Hypercrosslinked porous polymers hybridized with graphene oxide for water treatment: dye adsorption and degradation

Hypercrosslinked porous polymer hybridized graphene oxide with polymeric high internal phase emulsions (polyHIPEs/GO) were designed as versatile composites for water treatment. Morphologies, chemical composition and thermal stability of the composites were characterized by SEM, FTIR, XPS, XRD and TGA. Tunable adsorption properties and enhanced visible-light photocatalysis towards organic dyes were achieved by the manipulation of functional groups and the inclusion of Ag3PO4, respectively. The adsorption capacity of polyHIPEs/GO towards cationic methyl blue (MB) and rhodamine B (RB) is 1250.3 and 1054.1 μg g−1, respectively. Aminated polyHIPEs/GO (polyHIPEs(NH2)/GO) possesses an adsorption capacity of 1967.3 μg g−1 to anionic eosin Y (EY). The tandem columns of polyHIPEs(NH2)/GO and polyHIPEs/GO can successively and selectively remove the cationic and anionic dyes in a mixed dye solution. Furthermore, enhanced photodegradation ability was obtained after GO reduction and Ag3PO4 addition on polyHIPEs(NH2)/GO. Results show that 3.5 × 10−5 M of MB, RB and EY can be completely photodegraded by 20 mg of the novel photocatalyst within 20, 40 and 35 min, respectively. This work demonstrates that polyHIPEs/GO exhibits tunable properties for multiply progressive applications in water treatment and catalysis.


Introduction
Graphene has attracted signicant attention in many disciplines due to its impressive surface area, extreme mechanical strength, high thermal and high electron mobility. 1,2 Assembling 2D graphene derivatives into macroscopic 3D structures is an essential step to expand their practical applications in oil or contaminant treatment, 3-5 catalysis, 6,7 sensing 8-10 and energy storage. 3,11 However, many 3D macroporous graphene materials collapse easily, and the assembly of a robust structure is still a challenge. 2 In recent years, numerous attempts have been made to incorporate graphene oxide (GO) and/or reduced graphene oxide (RGO) into functional polymers for mechanical enhancement. Besides, the resultant graphene-polymer hybrids not only retain some physiochemical properties of the individual components, but also are endowed with some novel characters that are different from those individual components. On account of this, GO-PVP, GO-PVA, GO-PEO, etc. have been reported. [12][13][14][15] However, the preparation of these GO-polymer 3D networks relies on high quantities of GO matrix, which results in high fabricating costs.
As an effective method to fabricate macroporous polymers, high internal phase emulsions (HIPEs) have been intensively exploited. 16 Polymers obtained from HIPE templates which are known as polyHIPEs are oen endowed with high porosity, good permeability, and functional group tenability. 17,18 Fabricating GO hybridized polyHIPEs (polyHIPEs/GO) is seemingly a great option to improve the structural stability of macropores and to impart many organic groups to the resultant polymers. In addition, the preparation of polyHIPEs/GO can greatly reduce the GO quantity used, which is cost-effective. But surprisingly, polyHIPEs/GO has rarely been exploited in the past years.
In this work, we rstly report the synthetic and surface manipulation strategies of polyHIPEs/GO. Then, we study the morphologies, chemical composition and thermal stability of the polymers using scanning electron microscope (SEM), Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD) and thermo gravimetric analysis (TGA). Furthermore, we show that the asprepared polyHIPEs/GO is useful in dye adsorption and photodegradation.

GO preparation and modication
Graphite oxides were prepared according to the previous modied Hummer's method. 19 Aqueous GO suspension (5.0 mg mL À1 ) was obtained by ultrasonicating the graphite oxide dispersion for 1 h. Then, 667 mg PVP was added into 20 mL of GO suspension, followed by magnetic stirring for 12 h at room temperature, 20 then centrifuged at 15 000 rpm for 20 min to remove the unbound PVP. The collected slurry was redispersed in DI water with a nal volume of 20 mL.

Synthesis of polyHIPEs/GO
An oil phase consisting of 400 mL of EHA, 300 mL of DVB, 15 vol% (with respect to the whole volume of oil phase) of Span 80, and 1.5 wt% (with respect to the total mass of EHA and DVB) of initiator AIBN was added to a 3.2 mL of aqueous phase containing 100 mg of AAm and 2.0 mL of PVP-GO in a 10 mL polypropylene centrifuge tube. The mixtures were emulsied with an MS-3B homogenizer (IKA, Germany) at 3000 rpm for 5 min to form HIPEs (80% internal phase). HIPEs with identical internal volume ratio but different AAm, DVB, and PVP-GO quantity were prepared as described in Table S1. The HIPEs were polymerized at 70 C for 15 h to yield polyHIPEs/GO porous monoliths. Aer cooled to room temperature, the unreacted components were eliminated with ethanol and DI water, then the polymers were lyophilized at À50 C and <20 Pa for 24 h.

Amination and Ag 3 PO 4 decoration on polyHIPEs/GO
Hoffman reaction 21 was used for the transformation of amide groups to primary amine groups. Briey, 1.0 g polyHIPEs/GO monolith was added to 10 mL ice water. Subsequently, 2.0 mL of NaOH solution (1 M) and 2.0 mL of NaClO solution (5.5% effective chlorine) were added. The reaction was carried out at 0 C for 6.5 h, then heated at 70 C for 1.5 h with gently shaking. Aer the reaction, the excess NaOH and NaClO were eliminated with DI water, and the products were lyophilized. The obtained aminated polyHIPEs/GO is denoted as to polyHIPEs (NH 2 ) /GO. The monolithic polyHIPEs (NH 2 ) /GO was crushed into powders and the powders passing through the sieve (100 meshes per cm 2 ) were collected. The collected powders (50 mg) were coated with additional GO (5.0 mg) via self-assembly in 10 mL of water, then the GO was reduced by AA (50 mg) in a 25 mL Teon-lined stainless steel autoclave at 95 C for 3 h. 22 Aer the reaction, the autoclave was le to cool naturally to room temperature, then washed with DI water and collected by centrifugation. The obtained polyHIPEs (NH 2 ) /RGO was dispersed in 20 mL of water, and AgNO 3 aqueous solution (2.0 mL, 0.6 M) was added dropwise with magnetic stirring. Aer the addition, the suspension was kept stirring for further 12 h to ensure the adsorption of Ag + on the surface of polyHIPEs (NH 2 ) / RGO. Then Na 2 HPO 4 aqueous solution (2.0 mL, 0.2 M) was added dropwise to the mixture, and the mixture was kept stirring for 30 min. 23 PolyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 composites were obtained by collecting and freeze drying the centrifugal precipitates.

Characterizations
Atomic force microscope (AFM) images of PVP-GO were obtained using a Multimode 8 in the tapping mode. The PVP-GO sample was dispersed in water and spin coating onto freshly cleaved mica substrates before the test. FTIR spectra were recorded in KBr pellets using an IS10 FTIR spectrometer (Thermo Fisher Scientic Co., USA). Surface morphologies of the samples were observed by a SU5000 eld emission SEM (Hitachi Ltd, Japan). Powder XRD spectra were recorded using a PANalytical X 0 Pert 3 Powder diffractometer with Cu Ka radiation at 40 kV and 40 mA, and a scanning rate of 5 (2q)/min from 10 to 80 . TGA was carried out using TA Instruments (SDT Q 600) at a heating rate of 10 C min À1 from 25 C to 600 C in nitrogen atmosphere. XPS spectra were obtained with an Axis Ultra DLD (Kratos Ltd, U.K.) paired with a monochromatic Al Ka X-ray source (1486.6 eV).

Dye adsorption and desorption
Two cationic dyes (MB and RB) and one anionic dye (EY) were employed to investigate the adsorption behaviors of the polyHIPEs/GO and polyHIPEs (NH 2 ) /GO. Typically, the monolithic sorbent (0.1 g) was added into aqueous dye solution (20 mL of 10 mg mL À1 ), followed by gently shaking at room temperature. At predetermined time intervals, the dye concentration remaining in the solution was measured using a TU-1901 UV-Vis spectrophotometer at the maximum absorbance of each dye (MB: 664 nm; RB: 557 nm; EY: 515 nm). The adsorption amount at time t, q t (mg g À1 ), was calculated using the following equation: where V is the volume of dye solution (mL); C 0 and C t (mg mL À1 ) are the dye concentration initially and at time t, respectively; and m stands for the mass of the sorbent (g).
A mixed dye solution (50 mL) containing 10 mg mL À1 of MB, RB and EY was successively passed through a polyHIPEs (NH 2 ) / GO column and two polyHIPEs/GO columns under different pH values (Fig. 1). The columns used in all processes were 0.2 g in weight and $4 mm in length, and the ltration rate is controlled at 0.1 mL min À1 . The pH of the solution was adjusted using 0.1 MHCl and NaOH. UV-Vis spectra of the solution at different ltration stages were all recorded.

Photocatalytic study
A 350 W Xe lamp (XPA-4 Photoreactor) equipped with a cutoff lter (l > 420 nm) was used as irradiation source. Dye solutions (20 mL, 3.5 Â 10 À5 M) containing 20 mg of photocatalysts were put in a cylindrical glass beaker and stirred in the dark to reach absorption-desorption equilibrium. Next, Xe lamp was turned on to start the photocatalytic reaction. At regular time intervals, Xe lamp was turned off and 4 mL of the solution was taken out and centrifuged to separate the photocatalyst. Then UV-Vis spectra of the supernatant were recorded. Aer that, the solution was pour back in the beaker. The photocatalytic reaction and supernatant analysis processes were repeated until the dyes were degraded completely. The degradation efficiency was evaluated by C t /C 0 . Here, C t and C 0 are the concentration (M) of dyes at time t (min) and initially.

Characterization of polymers and composites
The dispersibility and sheet thickness of PVP-GO were collected using atomic force microscopy. As shown in Fig. S1, † PVP-GO sheets remain good dispersibility in water, and the thickness is less than 1 nm. The sheet edge tends to curl up, causing the signicant increase in vertical thickness. Morphologies of the HIPE-based polymers and Ag 3 PO 4 -based composites were observed by SEM. Fig. 2a and b reveal the open-cell porous structure of polyHIPEs/GO and polyHIPEs (NH 2 ) /GO. Spherical voids with tens of microns in diameter are derived from the emulsion droplets. Many circular windows existing in the void surface interconnect the adjacent voids. Besides, a myriad of nanoscale to submicroscale grooves are also found in the wall (Fig. S2 †). Because the relatively low PVP-GO content and the non-transparency of the polyHIPEs matrix, it is unfeasible to directly observe the distribution of PVP-GO sheets under SEM or TEM, unless the PVP-GO sheets are aggregated on the surface. According to Fig. 2a and b, no aggregated PVP-GO sheets can be seen, we can presume that the amphiphilicity of PVP-GO is uniformly located at the surface between oil phase and aqueous phase of an emulsion. TGA results (Fig. 2e) reveal that the incorporation of GO can apparently enhance the thermal stability of the polymers. The thermal degradation of polyHIPEs and polyHIPEs/GO happens mainly in the range of 230-430 C and 350-480 C, respectively. This in turn demonstrates our previous presumption, because GO can act as an excellent barrier to protect the inner polymer matrix from thermal degradation before 350 C only when GO is uniformly located at the surface. 24 The amination further increases the degradation temperature (37% of the weight for polyHIPEs (NH 2 ) / GO is remained even at 600 C), revealing the higher thermal stability of amines to amides. In Fig. 2c, as the RGO content is increased (compared with the GO content in Fig. 2a and b), the RGO layer is observable. Fig. 2c clearly shows the hierarchical structure of polyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 . The RGO layer avoids the direct contact of polyHIPEs (NH 2 ) and Ag 3 PO 4 . The particle size of Ag 3 PO 4 in polyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 hybrids has a distinct decrease compared with that of bared Ag 3 PO 4 (Fig. S3 †). This phenomenon is consistent with the precious report which revealed that GO has an obvious effect on the size of Ag 3 PO 4 particles. 23 The XRD pattern (Fig. 2f) for polyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 clearly shows all characteristic diffraction peaks that could be readily indexed to the bodycentered cubic structure of Ag 3 PO 4 (JCPDS no. 06-0505), demonstrating the same structure of Ag 3 PO 4 in these two samples despite their different particle sizes.
In Fig. 2d, the FTIR spectrum of GO shows absorption bands at 1722 cm À1 (n OC]O ), 1400 cm À1 (n C-OH ), and 1090 cm À1 (n C-O-C ), depicting the existence of the carboxyl, hydroxyl and epoxy groups. 25 Absorption peaks at 2970 cm À1 (n C-H ), 1663 cm À1 (n NC]O ) and 1289 cm À1 (n N-C ) 26 in the FTIR spectrum of PVP-GO conrms the PVP chains were successfully modied onto the GO sheets. According to the change in relative peak intensity at 1734 cm À1 (n OC]O ) and 1667 cm À1 (n NC]O ) in FTIR spectra of polyHIPEs and polyHIPEs/GO, we can presume the presence of PVP-GO in the polymer matrix. In the FTIR spectrum of polyHIPEs (NH 2 ) /GO, the decreased absorption at 1667 cm À1 (n NC]O ) and the appeared absorption at 1562 cm À1 (d N-H ) proves that amides have translated to amines. 21 The absorption peak corresponded to PO 4 3À (1076 cm À1 ) in the FTIR spectrum of polyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 demonstrates the presence of Ag 3 PO 4 in the hybrids. 27 To prove the GO has been reduced in polyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 , high resolution C1s XPS spectra of polyHIPEs (NH 2 ) /GO and polyHIPEs (NH 2 ) /RGO/ Ag 3 PO 4 samples were collected ( Fig. 2g and h). Peak intensity corresponding to carbons singly bonded to epoxy/hydroxyls (C-O, 286.5 eV) and carbons in carboxyl/ester groups (O-C]O, 288.9 eV) 22 in the C1s XPS spectrum of polyHIPEs (NH 2 ) /RGO/ Ag 3 PO 4 is decreased compared with that in polyHIPEs (NH 2 ) /GO, conrming the reduction of GO. The remaining intensity at 286.5 and 288.9 eV is ascribed to the C-O and O-C]O from the polymer matrix that cannot be reduced by ascorbic acid. The reduction of oxygen-containing groups on GO sheets forms C] C, leading to a slightly broadening to higher binding energy in the peak associated with C-C and C]C.

Dye adsorption
Dye removal behaviors of monolithic polyHIPEs/GO and polyHIPEs (NH 2 ) /GO were investigated using two cationic dyes (MB and RB) and an anionic dye (EY) as models. Since the adsorption of dyes on these sorbents is expected to have great connections with surface charge, effect of pH on the adsorption capacity was rstly investigated. As shown in Fig. 3a, the adsorption of MB and RB on polyHIPEs (without any GO) is rather weak and varies little as the pH is increased. Differently, the adsorption of EY on polyHIPEs (NH 2 ) decreases when the pH is increased from 1.0 to 5.0, then remains a constant. Aer GO is combined with polyHIPEs or polyHIPEs (NH 2 ) , the adsorption capacity to these dyes enhances signicantly. The adsorption of cationic MB on polyHIPEs/GO enhances with the increase in pH value, and the maximum adsorption amount is obtained when the pH $ 7.0, because the deprotonation of carboxyl and hydroxyl groups trigger the electrostatic interactions (Fig. S4 †). Differently, the maximum adsorption of RB appears at the pH range of 3.0-5.0. This is because the highest deprotonation of carboxyl groups in RB molecules at high pH range inversely leads to the weak positive charge of the whole molecule, weakening the electrostatic interactions. Strongest electrostatic interaction between RB molecule and polyHIPEs/GO is achieved under weakly acidic condition due to carboxyl groups are partially deprotonated. When we use polyHIPEs (NH 2 ) /GO to absorb the anionic EY, the highest adsorption amount is achieved when the pH # 3.0, where amine, hydroxyl and epoxy on the sorbent are protonated, leading to a strong electrostatic interaction with the anionic EY. Under the optimized pH, the maximum adsorption amount of polyHIPEs (NH 2 ) /GO to MB and RB is signicantly lower than that of polyHIPEs/GO, while the maximum adsorption amount of polyHIPEs (NH 2 ) /GO to EY is much higher than that of polyHIPEs/GO (Fig. 3b), well demonstrating the reasonability of employing polyHIPEs/GO and polyHIPEs (NH 2 ) /GO for the adsorption of cationic dyes and anionic dyes, respectively. To understand the adsorption mechanism in depth, the quantity of AAm, DVB and GO were investigate. It was found that the adsorption capacity of polyHIPEs (NH 2 ) /GO to EY heightens with the ascendent quantity of AAm, however, the increased AAm quantity has less effects on the adsorption capacity of polyHIPEs/GO to MB and RB (Fig. 4a). This prove that the amide groups have little effects on the adsorption of cationic dyes. When amide groups are transformed to amine groups, the protonated amine groups contribute to the electrostatic attraction with the anionic EY. The increase in DVB quantity ends with higher adsorption capacity of both polyHIPEs/GO and polyHIPEs (NH 2 ) /GO (Fig. 4b), revealing the pp interaction adsorption mechanism. Increasing in GO quantity can enhance both the electrostatic interaction and p-p interaction, which is supported by the result that the adsorption amount of MB, RB, and EY all ascends when the GO quantity rises from 0 to 10 mg (Fig. 4c). Nonetheless, further increase in   GO quantity cannot obtain higher adsorption capacity probably due to the aggregation of GO and some of the GO sheets are unable to show their functions.
The dye adsorption on polyHIPEs/GO and polyHIPEs (NH 2 ) /GO follows the pseudo-second-order kinetic model (Fig. S5 †). The saturated adsorption capacity (q e ) of polyHIPEs/GO to MB and RB, polyHIPEs (NH 2 ) /GO to EY are calculated to be 1250.3, 1054.1, 1967.3 mg g À1 , respectively. The adsorption rate constants (k) to MB, RB and EY are 2.30 Â 10 À4 , 2.38 Â 10 À4 and 1.49 Â 10 À4 (Table S2). In addition, the polymer sorbents adsorbed with dyes are ease of releasing the dyes in ethanol (desorption efficiencies > 87%), and thus possess good cycling performance (Fig. S6 † and S7 †). As pointed in Table S3, polyHIPEs/GO and polyHIPEs (NH 2 ) /GO exhibit superiorities including the broadspectrum adsorption ability, simple synthesis procedure, lowcost of the monolithic sorbent and good recycling performance when compared with the powder-like, aerogel, or hydrogel sorbents. The relatively low adsorption capacity of the as-prepared sorbents probably because the higher bulk density compared with the light weight aerogel and GO rich sorbents. When the adsorption capacity is evaluated by the dye quantity adsorbed in per gram sorbent, the light weight sorbent possessing a much larger volume benets a higher uptake of dyes. To address the low adsorption capacity, continuous research will be carried out in the future.
The removal of dyes from water were further studied using the polyHIPEs/GO and polyHIPEs (NH 2 ) /GO as column packing. As illustrated in Fig. 1, a mixed solution containing MB, RB and EY was successively passed through the polyHIPEs (NH 2 ) /GO column at pH 3.0, polyHIPEs/GO column at pH 4.0, and polyHIPEs/GO column at pH 7.0. Fig. 5 showed that 77% of EY, 55% of RB and 5% of MB were ltered by polyHIPEs (NH 2 ) /GO column at pH 3.0. Subsequently, the remaining 23% of EY, 45% of RB and 31% MB were further ltered aer passing the polyHIPEs/GO column at pH 4.0. Finally, the rest of 63% of MB were eliminated by polyHIPEs/GO column at pH 7.0. This result suggests that various dyes in the polluted water can be selective ltered via the tandem use of polyHIPEs (NH 2 ) /GO and polyHIPEs/GO columns. Moreover, the polluted solution can recover to neutral aer such treatment.
It is known that the bandgap of Ag 3 PO 4 semiconductor is 2.45 eV (conduction band, CB: +0.45 eV; valence band, VB: +2.9 eV) 33 The possible photodegradation processes are illustrated in   ow of electron-hole pairs. (3) Once the photogenerated electrons are captured by O 2 molecules, superoxide radicals (O 2 À c) are produced. At the same time, hydroxyl radicals ($OH) generates from the reaction of H 2 O and the active holes. 23,34 (4) Organic dyes are attracted to the surface of polyHIPEs (NH 2 ) /RGO/Ag 3 PO 4 due to the good adsorption behavior of the polymer, which accelerates the reaction of dye molecules and radicals. (5) By continuously working in the aforementioned manners, dye molecules are nally degraded into CO 2 , H 2 O and other small molecules. 30,35

Conclusions
We have successfully prepared a versatile polyHIPEs/GO porous polymer with tunable properties for multiply progressive applications. This presented polyHIPEs/GO shows good adsorption performance to cationic dyes. Aer simply tuning the surface groups, the resultant polyHIPEs (NH 2 ) /GO exhibits improved adsorption properties to anionic dyes. The adsorption mechanisms, adsorption kinetics, and cycling performance of these two sorbents have been systematically investigated and discussed. Furthermore, we reported a novel polyHIPEs (NH 2 ) / RGO/Ag 3 PO 4 with enhanced visible-light photocatalytic activity by RGO coating and Ag 3 PO 4 decoration on the surface of polyHIPEs (NH 2 ) /GO. Our work suggests that the new polyHIPEs/ GO porous polymer can be used as sorbents, lters, and photocatalysts.

Conflicts of interest
There are no conicts to declare.