Facile Fabrication of Magnetic Metal-Organic Framework Composites for the Highly Selective Removal of Cationic Dyes

In this work, we show a novel magnetic composite material Fe3O4@HPU-9 (HPU-9 = {[Cd(L)0.5(H2O)](DMA)(CH3CN)}n) (H4L = 1,1′-di(3,5-dicarbonylbenzyl)-2,2′bimidazoline, DMA = N,N-dimethylacetamide) constructed by in situ growth of HPU-9 on Fe3O4, which has excellent absorption of cationic dyes from aqueous solution. The Fe3O4@HPU-9 particle possesses a well-defined core-shell structure consisting of a Fe3O4 core (diameter: 190 nm) and a HPU-9 shell (thickness: 10 nm). In the composite, the HPU-9 shell contributes to the capsulation of cationic dyes through electrostatic attractions between HPU-9 and cationic dyes, while the Fe3O4 core serves as magnetic particle. The maximum absorption capacity of Fe3O4@HPU-9 for R6G was 362.318 mg·g−1. The absorption kinetics data were well described by a psedo-second-order model (R2 > 0.99), and the equilibrium data were also well fitted to Langmuir isotherm model (R2 > 0.99). Our data confirmed that the proposed magnetic composite could be recycled and reused several times without centrifugal separation, making it more convenient, economic and efficient than common adsorbents.


Introduction
Dyes are currently widely used in various industries, such as printing, wool, paper, nylon and silk [1,2]. However, the emission of dyes into the environment has raised widespread public concern relating to water pollution and human health [3][4][5][6]. Numerous conventional porous materials, such as zeolites, polymeric resins, carbon materials and mesoporous silica, have been considered for dye absorption but exhibit weak selectivity toward targeted dyes and are not cost-effective for practical application [7][8][9]. Previous reports have demonstrated that metal-organic frameworks (MOFs) show excellent performances in the recognition capability and selectivity toward a variety of organic dye pollutants because of their tunable chemical functionality, pore microenvironment, structural diversity and high surface areas [10,11]. Principally, the size dimension and property of the pores in the MOFs play important roles in the selective absorption of targeted dyes [12,13]. First, MOFs possessing a relatively large pore size are the necessary prerequisite to selective dye absorption. Second, the interactions between dye molecules and absorption sites on the pore walls of MOFs, which could improve the physical and/or chemical absorption capability of targeted dye molecules, are the guarantee of the success of removing targeted dyes. In this article, we demonstrated that the utilization of a large vacant pore space with anionic binding sites can improve cationic dye uptake through direct electrostatic attractions between adsorbent and adsorbate moieties.
Although MOFs have purely microporous structures that offer high-affinity binding sites for targeted dyes, they are limited by significant defects, such as powder morphology, difficulty of separation and poor stability, which reduce the efficiency and recyclability of adsorbents in practice. To address this problem and maximize the dye absorption capability of MOFs, nanoparticles with unique magnetic properties could be incorporated into MOFs [14,15]. Magnetic separation based on Fe 3 O 4 is a considerably convenient, economic and efficient approach [16,17]. Thus, we proposed a cationic dye absorption method with magnetic anionic MOF composites that can be magnetically separable and are conveniently reusable. To the best of our knowledge, few studies have reported on the use of magnetic MOF composites for dye removal from aqueous solution. Herein, we successfully synthesized a new porous anionic MOF, namely, {[Cd(L) 0.5 (H 2 O)](DMA)(CH 3 CN)} n (HPU-9), with a large 1D channel along the c axis. The anionic channel enables the encapsulation of cationic dyes through host-guest interactions. In addition, a magnetic Fe 3 O 4 @HPU-9 hybrid was prepared through a hydrothermal method by in situ growth of HPU-9 on Fe 3 O 4 . The crystalline structure and composition of Fe 3 O 4 and Fe 3 O 4 @HPU-9 were identified by TEM, SEM, PXRD, IR and ICP characterization. The maximum absorption capacity of Fe 3 O 4 @HPU-9 for R6G was 362.318 mg·g −1 . The absorption kinetics data were well described by a psedo-second-order model (R 2 > 0.99) and the equilibrium data were also well fitted to Langmuir isotherm model (R 2 > 0.99). As a result, the absorption and reusability of the magnetic Fe 3 O 4 @HPU-9 hybrid for dyes suggested that the magnetic hybrid could be employed as a convenient, economic and efficient adsorbent for the treatment of wastewater containing cationic dyes.

Materials and Methods
All chemicals were commercially available and used as purchased. The detailed information is listed in Table S1 in the Supplementary Materials. Infrared (IR) data were recorded on a BRUKER TENSOR 27 spectrophotometer (BRUKER OPTICS, Munich, Bavaria, Germany) with KBr pellets in the region of 400-4000 cm −1 . Elemental analyses (C, H and N) were carried out on a Flash EA 1112 elemental analyzer (Suzhou Orco Metrology instrument Co., Ltd., Suzhou, China). Powder X-ray diffraction (PXRD) patterns were recorded using CuKα radiation on a PANalytical X'Pert PRO diffractometer (PANalytical B.V., Almelo, The Netherlands). Thermogravimetric analysis (TGA) was recorded on a Netzsch STA 449C thermal analyzer (NETZSCH, Selb, Bavaria, Germany) between 30 and 800 • C at a heating rate of 10 • C·min −1 in atmosphere. The UV spectra were recorded on a Purkinje General TU-1800 spectropho-tometer (Beijing Purkinje General Instrument, Beijing, China). The morphologies and microstructures of the as-synthesized samples were characterized by field-emission scanning electron microscopy (SEM) (S-4800, Hitachi, Chiyoda, Japan) and transmission electron microscopy (TEM) (JEM-1200EX, JEOL Ltd., Tokyo, Japan). Surface elements of samples were mapped with electron disperse spectroscopy (EDS) equipped with SEM. anhydrous DMA. The mixture was stirred for 30 min, and 2 mg Fe 3 O 4 was added and then transferred to a 25 mL Teflon liner and kept at 80 • C for 72 h. Subsequently, the reaction mixture was centrifuged, and the composite materials were collected and washed several times with water.
Crystal Data Collection and Refinement. The crystallographic diffraction data for HPU-9 were obtained on a Siemens Smart CCD single-crystal X-ray diffractometer (Germany bruker Ltd., Karlsruhe, Germany) with a graphite monochromatic MoKα radiation (λ = 0.71073 Å) at 293 K. The structure was solved by direct methods using the SHELXS-2014 program of the SHELXTL package and refined on F 2 by full-matrix least-squares techniques with SHELXL-2014. All empirical absorption corrections were applied using the SADABS program. All non-hydrogen atoms in the crystal structure were refined with anisotropic thermal parameters. The crystallographic data and structural refinement parameters of the complex are summarized in Table S2. CCDC 1812922 for HPU-9.

Structural Description
Single crystal X-ray diffraction analysis reveals that HPU-9 crystallizes in the orthorhombic space group Pccn and displays a 3D porous network structure. The asymmetric unit of HPU-9 consists of one crystallographically independent Cd 2+ ion, half H 4 L ligand, as well as one H 2 O coordinated molecule, one CH 3 CN molecule and one DMA guest molecule. The Cd center is six-coordinated by five oxygen atoms and one nitrogen atom in a distorted octahedral [CdO 5 N] geometry. The Cd-O distances are in the range of 2.258-2.377 Å. Each L 4− ligand serves as a µ 6− bridge connecting with six Cd 2+ atoms, in which the four carboxylate groups exhibit the same µ 1 -η 1 :η 1 chelate coordination mode (Figure 1a). Through the inter-connnection of Cd 2+ and L 4− , a two dimensional layer is formed, as shown in Figure 1b. It is interesting to note that there are two different helical chains, left-and right-helical chains (Figure 1d), and due to the existence of the benzene and trizole ring, the two dimensional layers are connected with each other, forming a three-dimensional open framework imparting nanoscale quadrangle, as shown in Figure 1c. As evidenced from the single-crystal X-ray diffraction analyses, the quadrangle window in HPU-9 shows a pore size of ca. 8.3 × 10.9 Å 2 (Figure 1e,f). The PLATON program indicates that the vacant space in HPU-9 is approximately 40.2% (1446.5/3601.0 Å 3 ). The TG curve of HPU-9 is shown in Figure S1. A weight loss of 28.06% (calcd, 28.96%) is detected in the temperature range of 50-180 • C, corresponding to the loss of guest molecules and coordinated water molecules. Finally, the white CdO residue constitutes 26.11% (calcd, 25.36%). The BET surface area of HPU-9 and Fe 3 O 4 @HPU-9 is 312 and 299 m 2 ·g −1 .

The Hybrization of HPU-9 and Fe 3 O 4
Dyes are very widely used in various industries such as cosmetics, printing and paper [18,19]. Dye removal from wastewater has led to tremendous environmental pollution, which has raised public concern [7][8][9]. Nonetheless, defects such as the difficulty of separation exist in the dye removal process when HPU-9 is used as an absorbent. Magnetic hybrid materials used in wastewater treatment are very practical [20][21][22]. Magnetic separation based on the superparamagnetic Fe 3 O 4 was obviously more efficient, economic and convenient [23,24]. Thus, a magnetic Fe 3 O 4 @HPU-9 hybrid was synthesized through the hydrothermal method (Scheme 1). The crystalline structure and composition of Fe 3

The Hybrization of HPU-9 and Fe3O4
Dyes are very widely used in various industries such as cosmetics, printing and paper [18,19]. Dye removal from wastewater has led to tremendous environmental pollution, which has raised public concern [7][8][9]. Nonetheless, defects such as the difficulty of separation exist in the dye removal process when HPU-9 is used as an absorbent. Magnetic hybrid materials used in wastewater treatment are very practical [20][21][22]. Magnetic separation based on the superparamagnetic Fe3O4 was obviously more efficient, economic and convenient [23,24]. Thus, a magnetic Fe3O4@HPU-9 hybrid was synthesized through the hydrothermal method (Scheme 1). The crystalline structure and composition of Fe3O4 and Fe3O4@HPU-9 were identified by TEM, SEM, PXRD, IR and ICP characterization.
The TEM image of Fe3O4@HPU-9 showed that the Fe3O4 particles were wrapped with a HPU-9 layer consisting of numerous small crystals ( Figure 2). The Fe3O4@HPU-9 particles possessed a welldefined core-shell structure consisting of a Fe3O4 core (diameter: 190 nm) and a HPU-9 shell (thickness: 10 nm). The SEM results revealed that the Fe3O4@HPU-9 had spherical morphology ( Figure 3). The elemental mapping of Fe3O4@HPU-9 revealed the distribution of Fe, Cd, O within the structures. In particular, Fe and Cd were almost uniformly distributed in Fe3O4@HPU-9. The ICP result showed that the ratio of Fe:Cd was 1.83:1. The TGA curve of Fe3O4@HPU-9 is also given in Figure S1. The initial weight loss in the temperature range of 50-180 °C is attributed to the release of CH3CN, H2O and DMA (observed, 21.83, calcd, 22.69%). From 400 °C, the framework begins to collapse. The Fe2O3 and CdO residues of 42.33% (calcd, 41.03%) are observed.   The TEM image of Fe 3 O 4 @HPU-9 showed that the Fe 3 O 4 particles were wrapped with a HPU-9 layer consisting of numerous small crystals ( Figure 2). The Fe 3 O 4 @HPU-9 particles possessed a well-defined core-shell structure consisting of a Fe 3 O 4 core (diameter: 190 nm) and a HPU-9 shell (thickness: 10 nm). The SEM results revealed that the Fe 3 O 4 @HPU-9 had spherical morphology ( Figure 3). The elemental mapping of Fe 3 O 4 @HPU-9 revealed the distribution of Fe, Cd, O within the structures. In particular, Fe and Cd were almost uniformly distributed in Fe 3 O 4 @HPU-9. The ICP result showed that the ratio of Fe:Cd was 1.83:1. The TGA curve of Fe 3 O 4 @HPU-9 is also given in Figure S1. The initial weight loss in the temperature range of 50-180 • C is attributed to the release of CH 3    The PXRD patterns of Fe3O4, HPU-9 and Fe3O4@HPU-9 are shown in Figure 4a. The diffraction peaks of Fe3O4 and HPU-9 were observed in Fe3O4@HPU-9, thereby indicating the successful hybridization of Fe3O4@HPU-9. The IR spectra of Fe3O4, HPU-9 and Fe3O4@HPU-9 are also provided in Figure 4b. The carboxylate groups in H4L were observed at 1383 and 1409 cm −1 . All of the characteristics peaks of HPU-9 were also observed in Fe3O4@HPU-9, confirming that the structure of   The PXRD patterns of Fe3O4, HPU-9 and Fe3O4@HPU-9 are shown in Figure 4a. The diffraction peaks of Fe3O4 and HPU-9 were observed in Fe3O4@HPU-9, thereby indicating the successful hybridization of Fe3O4@HPU-9. The IR spectra of Fe3O4, HPU-9 and Fe3O4@HPU-9 are also provided in Figure 4b. The carboxylate groups in H4L were observed at 1383 and 1409 cm −1 . All of the characteristics peaks of HPU-9 were also observed in Fe3O4@HPU-9, confirming that the structure of All of the characteristics peaks of HPU-9 were also observed in Fe 3 O 4 @HPU-9, confirming that the structure of HPU-9 was preserved. Therefore, considering that the results were supported by the experiments, we concluded that the Fe 3 O 4 @HPU-9 composite that consists of Fe 3 O 4 and HPU-9 was successfully synthesized. HPU-9 was preserved. Therefore, considering that the results were supported by the experiments, we concluded that the Fe3O4@HPU-9 composite that consists of Fe3O4 and HPU-9 was successfully synthesized.

Dye Absorption and Separation
Three cationic dyes, namely, MB, RhB and R6G, with different sizes and charges, and three anionic dyes, namely, MO, OG and NGB, with different sizes and charges were selected. The molecular structures of dyes are shown in Figure S2. The selected dyes displayed characteristic peaks. In particular, the main absorption peaks of MB, RhB, R6G, MO, OG and NGB were at 650, 550, 525, 466, 490 and 720 nm, respectively. In brief, 10 mg of freshly prepared Fe3O4@HPU-9 crystals was immersed in the different aqueous solutions of the dyes (4 mg/L, 5 mL) at room temperature. All mixtures were placed in the dark and stirred for 24 h. UV-visible (UV-Vis) spectroscopy was performed to determine the absorption abilities of Fe3O4@HPU-9. As shown in Figure 5, the disappearance absorption peaks of the cationic dyes indicated that Fe3O4@HPU-9 preferred to absorb the cationic dyes (MB, RhB and R6G) over a period of time. By contrast, the absorption of the anionic dyes (OG, MO and NGB) was not detected. So Fe3O4@HPU-9 exhibited different absorption abilities toward different types of organic dyes (The self-changes of the cationic dye solution after 24 h were unchanged shown in Figure S3 Figures 6 and S4, after absorption for 24 h, the absorption peaks of the cationic dyes at their respective peaks were weakened, whereas those of the anionic dyes remain unchanged. It is observed that cationic dyes could be separated from other anionic dyes in aqueous solution by Fe3O4@HPU-9 (Fe3O4 has no selectivity). In addition, the selective absorption and separation of the different cationic dyes (MB, R6G or RhB) from RhB & MB, R6G & MB, R6G & RhB were also studied ( Figure S5). We found that Fe3O4@HPU-9 did not exhibit a significant selectivity toward different cationic dyes. Therefore, we concluded that the distinguishing characteristics of the absorption behavior of Fe3O4@HPU-9 toward dyes were possibly due to the structural traits of Fe3O4@HPU-9 instead of size effect, specifically, the strong electrostatic affinity of the anionic channel of Fe3O4@HPU-9 toward cationic dyes. PXRD confirmed that Fe3O4@HPU-9 almost retained its framework structure during dye absorption (Figure 7), implying that the successful absorption of cationic dyes did not influence the integrity of the crystalline structure of Fe3O4@HPU-9. Therefore, Fe3O4@HPU-9 can be developed as a potential adsorbent for removing cationic dyes in aquatic environments.

Dye Absorption and Separation
Three cationic dyes, namely, MB, RhB and R6G, with different sizes and charges, and three anionic dyes, namely, MO, OG and NGB, with different sizes and charges were selected. The molecular structures of dyes are shown in Figure S2. The selected dyes displayed characteristic peaks. In particular, the main absorption peaks of MB, RhB, R6G, MO, OG and NGB were at 650, 550, 525, 466, 490 and 720 nm, respectively. In brief, 10 mg of freshly prepared Fe 3 O 4 @HPU-9 crystals was immersed in the different aqueous solutions of the dyes (4 mg/L, 5 mL) at room temperature. All mixtures were placed in the dark and stirred for 24 h. UV-visible (UV-Vis) spectroscopy was performed to determine the absorption abilities of Fe 3 O 4 @HPU-9. As shown in Figure 5, the disappearance absorption peaks of the cationic dyes indicated that Fe 3 O 4 @HPU-9 preferred to absorb the cationic dyes (MB, RhB and R6G) over a period of time. By contrast, the absorption of the anionic dyes (OG, MO and NGB) was not detected. So Fe 3 O 4 @HPU-9 exhibited different absorption abilities toward different types of organic dyes (The self-changes of the cationic dye solution after 24 h were unchanged shown in Figure S3 Figure 6 and Figure S4, after absorption for 24 h, the absorption peaks of the cationic dyes at their respective peaks were weakened, whereas those of the anionic dyes remain unchanged. It is observed that cationic dyes could be separated from other anionic dyes in aqueous solution by Fe 3 O 4 @HPU-9 (Fe 3 O 4 has no selectivity). In addition, the selective absorption and separation of the different cationic dyes (MB, R6G or RhB) from RhB & MB, R6G & MB, R6G & RhB were also studied ( Figure S5). We found that Fe 3 O 4 @HPU-9 did not exhibit a significant selectivity toward different cationic dyes. Therefore, we concluded that the distinguishing characteristics of the absorption behavior of Fe 3 O 4 @HPU-9 toward dyes were possibly due to the structural traits of Fe 3 O 4 @HPU-9 instead of size effect, specifically, the strong electrostatic affinity of the anionic channel of Fe 3 O 4 @HPU-9 toward cationic dyes. PXRD confirmed that Fe 3 O 4 @HPU-9 almost retained its framework structure during dye absorption (Figure 7), implying that the successful absorption of cationic dyes did not influence the integrity of the crystalline structure of Fe 3 O 4 @HPU-9. Therefore, Fe 3 O 4 @HPU-9 can be developed as a potential adsorbent for removing cationic dyes in aquatic environments.

Absorption Kinetics and Absorption Isotherms
The absorption experiment was run as follows: 10 mg of Fe3O4@HPU-9 was immersed in 10 mL of R6G aqueous solutions. The data showed that the absorption capacities of the Fe3O4@HPU-9 increased with the increase in the concentration of R6G (20, 50, 100, 200, 300, 500 M). As shown in Figure 8a, the absorption capacity for R6G significantly increased in the initial 18 h, gradually reaching equilibrium. Then, a pseudo-second-order model was explored for further exploration of the absorption kinetics. The constants were calculated by using the following equation [25]: where qt and qe are the amounts of R6G adsorbed at a given time and at equilibrium, respectively, and k2 is the rate constant for pseudo-second-order kinetics of R6G. As can be seen in Table 1 and Figure 8b, the fitting of the experimental results shows that the values obtained from the model fitting

Absorption Kinetics and Absorption Isotherms
The absorption experiment was run as follows: 10 mg of Fe3O4@HPU-9 was immersed in 10 mL of R6G aqueous solutions. The data showed that the absorption capacities of the Fe3O4@HPU-9 increased with the increase in the concentration of R6G (20, 50, 100, 200, 300, 500 M). As shown in Figure 8a, the absorption capacity for R6G significantly increased in the initial 18 h, gradually reaching equilibrium. Then, a pseudo-second-order model was explored for further exploration of the absorption kinetics. The constants were calculated by using the following equation [25]: where qt and qe are the amounts of R6G adsorbed at a given time and at equilibrium, respectively, and k2 is the rate constant for pseudo-second-order kinetics of R6G. As can be seen in Table 1 and Figure 8b, the fitting of the experimental results shows that the values obtained from the model fitting

Absorption Kinetics and Absorption Isotherms
The absorption experiment was run as follows: 10 mg of Fe 3 O 4 @HPU-9 was immersed in 10 mL of R6G aqueous solutions. The data showed that the absorption capacities of the Fe 3 O 4 @HPU-9 increased with the increase in the concentration of R6G (20, 50, 100, 200, 300, 500 M). As shown in Figure 8a, the absorption capacity for R6G significantly increased in the initial 18 h, gradually reaching equilibrium. Then, a pseudo-second-order model was explored for further exploration of the absorption kinetics. The constants were calculated by using the following equation [25]: where q t and q e are the amounts of R6G adsorbed at a given time and at equilibrium, respectively, and k 2 is the rate constant for pseudo-second-order kinetics of R6G. As can be seen in Table 1 and Figure 8b Absorption isotherms are also applied to describe the interaction between R6G and Fe3O4@HPU-9. Figure 9a shows the absorption isotherms of R6G onto Fe3O4 and HPU-9 and Fe3O4@HPU-9. The absorption data were analyzed with the Langmuir equation [26]: where Ce, qe, Qm and KL are the equilibrium concentration of R6G, the equilibrium absorption capacity, the maximum absorption capacity and the Langmuir constant, respectively. Table 2 shows the detailed data obtained with this analysis. The maximum absorption capacity of Fe3O4@HPU-9 for R6G was 362.318 mg·g −1 . Ce/qe plotted against Ce yielded straight lines, as shown in Figures 9b and S6. The correlation coefficients, R 2 , of the Langmuir equation were found to be larger than 0.99, indicating the absorption of R6G follows the Langmuir absorption model. A reusability test was performed to analyse the recyclable and reusable properties of Fe3O4@HPU-9 ( Figure 10). Under room temperature, the regeneration of Fe3O4@HPU-9 was successfully achieved by immersing R6G@Fe3O4@HPU-9 in saturated ethanol and NaCl for 12 h. The removal efficiency during the fourth cycle exhibited an almost similar rate as that of the first absorption cycle, indicating good regeneration and reusability. In the entire recycling process, the solution was poured directly without centrifugal separation. Our data confirmed that the proposed magnetic composite could be recycled and reused several times, making it more convenient, economic and efficient than common adsorbents.  Absorption isotherms are also applied to describe the interaction between R6G and Fe 3 O 4 @HPU-9. Figure 9a shows the absorption isotherms of R6G onto Fe 3 O 4 and HPU-9 and Fe 3 O 4 @HPU-9. The absorption data were analyzed with the Langmuir equation [26]: where C e , q e , Q m and K L are the equilibrium concentration of R6G, the equilibrium absorption capacity, the maximum absorption capacity and the Langmuir constant, respectively. Table 2 shows the detailed data obtained with this analysis. The maximum absorption capacity of Fe 3 O 4 @HPU-9 for R6G was 362.318 mg·g −1 . C e /q e plotted against C e yielded straight lines, as shown in Figure 9b and Figure S6. The correlation coefficients, R 2 , of the Langmuir equation were found to be larger than 0.99, indicating the absorption of R6G follows the Langmuir absorption model. Absorption isotherms are also applied to describe the interaction between R6G and Fe3O4@HPU-9. Figure 9a shows the absorption isotherms of R6G onto Fe3O4 and HPU-9 and Fe3O4@HPU-9. The absorption data were analyzed with the Langmuir equation [26]: where Ce, qe, Qm and KL are the equilibrium concentration of R6G, the equilibrium absorption capacity, the maximum absorption capacity and the Langmuir constant, respectively. Table 2 shows the detailed data obtained with this analysis. The maximum absorption capacity of Fe3O4@HPU-9 for R6G was 362.318 mg·g −1 . Ce/qe plotted against Ce yielded straight lines, as shown in Figures 9b and S6. The correlation coefficients, R 2 , of the Langmuir equation were found to be larger than 0.99, indicating the absorption of R6G follows the Langmuir absorption model. A reusability test was performed to analyse the recyclable and reusable properties of Fe3O4@HPU-9 ( Figure 10). Under room temperature, the regeneration of Fe3O4@HPU-9 was successfully achieved by immersing R6G@Fe3O4@HPU-9 in saturated ethanol and NaCl for 12 h. The removal efficiency during the fourth cycle exhibited an almost similar rate as that of the first absorption cycle, indicating good regeneration and reusability. In the entire recycling process, the solution was poured directly without centrifugal separation. Our data confirmed that the proposed A reusability test was performed to analyse the recyclable and reusable properties of Fe 3 O 4 @HPU-9 ( Figure 10). Under room temperature, the regeneration of Fe 3 O 4 @HPU-9 was successfully achieved by immersing R6G@Fe 3 O 4 @HPU-9 in saturated ethanol and NaCl for 12 h. The removal efficiency during the fourth cycle exhibited an almost similar rate as that of the first absorption cycle, indicating good regeneration and reusability. In the entire recycling process, the solution was poured directly without centrifugal separation. Our data confirmed that the proposed magnetic composite could be recycled and reused several times, making it more convenient, economic and efficient than common adsorbents.

Conclusions
In summary, we successfully synthesized a novel magnetic Fe 3 O 4 @HPU-9 hybrid, which shows excellent absorption of cationic dyes from aqueous solution and could be easily separated and reused several times without degrading the absorption capacity. The Fe 3 O 4 @HPU-9 particle possessed a well-defined core-shell structure consisting of a Fe 3 O 4 core and a HPU-9 shell. In the composites, the anionic HPU-9 shell exhibits the selective absorption of cationic dyes by the utilization of the large pore space with anionic binding sites through direct electrostatic attractions between HPU-9 and cationic dyes, and the Fe 3 O 4 core serves as a magnetic particle. The maximum absorption capacity of Fe 3 O 4 @HPU-9 for R6G was 362.318 mg·g −1 . In the entire recycling process of dye absorption, the superabundant solution of dye was poured directly without centrifugal separation. Our data suggested that the magnetic Fe 3 O 4 @HPU-9 hybrid has great potential to be employed as a convenient, economic and efficient adsorbent for the treatment of wastewater containing cationic dyes.
Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1944/11/5/744/s1, Figure S1: The TG curve of HPU-9 and Fe 3 O 4 @HPU-9, Figure S2: The molecular structures of dyes, Figure S3: The self-changes of cationic dye solution after 24 h, Figure S4: The selective adsorption of cationic dye from the mixtures of cationic and anionic dye solutions by Fe 3 O 4 @HPU-9, Figure S5: The adsorption competition between two kinds of cationic dyes by Fe 3 O 4 @HPU-9., Figure S6: Langmuir plots of the isotherms for R6G adsorption onto Fe 3 O 4 and HPU-9. Table S1: The information of regents, Table S2: Crystal data and structure refinement for HPU-9 a .