Poly(4-styrenesulfonic acid-co-maleic acid)-sodium-modified magnetic reduced graphene oxide for enhanced adsorption performance toward cationic dyes

Abstract

By combining the advantages of poly(4-styrenesulfonic acid-co-maleic acid) sodium (PSSMA) with abundant anionic functional groups (–COO⁻ and –SO₃⁻), graphene oxide (GO) with high specific surface area and Fe₃O₄ nanoparticles with excellent magnetic responsiveness, a novel type of PSSMA-modified magnetic reduced graphene oxide nanocomposite (PSSMA/M-rGO) was synthesized via a simple and facile one-step solvothermal method and used for removing cationic dyes from aqueous solutions in this study. The as-synthesized PSSMA/M-rGO was characterized by Fourier transform infrared spectroscopy, UV-vis spectroscopy, scanning electron microscopy, transmission electron microscopy, thermogravimetric analysis, X-ray diffraction, vibrating sample magnetometry, dynamic light scattering and nitrogen adsorption-desorption technique. Three typical cationic dyes, basic fuchsin (BF), crystal violet (CV) and methylene blue (MB) were used as model dye pollutants to evaluate the adsorption performance of the resultant PSSMA/M-rGO. The adsorption of three cationic dyes onto both PSSMA/M-rGO and M-rGO without PSSMA modification on the surface were systematically investigated at different experiment conditions. The results indicate that the binding of PSSMA on M-rGO can significantly enhance the adsorption capacities and removal efficiencies of the three dyes. This is due to the rich –COO⁻ and –SO₃⁻ groups on PSSMA/M-rGO having strong electrostatic interactions with the positively charged dye molecules. The adsorption kinetics and isotherms of the three dyes onto both adsorbents demonstrate that the kinetics and equilibrium adsorptions can be well-described by pseudo-second-order kinetics and Langmuir model, respectively. Moreover, the PSSMA/M-rGO nanocomposites also demonstrate high removal efficiencies toward mixed dyes of BF, CV and MB. Such functional nanocomposites with high adsorption capacity, low production cost and excellent recyclability, are promising as candidate adsorbents for highly-efficient removal of cationic organic pollutants from aqueous solutions.

---

1. Introduction

Organic dyes have emerged as harmful pollutants in the environment because of their widespread applications in industrial manufacture of products such as textiles, leather, paper, plastics, printing and cosmetics.¹ Vast amounts of dyes together with industrial effluents are released into the aquatic environment, causing severe environmental pollution and harmful hazards to humans.² The presence of dyes in water reduces the water transparency, and weakens the light penetration and oxygen gas solubility in water, thus decreasing the photosynthetic efficiency of aquatic plants.³ Moreover, most dyes are highly toxic and can pose teratogenic, carcinogenic and mutagenic effects to aquatic life and humans even at a very low concentrations.⁴ Therefore, it is of vital importance to remove dye contaminants prior to their discharge into water. To date, various methods including biological oxidation,⁵ photocatalytic degradation,⁶ ion exchange,⁷ membrane process⁸ and adsorption^3,9 have been widely used to dispose dye wastewaters. Among those techniques, adsorption is considered as the most efficient and versatile approach to treat dyestuff wastewaters due to its simple design, low cost, wide adaptability and easy operation.^3,9 A variety of materials, such as natural zeolites,¹⁰ modified mesoporous clays,¹¹ biomass,⁹ industrial/agricultural wastes,^12,13 mesoporous silica or titania,^14,15 polymer microspheres,¹⁶ and carbon-based materials (activated carbon,³ carbon nanotubes,¹⁷ graphene or graphene oxide (GO)^18–22), have been used as adsorbents to remove dye pollutants from contaminated water. Among those materials, GO especially has attracted considerable attention for wastewater treatment due to its distinctive one-atom-thick two-dimensional structure with large surface area, high stability as well as abundant oxygen-containing functional groups on its surface, such as carboxyl, hydroxyl and epoxy groups.^18–20,22 Those unique properties of GO endow its high adsorption capacity toward cationic dyes with benzene ring structure via electrostatic, hydrogen-bonding or/and π–π stacking interactions.^18–20 However, the separation of GO suspensions from treated water, especially from a large volume of water, is time-consuming and expensive due to the involvement of complex filtration or/and centrifugation process,^2,18–20,22 and thus limiting the practical application of GO in wastewater treatment.

Through introducing magnetic nanoparticles (MNPs), typically Fe₃O₄ NPs, into GO sheets to impart its convenient magnetic separability is a feasible approach to solve the problem for separation.^23–26 The MNPs loaded on GO can also serve as stabilizers or separators to prevent the possible aggregation of graphene sheets.²³ Solvothermal and chemical co-precipitation methods are common techniques for fabricating magnetic graphene nanocomposites.^23–26 However, the introduction of MNPs into GO usually leads to decrease in adsorption capacities of the magnetic graphene nanocomposites,²⁷ due to that the active sites (–COOH and –OH) on GO sheets are preferentially occupied by MNPs or can be partially even completely reduced.^23–25 One possible solution to tackle this dilemma is to modify magnetic graphene nanocomposites with organic small molecules or polymers containing rich functional groups like –COOH, –OH and –NH₂, including citric acid (CA),²⁸ xanthate,²⁹ poly(acrylic acid) (PAA),²⁷ chitosan (CS),^4,30 cyclodextrin (CD)^26,31 and CS&CD copolymers.³² The functionalized magnetic graphene nanocomposites exhibit enhanced adsorption capacity toward various organic or inorganic contaminants through electrostatic and/or chelating interactions.^4,28–32 Poly(4-styrenesulfonic acid-co-maleic acid) sodium salt (PSSMA), an anionic polyelectrolyte, contains abundant carboxyl (–COO⁻) and sulfonic (–SO₃⁻) groups in the molecule (Fig. S1A†). The distinctive properties of PSSMA provide strong electrostatic affinity to various cationic pollutants, such as positively charged organic dyes and heavy metal ions. In practical application, if PSSMA can be effectively bound on MNPs-loaded graphene, the obtained nanocomposites will not only enhance the adsorption capacity toward cationic dyes, but also provide excellent magnetic separability, thus the problem of time-consuming and high-cost of GO when used as adsorbents to treat dyestuff wastewaters can be effectively solved. Our recent work also indicates that the introduction of PSSMA on the surface of Fe₃O₄ NPs can significantly enhance the adsorption performance of the functionalized MNPs toward a cationic dye, methylene blue (MB).³³ However, due to a relatively smaller specific surface area of Fe₃O₄ NPs for the grafting of PSSMA functional molecules (lower than 50 m² g⁻¹), the dye uptake of PSSMA-modified MNPs is still limited. The maximum equilibrium adsorption capacity of MB is only 52.2 mg g⁻¹ at an initial concentration of 250 mg L⁻¹.

In this study, Fe₃O₄ NPs-loaded GO nanosheets, which combines the large specific surface area of GO and the excellent magnetic responsiveness of Fe₃O₄ NPs, were used for the binding of PSSMA via simple and facile one-pot solvothermal method. The PSSMA-modified magnetic rGO nanocomposites (PSSMA/M-rGO) were used to remove three typical cationic dyes including basic fuchsin (BF), crystal violet (CV) and MB from model dye wastewaters, (Fig. S1B–D†). The effects of solution pH, contact time, and initial dye concentrations on the adsorption of three dyes were systematically investigated. Moreover, adsorption mechanisms of the three dyes were also discussed in detail. The results indicate that the PSSMA/M-rGO demonstrate higher adsorption capacity and improved removal efficiency as compared to that of M-rGO without PSSMA modification. The adsorption kinetics and isotherms of three dyes onto both PSSMA/M-rGO and M-rGO show that the kinetics and equilibrium adsorptions can be well-described by the pseudo-second-order kinetic and Langmuir model, respectively. Furthermore, the PSSMA/M-rGO also exhibit high removal efficiency toward mixed dyes of BF, CV and MB. The dye-loaded PSSMA/M-rGO can be easily recovered under an external magnetic field and regenerated using 2.0 wt% NaOH ethanol solutions. Such functional nanocomposites with high adsorption capacity, low production cost and excellent recyclability, show great potential as candidate adsorbents for highly-efficient removal of cationic dye pollutants from aqueous solutions.

2. Materials and experiment methods

2.1. Chemicals and materials

Graphite powders (325 mesh) were purchased from Qingdao Huatai Lubricating and Sealing Technology Co., Ltd. (Qingdao, China). Poly(4-styrenesulfonic acid-co-maleic acid, SS : MA = 3 : 1) sodium salt (PSSMA, M_w = 20 000) was obtained from Sigma-Aldrich. Ferric chloride hexahydrate (FeCl₃·6H₂O), potassium permanganate (KMnO₄), sodium acetate trihydrate (CH₃COONa·3H₂O, NaAc), basic fuchsin (BF), crystal violet (CV) and methylene blue (MB) were bought from Chengdu Kelong Chemicals (Chengdu, China). All chemicals were of analytical reagent grade. Deionized water was used throughout this study.

2.2. Preparation of GO

GO were prepared from natural graphite power based on the modified Hummers method.³⁴ Briefly, graphite power (1.0 g) and NaNO₃ (1.0 g) were mixed in 50 mL of H₂SO₄ (98 wt%) in a beaker at ice temperature (0–5 °C). Then KMnO₄ (6.0 g) was added in the mixture slowly under stirring over 1 h. Next, the temperature of mixture was kept at 40 °C for 2 h, and 100 mL of deionized water was added slowly. Then the reaction temperature was maintained at 90 °C for 0.5 h. After that, 200 mL of deionized water and 10 mL of hydrogen peroxide (H₂O₂) (30%) were added, and the color of mixture changed from dark brown to brilliant yellow. The obtained dispersion was washed using HCl (37%) to remove residual metal ions, and followed by deionized water until the pH of filtrate was neutral. After drying under vacuum at room temperature, GO solid was obtained.

2.3. Preparation of PSSMA/M-rGO

PSSMA/M-rGO were prepared via simple and facile one-pot solvothermal method with slight modification.²³ Typically, 0.1 g of GO was dispersed in 50 mL of ethylene glycol (EG) by ultrasonication for 2 h. Then 0.25 g of FeCl₃·6H₂O was added and the mixture was vigorously stirred for 2 h. After that, 0.6 g of PSSMA and 0.9 g of NaAc were added under vigorous stirring for 0.5 h. The brownish yellow mixture was transferred to a Teflon-lined stainless steel autoclave and heated at 200 °C for 10 h, and then cooled to room temperature. The obtained black product was washed using deionized water for 5 times, and dried under vacuum. The PSSMA/M-rGO serials synthesized at dosage of 0, 0.1, 0.25, 0.6 and 1.0 g were denoted as M-rGO, M-rGO-010, M-rGO-025, M-rGO-060 and M-rGO-100, separately. For comparison, 0.6 g of PSSMA-modified MNPs and rGO were also prepared using the same method as the M-rGO-060 except that no GO or FeCl₃·6H₂O was added during the solvothermal process, and were denoted as the Fe₃O₄-060 and rGO-060, respectively.

2.4. Characterization

The FT-IR spectra were recorded on an IR 200 spectrometer (Thermo Nicolet, USA). Field-emission scanning electron microscope (SEM) images were obtained with a JSM-7600F microscope (JEOL, Japan) at an accelerating voltage of 15 kV. The samples were prepared by mounting a drop of dispersion onto a glass sheet. Transmission electron microscope (TEM) images were taken on a JEM-2010 microscope (JEOL, Japan) at an accelerating voltage of 120 kV. The samples were prepared by mounting a drop of dispersion onto a carbon-coated copper grid. Powder X-ray diffraction (XRD) patterns were obtained on a DX-1000 diffractometer (Dandong Fangyuan Instrument Co., Ltd, China) with Cu Kα radiation operating at 40 kV and 36 mA. The thermogravimetric analysis (TGA) was performed on a Mettler TGA/SDTA851e (Switzerland) under a constant nitrogen flow at heating rate of 5 °C min⁻¹. Magnetic property was conducted by vibrating sample magnetometer (VSM) on a Model 6000 physical property measurement system (Quantum Design, USA) at room temperature. The Brunauer–Emmett–Teller (BET) specific surface area was determined by nitrogen adsorption–desorption isotherms (Quantachrome Instruments, USA). The zeta potentials of sample dispersions were measured on a dynamic light scattering (DLS) instrument (Zetasizer Nano-ZS, Malvern Instruments, UK).

2.5. Batch adsorption experiments

In this study, all the dye adsorption experiments were carried out using a serial of glass vials (30 mL) equipped with Teflon screw caps. Batch adsorption experiments were conducted to investigate the effects of solution pH, contact time, and initial dye concentrations on the adsorption of BF, CV and MB. The effect of pH on the dye adsorption was studied by adjusting solution pH in the range from 2.0 to 9.0. The pH values of dye solutions were regulated by adding negligible volume of 0.1 M HCl or NaOH. To investigate the effect of contact time on the dye adsorption and the adsorption kinetics, the initial concentrations of BF, CV and MB were fixed at 200, 150 and 125 mg L⁻¹, respectively, and the supernatants after magnetic separation were withdrawn at certain time intervals during the adsorption. To explore the effect of initial dye concentrations on the dye adsorption and adsorption isotherms, some preliminary experiments were performed in advance to determine the suitable initial dye concentrations in this study. It was found that the removal efficiencies of three dyes were low (less than 50%) as the initial concentrations of BF, CV and MB were higher than 600, 450 and 250 mg L⁻¹. Therefore, to obtain relative high dye removal efficiency, the dye concentrations were determined from 50 to 600 mg L⁻¹ for BF, 50 to 450 mg L⁻¹ for CV and 50 to 250 mg L⁻¹ for MB, respectively. After adsorption, the dye-loaded adsorbents were magnetically separated from dye solutions, and the dye concentrations were determined on a UV spectrophotometer (TU-1950, Persee, Beijing, China) at excitation wavelengths of 543, 585 and 664 nm for BF, CV and MB, respectively. The equilibrium adsorption capacity q_e (mg g⁻¹) and removal efficiency(%) of three dyes were calculated from eqn (1) and (2).where C₀ and C_e are the initial and equilibrium dye concentrations (mg L⁻¹) in solutions, respectively; q_e is the amount of dyes adsorbed on the adsorbents (mg g⁻¹) after equilibrium, V is the solution volume (L), m is the mass of adsorbents (g). All the adsorption experiments were performed at least three times for data analysis.

2.6. Desorption and regeneration

For the desorption and regeneration study, M-rGO-060 (10 mg) were added in 20 mL dye solutions with different initial concentrations (200, 150 and 125 mg L⁻¹ for BF, CV and MB, respectively), and oscillated in a temperature-controlled shaker at 25 °C for 5 min. After magnetic separation, the dye-adsorbed M-rGO-060 were rinsed five times using NaOH ethanol solution (0.2 M), and followed by deionized water and ethanol to remove excess NaOH. The regenerated M-rGO-060 were then reused for further adsorption experiments.

3. Results and discussion

3.1. Materials synthesis and characterization

Fig. 1 shows the schematic illustration for the synthesis of PSSMA/M-rGO. GO, FeCl₃·6H₂O and NaAc were first added in EG, and the Fe³⁺ would attach to the negatively charged surface of GO through electrostatic interactions and served as nucleation sites.^23,24,35 After an addition of PSSMA in the above mixture, the obtained homogeneous dispersion was suffered from solvothermal reaction at 200 °C for 10 h. In this case, EG was used as both reducing reagent and solvent. NaAc alters the alkalinity and assists the reduction of GO.²⁴ Also, in the alkaline condition, the Fe³⁺ in reaction solution was slowly hydrated to form Fe(OH)₃, and then partially reduced by EG to form Fe(OH)₂, and finally formed Fe₃O₄.³⁶ As the reaction proceeds, these Fe₃O₄ nuclei grew into spherical particles on the partially-reduced rGO.^23,24,35 Meanwhile, PSSMA functional molecules were bound on M-rGO simultaneously through a condensation reaction between the –COOH groups of PSSMA and the –OH groups of M-rGO.

Fig. 1 Schematic illustration for the synthesis of PSSMA/M-rGO.

The chemical structures of the functionalized graphene were characterized by FT-IR. Fig. 2 shows the FT-IR spectra of GO, M-rGO, M-rGO-060 and PSSMA. For GO (Fig. 2a), the peaks at 1745, 1632, 1403 and 1057 cm⁻¹ are respectively attributed to the C=O stretching vibrations of –COOH groups, the C=C stretching vibrations of aromatic skeletal, the C–OH deformation vibrations of phenolic groups, and the C–O–C stretching vibrations of epoxy groups.^18,25–27 When Fe₃O₄ NPs were loaded on graphene to obtain M-rGO, the peaks of the Fe–O stretching vibrations at 580 cm⁻¹ and the C=C stretching vibrations of aromatic skeletal at 1645 cm⁻¹ are observed. The peaks of the C=O stretching vibrations of –COOH groups shifted to 1728 cm⁻¹; the C–OH deformation vibrations of phenolic groups at 1404 cm⁻¹ and the C–O–C stretching vibrations of epoxy groups at 1055 cm⁻¹ are still observable but gradually weakened, indicating the partial reduction of the –COOH, –OH and C–O–C groups during the solvothermal process. The residual –OH groups on M-rGO may provide reactive sites for the binding of PSSMA functional molecules.²³ After further functionalization of the Fe₃O₄ NPs-loaded GO with PSSMA, the asymmetric stretching vibrations of –SO₃⁻ groups at 1119 and 1193 cm⁻¹, and the symmetric and asymmetric stretching vibrations of the C=O in PSSMA at 1409 and 1578 cm⁻¹ are all observed, suggesting a successful binding of PSSMA functional molecules on GO.^33,36 Besides, the peaks of the C=O stretching vibrations of –COOH groups at 1710 cm⁻¹, and the C–O–C stretching vibrations of epoxy groups at 1055 cm⁻¹ of GO are still observed but not as prominent as that of GO, indicating that GO was partially reduced during the solvothermal process. All above the results show that PSSMA functional molecules and Fe₃O₄ NPs have been successfully introduced to the partially-reduced rGO and resulted in the formation of PSSMA/M-rGO.

Fig. 2 FT-IR spectra of GO (a), M-rGO (b), M-rGO-060 (c) and PSSMA (d).

The reduction and surface functionalization of the Fe₃O₄ NPs-loaded GO were also confirmed by UV-vis in this study. Fig. S2† shows the UV-vis spectra of GO, M-rGO, M-rGO-060 and PSSMA. For GO, the representative characteristic peak at 231 nm corresponds to the π–π* electronic transitions of aromatic C–C bonds, and the shoulder at about 300 nm attributes to the n–π* transitions of C=O bonds.^37–40 After partial reduction of GO, the peak at 231 nm is red-shifted to 264 nm and no absorption peak at 300 nm is observed, indicating that the highly conjugated electronic structure is restored in M-rGO.^37–41 For M-rGO-060, two new adsorption peaks at 261 and 222 nm, representing the n–π* transitions of C=O and S=O bonds and the π–π* electronic transitions of aromatic C–C bonds are also found, giving the direct evidences that PSSMA functional molecules have been successfully attached to M-rGO.

The morphology and microstructure of M-rGO-060 were observed by SEM and TEM as shown in Fig. 3. The M-rGO-060 have crumpled and flake-like structure (Fig. 3A), and the graphene plane is sparely coated with black Fe₃O₄ NPs with mean diameter of about 50 nm (Fig. 3B). The Fe₃O₄ NPs loaded on the M-rGO-060 endow them excellent magnetic separability. Moreover, the thin layer structure of GO provides large specific surface area and abundant reactive sites for binding of PSSMA functional molecules. The rich –COOH groups in PSSMA facilitate the probability of attaching them to the Fe₃O₄ NPs-loaded rGO through a condensation interaction of the –COOH groups in PSSMA and the –OH groups of M-rGO, thus endowing the negatively charged surface of M-rGO-060.

Fig. 3 Typical SEM (A) and TEM (B) images of M-rGO-060.

The thermal behaviors of GO and M-rGO/PSSMA series prepared with different PSSMA dosages were characterized by TGA. Fig. 4A shows the TGA curves of GO, M-rGO, M-rGO-010, M-rGO-025, M-rGO-060 and M-rGO-100. For GO, the weight loss below 120 °C is due to the evaporation of physically adsorbed water,^27,35 and the weight loss from 120 °C to 300 °C is ascribed to the decomposition of oxygen-containing functional groups, such as –COOH, –OH and C–O–C.²⁷ After the deposition of Fe₃O₄ NPs on graphene to obtain M-rGO, a weight loss of about 29% in the temperature range between 25 °C and 600 °C is due to the decomposition of residual –OH, –COOH and C–O–C groups.^27,35 After simultaneous reduction and functionalization using PSSMA for the Fe₃O₄ NPs-loaded GO, the weight loss of M-rGO/PSSMA series exhibits a two-step process. The first weight loss below 120 °C is also attributable to the evaporation of physically adsorbed water on PSSMA/M-rGO,²⁷ and the second weight loss from 120 °C to 600 °C shows the decomposition of PSSMA functional molecules covalently bound on PSSMA/M-rGO.^33,36 The amount of PSSMA bound on M-rGO increase with increasing the dosage of PSSMA from 0.1 to 0.6 g during the solvothermal process, and no evident weight loss is observed for the dosage larger than 0.6 g. The binding amount of PSSMA on M-rGO-010, M-rGO-025, M-rGO-060 and M-rGO-100 are respectively 200.8, 247.4, 263.9 and 264.7 mg g⁻¹ based on the TGA results (Fig. 4A, inset). The abundant PSSMA functional molecules on PSSMA/M-rGO play a crucial role in enhancing its uptakes toward cationic pollutants, such as cationic dyes or heavy metal ions.

Fig. 4 (A) TGA curves of GO (a), M-rGO (b), M-rGO-010 (c), M-rGO-025 (d), M-rGO-060 (e) and M-rGO-100 (e); (B) XRD patterns of M-rGO (a) and M-rGO-060 (b); (C) room temperature magnetic hysteresis curve of M-rGO-060. The inset shows the photographs of M-rGO-060 in the absence (a) and presence (b) of an external magnetic field, and (D) pH-dependent zeta potentials of M-rGO (a) and M-rGO-060 (b). The concentration of dispersions is 0.1 mg mL⁻¹.

XRD was employed to characterize the crystalline structure of Fe₃O₄ NPs on M-rGO and PSSMA/M-rGO. The XRD patterns of M-rGO and M-rGO-060 are shown in Fig. 4B, which are in accordance with the standard XRD spectra of Fe₃O₄ (JCPDS no. 89-4319). This further confirms the successful deposition of Fe₃O₄ NPs on rGO, and the binding of PSSMA functional molecules on M-rGO does not change the crystalline structure of the PSSMA/M-rGO. The magnetic responsiveness of the PSSMA/M-rGO enables them to be easily separated from dye solutions under an external magnetic field. The magnetism property of the M-rGO-060 was studied using VSM at room temperature by cycling the magnetic field between −30 kOe to 30 kOe as shown in Fig. 4C. The hysteresis and coercivity are almost undetectable, suggesting a high superparamagnetism of the M-rGO-060. The superparamagnetic property of the M-rGO-060 is critical for their practical applications, which can prevent them from occurring aggregation and make them redisperse rapidly after removing the external magnetic field (Fig. 4C, inset). The magnetic saturation value of the M-rGO-060 is 27.41 emu g⁻¹, lower than that of Fe₃O₄-060 NPs (about 62.1 emu g⁻¹). The Fe₃O₄-060 was synthesized using the same method as the M-rGO-060. Such high saturation magnetization allows them for effectively magnetic manipulation. Moreover, the M-rGO-060 have BET specific surface area of 426 m² g⁻¹, smaller than that of bare GO (the theoretically calculated value is 2630 m² g⁻¹), which is due to the occupation of the immobilized Fe₃O₄ NPs and PSSMA functional molecules on GO. However, the rich anionic functional groups (–SO₃⁻ and –COO⁻) on the surface and relatively large specific surface area endow them high adsorption capacities toward cationic dyes.

Surface charges of an adsorbent significantly affect its uptakes toward oppositely charged contaminants through electrostatic interactions. The pH-dependent zeta potentials of M-rGO and M-rGO-060 in aqueous solutions (0.1 mg mL⁻¹) over a pH range of 2–10 were measured by DLS as shown in Fig. 4D. The solution pH values were adjusted by adding 0.1 M HCl or NaOH. The zeta potentials of M-rGO-060 and M-rGO decrease with increasing pH from 2.0 to 10.0. All the zeta potentials of M-rGO-060 are negative over the experimental pH range, and lower than those of M-rGO. The binding of PSSMA with rich –COO⁻ and –SO₃⁻ functional groups on the M-rGO-060 facilitate their negatively charged surface over the investigated pH range. For M-rGO, an isoelectric point (PI) of about 4.5 is observed. Meanwhile, the M-rGO shows positively charged surface at pH lower than the PI, and negatively charged surface at pH higher than the PI. Li and co-workers also observed a similar phenomenon during the preparation of rGO nanosheets using hydrazine as the reducing reagent.³⁸ This can be explained that the partially unreduced –COOH and –OH groups on M-rGO were protonized under an acidic condition, giving rise to positively charged surface of M-rGO. While as the pH increases, the protonized –COOH and –OH groups deprotonized gradually, therefore leading to negatively charged surface of M-rGO.

Colloidal stability of an adsorbent is also crucial for its practical application in wastewater treatment. In this study, the binding of hydrophilic PSSMA molecules on M-rGO can significantly improve the colloidal stability of PSSMA/M-rGO (Fig. S3†). As observed in Fig. S3,† the M-rGO-060 still exhibit excellent dispersion stability in water after standing for four days at room temperature (Fig. S3a–f†). Electrostatic repulsion between the –COO⁻ and –SO₃⁻ groups on M-rGO plays a significant role in stabilizing the M-rGO-060 colloids in aqueous solution. The negatively charged surface of M-rGO-060 is also greatly beneficial to their practical applications in cationic dye wastewater treatment. However, on the other hand, due to the absence of PSSMA functional molecules on the surface, evident aggregation and precipitation are observed for the M-rGO after standing for less than 30 s (Fig. S3a′–f′†). Therefore, M-rGO exhibits limited adsorption capacity when used as an adsorbent.

3.2. Adsorption performance test

To confirm the adsorption of cationic dyes onto PSSMA/M-rGO is mainly through electrostatic interactions between the positively charged dye molecules and the negatively charged surface of PSSMA/M-rGO, the M-rGO/PSSMA serials (M-rGO, M-rGO-010, M-rGO-025, M-rGO-060 and M-rGO-100) were respectively used as adsorbents to remove three typical cationic dyes (BF, CV and MB) from aqueous solutions. As shown in Fig. 5, the adsorption capacities and removal efficiencies of three dyes onto PSSMA/M-rGO increase with increasing the PSSMA dosage, and reach a plateau as the dosage larger than 0.6 g. This indicates that the binding of PSSMA functional molecules on M-rGO plays a crucial role in enhancing the adsorption capacities of three dyes. A higher PSSMA dosage contributes to a higher binding amount of PSSMA on M-rGO during the solvothermal process, thus resulting in larger amount of negative charges on PSSMA/M-rGO for adsorption of dye molecules through strong electrostatic interactions, and finally giving rise to higher dye adsorption uptakes.

Fig. 5 The effect of PSSMA dosage during the solvothermal process on the adsorption capacities (A) and removal efficiencies (B) of BF, CV and MB. Condition: C_BF = C_CV = 200 mg L⁻¹, C_MB = 150 mg L⁻¹, m/V = 10 mg/20 mL, T = 25 °C and contact time = 30 min.

Moreover, due to larger specific surface area of M-rGO compared with that of Fe₃O₄ NPs, larger amount of PSSMA can be bound on M-rGO. The binding amount of PSSMA on M-rGO-010 and Fe₃O₄-010 are 200.8 and 99.7 mg g⁻¹, respectively. Therefore, the M-rGO/PSSMA series are expected to possess higher dye uptakes. To verify our hypothesis, Fe₃O₄-060, synthesized using the same method as the M-rGO-060 except that no GO was added during the solvothermal process, was also used to adsorb three cationic dyes. As observed in Fig. S4,† the Fe₃O₄-060 demonstrates very lower adsorption capacities and removal efficiencies toward three dyes than those of the M-rGO-060. The relatively smaller specific surface area of the Fe₃O₄ NPs compared with that of the M-rGO, providing relatively lower binding amount of PSSMA on the surface, may be responsible for this phenomenon. However, due to the larger specific surface area of GO for the binding of PSSMA, 0.6 g PSSMA-modified GO (rGO-060) was also prepared using the same method as the M-rGO-060 expect that no iron source (FeCl₃·6H₂O) was added in this study. The as-prepared rGO-060 shows higher adsorption capacities and removal efficiencies toward three dyes than those of the M-rGO-060 (Fig. S4†) due to larger binding amount of PSSMA. The FT-IR spectrum of the rGO-060 also confirms the successful attachment of PSSMA to GO (Fig. S5†). The binding amount of PSSMA on GO is about 350 mg g⁻¹ as calculated from the TGA results. However, the separation of dye-loaded rGO-060 usually involves high-speed centrifugation or/and complex filtration process, which is usually time-consuming and uneconomic. Therefore, the practical applications of rGO-060 in dye wastewater treatment are limited. The deposition of Fe₃O₄ NPs on GO will inevitably occupy some active sites for the binding of PSSMA functional molecules, thus resulting in decrease of dye uptakes.²³ However, considering the relatively high adsorption capacities and excellent magnetic separability of the M-rGO-060 (TOC and Fig. 4C, insets), M-rGO-060 were used to study the adsorption performance of three cationic dyes in the following investigation.

3.2.1. Effect of initial solution pH. Initial solution pH value is an important factor that affects not only the surface charges of an adsorbent, but also the dissociation degree of functional groups on the adsorbent and the structure of dye molecules.^23,26,42 Since many cationic dyes are unstable under strong acidic or strong alkalic conditions for easy denaturation, aggregation and precipitation, thus their adsorption behavior are difficult to interpret.^23,43 Therefore, the effect of initial solution pH on the adsorption of three cationic dyes onto M-rGO-060 and M-rGO was studied at pH value ranged from 3.0 to 9.0 for BF and CV, and from 2.0 to 10.0 for MB in this study. As shown in Fig. 6, M-rGO-060 show much higher adsorption capacities and removal efficiencies toward BF, CV and MB than those of M-rGO over the whole pH range investigated. The binding of PSSMA with abundant anionic functional groups (–COO⁻ and –SO₃⁻) on M-rGO plays a crucial role in enhancing the cationic dye uptakes.³³ Moreover, the adsorption capacities and removal efficiencies of BF, CV and MB onto M-rGO-060 increase slightly at pH lower than 4.0, and keep nearly invariable at pH higher than 4.0. Changing the pH over the range of 2.0–10.0 can significantly affect the zeta potentials of the M-rGO-060 (Fig. 4D), thus affecting the electrostatic interactions between the M-rGO-060 and cationic dyes,^18,23,27,29 and finally having an impact on the cationic dye adsorption. In strong acidic solutions, both the –COOH and –SO₃H groups in PSSMA and the residual –COOH and –OH groups on M-rGO without reduction are protonized as pH increases, resulting in more positively charged surface of the M-rGO-060.^18,23,27 In this case, active sites available for adsorption of cationic dyes decrease due to smaller amount of negative charges on M-rGO-060, thus leading to slightly low dye uptakes. On the other hand, as dye solution pH increases, the M-rGO-060 are provided with more negatively charged surface due to the gradual deprotonization of the protonized –COOH, –SO₃H and –OH groups.^18,23,27 This enhances the electrostatic interactions significantly between the adsorbent and cationic dye molecules, thus leading to higher dye adsorption capacities and removal efficiencies.

Fig. 6 The effect of initial solution pH on the adsorption capacities (A) and removal efficiencies (B) of BF, CV and MB onto M-rGO-060 (solid) and M-rGO (hollow). Condition: C_BF = 200 mg L⁻¹ C_CV = 150 mg L⁻¹ C_MB = 125 mg L⁻¹, m/V= 10 mg/20 mL, T = 25 °C and contact time = 30 min.

For M-rGO adsorbent, due to the absence of PSSMA functional molecules on the surface, available active sites for dye adsorption are extremely limited and therefore, dye uptakes onto M-rGO are much lower than those of M-rGO-060. π–π Stacking interactions between dye molecules and the aromatic rings of M-rGO may have contributed to the adsorption of three dyes onto M-rGO.^23,27 Additionally, the uptakes of three dyes onto M-rGO-060 follow the order: BF > CV > MB. The different initial dye concentrations and interactions may account for this phenomenon. The initial concentrations of BF, CV and MB are respectively 200, 150 and 125 mg L⁻¹. Higher initial dye concentrations provide higher driving force to overcome the mass transfer resistance of dye molecules between the aqueous phase and the solid phase, resulting in more collision between dye molecules and adsorption sites (COO⁻ and SO₃⁻ groups) on M-rGO-060.^20,23 Moreover, the difference in interactions existed between dye molecules and M-rGO-060 also contributes to the low dye uptakes onto M-rGO. For BF adsorption, except for the dominant strong electrostatic interactions between BF molecules and the –COO⁻ and –SO₃⁻ groups on M-rGO-060, the hydrogen-bonding interactions between the –NH₂ groups of BF and the –COOH and –SO₃H groups on M-rGO-060, and π–π stacking interactions between BF and the aromatic rings of graphene on M-rGO-060 also exist, thus giving rise to higher dye uptakes. For CV adsorption, apart from the electrostatic interactions between CV molecules and the –COO⁻ and –SO₃⁻ groups on M-rGO-060, π–π stacking interactions between CV and the aromatic rings of graphene also cause dye adsorption. However, for MB adsorption, only electrostatic interactions between MB molecules and the –COO⁻ and –SO₃⁻ groups on M-rGO-060 exist. Since the adsorption capacities and removal efficiencies of three cationic dyes onto both M-rGO-060 and M-rGO are almost unchanged at pH larger than 5.0, the solution pH values were fixed at 7.0 in the subsequent adsorption study.

3.2.2. Effect of contact time and adsorption kinetics. Adsorption kinetics is another fundamental and significant aspect for evaluating the adsorption behavior of an adsorbent. The effect of contact time on dye adsorption onto both M-rGO-060 and M-rGO, and their adsorption kinetics were studied at different initial dye concentrations (200, 150 and 125 mg L⁻¹ for BF, CV and MB, respectively) as shown in Fig. 7. Both two adsorbents demonstrate fast adsorption for three cationic dyes and the adsorption equilibrium are achieved within 5 min (Fig. 7A). Besides, M-rGO-060 show higher dye adsorption capacities and removal efficiencies compared with those of M-rGO, and the removal efficiencies of three dyes are nearly 100% (Fig. 7B). Strong electrostatic interactions between the –COO⁻ and –SO₃⁻ groups on M-rGO-060 and positively charged BF, CV and MB molecules give rise to higher dye adsorption capacities and removal efficiencies. Moreover, both the higher initial concentration of BF and more interaction forces between positively charge BF molecules and negatively charged functional groups on M-rGO-060 contribute to the higher dye uptakes.

Fig. 7 The effect of contact time on the adsorption capacities (A) and removal efficiencies (B) of BF, CV and MB onto M-rGO-060 (solid) and M-rGO (hollow). Conditions: C_BF = C_CV = C_MB = 200 mg L⁻¹, pH = 7.0, m/V= 10 mg/20 mL and T = 25 °C.

To well-understand the adsorption mechanisms of three cationic dyes onto both M-rGO-060 and M-rGO, three well-known kinetic models (the Lagergren pseudo first-order,⁴⁴ Lagergren pseudo-second-order⁴⁵ and Webber–Morris intraparticle diffusion⁴⁶) were used to study the adsorption kinetics.

The Lagergren pseudo-first-order kinetic model describes the adsorption of liquid–solid system based on solid capacity,⁴⁴ which can be expressed as follows:

ln(q_e − q_t) = ln q_t − k₁t (3)

where q_t (mg g⁻¹) and q_e (mg g⁻¹) are respectively the amount of dyes adsorbed on the adsorbent surface at time t (min) and equilibrium; k₁ (min⁻¹) is the pseudo-first-order rate constant for the dye adsorption. The values of q_e and k₁ can be determined experimentally by plotting ln(q_e − q_t) versus t and extracting information from the least squares analysis of slope and intercept and substituting into eqn (3)

The Lagergren pseudo-second-order kinetic model consists of all the steps of adsorption including external film diffusion, adsorption, and internal particle diffusion, which can be expressed as follows:⁴⁵

where k₂ (g (mg min)⁻¹) is the pseudo-second-order rate constant for the dye adsorption. The values of k₂ and q_e can be calculated from the intercepts and the slopes of the linear relationship between (t/q_t) and t (eqn (4)).

The Webber–Morris intraparticle diffusion model describes the adsorption process for adsorbate transport from the solution phase to the surface of an adsorbent occurs in several steps.⁴⁶ The overall adsorption process may be controlled by any one of several steps, e.g. film or external diffusion, surface diffusion, pore diffusion and adsorption on the pore surface, or a combination of several steps. It can be expressed by the following equation:

q_t = k_idt^0.5 + C_i (5)

where k_id (mg g⁻¹ min^−1/2) is the intraparticle diffusion constant, C_i is the constant that describes the boundary layer affects. When C_i = 0, the adsorption kinetics are controlled only by intraparticle diffusion; if C_i ≠ 0, the adsorption process is quite complex. The values of k_id and C_i can be calculated from the intercepts and the slopes of q_t versus t^0.5.

By regressing the kinetic data using above the three models, the kinetic parameters and the correlation coefficients (R²) were calculated and listed in Table 1. From the R² values shown in Table 1, the pseudo-second-order model fits the adsorption kinetics much better than the other two models. Moreover, the q_e values calculated (q_e,cal) from the pseudo-second-order model are more consistent with the experimental q_e values (q_e,exp) than those calculated from the pseudo-first-order model, indicating that the adsorption kinetics of BF, CV and MB onto both M-rGO-060 and M-rGO follow the pseudo-second-order model. Moreover, the equilibrium adsorption capacities of M-rGO-060 for each kind of investigated dyes are nearly three times larger than those of M-rGO. This further indicates that the binding of PSSMA functional molecules on M-rGO can dramatically enhance the adsorption capacities of cationic dyes through strong electrostatic interactions between dye molecules and anionic functional groups on the adsorbent.

Table 1 Kinetic parameters of three models (pseudo-first-order, pseudo-second-order and intraparticle diffusion)

Adsorbent	Dye	qe,exp (mg g^−1)	Pseudo-first-order			Pseudo-second-order			Intraparticle diffusion
			k1	qe,cal (mg g^−1)	R^2	k2	qe,cal (mg g^−1)	R^2	kid	Ci	R^2
M-rGO-060	BF	388.2	0.0682	7.3	0.6005	0.038	392.4	0.9999	1.307	384.3	0.7854
	CV	295.5	0.0836	1.7	0.7426	0.262	295.5	0.9999	0.245	294.8	0.3450
	MB	241.5	0.0474	1.6	0.4796	0.289	242.2	0.9999	0.254	240.9	0.3272
M-rGO	BF	161.2	0.1587	48.0	0.8228	0.007	175.4	0.9989	9.322	125.3	0.5724
	CV	117.5	0.0334	3.7	0.4584	0.336	122.0	0.9999	0.535	119.4	0.2628
	MB	90.5	0.0456	3.9	0.6838	0.073	92.6	0.9995	0.626	88.7	0.3313

---

3.2.3. Effect of initial dye concentrations and adsorption isotherms. The effect of initial dye concentrations on the adsorption of BF, CV and MB onto both M-rGO-060 and M-rGO were studied at 25 °C and pH 7 for further investigating the adsorption mechanisms, as shown in Fig. 8. The adsorption capacities and removal efficiencies of three cationic dyes onto both two adsorbents are highly dependent on initial dye concentrations. Dye uptakes increase with increasing the initial dye concentrations at the beginning, and then reach to surface saturation at high concentrations. Moreover, the removal efficiencies of three dyes are very close to 100% at low dye concentrations for M-rGO-060, indicating complete dye removal. This reveals that the active adsorption sites (–COO⁻ and –SO₃⁻) on M-rGO-060 are sufficient and have strong electrostatic interactions with three cationic dyes.^18,27,29 However, at higher dye concentrations, the removal efficiencies of dyes decrease with increasing the initial concentrations. This is due to that the adsorption sites on M-rGO-060 have been saturated by dye molecules for equilibrated adsorption, thus resulting in the decrease of removal efficiencies.¹⁸ Additionally, the adsorption capacities and removal efficiencies of three dyes onto M-rGO-060 are all much higher than those of M-rGO. This indicates again that the binding of PSSMA functional molecules on M-rGO can dramatically enhance the cationic dye uptakes through strong electrostatic interactions between dye molecules and anionic functional groups on M-rGO-060.

Fig. 8 The effect of initial dye concentrations on the adsorption capacities (A) and removal efficiencies (B) of BF, CV and MB onto M-rGO-060 (solid) and M-rGO (hollow). Conditions: pH = 7.0, m/V = 10 mg/20 mL, T = 25 °C and contact time = 5 min.

Equilibrium adsorption isotherm is one of the most important parameters that describe the relationship between the amount of adsorbate uptaken by the adsorbent and the adsorbate concentration remaining in solution. By determining the adsorption capacity of an adsorbent and modeling of isotherms by different equilibrium models, an insight into both the adsorption mechanism and the adsorbent affinity can be elucidated.^12,23,42,47 Fig. 9 shows the equilibrium adsorption isotherms of three dyes onto M-rGO-060 and M-rGO. The adsorption capacities of BF, CV and MB onto both two adsorbents increase dramatically at first, suggesting a high driving force for three dyes adsorption.²³ Then, the amount of adsorbed dyes reach a plateau at high equilibrium solution concentrations, reflecting the saturated dye adsorption.^18,23 Moreover, after the binding of PSSMA functional molecules on M-rGO, the maximal equilibrium dye uptakes (q_e,max) are raised from 171.6 to 589.5 mg g⁻¹ for BF, from 123.4 to 379.8 mg g⁻¹ for CV, and from 87.9 to 269.0 mg g⁻¹ for MB, enhancing 3.43, 3.08 and 3.06 times, respectively. Those results indicate that the obvious enhancement of dye uptakes is again ascribed to the binding of PSSMA on M-rGO, which provides the strong electrostatic interactions between cationic dyes and negatively charged surface of the M-rGO-060.

Fig. 9 Adsorption isotherms of BF, CV and MB onto M-rGO-060 (solid) and M-rGO (hollow). Condition: pH = 7.0, m/V = 10 mg/20 mL, T = 25 °C and contact time = 5 min.

To further study the adsorption mechanisms of three cationic dyes onto M-rGO-060 and M-rGO, the obtained equilibrium adsorption data are fitted by the Langmuir and Freundlich isothermal models,^48–50 respectively. The Langmuir isotherm model is based on monolayer adsorption with uniform energies of adsorption on the surface,⁴⁸ while the Freundlich isotherm model is based multilayer adsorption with the adsorption energy decreases with the surface coverage,⁴⁹ which can be expressed as follows:

where C_e (mg L⁻¹) and q_e (mg g⁻¹) are the concentration and adsorption capacity of dyes at equilibrium, respectively, K_L is the Langmuir constant (L mg⁻¹), and q_m (mg g⁻¹) is the Langmuir monolayer adsorption capacity. The values of q_m and K_L can be determined from the slopes and intercepts of the linear plots of C_e/q_e versus C_e. K_F (mg g⁻¹) and 1/n_F are the Freundlich constant measuring the adsorption capacity and the adsorption intensity, respectively, and can be obtained from the slopes and intercepts of the linear plots of ln q_e versus ln C_e.

Fig. 10 shows the linear plots fitting by the above models for three dyes adsorption onto M-rGO-060 and M-rGO, with their relevant parameters shown in Table 2. The linear correlation coefficients (R²) for both two adsorbents obtained from the Langmuir model are much higher than those from the Freundlich model (>99%). This indicates that the adsorption isotherms of three dyes onto both two adsorbents meet Langmuir model better than Freundlich model and assume a monolayer adsorption process. The maximum Langmuir monolayer adsorption capacities (q_m) of BF, CV and MB calculated from the Langmuir model are 558.2, 384.6 and 270.3 mg g⁻¹ for M-rGO-060, and 169.5, 120.5, and 88.5 mg g⁻¹ for M-rGO, respectively, which are all consistent with the experimental results well. Furthermore, the adsorption process of three dyes onto both two adsorbents was also characterized by Vermeulan criteria associated with the Langmuir isotherm. The Vermeulan criteria can be expressed by a dimensionless constant separation factor (R_L) given as followings:⁴

Fig. 10 Linearized Langmuir (A and B) and Freundlich (C and D) isotherms for BF, CV and MB onto M-rGO-060 (A and C) and M-rGO (B and D). Condition: pH = 7.0, m/V = 10 mg/20 mL, T = 25 °C and contact time = 5 min.

Table 2 Parameters of the adsorption isotherm models

Langmuir isotherm model
Dye	M-rGO-060				M-rGO
	qm (mg g^−1)	KL (L mg^−1)	R^2	RL	qm (mg g^−1)	KL (L mg^−1)	R^2	RL
BF	588.2	0.2152	0.9988	0.085–0.008	169.5	0.1329	0.9968	0.131–0.015
CV	384.6	0.6667	0.9999	0.030–0.003	120.5	0.5287	0.9988	0.036–0.004
MB	270.3	1.480	0.9999	0.013–0.003	88.50	0.6975	0.9984	0.028–0.003

---

Freundlich isotherm model
Dyes	M-rGO-060			M-rGO
	nF	KF ((mg^1−1/n L^1/n) g^−1)	R^2	nF	KF ((mg^1−1/n L^1/n) g^−1)	R^2
BF	4.96	210.3	0.8279	11.1	99.32	0.8031
CV	6.60	182.0	0.6662	15.4	84.94	0.6721
MB	6.91	151.3	0.7800	30.9	74.70	0.4980

---

The value of R_L shows the adsorption process to be either unfavorable (R_L > 1), linear (R_L = 1), favorable (0 < R_L < 1) or irreversible (R_L = 0). In this study, all the R_L values lie in between 0 and 1, indicating that the adsorption of three dyes onto both M-rGO-060 and M-rGO are favorable processes. Moreover, the R_L value of the M-rGO-060 for each dye is smaller than that of the M-rGO, implying more favorable adsorption after the binding of PSSMA functional molecules on M-rGO. The favorability for three dyes onto M-rGO-060 and M-rGO follows the order: BF > CV > MB according to the R_L results, which is fully in accordance with the actual adsorption experiment based on the adsorption capacities as mentioned above.

3.2.4. Adsorption of mixed dyes onto M-rGO-060. Industrial dye wastewater usually contains dyes with multiple kinds. Therefore, in this study, a series of mixed-dye solutions containing BF, CV and MB with different combinations of their concentrations (BF 40 mg L⁻¹, CV 40 mg L⁻¹, and MB 40 mg L⁻¹; BF 80 mg L⁻¹, CV 20 mg L⁻¹, and MB 20 mg L⁻¹; BF 20 mg L⁻¹, CV 80 mg L⁻¹, and MB 20 mg L⁻¹; BF 20 mg L⁻¹, CV 20 mg L⁻¹, and MB 80 mg L⁻¹), were used as model industrial dye wastewater to investigate the removal efficiencies of M-rGO-060 for mixed dyes. As observed from Fig. 11, M-rGO-060 demonstrate relatively higher removal efficiency toward MB (larger than 98%) than those of BF and CV among the mixed-dye solutions with four different combined concentrations. The removal efficiencies of BF and CV onto M-rGO-060 are varied from 80.9% to 98.3%, depending on the initial concentrations of dyes in the mixed-dye solutions. The larger the initial dye concentrations are, the higher the dye removal efficiencies onto M-rGO-060 become. Larger initial dye concentrations can provide higher driving force to overcome the mass transfer resistance of dyes between the aqueous phase and the solid phase, thus resulting in more collisions between dye molecules and the active sites on M-rGO-060.²³ Moreover, the residual BF and CV color can still be observed in the residual mixed-dye solutions after adsorption. This shows that the adsorption of MB may hamper the BF and CV adsorption possibly by steric effect or superior occupation of active sites on M-rGO-060,^43,51 resulted from a relative larger molecular weight of BF and CV compared with that of MB and their different structure features (Fig. S1B–D†).

Fig. 11 Removal efficiencies of BF, CV and MB in mixed-dye solutions with different concentration combinations by M-rGO-060. Condition: pH = 7.0, m/V = 4 mg/15 mL, T = 25 °C and contact time = 5 min. The four concentration combinations of the mixed-dye solutions as: (a) C_BF = 40 mg L⁻¹, C_CV = 40 mg L⁻¹, and C_MB = 40 mg L⁻¹; (b) C_BF = 80 mg L⁻¹, C_CV = 20 mg L⁻¹, and C_MB = 20 mg L⁻¹; (c) C_BF = 20 mg L⁻¹, C_CV = 80 mg L⁻¹, and C_MB = 20 mg L⁻¹; (d) C_BF = 20 mg L⁻¹, C_CV = 20 mg L⁻¹, and C_MB = 80 mg L⁻¹.

3.2.5. Desorption and reusability. The recycling and regenerating ability of an adsorbent is crucial for its practical application. Fe₃O₄ NPs loaded on PSSMA/M-rGO endow them excellent separability from dye solutions under an external magnetic field (Fig. 4C, inset). Accordingly, the dye-saturated adsorbent can be easily recovered and regenerated using NaOH ethanol solution as desorption reagent. To test its adsorption stability, the regenerated M-rGO-060 were used again to adsorb three dyes solutions with the same initial concentrations. As shown in Fig. 12, the removal efficiencies of M-rGO-060 toward three dyes are still higher than 80% after regeneration for 5 times. The excellent desorption performance of the dye-adsorbed M-rGO-060 at alkalic solution is due to that excessive OH⁻ ions compete with positively charged dye molecules for the active adsorption sites on M-rGO-060, and replace the active adsorption sites through substance exchange resulting in desorbing dye molecules from M-rGO-060, thus causing the regeneration of the M-rGO-060.⁵² All the results reveal the feasibility of recycling and regenerating of the PSSMA/M-rGO by using an external magnetic field and an alkalic solution in their practical wastewater treatment applications.

Fig. 12 Removal efficiencies of BF, CV and MB onto M-rGO-060 in five successive cycles of desorption/adsorption. Condition: C_BF = 200 mg L⁻¹, C_CV = 150 mg L⁻¹ C_MB = 125 mg L⁻¹, pH = 7.0, m/V = 10 mg/20 mL and T = 25 °C.

4. Conclusion

In summary, a novel type of PSSMA-modified magnetic rGO nanocomposites (PSSMA/M-rGO) have been successfully prepared via simple one-pot solvothermal method and used for removal of three typical cationic dyes (BF, CV and MB) from aqueous solutions. The resultant PSSMA/M-rGO demonstrate much higher adsorption capacities and removal efficiencies toward three cationic dyes than those of M-rGO without PSSMA modification, which is due to the strong electrostatic interactions between the positively charged dye molecules and the negatively charged surface of PSSMA/M-rGO. The adsorption of BF, CV and MB onto M-rGO-060 follows the Langmuir isotherm model with maximum monolayer adsorption capacities of 588.2 mg g⁻¹ for BF, 384.6 mg g⁻¹ for CV and 270.3 mg g⁻¹ for MB, which are three times larger than those of M-rGO. The kinetics of the adsorption process fit the Lagergren pseudo-second-order kinetic model. Moreover, the M-rGO-060 also exhibit high removal efficiencies toward mixed dyes of BF, CV and MB. The saturatedly-adsorbed M-rGO-060 can be easily recovered under an external magnetic field and effectively regenerated using NaOH ethanol solution. Furthermore, after regeneration for 5 times, the M-rGO-060 still show high adsorption performances toward BF, CV and MB (>80%). The PSSMA/M-rGO developed in this work are expected to be promising adsorbents for highly-efficient removal of cationic organic pollutants from aqueous solutions, due to their high separation efficiency, low production cost and excellent recyclable property.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (21106116), the project of Science & Technology department of Sichuan Province (2014GZ0012) and the project of postgraduate degree construction, Southwest University for Nationalities (2015XWD-S0703).

Notes and references

1. Y.-Q. Hu, T. Guo, X.-S. Ye, Q. Li, M. Guo, H.-N. Liu and Z.-J. Wu, Chem. Eng. J., 2013, 228, 392–397.
2. Z. Dong, D. Wang, X. Liu, X. Pei, L. Chen and J. Jin, J. Mater. Chem. A, 2014, 2, 5034–5040.
3. E. N. E. Qada, S. J. Allen and G. M. Walker, Chem. Eng. J., 2008, 135, 174–184.
4. L. Fan, C. Luo, X. Li, F. Lu, H. Qiu and M. Sun, J. Hazard. Mater., 2012, 215–216, 272–279.
5. S. M. Ghoreishi and R. Haghighi, Chem. Eng. J., 2003, 95, 163–169.
6. C. H. Wu, Y. Z. Zhang, S. Li, H. J. Zheng, H. Wang, J. B. Liu, K. W. Li and H. Yan, Chem. Eng. J., 2011, 178, 468–474.
7. S. Raghu and C. Ahmed Basha, J. Hazard. Mater., 2007, 149, 324–330.
8. S. Chakraborty, M. K. Purkait, S. DasGupta, S. de and J. K. Basu, Sep. Purif. Technol., 2003, 31, 141–151.
9. M. Rafatullah, O. Sulaiman, R. Hashim and A. Ahmad, J. Hazard. Mater., 2010, 177, 70–80.
10. M. Auta and B. H. Hameed, Chem. Eng. J., 2012, 198–199, 219–227.
11. R. Han, J. Zhang, P. Han, Y. Wang, Z. Zhao and M. Tang, Chem. Eng. J., 2009, 145, 496–504.
12. B. Royer, N. F. Cardoso, E. C. Lima, J. C. P. Vaghetti, N. M. Simon, T. Calvete and R. C. Veses, J. Hazard. Mater., 2009, 164, 1213–1222.
13. B. H. Hameed and A. A. Ahmad, J. Hazard. Mater., 2009, 164, 870–875.
14. C. H. Huang, K. P. Chang, H. D. Ou, Y. C. Chiang and C. F. Wang, Microporous Mesoporous Mater., 2011, 141, 102–109.
15. A. Mills, M. Sheik, C. ÓRourke and M. McFarlane, Appl. Catal., B, 2009, 89, 189–195.
16. J. Fu, Z. Chen, M. Wang, S. Liu, J. Zhang, J. Zhang, R. Han and Q. Xu, Chem. Eng. J., 2015, 259, 53–61.
17. S. Wang, C. W. Ng, W. Wang, Q. Li and Z. Hao, Chem. Eng. J., 2012, 197, 34–40.
18. H. Yan, X. Tao, Z. Yang, K. Li, H. Yang, A. Li and R. Cheng, J. Hazard. Mater., 2014, 268, 191–198.
19. G. K. Ramesha, A. Vijaya Kumara, H. B. Muralidhara and S. Sampath, J. Colloid Interface Sci., 2011, 361, 270–277.
20. P. Sharma and M. R. Das, J. Chem. Eng. Data, 2013, 58, 151–158.
21. T. Liu, Y. Li, Q. Du, J. Sun, Y. Jiao, G. Yang, Z. Wang, Y. Xia, W. Zhang, K. Wang, H. Zhu and D. Wu, Colloids Surf., B, 2012, 90, 197–203.
22. S.-T. Yang, S. Chen, Y. Chang, A. Cao, Y. Liu and H. Wang, J. Colloid Interface Sci., 2011, 359, 24–29.
23. L. Ai, C. Zhang and Z. Chen, J. Hazard. Mater., 2011, 192, 1515–1524.
24. S. Bai, X. Shen, X. Zhong, Y. Liu, G. Zhu, X. Xu and K. Chen, Carbon, 2012, 50, 2337–2346.
25. Y. Yao, S. Miao, S. Liu, L. P. Ma, H. Sun and S. Wang, Chem. Eng. J., 2012, 184, 326–332.
26. X. Liu, L. Yan, W. Yin, L. Zhou, G. Tian, J. Shi, Z. Yang, D. Xiao, Z. Gu and Y. Zhao, J. Mater. Chem. A, 2014, 2, 12296–12303.
27. J. Zhang, M. S. Azam, C. Shi, J. Huang, B. Yan, Q. Liu and H. Zeng, RSC Adv., 2015, 5, 32272–32282.
28. M. Namvari and H. Namaz, Polym. Int., 2014, 63, 1881–1888.
29. L. Cui, X. Guo, Q. Wei, Y. Wang, L. Gao, L. Yan, T. Yan and B. Du, J. Colloid Interface Sci., 2015, 439, 112–120.
30. L. Fan, C. Luo, M. Sun, X. Li, F. Lu and H. Qiu, Bioresour. Technol., 2012, 114, 703–706.
31. D. Wang, L. Liu, X. Jiang, J. Yu and X. Chen, Colloids Surf., A, 2015, 466, 166–173.
32. L. Fan, C. Luo, M. Sun, H. Qiu and X. Li, Colloids Surf., B, 2013, 103, 601–607.
33. Y.-B. Song, S.-N. Lv, C.-J. Cheng, G.-L. Ni, X.-W. Xie, W. Huang and Z.-G. Zhao, Appl. Surf. Sci., 2015, 324, 854–863.
34. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
35. K.-F. Zhou, Y.-H. Zhu, X.-L. Yang and C.-Z. Li, New J. Chem., 2010, 34, 2950–2955.
36. J. Gao, X. Ran, C. Shi, H. Cheng, T. Cheng and Y. Su, Nanoscale, 2013, 5, 7026–7033.
37. J. I. Paredes, S. Villar-Rodil, A. Martínez-Alonso and J. M. D. Tascón, Langmuir, 2008, 24, 10560–10564.
38. D. Li, M. B. Müller, S. Gilje, R. B. Kaner and G. G. Wallace, Nat. Nanotechnol., 2008, 3, 101–105.
39. S. Villar-Rodil, J. I. Paredes, A. Martínez-Alonso and J. M. D. Tascón, J. Mater. Chem., 2009, 19, 3591–3593.
40. H. Peng, L. Meng, L. Niu and Q. Lu, J. Phys. Chem. C, 2012, 116, 16294–16299.
41. F. He, J. Fan, D. Ma, L. Zhang, C. Leung and H. L. Chan, Carbon, 2010, 48, 3139–3144.
42. L. Ai, C. Zhang, F. Liao, Y. Wang, M. Li, L. Meng and J. Jiang, J. Hazard. Mater., 2011, 198, 282–290.
43. H. Yan, H. Li, H. Yang, A. Li and R. Cheng, Chem. Eng. J., 2013, 223, 402–411.
44. S. Largegren, K. Sven. Vetenskapsakad. Handl., 1898, 24, 1–39.
45. Y. S. Ho and G. M. Mckay, Process Biochem., 1999, 34, 451–465.
46. W. J. Weber and J. C. Morris, J. Sanit. Eng. Div., Am. Soc. Civ. Eng., 1963, 89, 31–59.
47. B. S. Inbaraj and B. H. Chen, Bioresour. Technol., 2011, 102, 8868–8876.
48. I. Langmuir, J. Am. Chem. Soc., 1916, 38, 2221–2295.
49. H. M. F. Freundlich, J. Phys. Chem., 1906, 57, 385–471.
50. G. McKay, M. El Guendi and M. Nassar, Water Res., 1987, 21, 1513–1520.
51. T. Yao, S. Guo, C. Zeng, C. Wang and L. Zhang, J. Hazard. Mater., 2015, 292, 90–97.
52. H. Zhang, X. Li, G. He, J. Zhan and D. Liu, Ind. Eng. Chem. Res., 2013, 52, 16902–16910.

---

Footnote

† Electronic supplementary information (ESI) available: Molecular structures of PSSMA and three cationic dyes: basic fuchsin (BF), crystal violet (CV) and methylene blue (MB) (Fig. S1); UV-vis spectra of GO, M-rGO, M-rGO-060 and PSSMA dispersed in water (Fig. S2); digital photographs of M-rGO-060 and M-rGO aqueous dispersions upon storing different time at room temperature (Fig. S3); adsorption capacities and removal efficiencies of BF, CV and MB onto Fe₃O₄-060, M-rGO-060 and rGO-060 and FT-IR spectrum of rGO-060. See DOI: 10.1039/c5ra18255g

---