Synthesis of magnetic graphene oxide doped with strontium titanium trioxide nanoparticles as a nanocomposite for the removal of antibiotics from aqueous media

Abstract

In this study, strontium titanium trioxide (SrTiO₃) nanoparticles were synthesized and doped onto graphene oxide (GO) based magnetic nanoparticles (MNPs) simply via ultrasound. The newly synthesized GO/MNPs–SrTiO₃ magnetic nanocomposite was characterized by using Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and X-ray diffraction (XRD). As prepared GO/MNPs–SrTiO₃ was applied for the removal/adsorption of antibiotics namely tetracycline and cefotaxime from water samples. The effective parameters on the adsorption process including adsorbent dosage (mg), NaCl salt (% w/v), and solution pH were studied and their optimum values were obtained. The GO/MNPs–SrTiO₃ nanocomposite showed high adsorption capacities of 65.78 mg g⁻¹ and 18.21 mg g⁻¹ toward tetracycline and cefotaxime, respectively. A comparison study of Langmuir, Freundlich and Dubinin-Radushkevich isotherm models along with a kinetic study showed that the adsorption experimental data were well fitted to the Langmuir isotherm (R² > 0.996) and pseudo-second-order rate model (R² > 0.995). An intra-particle diffusion model demonstrated that the adsorption was not controlled by diffusion process. The thermodynamic studies (+ΔH) of the antibiotics adsorption onto GO/MNPs–SrTiO₃ revealed an endothermic nature. However, adsorption isotherms, free energy, kinetic models and thermodynamic studies suggested a monolayer sorption model followed by physisorption process for the selected antibiotics.

---

1. Introduction

Recently, great concern is mounting toward environmental contamination regarding the use of pharmaceutical substances.¹ Antibiotics are being discharged into water resources through municipal wastewater treatment plants and pharmaceutical manufacturing processes.² Antibiotics are the most common pharmaceutical substances that have been widely used to treat infectious diseases of humans and animals.^3–5 Tetracycline is the second most-widely used antibiotic in the world, thus it would be distributed in the environment and create environmental problems.^6,7 Although, its side effects are not common excess use of it can cause vomiting, nausea, diarrhea and sore throat.⁸ Cefotaxime is another antibiotic that is illegally used in food preservation and to treat respiratory and urinary infections, since it is a highly active agent against bacteria.^9,10 The excessive use of cefotaxime causes possible side effects i.e. skin rash, fever, body aches, and bruising.^9,11 Antibiotic residues in the environment can lead to the development of microorganism resistance and bacterial resistance.^3,12 However, chemical and/or physical treatment is needed to prevent the undesired antibiotics entering water resources.

Advanced approaches such as solid phase extraction, microextraction, ozonation, electrochemical, electrodegradation, biodegradation, and adsorption process have widely been developed for antibiotics monitoring.^13–15 Among these, adsorption process based on synthetic or biomass materials has been introduced as an effective, economic, simple, environmental friendly, high efficiency, easy recovery and renewable method.^16,17 However, adsorbent materials with high sorption capacity and high affinity toward antibiotics is still under development to overcome disposal problems and reduce price.¹⁸ Recently, carbon-based materials have been successfully used for antibiotics adsorption/removal due to their low cost, easy disposal, high surface area and large π-stacking systems.^18–22 Graphene oxide (GO) has been individually considered to be a promising adsorbent to antibiotics decontamination, since GO owing to its immense high surface area, two-dimensional structure, chemical stability, environmental friendly, biocompatibility and having massive hydroxyl, carboxyl and epoxide polar functional groups.^18,23–25 Due to decreased removal efficiency of GO followed by GO-sheets aggregation and self H-bonding (inter-layer and/or intra-layer), the researchers are interested in development of novel GO-doped nanocomposites.²⁶ Strontium titanium trioxide (SrTiO₃) nanoparticles as an interesting nanoparticles have been exploited extensively for water decontamination due to their thermal and chemical stability, high surface area, photocatalytic activity and large carrier effective mass.^27–30 It has been reported that graphene oxide doped with SrTiO₃ nanoparticles provided an environmental friendly, low cost nanocomposite with a good thermal stability.³⁰ A survey of literatures showed that SrTiO₃ based nanocomposite has been rarely considered as adsorption material.

In this study, magnetic graphene oxide doped with SrTiO₃ nanoparticles (GO/MNPs–SrTiO₃) was synthesized and characterized for the first time. The existence of magnetic nanoparticles (MNPs) in the nanocomposite provide fast separation of ultra-light materials (GO and SrTiO₃) from aqueous media without applying a filtration or centrifugation step. The SrTiO₃ nanoparticles doped on GO synergistically prevent effectively the GO sheets aggregation and self H-bonding. The newly synthesized GO/MNPs–SrTiO₃ could be successfully used for adsorption of tetracycline and cefotaxime antibiotics from water sample. We predicted that the synthesized GO/MNPs–SrTiO₃ indicate a high adsorption capacity toward tetracycline and cefotaxime. This property can be attributed to the van der Waals interaction (π–π and electrostatic interactions), n–π interaction, and H-bonding via O–H/N–H terminal. The validity of the experimental data will be investigated by testing different adsorption isotherms, kinetic rate and thermodynamic models to distinguish the mechanism and type of the adsorption process.

2. Experimental

2.1 Reagents and materials

Antibiotics namely tetracycline hydrochloride and cefotaxime sodium salt (Table 1) were purchased from Solarbio Science & Technology Co., Ltd (Beijing, China). Strontium nitrate, titanium dioxide (TiO₂), methanol, hydrochloric acid (HCl), sodium hydroxide (NaOH), sodium chloride (NaCl), ammonia solution (25%), ferrous chloride tetrahydrate (FeCl₂·4H₂O), ferric chloride hexahydrate (FeCl₃·6H₂O) and other chemicals were obtained from Merck Chemicals (Darmstadt, Germany).

Table 1 Chemical structure of the selected antibiotics and their log P and pK_a values

a Source: Drug Bank (http://www.drugbank.ca/drugs/DB00759).
Name	Tetracycline hydrochloride	Cefotaxime sodium salt
Chemical structure
log P^a	−0.56	−0.14
pKa^a	3.3	3.18

---

2.2 Instruments

The infrared spectra were recorded on a Bruker Equinox 55 FT-IR spectrometer (Bremen, Germany), with the wavenumber set in the range from 450 to 4000 cm⁻¹. A KYKY-EM3200 scanning electron microscope (Beijing, China) was used for surface morphology study of the synthesized adsorbent. A Bruker X-ray diffractometer (Bremen, Germany) was used for crystalline analysis with its angle range (2θ) was set from 10° to 90° using CuK radiation (λ = 1.5418 Å). The magnetic properties of the adsorbent were investigated using a vibrating sample magnetometer (VSM/AGFM Meghnatis Daghigh Kavir Co., Kashan, Iran). An ICP-AES (Vista-MPX, Varian Inc., Melbourne, Victoria 3170, Australia) equipped with a slurry nebulizer and a charge coupled device detector was used for determination of metal ion.

2.3 Synthesis of GO/MNPs–SrTiO₃ nanocomposite

Strontium titanium (SrTiO₃) nanoparticles were synthesized as reported previously.²⁷ Briefly, solution A was prepared by addition of 0.4 g of TiO₂ nanoparticles in 20 mL aqueous solution of NaOH (3 mol L⁻¹). Solution B obtained by dissolving 1.4 g of strontium hydroxide in 20 mL NaOH (3 mol L⁻¹). Then, solution A was added dropwise into solution B with vigorous stirring for 30 min. Next, the mixture was transferred into a glass vial (capped with aluminum seal) and kept for 72 h at 120 °C. Afterward, the product was washed several times using excess deionized water and methanol, then oven dried at 80 °C for 24 h.

In order to synthesize GO/MNPs–SrTiO₃, the freshly prepared SrTiO₃ nanoparticles were physically deposited on GO surface simply by sonication (Fig. 1). Typically, 1 g of synthesized graphene oxide based magnetic nanoparticles (GO/MNPs), as reported previously,³¹ were added into 10 mL distilled water : methanol (10 : 1) and sonicated for 1 h. In other glass vial, 0.5 g of as prepared SrTiO₃ was dispersed in 10 mL distilled water followed by sonication for 1 h. Then, SrTiO₃ solution was mixed by GO/MNPs solution, and thereafter the mixture was sonicated for 30 min. Finally, the produced GO/MNPs–SrTiO₃ was washed by excess of deionized water and methanol, and then oven dried at 80 °C for 24 h.

Fig. 1 Schematic process for the synthesis of GO/MNPs–SrTiO₃.

2.4 Adsorption procedure

Fig. 2 illustrates the schematic procedure for the removal of the antibiotics from aqueous media. The experiments were carried out by shaking the aliquot of adsorbent (GO/MNPs–SrTiO₃) with 20 mL of the water sample (include 20 mg L⁻¹ of each antibiotics). The adsorption performance of GO/MNPs–SrTiO₃ was studied under batch adsorption method by varying solution pH (2–10), adsorbent dosage (5–100 mg), salt effect (0.01–20%, w/v), contact time (1–180 min), initial antibiotic concentration (10–100 mg L⁻¹), and temperature (293–313 K). After each adsorption (batch) process, GO/MNPs–SrTiO₃ was separated from aqueous media using an external magnet. The residual concentration of antibiotics were measured by UV-Vis spectrophotometry at the corresponding wavelengths (Fig. 2). Regarding the removal performance and adsorption capacity, the antibiotics concentrations were recorded for both before and after adsorption process. The removal efficiency (E%) and adsorption capacity (q_e) were calculated using eqn (1) and (2), respectively.³²where, q_e (mg g⁻¹) is the equilibrium adsorption capacity, C₀ (mg L⁻¹) is the initial antibiotics concentration (before adsorption), C_e (mg L⁻¹) is the equilibrium antibiotics concentration (after adsorption), V (L) is the aqueous solution volume, and m (g) is the adsorbent dosage.

Fig. 2 Graphical procedure for removal of antibiotics from aqueous media using GO/MNPs–SrTiO₃.

3. Results and discussion

3.1 Characterization

The newly synthesized adsorbent was characterized using FT-IR, XRD, SEM, and TEM techniques. Fig. 3A shows the FT-IR spectrum of SrTiO₃ nanoparticles with the transmission peaks at 3412 cm⁻¹, 1622 cm⁻¹, 1475 cm⁻¹, 856 cm⁻¹ and 607 cm⁻¹ that probably correspond to O–H stretching, H–O–H bending, C=O stretching (mode in Sr–OOC), Sr–O/Ti–O (octahedron bending) and Ti–O (stretching vibration), respectively.²⁷ Therefore, the formation of SrTiO₃ is confirmed with two main characteristic peaks at 856 cm⁻¹ and 607 cm⁻¹ that respectively correspond to the stretching vibrations Sr–O/Ti–O and Ti–O in SrTiO₃.³³ Fig. 3B exhibits the IR spectrum of GO/MNPs in which the bands at 3413 cm⁻¹ and 584 cm⁻¹ can be ascribed to the allocate O–H stretching vibration in GO and Fe–O stretching vibration in MNPs (Fe₃O₄) nanoparticles, respectively. Besides, the IR signals at 2919/2857 cm⁻¹, 1734 cm⁻¹, 1572 cm⁻¹, 1467 cm⁻¹, 1204 cm⁻¹, and 1032 cm⁻¹, could be attributed to C–H (stretching vibration), C=O, C=C, C–C, C–O (epoxy), C–OH on the surface of graphene oxide.³⁴ The peak at 584 cm⁻¹ (Fe–O) is imply that the Fe₃O₄ nanoparticles (MNPs) are properly attached onto the GO surface. Fig. 3C indicates that the successful synthesis of GO/MNPs–SrTiO₃, since two extra peaks appeared at 1471 cm⁻¹ and 609 cm⁻¹ (Fig. 3C) as compared with GO/MNPs. In addition, these two main peaks are demonstrates the presence of strontium and titanium oxide on the surface of GO.

Fig. 3 FT-IR spectra of (A) SrTiO₃ nanocubes, (B) GO/MNPs, and (C) GO/MNPs–SrTiO₃.

X-ray diffractometry (XRD) was used for crystal structure analysis of the synthesized magnetic adsorbent. Fig. 4A–C illustrate the XRD patterns for MNPs (Fe₃O₄), SrTiO₃ nanoparticles, and GO/MNPs–SrTiO₃, respectively. The sharp XRD signals indicate that the as prepared nanoparticles are highly crystalline materials. Several characteristic peaks of Fe₃O₄ MNPs were detected at 2 theta (2θ) 30.15°, 35.52°, 43.17°, 53.56°, 57.09° and 62.71° (Fig. 4A). The XRD pattern for Fe₃O₄ MNPs is relatively broad that shows the particles size are small. The resulted signals are similar to the Fe₃O₄ MNPs pattern (XRD) that investigated previously.^35,36 Fig. 4B, consisting of sharp and clear diffraction peaks at 2θ = 32°, 40°, 45°, 56° and 67°, shows a good crystalline structure for the synthesized SrTiO₃ nanoparticles. These signals indicate the cubic perovskite structure (space group: Pm3m) of SrTiO₃ corresponds to JCPDS (joint committee on powder diffraction standards) card number no. 35-0734.²⁷ The SrTiO₃ spectrum shows that there is no extra peaks related to TiO₂ and/or other residuals, and thus confirms the successful preparation of pure SrTiO₃. Fig. 4C displays the presence of both Fe₃O₄ MNPs and SrTiO₃ in the GO/MNPs–SrTiO₃ nanocomposite. It is clear that the main sharp XRD signals for Fe₃O₄ marked by their 2 theta (35.52°, 57.09° and 62.71°) and SrTiO₃ (32°, 45°, 56° and 67°) were appeared in the final product (Fig. 4C), also, Fig. 4C shows a broad peak at 10–20° corresponding to the amorphous graphene-based structure. However, the position of peaks for Fe₃O₄ and SrTiO₃ nanoparticles did not shift but showed the signals intensity decreased as compared with the original signals in Fig. 4A and B. Additionally, these observations confirmed that the MNPs and SrTiO₃ nanoparticles have been successfully fabricated onto the amorphous graphene to produce GO/MNPs–SrTiO₃.

Fig. 4 The XRD analysis of: (A) magnetic nanoparticles (Fe₃O₄), (B) SrTiO₃ nanocubes, and (C) GO/MNPs–SrTiO₃.

The surface morphology of the synthesized nanocomposite was studied using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Fig. 5A and B illustrates the SEM micrographs of the newly synthesized nanocomposite. Fig. 5A shows that the SrTiO₃ nanoparticles with the size of 20–100 nm have been successfully synthesized by hydrothermal method. Fig. 5B represents the GO/MNPs–SrTiO₃ micrograph in which the as prepared GO sheets are clearly observed with the size of 1 μm. As can be seen, the MNPs and SrTiO₃ nanoparticles have been uniformly anchored on the graphene sheets.

Fig. 5 The SEM micrographs for (A) the SrTiO₃ particles, and (B) GO/MNPs–SrTiO₃ nanocomposite. The TEM images of GO/MNPs–SrTiO₃ with magnification scale of (C) 200 nm, and (D), (E) 25 nm, and (F) histogram for particles size distribution.

Fig. 5C and D depict the TEM images of GO/MNPs–SrTiO₃ nanocomposite at different magnification scales. As Fig. 5C shows, a large amount of nanoparticles has been distributed on the surface of transparent flake-like GO sheets. The presence of these nanoparticles effectively prevents GO aggregation. Fig. 5D at a higher magnification level, displays the presence of Fe₃O₄ magnetic nanoparticles (as small dark spots) and a SrTiO₃ nanoparticles particle stuck on GO graphene sheet. Therefore, the TEM micrographs reflect successful combination of Fe₃O₄ and SrTiO₃ nanoparticles with GO to produce the GO/MNPs–SrTiO₃ composite. Finally, the size distribution of SrTiO₃ particles were calculated using ImageJ software (developed by National Institutes of Health, Maryland, USA), and illustrated as a histogram in Fig. 5F. As can be seen, 75% of the particles are within the range of 5 to 15 nm.

The magnetic properties of as prepared Fe₃O₄ MNPs and GO/MNPs–SrTiO₃ were characterized using VSM technique at room temperature. Based on the VSM curves illustrated in Fig. 6A magnetic behavior of materials can be described. For superparamagnetic particles, the magnetization become zero when the external field approached zero. Saturation magnetization (M_s) can also be determined from the plateau part of the VSM curve. Fig. 6A shows that the magnetic hysteresis curve is S-like and passes through the zero point of magnetization with no remanence nor coercivity, suggesting superparamagnetic property of Fe₃O₄ MNPs and GO/MNPs–SrTiO₃. Furthermore, the high saturation magnetization amount of 75 and 30 emu g⁻¹ was observed for Fe₃O₄ MNPs and GO/MNPs–SrTiO₃ respectively, indicate superparamagnetic properties of them. The decrease in saturation magnetization of GO/MNPs–SrTiO₃ can be due to contribution of the nonmagnetic GO layers and SrTiO₃ to the total magnetization. However, it is well enough for magnetic separation of particles from solution by a magnet.

Fig. 6 (A) Magnetization curves for magnetic materials, (B) influence of solution pH, (C) adsorbent dosage, and (D) NaCl salt on the adsorption of antibiotics onto GO/MNPs–SrTiO₃.

3.2 Effect of different parameters on adsorption capacity

3.2.1 Effect of pH. The solution pH can affect the chemical structure of analytes (antibiotics) and surface charge of adsorbent.¹⁹ Solubility of the antibiotics also can be strongly affected by pH due to the various species of antibiotics at different pH levels.²⁰ Fig. 6B illustrates the effect of pH on the adsorption behavior of tetracycline and cefotaxime antibiotics onto GO/MNPs–SrTiO₃. The adsorption capacity (q_e) for tetracycline increased with increasing pH from 2 to 5, and thereafter, q_e values decreased gradually until pH 10. Regarding the pK_a value (3.3) of tetracycline, and zeta potential of graphene oxide-based adsorbent (pH_zpc ∼4),^37,38 tetracycline and the adsorbent are expected to be protonated. At various pH values, tetracycline has four different species (TCH₃⁺, TCH₂^±, TCH⁻ and TC²⁻) at various pH values.⁴ At pH < pK_a and pH < pH_zpc the q_e value declined due to the repulsion between protonated tetracycline (TCH₃⁺) and positive surface charge of the adsorbent. However, at low pH (2–4), the proton (H⁺) adsorption onto the adsorbent is more probable than tetracycline (TCH₃⁺) adsorption. Furthermore, electrostatic interaction between adsorbent and tetracycline can occur at pH 5–7 due to existence of TCH₂^± and TCH⁻ ions. At pH > pH_zpc decrease of q_e values is due to the repulsion forces between TC²⁻ and the negatively charged adsorbent. Fig. 6B shows that low pH values are more suitable for the cefotaxime adsorption onto GO/MNPs–SrTiO₃, and thus the highest q_e value was obtained at pH 2–4. In addition, determination of cefotaxime in water samples has been previously investigated at low pH (=2).^9,10 The significant decrease of the q_e value at high pH (7–10) is probably due to the degradation of cefotaxime followed by hydrolysis of the β-lactam ring and the acetoxy ester.¹⁰ Also, increasing pH promoted the deprotonation of cefotaxime followed by the repulsion between analyte and negatively charged adsorbent. The highest q_e value at low pH can be ascribed to maximum stability of cefotaxime (at pH range 4–6),³⁹ cefotaxime pK_a (3.1) and graphene oxide pH_zpc (∼4). These factors enhanced the formation of possible electrostatic interactions, H-bonding, and π–π interaction at pH 2–4. Thus, pH 5 and 4 are selected for further adsorption studies of tetracycline and cefotaxime, respectively.

3.2.2 Effect of adsorbent dosage. The amount of adsorbent is a key parameter in the adsorption process and directly affects the adsorption performance and cost of treatment.¹⁴ The influence of adsorbent dosage on the efficiency of method was investigated by varying amount of GO/MNPs–SrTiO₃ from 5 to 100 mg at pH 5 and 4 for tetracycline and cefotaxime, respectively (50 mg mL⁻¹ initial concentration and 30 min adsorption time). Fig. 6C indicates that by increasing adsorbent dosage from 5 mg to 40 mg, the adsorption percent increased from 20% to 95% for tetracycline. However, the adsorption (%) for cefotaxime was obtained >86% using 80 mg of the adsorbent. The increase of adsorption efficiency with an increase in adsorbent dosage (mg) is probably corresponds to the increase of available adsorption active sites. In contrast, adsorption capacity decreases with increase of adsorbent dosage, because, amount of analytes per unit mass of the adsorbent decreases considerably.^40–42 Thus, 40 mg and 80 mg of GO/MNPs–SrTiO₃ were used for tetracycline and cefotaxime adsorption, respectively.

3.2.3 Effect of salt concentration. The influence of NaCl salt (ionic strength) on the adsorption of tetracycline and cefotaxime by GO/MNPs–SrTiO₃ was studied from 0.0 to 20% (w/v) at the optimized pH (initial concentration of 50 mg mL⁻¹ and adsorption time of 30 min). Fig. 6D shows a slight decline for tetracycline adsorption (%) from 94.2% to 83.6%, as NaCl% increased from 0.0–0.5. Thereafter, adsorption (%) did not changed significantly by increasing NaCl salt to 20% (w/v). However, Fig. 6D indicates that by increasing NaCl% from 0.0 to 20, the adsorption efficiency decreased from 96% to 89% for tetracycline. The same trend was also observed for tetracycline adsorption using GO, that reported by Gao et al., (2012).⁴³ This trend can be explained by existence of electrostatic interactions (deprotonated carboxyl group of graphene oxide and charged amino group of tetracycline), weakened the π–cation interaction, and Na⁺ ions competition with TCH₃⁺.^43–45 Therefore, the NaCl effects on tetracycline adsorption was ignored for further studies. Moreover, Fig. 6D shows that the cefotaxime adsorption (%) increased from 50% to 84% by increasing the NaCl salt from 0.0% to 1.0%, and thereafter adsorption (%) became constant until 20% (w/v) of salt. This trend can be explained by existing Na atom (Na–C=O) in the structure of cefotaxime (Table 1) thus cefotaxime stability is increased as the NaCl salt increases up to 20% (w/v). Thus, 1.0% (w/v) NaCl was used for further studies.

3.3 Effect of contact time and kinetic studies

The contact time influence on the adsorption of tetracycline and cefotaxime onto GO/MNPs–SrTiO₃ was investigated by varying of time in the range of 5 to 180 min (Fig. 7A). It was observed that the adsorption of both tetracycline and cefotaxime reached to equilibrium within 30 min. This fast adsorption process is probably due to the presence of adequate available active sites on the surface of adsorbent. Fig. 7A also indicates that the tetracycline adsorption on GO/MNPs–SrTiO₃ is faster than the cefotaxime adsorption, since at similar contact time (30 min) the adsorption capacities were 40 mg g⁻¹ and 7.1 mg g⁻¹ for tetracycline and cefotaxime, respectively.

Fig. 7 (A) Effect of contact time, (B) pseudo-first-order model, (C) pseudo-second-order model, and (D) intra-particle diffusion.

The adsorption kinetic to describe the rate and mechanism of adsorption were also studied.⁴⁶ The potential kinetic models such as pseudo-first-order, pseudo-second-order and intra-particle diffusion expressed by eqn (3)–(5), respectively, were investigated.^47,48

ln(q_e − q_t) = ln q_e − k₁t (3)

q_t = k_idt^1/2 + C_i (5)

where q_e and q_t are the maximum adsorption capacity (mg g⁻¹) and equilibrium capacity (mg g⁻¹) at the time t, respectively. k₁ (min⁻¹) and k₂ (g mg⁻¹ min⁻¹) are the pseudo-first-order and pseudo-second-order rate constants, respectively. k_id is intra-particle diffusion constant and C_i is thickness of boundary layer. The slope and intercept of eqn (3) correspond to 1/q_e (theoretical) and adsorption rate (1/k₁), respectively (Fig. 7B). In eqn (4) the slope and intercept are respectively equal to q_e and k₂ (Fig. 7C). The linier form of intra-particle diffusion is shown in Fig. 7D.

The validity of the proposed kinetic models were studied by using the linearized eqn (3)–(5), and the calculated values of kinetic parameters were listed in Table 2. Accordingly, it was found that the pseudo-second-order with a higher coefficient of determination (R² > 0.995) and better q_e (theoretical) for both tetracycline and cefotaxime was well fitted with experimental contact time, since the theoretical q_e (47.61 mg g⁻¹) is close to the experimental q_e (46 mg g⁻¹) at the time t.⁴⁹ In addition, according to the results tetracycline and cefotaxime adsorption processes might be controlled by different interactive forces. The presence of two slopes in intra-particle diffusion linearized model (Fig. 7D) indicates that the adsorption process proceeds through two steps. The first step with a sharp slope (m = 1.87) demonstrates a fast adsorption process within the time range of 5–30 min. The second step with low slope (m = 0.83) provided the slower adsorption process in the time ranging from 30 to 180 min. Consequently, it was indicated that the antibiotics adsorption onto GO/MNPs–SrTiO₃ are not controlled by intra-particle diffusion.

Table 2 The pseudo-first-order, pseudo-second-order and intra-particle diffusion models for the adsorption kinetics of antibiotics (the experimental q_e were 46 mg g⁻¹ and 9.9 mg g⁻¹ for tetracycline and cefotaxime, respectively)

Model	Parameters	Antibiotics
		Tetracycline	Cefotaxime
Pseudo first order	qe (mg g^−1)	9.84	4.40
	k1 (min^−1)	0.022	0.012
	R^2	0.987	0.926
Pseudo second order	qe (mg g^−1)	47.61	9.42
	k2 (g mg^−1 min^−1)	0.0004	0.0111
	R^2	0.999	0.995
Intra-particle diffusion	Kid,1	1.87	0.83
	Ci1	30.92	2.63
	R1^2	0.924	0.974
	Kid,2	0.44	0.21
	Ci2	40.06	6.51
	R2^2	0.902	0.922

---

3.4 Effect of antibiotics initial concentration and adsorption studies

Fig. 8A depicts the antibiotics concentration effect on the adsorption behavior of the synthesized GO/MNPs–SrTiO₃. It is clear that q_e for tetracycline increased from 4.5 mg g⁻¹ to 62.05 mg g⁻¹, as the initial concentration increased from 10 mg L⁻¹ to 200 mg L⁻¹. For cefotaxime, the q_e value increased from 1.12 mg g⁻¹ to 20.73 mg g⁻¹ by increasing the concentration. This adsorption process is matched to the type I adsorption isotherm (set by IUPAC 1985), and demonstrates the monolayer adsorption process followed a Langmuir model.⁵⁰

Fig. 8 (A) Effect of initial concentration of antibiotics on the adsorption process, (B) Langmuir isotherm, (C) Freundlich isotherm, and (D) Dubinin-Radushkevich isotherm.

The experimental adsorption procedure for tetracycline and cefotaxime adsorption onto GO/MNPs–SrTiO₃ were investigated by adsorption isotherm models namely Langmuir, Freundlich, and Dubinin-Radushkevich. The linear equations of these models are expressed by eqn (6)–(8), respectively:

ln q_e = ln q_s − K_ad(ε²) (8)

where q_e (mg g⁻¹) is the equilibrium adsorption capacity, q_m (mg g⁻¹) is the maximum adsorption capacity, C_e (mg L⁻¹) is the equilibrium concentration for antibiotics after adsorption, and k_L is the Langmuir constant. The K_F [(mg g⁻¹) (L mg⁻¹)^1/n] is Freundlich constant that describes adsorption of antibiotics, and 1/n is the intensity of the adsorption. The q_s (mg g⁻¹) is the theoretical adsorption capacity, K_ad (mol² kJ⁻²) is a constant corresponding to energy of adsorption, ε is the isotherm constant, R (0.008314 kJ mol⁻¹ K⁻¹) is universal gas constant, and T (kelvin) is the temperature. Furthermore, the Langmuir model explains the monolayer adsorption process, but the Freundlich isotherm model represents multilayer sorption onto the heterogeneous surface.⁵¹ Dubinin-Radushkevich corresponds to multilayer adsorption with van der Waals forces onto heterogeneous surfaces.^51,52 However, free energy (E) of adsorption process is essential for exact prediction of sorption mechanism (eqn (9)).

E = (2K_ad)^−1/2 (9)

where E (kJ mol⁻¹) is free energy and K_ad is adsorption constant from eqn (8). If the obtained value of E is lower than −40 kJ mol⁻¹, adsorption mechanism follows a physisorption process.⁵³

Fig. 8B–D show respectively the linear plot of Langmuir, Freundlich, and Dubinin-Radushkevich models for antibiotics adsorptions. Plotting C_e/q_e versus C_e shows the Langmuir model linearity in that 1/q_m and 1/k_Lq_m correspond to the slope and intercept, respectively (Fig. 8B). Furthermore, plot of log  q_e versus the log  C_e displays the Freundlich model linearity in which log  K_F and 1/n are the intercept and slope, respectively (Fig. 8C). Similarly, in Dubinin-Radushkevich linear model, the ln q_s and K_ad values are equal to intercept and slope when plotting ln  q_e versus ε² (Fig. 8D).

Table 3 shows that the adsorption data are fitted well with the Langmuir model (Fig. 8B) due to the high coefficient of determination (R² > 0.995) as compared with Freundlich and Dubinin-Radushkevich models. This indicates that tetracycline and cefotaxime adsorption onto the adsorbent is monolayer. The maximum adsorption capacity (q_m) for tetracycline and cefotaxime was 65.78 mg g⁻¹ and 18.21 mg g⁻¹, respectively. In the Freundlich model, the values of n between 1 and 10, indicate that the adsorption process is favorable.⁵⁴ The Dubinin-Radushkevich isotherm model was examined for the experimental process, and the insufficient R² and low q_s values, indicate that the adsorption process does not follow a multilayer adsorption. Therefore, the results given in Table 3, suggest that the adsorption of tetracycline and cefotaxime onto GO/MNPs–SrTiO₃ follow a monolayer model of adsorption.

Table 3 Isotherm constants and parameters for the as prepared adsorbent

Model	Parameters	Antibiotics
		Tetracycline	Cefotaxime
Langmuir	qm (mg g^−1)	65.78	18.21
	kL (L mg^−1)	0.506	0.135
	R^2	0.997	0.996
Freundlich	KF [(mg g^−1) (L mg^−1)^1/n]	2.164	8.933
	n	2.041	2.169
	R^2	0.946	0.951
Dubinin-Radushkevich	qs (mg g^−1)	26.94	7.81
	Kad	0.263	0.312
	R^2	0.737	0.742
Free energy	E (kJ mol^−1)	1.37	1.26

---

3.5 Effect of temperature and thermodynamic study

The effect of temperature on the adsorption process was studied from 293 to 313 K, and the q_e (mg g⁻¹) values were calculated at different temperatures. Fig. 9A shows that the q_e values for both tetracycline and cefotaxime raised with increasing temperature from 293 to 313 K. This indicates that the adsorption process has an endothermic nature.³²

Fig. 9 (A) Temperature effect on tetracycline and cefotaxime adsorption process, and (B) thermodynamic linearity by plotting ln (K_C) versus 1/T (K).

In addition, thermodynamic model was investigated to explain the adsorption mechanism. Besides, van't Hoff's eqn (10) was used for calculation of thermodynamic parameters namely enthalpy changes (ΔH) and entropy changes (ΔS).²⁰ Gibb's free energy (ΔG) was estimated using eqn (11).

−ΔG = −RT ln K_C (11)

where K_C (=q_e/C_e) is the thermodynamic equilibrium constant,³⁴ T (K) and R (0.0083145 kJ mol⁻¹ K⁻¹) are temperature and the universal gas constant, respectively (Fig. 9B). The values of ΔH, ΔS, and ΔG were listed in Table 4. Negative values of ΔG for the selected antibiotics confirmed the spontaneous nature of adsorption process. However, the obtained ΔG values between 0 and −20 kJ mol⁻¹, indicate that the antibiotics adsorption process follows a physisorption mechanism. Typically, the ΔG values between −80 and −400 kJ mol⁻¹ correspond to a chemisorption process.⁴⁹ The positive values of ΔH show that the selected antibiotics adsorption process is endothermic. The positive ΔS values correspond to the random adsorption of antibiotics by GO/MNPs–SrTiO₃.

Table 4 Thermodynamic parameters for the adsorption of tetracycline and cefotaxime onto GO/MNPs–SrTiO₃

Analyte	Temperature (K)	qe (mg g^−1)	ΔG (kJ mol^−1)	ΔH (kJ mol^−1)	ΔS (J mol^−1 K^−1)
Tetracycline	293	32.86	−1.58	33.67	0.012
	298	35.41	−2.19
	308	39.88	−3.51
	313	40.96	−3.93
Cefotaxime	293	20.55	−0.17	24.41	0.082
	298	21.74	−0.55
	308	24.25	−1.47
	313	25.08	−1.81

---

3.6 Adsorption ability of the synthesized nanocomposites

The adsorption capability of SrTiO₃, MNPs/GO and MNPs/GO–SrTiO₃ toward the selected antibiotics were also studied. The q_e values were calculated for these three nanocomposites under the same conditions (Section 3.2). For tetracycline, the q_e values were 69.65 mg g⁻¹, 52.17 mg g⁻¹ and 71.61 mg g⁻¹ using SrTiO₃, MNPs/GO and MNPs/GO–SrTiO₃, respectively. However, in the case of cefotaxime, the q_e values were found to be 17.63 mg g⁻¹, 6.70 mg g⁻¹ and 25.08 mg g⁻¹. The results showed that q_e for the SrTiO₃ nanoparticles were higher than that of MNPs/GO. However, the q_e values for MNPs/GO–SrTiO₃ were further improved. Consequently, the as prepared MNPs/GO–SrTiO₃ nanocomposite can be used as an efficient adsorbent for adsorption of antibiotics from aqueous media.

3.7 Adsorbent stability

The stability of the MNPs/GO–SrTiO₃ nanocomposite was tested using a leaching test for strontium. Aliquot of adsorbent (50 mg) in 20 mL aqueous solution was shaken for 24 h at different pH (2, 4, 5, 7 and 10). The supernatants were collected, then strontium concentration were analyzed using ICP-OES technique. Fig. 10A shows the percent leaching (desorbed) of strontium from adsorbent at pH 5 and 7 is much lower than pH 1 and 10. This indicates that the as prepared adsorbent is more stable at pH 5 to 7 and thus it is applicable for removal methods.

Fig. 10 (A) Leaching test for strontium desorption from adsorbent, and (B) proposed mechanism for the adsorption of salty tetracycline and cefotaxime using as prepared GO/MNPs–SrTiO₃.

3.8 Mechanism of antibiotics adsorption onto GO/MNPs–SrTiO₃

The as prepared GO/MNPs–SrTiO₃ showed a strong adsorption affinity toward tetracycline with a high adsorption capacity (65.78 mg g⁻¹) as compared with that of cefotaxime (18.21 mg g⁻¹). This can be explained by the presence of possible interactions between the selected antibiotics and the GO-based adsorbent. Fig. 10B shows different possible interactions for tetracycline and cefotaxime adsorption process. The π–π interaction (staking) between GO and tetracycline aromatic rings; H-bonding between free hydrogen (–H) atoms and electronegative atoms namely oxygen and nitrogen; n–π interactions between n–donor electrons atoms (of oxygen and nitrogen), and π-stacking of GO.¹⁹ Electrostatic and Lewis acid–base interactions are also possible between terminal –NH₂ group and COOH (–COO–H₃N–R)¹⁹ as well as interaction of Sr³⁺ with the atoms having lone pair electrons. However, tetracycline provided a strong π–π interaction as compared to cefotaxime due to the presence of aromatic ring in its structure. Strongly adsorption of tetracycline on the graphene oxide surface via π–π interaction has been confirmed previously.⁴³ Thus, the adsorption affinity via π–π interaction is the key parameter for higher adsorption capacity of tetracycline.

3.9 Regeneration and recovery of the adsorbent

Regeneration is the main parameter in adsorption studies that directly affects the energy and cost of process. Reusability of the newly synthesized GO/MNPs–SrTiO₃ was evaluated based on antibiotics adsorption–desorption processes. Aliquots of adsorbent were used for adsorption of 10 mg L⁻¹ of tetracycline and cefotaxime at pH 5 and 4, respectively. Then, the target compounds were desorbed using methanol/ammonia (10 : 0.5 v/v) for 10 min shaking time. However, after repeating the adsorption–desorption of antibiotics cycle for 5 times, adsorptions (%) for tetracycline and cefotaxime were 93% and 89%, respectively. Therefore, the adsorbent can be reused at least 5 times without a significant loss of the adsorption capacity and desorbed amount of the antibiotics.

3.10 Comparison with other published studies

Table 5 shows the comparison between different adsorbents that have been used for the adsorption of antibiotics in terms of adsorption time and maximum adsorption capacity (q_m). The GO/MNPs–SrTiO₃ nanocomposite provided higher q_m values for tetracycline compared to some adsorbents i.e., vine wood, pumice stone, nano MCN@MIPs, CdS–MWCNT, and GO–MNPs. In addition, all three carbon based adsorbents (GO/MNPs–SrTiO₃, CdS–MWCNT⁵⁵ and GO–MNPs⁵⁶) provided good adsorption capacities towards antibiotics. This is due to the governing π–π interaction between antibiotics and CNT/graphene large π-staking system. However, higher q_m obtained for GO/MNPs–SrTiO₃ probably is due to the presence of SrTiO₃ nanoparticles on adsorbent that increase interactions via Sr³⁺. The high q_m (310 mg g⁻¹) for graphene–MoS₂ was attributed the strong π–π interactions and high surface area.⁵⁷ High surface area of NaOH-activated carbon is the main factor to give high q_m (110 mg g⁻¹) for tetracycline. Meanwhile, both graphene–MoS₂ and NaOH-activated carbon having long equilibrium times of 24 h and 1 h, respectively, as compared with GO/MNPs–SrTiO₃. The CdS–MWCNT and polymer@ion exchange resins provided higher q_m for cefotaxime, probably due to the presence of S atom and ion exchange that are the key factors for adsorption process. However, the newly synthesized adsorbent is comparable with other adsorbents toward antibiotics adsorption. Also, the GO/MNPs–SrTiO₃ adsorbent provided fast adsorption/removal process (30 min) as compared with graphene–MoS₂ (24 h), polymer@ion exchange resin (15 h) and vine wood (8 h).

Table 5 Comparison of equilibrium time and maximum adsorption capacity (q_m) of various materials as antibiotic adsorbent

Material	Target compound	Equilibrium time	qm (mg g^−1)	Ref.
GO/MNPs–SrTiO3	Tetracycline	30 min	65.78	Current study
	Cefotaxime	30 min	18.10
GO–MNPs	Tetracycline	10 min	39.81	56
Graphene–MoS2	Doxycycline	24 h	310	57
Vine wood	Tetracycline	8 h	1.98	58
	Amoxicillin	8 h	2.69
Pumice stone	Tetracycline	90 min	20	59
NanoMCN@MIPs	Norfloxacin	60 min	40.98	60
NaOH-activated carbon	Ciprofloxacin	60 min	108.74	61
CdS–MWCNT	Cefotaxime	20 min	37.74	55
Polymer@ion exchange resin	Cefotaxime	15 h	962.07	11

---

4. Conclusion

The new type of GO/MNPs–SrTiO₃ nanocomposite was synthesized and used for the adsorption of tetracycline and cefotaxime antibiotics from water sample. The magnetic graphene oxide was successfully doped with SrTiO₃ nanoparticles to improve the adsorption capacity and prevent aggregation of the GO nanosheets. The selected antibiotics were loaded on the proposed nanocomposite at pH 2–10. Then, the highest q_e values were obtained at pH 5 and 4 for tetracycline and cefotaxime, respectively. The antibiotics adsorption process was investigated using Langmuir isotherm, Freundlich isotherm and Dubinin-Radushkevich model. Thus, the Langmuir isotherm provided monolayer adsorption process followed by high adsorption capacities of 65.78 mg g⁻¹ and 18.21 mg g⁻¹ for newly synthesized adsorbent towards tetracycline and cefotaxime, respectively. In order to describe the proposed adsorption mechanism, the free energy, pseudo-first-order rate, pseudo-second-order rate, intra-particle diffusion and thermodynamic models were investigated. These suggest that the adsorption of tetracycline and cefotaxime onto GO/MNPs–SrTiO₃ is an endothermic nature followed by a physisorption. Moreover, adsorption process was not effected by intra-particle diffusion, and the GO/MNPs–SrTiO₃ exhibited a promising alternative adsorbent for the removal of various antibiotics from aqueous samples using magnetic separation.

Acknowledgements

The authors would like to thank the University of Tehran and Iran's National Elites Foundation for the financial support through the Research Grants.

References

1. N. Dorival-García, A. Zafra-Gómez, S. Cantarero, A. Navalón and J. L. Vílchez, Microchem. J., 2013, 106, 323–333.
2. S. Shi, Y. Fan and Y. Huang, Ind. Eng. Chem. Res., 2013, 52, 2604–2612.
3. A. V. Herrera-Herrera, J. Hernández-Borges, M. M. Afonso, J. A. Palenzuela and M. Á. Rodríguez-Delgado, Talanta, 2013, 116, 695–703.
4. J. He, P. Ma, A. Xie, L. Gao, Z. Zhou, Y. Yan, J. Dai and C. Li, J. Taiwan Inst. Chem. Eng., 2016, 63, 295–302.
5. J. L. M. Vidal, M. del M. Aguilera-Luiz, R. Romero-Gonzalez and A. G. Frenich, J. Agric. Food Chem., 2009, 57, 1760–1767.
6. R. Acosta, V. Fierro, A. M. de Yuso, D. Nabarlatz and A. Celzard, Chemosphere, 2016, 149, 168–176.
7. L. Yue, S. Wang, G. Shan, W. Wu, L. Qiang and L. Zhu, Appl. Catal., B, 2015, 176, 11–19.
8. K. Gritsenko, I. Muse, and A. Vydyanathan, in Substance Abuse, Springer, 2015, pp. 165–178.
9. G. Yang, F. Zhao and B. Zeng, Biosens. Bioelectron., 2014, 53, 447–452.
10. S. Shahrokhian and S. Rastgar, Analyst, 2012, 137, 2706–2715.
11. S. Vasiliu, I. Bunia, S. Racovita and V. Neagu, Carbohydr. Polym., 2011, 85, 376–387.
12. S. Blasioli, A. Martucci, G. Paul, L. Gigli, M. Cossi, C. T. Johnston, L. Marchese and I. Braschi, J. Colloid Interface Sci., 2014, 419, 148–159.
13. C.-W. Yang, W.-C. Hsiao and B.-V. Chang, Chemosphere, 2016, 150, 559–565.
14. P. Singla, N. Goel and S. Singhal, Chem. Eng. J., 2016, 299, 403–414.
15. Z. Liang, Z. Zhaob, T. Sun, W. Shi and F. Cui, J. Hazard. Mater., 2016, 305, 8–14.
16. M. Ghaedi, S. Hajjati, Z. Mahmudi, I. Tyagi, S. Agarwal, A. Maity and V. K. Gupta, Chem. Eng. J., 2015, 268, 28–37.
17. J. Fu, Z. Chen, M. Wang, S. Liu, J. Zhang, J. Zhang, R. Han and Q. Xu, Chem. Eng. J., 2015, 259, 53–61.
18. F. Yu, J. Ma and D. Bi, Environ. Sci. Pollut. Res., 2015, 22, 4715–4724.
19. H. Zhao, X. Liu, Z. Cao, Y. Zhan, X. Shi, Y. Yang, J. Zhou and J. Xu, J. Hazard. Mater., 2016, 310, 235–245.
20. H. Li, D. Zhang, X. Han and B. Xing, Chemosphere, 2014, 95, 150–155.
21. A. C. Martins, O. Pezoti, A. L. Cazetta, K. C. Bedin, D. A. S. Yamazaki, G. F. G. Bandoch, T. Asefa, J. V. Visentainer and V. C. Almeida, Chem. Eng. J., 2015, 260, 291–299.
22. Y. Zhuang, F. Yu, J. Ma and J. Chen, RSC Adv., 2015, 5, 27964–27969.
23. D. Liu, N. Song, W. Feng and Q. Jia, RSC Adv., 2016, 6, 11742–11748.
24. H. Chen, B. Gao and H. Li, J. Hazard. Mater., 2015, 282, 201–207.
25. Y. Tang, X. Liu, C. Ma, M. Zhou, P. Huo, L. Yu, J. Pan, W. Shi and Y. Yan, New J. Chem., 2015, 39, 5150–5160.
26. N. V. Medhekar, A. Ramasubramaniam, R. S. Ruoff and V. B. Shenoy, ACS Nano, 2010, 4, 2300–2306.
27. S. Shahabuddin, N. Muhamad Sarih, S. Mohamad and J. Joon Ching, Polymers, 2016, 8, 27.
28. S.-T. Huang, W. W. Lee, J.-L. Chang, W.-S. Huang, S.-Y. Chou and C.-C. Chen, J. Taiwan Inst. Chem. Eng., 2014, 45, 1927–1936.
29. P. Shen, J. C. Lofaro, W. R. Woerner, M. G. White, D. Su and A. Orlov, Chem. Eng. J., 2013, 223, 200–208.
30. Y. Lin, C. Norman, D. Srivastava, F. Azough, L. Wang, M. Robbins, K. Simpson, R. Freer and I. A. Kinloch, ACS Appl. Mater. Interfaces, 2015, 7, 15898–15908.
31. H. R. Nodeh, W. A. W. Ibrahim, M. M. Sanagi and H. Y. Aboul-Enein, RSC Adv., 2016, 6, 24853–24864.
32. M. C. Ncibi and M. Sillanpää, J. Hazard. Mater., 2015, 298, 102–110.
33. R. Ashiri and A. Moghtada, Metall. Mater. Trans. B, 2014, 45, 1979–1986.
34. H. Rashidi Nodeh, W. A. Wan Ibrahim, I. Ali and M. M. Sanagi, Environ. Sci. Pollut. Res., 2016, 23, 9759–9773.
35. R. Ghosh, L. Pradhan, Y. P. Devi, S. S. Meena, R. Tewari, A. Kumar, S. Sharma, N. S. Gajbhiye, R. K. Vatsa and B. N. Pandey, J. Mater. Chem., 2011, 21, 13388–13398.
36. B. Kaboudin, F. Kazemi and F. Habibi, J. Iran. Chem. Soc., 2015, 12, 469–475.
37. Y. Bian, Z.-Y. Bian, J.-X. Zhang, A.-Z. Ding, S.-L. Liu and H. Wang, Appl. Surf. Sci., 2015, 329, 269–275.
38. X. Sun, B. Guo, L. He, P. Xia and S. Wang, AIChE J., 2016, 62, 2154–2162.
39. S. S. N. Ling, K. H. Yuen and S. A. Barker, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2003, 783, 297–301.
40. B. Kakavandi, R. R. Kalantary, A. J. Jafari, S. Nasseri, A. Ameri, A. Esrafili and A. Azari, Clean: Soil, Air, Water, 2015, 43, 1157–1166.
41. H. Aydın, Y. Bulut and Ç. Yerlikaya, J. Environ. Manage., 2008, 87, 37–45.
42. E. Mekonnen, M. Yitbarek and T. R. Soreta, S. Afr. J. Chem., 2015, 68, 45–52.
43. Y. Gao, Y. Li, L. Zhang, H. Huang, J. Hu, S. M. Shah and X. Su, J. Colloid Interface Sci., 2012, 368, 540–546.
44. C. Gu, K. G. Karthikeyan, S. D. Sibley and J. A. Pedersen, Chemosphere, 2007, 66, 1494–1501.
45. P.-H. Chang, Z. Li, W.-T. Jiang, C.-Y. Kuo and J.-S. Jean, Desalin. Water Treat., 2015, 55, 1380–1392.
46. P. Wang, M. Du, H. Zhu, S. Bao, T. Yang and M. Zou, J. Hazard. Mater., 2015, 286, 533–544.
47. L. Jiang, Y. Liu, G. Zeng, F. Xiao, X. Hu, X. Hu, H. Wang, T. Li, L. Zhou and X. Tan, Chem. Eng. J., 2016, 284, 93–102.
48. M. A. Kamboh, W. A. W. Ibrahim, H. R. Nodeh, M. M. Sanagi and S. T. H. Sherazi, New J. Chem., 2016, 40, 3130–3138.
49. Y. Li, Q. Du, T. Liu, J. Sun, Y. Wang, S. Wu, Z. Wang, Y. Xia and L. Xia, Carbohydr. Polym., 2013, 95, 501–507.
50. K. S. W. Sing, Pure Appl. Chem., 1985, 57, 603–619.
51. H. K. Boparai, M. Joseph and D. M. O'Carroll, J. Hazard. Mater., 2011, 186, 458–465.
52. S. Luo, X. Xu, G. Zhou, C. Liu, Y. Tang and Y. Liu, J. Hazard. Mater., 2014, 274, 145–155.
53. W. Wang, M. Li and Q. Zeng, Trans. Nonferrous Met. Soc. China, 2012, 22, 2831–2839.
54. B. Yu, J. Xu, J.-H. Liu, S.-T. Yang, J. Luo, Q. Zhou, J. Wan, R. Liao, H. Wang and Y. Liu, J. Environ. Chem. Eng., 2013, 1, 1044–1050.
55. A. Fakhri, S. Rashidi, M. Asif, I. Tyagi, S. Agarwal and V. K. Gupta, J. Mol. Liq., 2016, 215, 269–275.
56. Y. Lin, S. Xu and J. Li, Chem. Eng. J., 2013, 225, 679–685.
57. Y. Chao, W. Zhu, X. Wu, F. Hou, S. Xun, P. Wu, H. Ji, H. Xu and H. Li, Chem. Eng. J., 2014, 243, 60–67.
58. H. R. Pouretedal and N. Sadegh, J. Water Process Eng., 2014, 1, 64–73.
59. U. A. Guler and M. Sarioglu, J. Environ. Health Sci. Eng., 2014, 12, 1.
60. F. Tan, D. Sun, J. Gao, Q. Zhao, X. Wang, F. Teng, X. Quan and J. Chen, J. Hazard. Mater., 2013, 244, 750–757.
61. M. J. Ahmed and S. K. Theydan, J. Taiwan Inst. Chem. Eng., 2014, 45, 219–226.

---