH_f − pH_i on pH_i in 0.1 M and 0.01 M NaCl solutions intersect the horizontal axis at pH values of 5.9 and 5.8, respectively. The two intersections obtained are quite close to each other, so pH_PZC can be chosen at 5.85. At a pH less than 5.85, anions are adsorbed more strongly than cations (adsorbates). Conversely, at a pH higher than pHpzc, cations are more readily adsorbed.

The morphology of materials as presented in Figure 8 shows that spherical Fe₃O₄ nanoparticles with a uniform size of about 10 nm are distributed on the surface of the AC material which is highly porous.

3.2. Simultaneous Determination of CIP and TET Using UV–Vis Absorption Spectroscopy

The UV–Vis spectra of CIP and TET in the range wavelength from 200 to 400 nm are presented in Figure 1. In the single-component solution with the desired concentrations of CIP of 1.5 mg/L and TET of 2.0 mg/L, the experimental concentrations measured by UV–Vis absorption spectroscopy are 1.45 and 1.97 mg/L. These results are quite consistent with the concentrations of CIP and TET determined in the binary-component solution which are 1.57 and 2.08 mg/L, respectively. As can be seen, the determined concentrations of CIP and TET in the binary-component solution increase slightly compared to the single-component solution. Therefore, it can be considered that the effect of each antibiotic on each other’s UV–Vis spectrum is negligible.

Three experimental turns are set with prespiked concentrations of 1.07, 0.53, and 1.03 mg/L for CIP and 1.02, 2.02, and 1.46 mg/L for TET, respectively. The spiked concentrations of CIP and TET in the two first turns are 0.50 and 1.00 mg/L, while, for the third turn, both CIP and TET are spiked with the content of 1.00 mg/L. The postspiked concentrations are found at 1.61, 1.01, and 2.02 mg/L for CIP and 2.07, 2.94, and 2.43 mg/L for TET. These results demonstrate the values of recovery in 3 turns of the experiment are in the required range which is from 90% to 110%, which confirms the accuracy and precision of the method.

The t_exp value is compared with the theory value (t_tab) looked up from the table at reliability P of 0.05 and degrees of freedom f = n − 1 (n is the number of experimental values measured by the UV–Vis method). The results as shown in Table 2 show that all of t_exp values are lower than t_tab at P of 0.05 and f of 2 (2.920), which confirms that the experimental values of the UV–Vis absorption method are reliable. In other words, the UV–Vis spectroscopy exhibits high accuracy in simultaneously determining CIP/TET content.

3.3. Simultaneous Adsorption of CIP and TET Onto AC/Fe₃O₄ Composite

3.3.1. Effect of pH

The results show that with the initial CIP and TET concentrations of 20 mg/L and the AC/75Fe₃O₄ composite dosage of 0.4 g/L, the simultaneous adsorption efficiencies of CIP and TET strongly increase (from 53.4% to 88.3% for CIP and from 52.3% to 94.4% for TET) according to the rising of pH from 2 to 7 and 8, then gradually decrease from pH 8 to 12. CIP and TET (Figure 9) exhibit two and three pK_a values including 5.9 and 8.9 for CIP [77] and 3.3, 7.8, and 9.6 for TET [78]. CIP shows a zwitterionic functional group (CIPH⁺⁻) in the pH from 5.9 to 8.9. In a solution with pH lower than 5.9 and higher than 8.9, CIP molecules exist as CIPH₂⁺ and CIP⁻ ions, respectively. TET molecules have a neutral charge (TETH₂⁺⁻) at pH from 3.3 to 7.8, and in the case of pH from 7.8 to 9.6 and higher than 9.6, TETH⁻ and TET²⁻ are dominant, respectively. In contrast, TETH₃⁺ ions are prevalent. The species of CIP and TET affect their adsorption capacity on the material.

With a pH lower than 5.9 (pH_PZC of the adsorbent) (Figure 7(b)), the protonated surface of AC/75Fe₃O₄ composite hinders the accessibility of positive ions such as CIPH₂⁺ and TETH₃⁺, resulting in low adsorption ability. However, the increase of pH from 2 to 6 tends to the downtrend of positive charge on the surface of the material which enhances the attraction of positive ions to the surface of the adsorbent. That means the adsorption of CIP and TET is more favorable. Then, the adsorption efficiency of CIP and TET decreases at a pH higher than 8. The reason is that the negative surface of material creates a repulsive interaction with negative ions including CIP⁻, TETH⁻, and TET²⁻. It is found that the higher the pH is, the stronger the repulsive force is. In a neutral (pH of 7) or slightly alkaline (pH of 8) solution, the adsorption is maximum due to the negligible resistance of the material surface to zwitterionic functional groups (CIPH⁺⁻ and TETH₂⁺⁻). The highest adsorption efficiency reached about 86%–88% for CIP and 93%–94% for TET.

3.3.4. Adsorption Kinetic Study

With CIP and TET concentrations of 20 mg/L and an adsorbent dosage of 0.4 g/L, the linear HSO kinetic equation exhibits higher R² values (0.995 for CIP and 0.991 for TET) than those of the linear LFO kinetic equation (0.957 for CIP and 0.941 for TET). The adsorption capacity calculated from LFO and HSO equations is 32.16 mg·g⁻¹ and 40.32 mg/g for CIP and 57.45 and 47.39 mg/g for TET, respectively, while the experimental adsorption capacity is 35.73 mg/g for CIP and 42.15 mg/g for TET. These results indicate the compatibility between the HSO model with experimental data more than the LFO model. The rate of the adsorption which is the chemisorption process is a function of both adsorbent and adsorbate concentrations.

Weber–Morris IPD study depicts the diffusion of CIP/TET through two stages as follows: (i) the first is from the solution to the surface of AC/Fe₃O₄ particle causing a solution layer (L) which surrounds the particle (k_dif 1) and (ii) the second is from the particle surface to its intraparticle capillary (k_dif 2). C value in equation (7) is a constant relating to diffusion resistance and characteristic for the thickness of the L layer. A higher value of C means a thicker L layer which will hinder the diffusion of the CIP/TET into the intraparticle capillary of AC/Fe₃O₄ particle [48].

Figure 10 indicates the C value is very close to zero during the first 60 min of the adsorption process for both CIP (3.639) and TET (0.595), which means a thin layer of CIP and TET surrounding the composite results in the fast movement of CIP/TET molecules passing through this layer. Therefore, a strong uptrend of the adsorption capacity of CIP/TET is observed, while after 60 min, the strong increase in C value (36.053 for CIP and 41.739 for TET) demonstrates the enhancement of the thickness of the L layer, hindering the diffusion of CIP/TET inside the composite capillary. This stage is predicted as a step determining the rate of CIP/TET adsorption. The fact that k_dif1 is greater than k_dif2 refers to a quick external mass transfer for the first 60 min, after that, a slower internal diffusion takes place.

The mechanism of CIP and TET adsorption onto AC/Fe₃O₄ material can be proposed as shown in Figure 11 [57, 70]. Possible interactions between CIP/TET and the surface of the composite are (i) electrostatic interactions between the amine groups of CIP/TET and negatively charged oxygen atoms of Fe₃O₄, COO⁻ groups of CIP, and the metal ion of Fe₃O₄; (ii) H-bonding between OH groups of CIP/TET and O atoms of Fe₃O₄, OH groups on the surface of AC and N, F atoms of CIP/TET; (iii) π–π interactions between aromatic rings of CIP/TET and the C=C group of AC; and (iv) pore filling of AC with CIP/TET molecules.