Graphene oxide-based polyethersulfone core–shell particles for dye uptake

Abstract

Graphene oxide (GO), a graphene nanomaterial with great application potential, possesses promising adsorption abilities towards various water contaminants due to the ultra-large surface area and the nature of electric charge on the surface. However, ultrahigh centrifugation for a prolonged time is strongly needed to collect the highly dispersed GO in the recovery process. In this study, a GO-based polymeric composite particle with core–shell structure was fabricated by a facile method. Polyethersulfone (PES) was chosen as the shell to enwrap GO through a liquid–liquid phase inversion process, since the PES shell presented high porosity, good mechanical property and easily modified ability. Methylene blue (MB), a cationic dye, was chosen as the adsorbate to investigate the adsorption capabilities, kinetics and isotherms of the prepared particles. The PES@GO core–shell particles displayed an adsorption capacity as high as 352.11 mg g⁻¹ for MB dye, and the adsorption rates could be improved by modifying the PES shells with hydrophilic fillers. The MB adsorption behavior fitted the pseudo-second-order kinetic model and the Langmuir isotherm very well, and the adsorption process was controlled by the intra-particle diffusion. In addition, a particle column was used to further study the removal ability of environmental toxins, and the results revealed that the composite particles had great potential to remove cationic dyes for wastewater treatment on an industrial scale.

---

1. Introduction

Recently, water pollution by dyes has become a serious environmental issue worldwide. A large amount of dyes are produced annually and used to color industrial products such as textiles, leather, paper, painting and so on.¹ During the dyeing process, approximately 10–15% of the dye is lost in the effluent.^2,3 The immoderate release of wastewater containing dyes can color water resources, reduce light penetration, thus endangering the lives of aquatic life⁴; some synthetic dyes may also cause severe health threats to human beings.⁵ As a result, many efforts and attempts have been made towards clearing such contaminants from water and a variety of technologies have been exploited. Among them, the adsorption method is considered to be better because of its convenience, ease of operation and simplicity of design.⁶ Many types of materials were used as adsorbents to remove dyes in water with considerable adsorption capacities, such as inorganic clays and carbon-based materials,^7–11 organic polymer materials,^12–14 and organic–inorganic hybrid materials.^15–17 In general, inorganic materials are abundant in reserves, and usually have high adsorption capacities and small sizes; organic polymers have good elasticity, toughness, formability and low density; while organic–inorganic hybrid materials can integrate the advantages of the former two into one material.

As a new type of carbon-based material, graphene has attracted a great deal of scientific interesting in many fields due to its excellent performance since it was first reported in 2004.¹⁸ And as a precursor for graphene preparation, graphene oxide (GO) has also gained considerable attention. Compared with graphene, GO contains reactive oxygen functional groups including hydroxyl, epoxy and carboxylic acid groups on its large surfaces.¹⁹ Based on these chemical properties, GO has been recognized as a superior sorbent for removing environmental contaminants, such as heavy metal ions,^20,21 endocrine disruptors²² and cationic dyes.²³ However, some characters of GO such as the hydrophilic and small size makes it inconvenient for manipulation and collection in use as the adsorbent; the GO emitted into water may cause secondary pollution and exhibit some fatal biological defects.^24,25

Therefore, many researchers have developed various GO-based hybrid adsorbents by the aid of organic polymers or organic poly-reactions; the integrated morphology of the GO-based monolith gives the adsorbents specific functions.²⁶ It was reported that various synthetic polymers and biopolymers, such as chitosan,²⁷ konjac glucomannan,²⁸ polyethylenimine,²⁹ DNA and bovine serum albumin,²⁵ could be used to fabricated 3-dimentional GO-based hydrogels by hydrogen bonding and other interactions between GO and polymers. Such GO-based hydrogels could also be prepared by adding GO into reaction systems of some monomers (such as acrylic acid, acrylamide and sodium styrene sulfonate), and followed by a polymerization and cross-linking; in this case, GO could be fixed in the networks of polymers or served as a micro-crosslinker.^30–32 In addition, GO-based polymeric particles,⁴ fibers,²⁴ foams,³³ films^34,35 and membranes^36,37 could be designed by dispersing GO into relevant raw materials of engineering plastics during their forming processes; after the formation, GO was embedded and immobilized in the polymer materials.

The above methods employed blends as the main strategy; and the prepared hydrogels and plastics could prevent the release of GO, meanwhile showed outstanding adsorption capacities and gained some special functions. However, the above mentioned materials also had some inevitable flaws, such as the mechanical strengths of hydrogel materials were often low, which would cause their breaking during use. Although plastic materials had better mechanical strengths, the embedment of GO would reduce the adsorption area and oxygen-containing functional groups, and then decrease the adsorption capacity. As a result, the development of an effective strategy to prepare GO-based adsorbents with high adsorption capacity and acceptable mechanical property is still desired.

In this paper, we demonstrated a facile approach to fabricate a novel GO/polyethersulfone (PES) composite adsorbent: dropping GO aqueous solution into PES solution, and then taking the GO “droplets” out. It is known that PES-based materials showed outstanding oxidative, thermal and good mechanical property and could be prepared by a phase inversion method in a very short period of time.³⁸ Based on the rapid phase inversion, a thin, porous PES film could form around the drop of GO aqueous solution as soon as it was dropped into the PES solution; when the GO droplets were took out, a core–shell structured composite particle could be obtained. The PES shell could effectively prevent the GO emitting, contribute good mechanical properties; and more importantly, the adsorption capacity of GO would not sacrifice. Herein, the GO-based core–shell particles with three different PES shells were fabricated: (i) PES@GO particles: the shells were unmodified PES; (ii) PES/PAA@GO particles, the particles with enhanced adsorption rates and capacities: the shells were modified PES by in situ polymerization of acrylic acid (AA); (iii) PES/GO@GO particles, the particles were produced for the same purpose as (ii), and the shells were modified PES by blending with GO. The morphologies and physicochemical properties of the particles were characterized, and a cationic dye methylene blue (MB) was chosen as the adsorbate to investigate the adsorption capacities of these composite particles.

2. Experimental

2.1. Materials

Graphite flakes were obtained from Sigma-Aldrich. Polyethersulfone (PES, Ultrason E6020P) was purchased from BASF chemical Co. (Germany). Methyl methacrylate (MMA) and acrylic acid (AA) were purchased from Aladdin reagent Co. Ltd. (China) with analytical grade and used as received. N,N′-Methylenebis(acrylamide) (MBA), azobis(isobutyronitrile) (AIBN) and methylene blue (MB) were received from Aladdin reagent Co. Ltd. (China). N,N-Dimethylacetamide (DMAc) was obtained from Chengdu Kelong Chemical Reagent Co. Ltd. (China). Unless otherwise stated, other reagents were received from Aladdin and used without any further purification. Deionized water was obtained from a pure water production system and used throughout this study.

2.2. Preparation of P(MMA-AA) and GO blended PES (PES/PAA and PES/GO) casting solutions

As for PES/PAA solution, the copolymerization of MMA and AA (mass ratio 1 : 1) was carried out in PES solution by an in situ cross-linking polymerization. AIBN was introduced as initiator and MBA as cross-linking reagent. Primarily, PES was completely dissolved in DMAc in a 250 mL three-necked round bottom flask. After passing nitrogen for 15 min, a mixture of monomers, AIBN (1 mol% of the monomers) and MBA (2 mol% of the monomers) were added into the PES solution under nitrogen atmosphere. The mass ratio of the monomers to PES is 30 wt%. The copolymerization was carried out at 75 °C with a stirring of 330 rpm for 24 hours. Aqueous GO dispersion was prepared by a modified Hummers' method from the natural flake graphite,³⁹ the details and the characterization methods were shown in ESI.† With regard to the preparation of PES/GO solution, the GO solution was firstly freeze-dried and dispersed in DMAc with a required concentration by vigorous agitation and ultrasonic for about 24 hours, and then the dried GO was dispersed in DMAc. Then PES was added and the mixture was stirred by a magnetic stirrer for 12 hours to get homogeneous solution. The mass ratio of GO to PES is 10 wt%.

2.3. Preparation and characterization of GO-based core–shell particles

GO-based core–shell particles were prepared by a phase inversion method, which has been described in detail in our earlier report.³⁸ Firstly, a GO aqueous solution with a concentration of 10 mg g⁻¹ was injected into PES solution using a 0.8 mm diameter syringe needle at room temperature. After about 3 seconds, the particles were removed from the polymer solution, exposed to air for certain seconds, and then placed into water. Eventually, the particles were incubated in deionized water by changing fresh deionized water frequently to remove the unreacted molecules and residual solvent thoroughly. With the purpose of investigating the effect of the PES shell on adsorption capacities, a series of PES@GO particles with different concentrations of PES were prepared, which were controlled at 8%, 10%, 12% and named as PES8@GO, PES10@GO and PES12@GO, respectively. The particles with the P(MMA-AA) and GO modified shells (PES/PAA@GO and PES/GO@GO) were prepared using the same method by replacing PES solution with PES/PAA solution and PES/GO solution.

The prepared particles were characterized by scanning electron microscopy (SEM), Fourier transform infrared (FTIR) spectroscopy analysis and water contact angle. For the SEM (JSM-7500F, JEOL) observation, the particles were freeze-dried overnight, and then cut by a single-edged razor blade after immersed into liquid nitrogen. Afterwards, the prepared samples were attached to the sample supports and coated with a gold layer prior to observation. The Fourier transform infrared (FTIR) spectroscopy analysis of the particles was obtained by using a FTIR spectrometer (Nicolet 560, USA), before which the particles were completely freeze-dried for 12 hours. The hydrophilicity of the shells of the particles was characterized on the basis of contact angle measurement using a contact angle goniometer (OCA20, Dataphysics, Germany) equipped with a video capture. For convenience, the membranes fabricated by the corresponding solutions as those used to form the shells were used to measure the hydrophilicity of the shells. For the static contact angle measurements, 3 μL of water was dropped on the surface of the membrane with an automatic piston syringe and photographed. At least three measurements were averaged to get a reliable value.

2.4. Adsorption experiments

For the adsorption experiments, 10 particles were applied in 10 mL of 250 μmol L⁻¹ MB solution in conical flasks at room temperature with the stirring speed of 100 rpm, and the pH value of the solution was 7. The MB concentrations were determined at different time intervals using the UV-vis spectrophotometer 756PC at the wavelength of 631 nm.⁴⁰ The data were also used for adsorption kinetics study. To study the influence of the initial concentrations, the particles were applied to 10 mL MB solutions with the initial concentrations of 100, 200, 300, 400, and 500 μmol L⁻¹, respectively. The test conditions and methods were the same as mentioned above. The concentrations were also determined at different time intervals and the data were also used for analyzing adsorption isotherm.

The adsorption capacity of pure GO was also studied, and used for comparison with the GO in the particles; the GO aqueous solution was used to adsorb MB at room temperature and pH 7. Generally, 5 mL of 1 mg L⁻¹ GO aqueous solution was added to 20 mL of MB solution (0.1–2 mmol L⁻¹) in conical flasks with stirring by a magnetic stirrer for 30 min to reach equilibrium. Then, the mixture was centrifuged at 12 000 rpm for 10 min to remove the GO, and the concentrations of the supernatant containing residual MB were determined with the UV-vis spectrophotometer 756PC at the wavelength of 631 nm.

2.5. Removal of MB by particle column

For the removal of MB by particle column, a polypropylene syringe (5 mL) was used to place the three particles (PES8@GO, PES/PAA@GO and PES/GO@GO), and the mobile phase had a length of about 20 mm. The solution containing MB passed through the particle column with a flow rate of 1 mL min⁻¹, and then the concentration of the MB in the eluted solution was determined. The process was repeated for three times by reusing the eluted solution.

3. Results and discussions

3.1. Characterization of GO

AFM image (Fig. S1a†) and SEM image (Fig. S1b†) directly revealed the external morphology of GO, as well as the apparent thickness of GO, verifying the characteristic single-layered two-dimensional structure of GO nanosheets. The infrared spectrum (Fig. S1c†) certified to the chemical compositions of GO and the oxygen functional groups in GO.

3.2. Preparation and characterization of GO-based core–shell particles

The preparation process of the core–shell particles is shown in Fig. 1a. Firstly, high concentration of GO aqueous solution was dripped into PES/DMAc casting solution. Since the concentrated GO solution was viscous, a uniform sphere of the GO droplets could be remained after they were dripped into the PES/DMAc solution. Meanwhile, to ensure the GO droplets could sink into the PES/DMAc solution successfully, a low viscosity of PES solution was needed, and the concentration was adjusted to 8 to 12 wt% (usually the concentrations of PES casting solutions were higher than 14 wt% (ref. 38)). As soon as the GO droplets immerged into the PES/DMAc solution, phase inversion would occur immediately, and a PES shell was formed around the GO droplets. Then, the nascent GO-based core–shell particles were removed from the PES/DMAc solution and exposed to air, to avoid the PES shell being re-dissolved. When the nascent particles were placed into water, second phase inversion occurred. After the completion of the exchange between water and the solvent of DMAc, GO-based core–shell particles were prepared. The digital photos of PES8@GO and its cross-section are shown in Fig. 1b and c, respectively. The spherical and uniform exterior PES surfaces could be observed from Fig. 1b. And seen from Fig. 1c, the core–shell structure of the particle could be observed roughly (for the observation of the cross-section, the particle was first frozen in liquid nitrogen, and then sectioned into two halves with a blade): the brown GO core was enwrapped in the thin, white PES shell.

Fig. 1 The schematic illustration for preparing particles (a); the digital photos of panorama (b) and the cross-section (c) of PES8@GO.

The SEM images of the cross-section of PES8@GO are shown in Fig. 2. Seen from the panorama of the cross-section (Fig. 2a), a round sphere with the core–shell structure was easy to be observed. The thin, porous PES film in the edge of the sphere served as the shell (as shown in Fig. 2b). Due to the two times phase inversion, dense skin layers were appeared both in the internal and external surface of the PES film, which could effectively prevent the GO sheets eluting. The interior of the film was composed by relatively larger pores with finger-like structure. The enwrapped core was a kind of porous network which was composed with gossamer substance (as shown in Fig. 2c), which could be identified as the GO network according to previous study.²⁶ It is worth mentioning that since the particles were freeze-dried before SEM observation, the GO sheets were immobilized and self-assembled into a whole porous network. However, based on our design, the particles did not need drying so the cores of them were composed by GO aqueous solution. In fact, GO sheets were always dispersed in water independently and freely during the adsorption process.

Fig. 2 The SEM images of PES8@GO: the cross-section of the particle (a); the magnified view for the core–shell structure (b); and the magnified view for the inside core (c).

The dense PES shell could prevent the GO sheets eluting, and it might also impede the contacting of adsorbates and GO sheets. In order to study the effect of the shell on adsorption capacities, three kinds of PES@GO particles were prepared, i.e., PES8@GO, PES10@GO and PES12@GO, and the concentrations of PES casting solutions were 8, 10 and 12 wt%, respectively. The three particles have been observed by SEM, and the results showed no significant differences in the thickness or pore structure of their shells (seen in Fig. S2†). However, the dry weights of the three particles were increased with the increase of the PES concentrations; and the values were 3.2 ± 0.3 mg, 4.0 ± 0.3 mg and 5.0 ± 0.3 mg for the PES8@GO, PES10@GO and PES12@GO, respectively. The dosages of GO aqueous solution were the same during the preparation, thus the difference in the quality of the three particles was caused by the differences in PES concentrations; the higher concentration led to the denser shell, which would affect the adsorption capacity as discussed in the following sections.

As we known, various methods could be applied for PES membrane modification; some of them were simple and effective, such as physical blending and in situ cross-linking.^41–44 In order to reduce the obstruction to adsorption by the GO core which caused by the PES shell, two kinds of particles with modified PES shells were prepared: (1) by in situ cross-linking copolymerization with P(MMA-AA) into PES casting solution, a core–shell particle with the PES/PAA hybrid shell could be obtained (PES/PAA@GO); (2) by blending GO into PES casting solution, another core–shell particle with the PES/GO hybrid shell could be obtained (PES/GO@GO). The digital photos of panoramas and cross-sections of PES8@GO, PES/PAA@GO and PES/GO@GO are shown in Fig. 3a. All the three particles displayed spherical and uniform exterior surfaces; the colors of PES8@GO and PES/PAA@GO shells were white, while the PES/GO@GO shell was brown due to the blending of GO in the shell. As seen from the cross-section views, the three particles all showed core–shell structures. The FTIR of the shells of PES8@GO and PES/PAA@GO are shown in Fig. 3b. Compared with PES8@GO, the FTIR spectrum of the shell of PES/PAA@GO showed an adsorption peak at 1729.5 cm⁻¹, which was attributed to the carboxyl and carbonyl groups in the P(MMA-AA) copolymer, verifying the existence of P(MMA-AA) in the shell. However, the FTIR results showed no obvious differences between the shells of PES/GO@GO and PES8@GO (the results of FTIR of PES/GO@GO shell were not given) due to the lower GO content.

Fig. 3 The digital photos of PES8@GO, PES/PAA@GO and PES/GO@GO (a); the FTIR spectrum of PES8@GO and PES/PAA@GO (b); the water contact angles of PES8@GO, PES/PAA@GO and PES/GO@GO (c); the SEM images of the interior structures of shells: PES8@GO (d), PES/PAA@GO (e), and PES/GO@GO (f).

In order to characterize the water contact angle, three flat membranes with the same contents as the shells of PES8@GO, PES/PAA@GO and PES/GO@GO were prepared respectively, and the results are shown in Fig. 3c. The water contact angle of the shell of PES8@GO was about 90 ± 5°; while the water contact angles of the shells of the PES/PAA@GO and PES/GO@GO were 70 ± 5° and 80 ± 5°, respectively. Besides, the water droplets on the PES/PAA@GO and PES/GO@GO were absorbed by the membranes after a short time, whereas the shapes of water droplets on PES8@GO were maintained. These results indicated that the hydrophilicity of the shells for the PES/PAA@GO and PES/GO@GO were improved. Thus, the adsorption capacities of the particles with the modified shells might be improved.

The SEM observations of PES8@GO, P-PAA/PES@GO and PES/GO@GO were carried out, and the results indicated that the cross-section morphologies of the three particles were the same in general; however, some differences could be observed in the interior of the shells, as shown in Fig. 3d–f. By comparing Fig. 3d and e, the gauze-like GO sheets could be found depositing and covering in some of the internal surface of the shell of PES/PAA@GO (as indicated by red arrows); these results were similar with our previous work.⁴⁵ As the GO sheets did not exist in the casting solution of PES/PAA@GO, the appearance of GO sheets could be attributed to the diffusion behavior of GO drops. The existence of P(MMA-AA) made the PES casting solution more hydrophilic, and had a certain compatibility with GO aqueous solution. So the GO could diffuse into the casting solution during the phase inversion. Meanwhile, the hydrophilic P(MMA-AA) polymer would also diffuse into GO drops, this provided spaces for GO diffusion. For the shell of PES/GO@GO, the SEM image (Fig. 3f) indicated that most of the internal surface of the shell was covered by GO sheets; compared with Fig. 3e, the GO sheets covered on the surface appeared smoother; and the appeared GO sheets came from the PES/GO casting solution.

3.3. Adsorption kinetics study

In this study, a cationic dye methylene blue (MB) was chosen as the model adsorbate to investigate the adsorption capacities of the core–shell particles. The relationship curves between contacting times and the adsorbed amounts per unit mass of the particles are shown in Fig. 4a. The results indicated that all the particles showed the similar adsorption equilibrium times, and the adsorption equilibriums were reached after about 72 hours. The long equilibrium time might be caused by the dense PES shells, which led to the long diffusion time of the adsorbate molecules. In addition, the contents of GO cores were the same among all the particles and the major adsorption capacities for MB were contributed by the cores; as a result, the equilibrium adsorption capacities of the particles were close too (the values were 261.16, 257.07, 256.44, 260.12 and 260.99 mg g⁻¹ for the PES8@GO, PES10@GO, PES12@GO, PES/PAA@GO and PES/GO@GO, respectively). However, with the different shells, the adsorption rates were different in the earlier stage of the adsorption processes. Comparing the curves of PES8@GO, PES10@GO and PES12@GO, the adsorption rates were decreased with the increase of the PES concentrations in the casting solutions. As mentioned above, the density of the shells would increase with the increase of PES concentrations, and thus, the porosity and permeability decreased;³⁸ such a decrease might restrict the diffusion of the adsorbates, and lead to the decrease of the adsorption rates. When comparing to PES@GO particles, the adsorption rates of PES/PAA@GO and PES/GO@GO were increased. There might be two reasons for this result: firstly, modifying with P(PMMA-AA) or GO could improve the hydrophilicity of the PES shells; and the modifications could also give the shells adsorption capacities for the adsorbates since the P(PMMA-AA) or GO in the shells could themselves adsorb dyes; as a result, the impeding effect of the shells were reduced; secondly, the porosities of the PES shells would increase after blending the hydrophilic fillers or macromolecules into the shells, which would improve the permeability of the adsorbates.

Fig. 4 The adsorbed amounts per unit mass for PES8@GO, PES10@GO, PES12@GO, PES/PAA@GO and PES/GO@GO. 10 particles are applied in 10 mL of 250 μmol L⁻¹ MB solution (a); applications of the pseudo-first-order adsorption model (b), the pseudo-second-order adsorption model (c) and the intra-particle diffusion model (d).

From the results of Fig. 4a, the adsorption ratios of MB were 97.98%, 96.44%, 96.21%, 97.59% and 97.92% for the PES8@GO, PES10@GO, PES12@GO, PES/PAA@GO and PES/GO@GO, respectively. The high adsorption ratio of MB portended the efficient adsorption capacities against toxic cationic molecules, and betokened the great potential in purification fields. Accordingly, the mechanism of the adsorption process on the GO-based core–shell particles became critical and was studied in detail. In order to investigate the kinetics of the adsorption processes of the particles, three kinetic models were used to analyze the experimental data, i.e. the pseudo-first-order equation, the pseudo-second-order equation and the intra-particle diffusion equation.

3.3.1. The pseudo-first-order and pseudo-second-order kinetic model. The pseudo-first-order equation, also known as the Lagergren rate equation, was the first rate equation for the adsorption of liquid–solid system based on solid capacity. The linear form is formulated as:

ln(q_e − q_t) = ln q_e − k₁t (1)

where k₁ is the rate constant of pseudo-first-order equation (min⁻¹), q_e and q_t are the adsorption amounts of MB at equilibrium and at time t (mg g⁻¹). The slope −k₁ and intercept ln q_e were evaluated from the linear regression of ln(q_e − q_t) versus t (Table 1). The correlation coefficients of all the particles for the pseudo-first-order adsorption model were higher than 0.99, which meant that the adsorption process fitted the pseudo-first-order model well; however, it was found that the calculated equilibrium adsorption capacities (q_e) were slightly lower than the experimental ones.

Table 1 The parameters of pseudo-first-order model for the adsorption of MB by the particles

Particles	k1 (min^−1)	qe(cal) (mg g^−1)	R^2
PES8@GO	0.0004	222.3	0.996
PES10@GO	0.0004	241.1	0.999
PES12@GO	0.0003	240.3	0.998
PES/PAA@GO	0.0004	214.6	0.996
PES/GO@GO	0.0004	202.3	0.993

---

The pseudo-second-order equation based on the adsorption equilibrium capacity of the dye molecules on the surface of the adsorbent was also used. It can be expressed in the following form:

where k₂ is the rate constant of pseudo-second-order adsorption; q_t (mg g⁻¹), q_e (mg g⁻¹), and t (min) have the same meaning as that in pseudo-first-order equation. As shown in Fig. 4c, the plot of t/q_t against t should give a straight line with slope 1/q_e and intercept 1/(k₂q_e²). As seen in Table 2, all the correlation coefficients for the pseudo-second-order adsorption model (R² > 0.99) were very high and close to unity, which meant that the contaminants adsorption to the particles agreed with the pseudo-second-order model very well. Seen from Table 2, the rate constants of PES8@GO, PES10@GO and PES12@GO decreased gradually; while those of PES/PAA@GO and PES/GO@GO increased when compared to PES@GO particles. The results indicated that the adsorption rates of the PES@GO particles decreased with the increase of PES concentrations in the casting solutions. Meanwhile, the adsorption rates of the particles with modified shells were faster than those of the PES@GO particles. The calculated equilibrium adsorption capacities q_e were closer to the experimental ones in pseudo-second-order equation (except PES12@GO) and larger than those in pseudo-first-order equation, which meant that the pseudo-second-order model described the adsorption process better than the pseudo-first-order model. It could be inferred that chemical process played a dominant role in the adsorption process and the rate-limiting step was chemisorption.

Table 2 The parameters of pseudo-second-order model for the adsorption of MB by the particles

Particles	k2 (× 10^−6 g mg^−1 min^−1)	qe(cal) (mg g^−1)	R^2
PES8@GO	3.3	277.8	0.999
PES10@GO	2.1	285.7	0.999
PES12@GO	1.5	294.1	0.998
PES/PAA@GO	3.7	277.8	0.999
PES/GO@GO	4.5	277.8	0.999

---

3.3.2. The intra-particle diffusion model. In this study, the adsorption of MB for the particles took very long time to reach equilibrium, and the adsorption rate might be governed by the intra-particle diffusion. The equation could be written in the following form:

q_t = k_pt^1/2 + C (3)

where k_p is the rate constant of intra-particle diffusion model, C is a constant for any experiment (mg g⁻¹), q_t and t have the same meaning as before. As seen in Fig. 4d, the curves by plotting q_t versus t^1/2 presented three steps and the slopes of the linear part of each step could give the rate constants. The values of C could be determined from the intercepts of the linear regression of the first step which is related to the thickness of the boundary layer. It could be seen in Table 3 that the intercepts C were less than zero, which represented the negative effect of the PES shells on MB molecule diffusion; meanwhile, the intercepts of PES/PAA@GO and PES/GO@GO were higher, which indicated that after the modifications, the impeding effect of the shells were reduced.

Table 3 The parameters of intraparticle diffusion model for the adsorption of MB by the particles

Particles	C (mg g^−1)	Step I		Step II		Step III
		k1 (mg g^−1 min^−1/2)	R1^2	k2 (mg g^−1 min^−1/2)	R2^2	k3 (mg g^−1 min^−1/2)	R3^2
PES8@GO	−21.54	4.732	0.998	2.560	0.987	0.525	0.950
PES10@GO	−21.65	3.822	0.992	2.747	0.990	0.891	0.963
PES12@GO	−21.70	3.410	0.989	2.709	0.995	0.980	0.987
PES/PAA@GO	−18.74	4.938	0.997	2.273	0.981	0.635	0.976
PES/GO@GO	−16.81	5.270	0.999	2.261	0.962	0.512	0.952

---

In addition, the three linear sections with different slopes of each curve indicated that during the adsorption process there were at least three rate-limited steps that occurred. For the particles with porous shells, the diffusion process of contaminants may follow three sequential steps: cross-film diffusion, intra-particle diffusion, and final equilibrium step. In the first step, the rate constants of PES/PAA@GO and PES/GO@GO were higher than the particles with unmodified shells, indicating that the PES shells played important role in the adsorption rates. However, in the second and third steps, the rate constants of PES@GO particles were higher. As mentioned above, the hydrophilicity and porosity of the PES shells increased after modifying with P(MMA-AA) and GO, so the adsorption rates were higher. Meanwhile, the PES shells also had adsorption sites after the modification, as the MB molecules were adsorbed in the pores of PES shells, the pore size of the PES shells gradually decreased, which led to the decrease of adsorption rates in the following steps.

3.4. Adsorption isotherms

In order to better understand the mechanism of the adsorption process, the effect of initial concentrations on the adsorbed MB amounts to the particles was also investigated by soaking the particles in 10 mL MB solutions with five different initial concentrations. The adsorbed amounts to the particles at various initial concentrations are shown in Fig. 5a. With the increase of the initial concentrations, the adsorbed amounts to the particles enhanced. As is known to all, higher initial concentration of MB resulted in higher contacting probability between the MB molecules and the adsorption sites in the particles, which led to higher adsorbed amounts. However, when the initial concentration reached a point, the adsorbed amounts to the particles slowed down; and it might be attributed to the adsorption saturation of the particles.

Fig. 5 The adsorbed amounts per unit mass of the particles at various initial concentrations. 10 particles are applied in 10 mL MB solution (a); application of the Langmuir adsorption isotherm (b).

The equilibrium sorption isotherm, which could explicate how pollutants interact with adsorbent materials, was of momentous practical significance when designing an adsorption system. In order to investigate the adsorption isotherm of MB, the Langmuir and Freundlich sorption isotherms were used to explore the adsorption process.

3.4.1. Langmuir isotherm. As the most widely used adsorption isotherm, the Langmuir adsorption isotherm was on the basis of several hypotheses: a monolayer adsorption for adsorbates, identical adsorption sites, no interaction between adsorbate molecules and no further adsorption for the adsorbed sites. One of the linear forms of the Langmuir equation is shown as:where q_e is the mass of the MB adsorbed by the unit mass of the particles after the concentration reaches equilibrium (mg g⁻¹); C_e is the concentration of the MB in the solutions at the equilibrium (mg L⁻¹); q_max is the maximal adsorbed amount of the particles (mg g⁻¹); k_L is the Langmuir adsorption constant. A plot of C_e/q_e versus C_e shown in Fig. 5b gives a straight line with the slope of 1/q_max and the intercept 1/(q_maxk_L), and the values are shown in Table 4. All the correlation coefficients (R²) were higher than 0.99, which revealed the adsorption data fitted the Langmuir isotherm very well and the adsorption was monolayer adsorption. The calculated maximal adsorbed amounts of PES8@GO, PES10@GO, PES12@GO, PES/PAA@GO and PES/GO@GO were 352.11, 334.45, 328.95, 363.64 and 425.53 mg g⁻¹, respectively.

Table 4 The parameters of Langmuir isotherm for the adsorption of MB by the particles

Particles	Langmuir
	qmax (mg g^−1)	kL (L mg^−1)	RL^2
PES8@GO	352.11	1.05	0.999
PES10@GO	334.45	1.00	0.999
PES12@GO	328.95	0.84	0.999
PES/PAA@GO	363.64	0.84	0.999
PES/GO@GO	425.53	0.74	0.999

---

3.4.2. Freundlich isotherm. As an empirical equation, the Freundlich equation is employed to describe heterogeneous systems and reversible adsorption ignoring the formation of monolayer. The equation may be expressed in the following form:

ln q_e = ln k_F + (1/n)ln C_e (5)

where k_F is the Freundlich adsorption constant; 1/n is the heterogeneity factor; q_e (mg g⁻¹) and C_e (mg L⁻¹) have the same meanings as those in the Langmuir equation. In this study, all the correlation coefficients (R²) for the Freundlich model were less than 0.90, which indicated that the adsorption processes of the particles did not fit this model (the fitting results were given in ESI, Table S1 and Fig. S3†).

3.5. Comparison of adsorption capacity between GO in aqueous solution and in particles

As mentioned above, GO demonstrated ultra-high adsorption capacity due to the single-layered structure and the large surface area with adsorption sites, yet the application of GO was limited by the complicated recovery process and the potential secondary pollution. Fortunately, this matter was commendably solved by GO-based core–shell particles in this study, for which the GO was enwrapped with PES. However, whether the adsorption capacity of GO was influenced by the PES shell still needed to be studied; and the adsorbed amounts of MB by GO aqueous solution was measured (seen ESI, Fig. S4†). Besides, the average mass of GO in each particle was measured to calculate the adsorption capacity per unit mass of GO in the particles. A brief formula was used for the calculation:where q_GO is the adsorbed amounts of pure GO in each particle (mg g⁻¹); q_max, which was obtained from the calculation of the Langmuir model, is the maximal adsorbed amount of the particle (mg g⁻¹); m_particle is the average mass of each particle; m_GO is the average mass of GO in each particle. As known from Table 5, the pure GO had the highest adsorption capacity definitely; after the enwrapping process, the adsorption capacity decreased slightly. The PES shell could be considered as a barrier to the diffusion of MB molecules, which would reduce the adsorption rate and the adsorption capacity; and it is obvious that the adsorbed amount decreased with increasing the PES concentration in casting solution. In addition, when compared with other adsorbents for the adsorption of MB,^40,46,47 the enwrapped GO still revealed outstanding adsorption capacity.

Table 5 The maximal adsorbed amount of GO aqueous solution and GO in the particles

GO situation	Aqueous solution	PES8@GO	PES10@GO	PES12@GO
Adsorbed amount (mg g^−1)	1136.47	1056.33	1003.35	986.85

---

3.6. Removal of MB by particle column

Removal of pollutants from water by an adsorption column is a practical technique which could be used in industrial field due to its operation convenience. Owing to the good mechanic property, we found that the GO-based core–shell particles could be used to fabricate a column for water purification. Besides, the porous structure in the particles allowed the penetration of water when the pollutants flowed past the column and the tiny volume of the particles enlarged specific surface area for adsorption. In this study, the PES8@GO, PES/GO@GO and PES/PAA@GO were further used to prepare GO-based core–shell particle columns for the adsorption of MB.

As shown in Fig. 6, the color of the shell of the PES8@GO almost unchanged after the adsorption, whereas the color of the shells of the PES/PAA@GO and PES/GO@GO changed obviously. It could be divined from this phenomenon that the PES shell possessed adsorption capacity after the modification. The initial concentration of MB was 250 μmol L⁻¹, and the flow rate of MB was 1 mL min⁻¹. When the MB solutions were separately applied to the columns for the first time, the concentrations of the MB in the eluted solutions were decreased to 134.16 μmol L⁻¹, 109.08 μmol L⁻¹ and 57.70 μmol L⁻¹ for the PES8@GO, PES/PAA@GO and PES/GO@GO, respectively. Once again, it meant that the adsorption rate of the modified particles was faster than that of the unmodified particles. After the eluted solutions were reapplied for another two times, the total removal ratios for the columns were increased to 86.14%, 88.91% and 96.25%, respectively. These results indicated that the GO-based core–shell particles had great potential for further scale-up application in industry for water purification by the particle column.

Fig. 6 Removal of MB by three particle columns: PES8@GO, PES/PAA@GO and PES/GO@GO.

4. Conclusions

In this study, different GO/PES composite particles with core–shell structure were successfully prepared by means of a phase inversion technique, which meant that the complex recovery process for GO was thoroughly eliminated; and the adsorption capacity of GO was not reduced. The adsorption rates of the particles were improved by modifying the PES shells with hydrophilic fillers. It was found that the PES concentration used to prepare shells, contacting time and the initial concentration of MB had great effect on the adsorption capacity. The adsorption experiments revealed that the adsorption process consisted with the pseudo-first-order model, the pseudo-second-order model and the Langmuir isotherm except the Freundlich isotherm. The intra-particle diffusion kinetics suggested that the adsorption process fitted the intra-particle diffusion model well, and the adsorption rate was mainly controlled by the dense PES shell. To further study the application value in industry, a particle column was fabricated, and the results demonstrated its potential application in the fields of wastewater treatment on an industrial scale. Thus, we believe that this facile method of preparing GO/PES composite particles opens a new route for the preparation of particles used for wastewater treatment.

Acknowledgements

This work was financially sponsored by the National Natural Science Foundation of China (No. 51225303 and 51433007), the State Key Laboratory of Polymer Materials Engineering (Grant No. sklpme2015-1-03), and the Sichuan Province Youngth Science and Technology Innovation Team (No. 2015TD0001). We would also thank our laboratory members for their generous help, and gratefully acknowledge the help of Ms Hui Wang, of the Analytical and Testing Center at Sichuan University, for SEM observation.

Notes and references

1. M. Rafatullah, O. Sulaiman, R. Hashim and A. Ahmad, J. Hazard. Mater., 2010, 177, 70–80.
2. U. G. Akpan and B. H. Hameed, J. Hazard. Mater., 2009, 170, 520–529.
3. A. Demirbas, J. Hazard. Mater., 2009, 167, 1–9.
4. X. Zhang, C. Cheng, J. Zhao, L. Ma, S. Sun, C. Zhao, X. Zhang, J. Zhao, L. Ma and S. Sun, Chem. Eng. J., 2013, 215–216, 266–272.
5. Y. Xie, B. Yan, H. Xu, J. Chen, Q. Liu, Y. Deng and H. Zeng, ACS Appl. Mater. Interfaces, 2014, 6, 8845–8852.
6. A. Bhatnagar and M. Sillanpää, Chem. Eng. J., 2010, 157, 277–296.
7. R. Elmoubarki, F. Z. Mahjoubi, H. Tounsadi, J. Moustadraf, M. Abdennouri, A. Zouhri, A. E. Albani and N. Barka, Water Resources & Industry, 2015, 9, 16–29.
8. A. Gil, F. C. C. Assis, S. Albeniz and S. A. Korili, Chem. Eng. J., 2011, 168, 1032–1040.
9. G. Mezohegyi, F. P. V. D. Zee, J. Font, A. Fortuny and A. Fabregat, J. Environ. Manage., 2012, 102, 148–164.
10. H. W. Liang, X. Cao, W. J. Zhang, H. T. Lin, F. Zhou, L. F. Chen and S. H. Yu, Adv. Funct. Mater., 2011, 21, 3851–3858.
11. S. Wang, C. W. Ng, W. Wang, Q. Li and L. Li, J. Chem. Eng. Data, 2012, 57, 1563–1569.
12. J. Yan, Y. Huang, Y. E. Miao, W. T. Weng and T. Liu, J. Hazard. Mater., 2014, 283, 730–739.
13. H. Gao, T. Kan, S. Zhao, Y. Qian, X. Cheng, W. Wu, X. Wang and L. Zheng, J. Hazard. Mater., 2013, 261, 83–90.
14. C. Zhang, P. C. Zhu, L. Tan, J. M. Liu, B. Tan, X. L. Yang and H. B. Xu, Macromolecules, 2015, 48, 8509–8514.
15. K. L. Yeap, T. T. Teng, B. T. Poh, N. Morad and K. E. Lee, Chem. Eng. J., 2014, 243, 305–314.
16. P. S. Suchithra, L. Vazhayal, A. P. Mohamed and S. Ananthakumar, Chem. Eng. J., 2012, 200–202, 589–600.
17. Y. Pan, J. Wang, C. Sun, X. Liu and H. Zhang, J. Hazard. Mater., 2016, 309, 253–255.
18. V. Georgakilas, M. Otyepka, A. B. Bourlinos, V. Chandra, N. Kim, K. C. Kemp, P. Hobza, R. Zboril and K. S. Kim, Chem. Rev., 2012, 112, 6156–6214.
19. D. Chen, H. Feng and J. Li, Chem. Rev., 2012, 112, 6027–6053.
20. X. Yuan, Y. Wang, W. Jian, Z. Chao, T. Qi and X. Rao, Chem. Eng. J., 2013, 221, 204–213.
21. X. J. Hu, Y. G. Liu, H. Wang, A. W. Chen, G. M. Zeng, S. M. Liu, Y. M. Guo, X. Hu, T. T. Li and Y. Q. Wang, Sep. Purif. Technol., 2013, 108, 189–195.
22. X. Jing, W. Li and Y. Zhu, Langmuir, 2012, 28, 8418–8425.
23. W. Zhang, C. Zhou, W. Zhou, A. Lei, Q. Zhang, Q. Wan and B. Zou, Bull. Environ. Contam. Toxicol., 2011, 87, 86–90.
24. Q. Du, J. Sun, Y. Li, X. Yang, X. Wang, Z. Wang and L. Xia, Chem. Eng. J., 2014, 245, 99–106.
25. C. Chong, D. Jie, L. Bei, H. Ai, Z. Xiang, M. Lang, L. Shuang and C. Zhao, J. Hazard. Mater., 2013, 263, 467–478.
26. Y. Shen, Q. Fang and B. Chen, Environ. Sci. Technol., 2015, 49, 67–84.
27. Y. Chen, L. Chen, H. Bai and L. Li, J. Mater. Chem. A, 2013, 1, 1992–2001.
28. L. Gan, S. Shang, E. Hu, C. W. M. Yuen and S. Jiang, Appl. Surf. Sci., 2015, 357, 866–872.
29. Z. Y. Sui, Y. Cui, J. H. Zhu and B. H. Han, ACS Appl. Mater. Interfaces, 2013, 5, 9172–9179.
30. C. He, Polym. Chem., 2015, 6, 7893–7901.
31. G. C. Li, Y. X. Zhao, L. Z. Zhang, M. Gao, Y. Kong and Y. Yang, Colloids Surf., B, 2016, 143, 547–556.
32. C. Cheng, RSC Adv., 2014, 4, 42346–42357.
33. Z. S. Ma, D. G. Liu, Y. Zhu, Z. H. Li, Z. X. Li, H. F. Tian and H. Q. Liu, Carbohydr. Polym., 2016, 144, 230–237.
34. Y. Wang, C. Shen, W. Lou and F. Shentu, Opt. Commun., 2016, 372, 229–234.
35. M. E. A. Ali, L. Wang, X. Wang and X. Feng, Desalination, 2016, 386, 67–76.
36. F. Jin, RSC Adv., 2013, 3, 21394–21397.
37. X. Zhang, RSC Adv., 2016, 5, 100651–100662.
38. C. S. Zhao, J. M. Xue, F. Ran and S. D. Sun, Prog. Mater. Sci., 2013, 58, 76–150.
39. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
40. A. Rodriguez, G. Ovejero, M. Mestanza and J. Garcia, Ind. Eng. Chem. Res., 2010, 49, 3207–3216.
41. H. Li, J. Liao, X. Tao, W. Rui, D. Wang, S. Sun and C. Zhao, Desalination, 2013, 309, 1–10.
42. C. Nie, L. Ma, Y. Xia, C. He, J. Deng, L. Wang, C. Cheng, S. Sun and C. Zhao, J. Membr. Sci., 2015, 475, 455–468.
43. Z. Han, C. Chong, L. Zhang, C. Luo, C. Nie, D. Jie, X. Tao and C. Zhao, Desalination, 2014, 349, 80–93.
44. Q. Hui, C. Sun, H. Chao, D. Wang, C. Chong, S. Nie, S. Sun and C. Zhao, J. Membr. Sci., 2014, 468, 172–183.
45. X. Zhang, Y. Liu, C. Sun, H. Ji, W. Zhao, S. Sun and C. Zhao, RSC Adv., 2015, 5, 100651–100662.
46. F. Liu, S. Chung, G. Oh and T. S. Seo, ACS Appl. Mater. Interfaces, 2012, 4, 922–927.
47. E. N. E. Qada, S. J. Allen and G. M. Walker, Chem. Eng. J., 2006, 124, 103–110.

---

Footnote

† Electronic supplementary information (ESI) available: Preparation and characterization of GO; SEM images of the cross-section of PES@GO particles; Freundlich isotherm; the adsorbed amount of MB by GO aqueous solution. See DOI: 10.1039/c6ra18950d

---