zeolitic imidazolate framework self-assembled nanocomposite

Applied Surface Science 361 (2016) 114–121

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Highly efficient removal of Malachite green from water by a magnetic reduced graphene oxide/zeolitic imidazolate framework self-assembled nanocomposite Kun-Yi Andrew Lin ∗ , Wei-Der Lee Department of Environmental Engineering, National Chung Hsing University, 250 Kuo-Kuang Road, Civil Eng Build, Taichung 402, Taiwan, ROC

a r t i c l e

i n f o

Article history: Received 12 August 2015 Received in revised form 19 October 2015 Accepted 10 November 2015 Available online 29 November 2015 Keywords: Zeolitic imidazolate frameworks (ZIFs) Magnetic reduced graphene oxide Malachite green

a b s t r a c t Compared to the relatively low adsorption capacities of conventional adsorbents for Malachite Green (MG) (i.e., ∼500 mg g−1 ), zeolitic imidazolate framework (ZIF) appears to be a promising adsorbent considering its significantly high adsorption capacity (i.e., >2000 mg g−1 ). Nevertheless, using such a nano-scale ZIF material for adsorption may lead to secondary contamination from the release of nanomaterials to the environment. Thus, ZIF has to be recovered conveniently to prevent the secondary contamination and facilitate the separation of adsorbent from water after adsorption. To this end, in this study ZIF nanocrystals were loaded on the sheet-like magnetic reduced graphene oxide (MRGO) to form a self-assembled MRGO/ZIF. The self-assembly of MRGO/ZIF was achieved possibly via the electrostatic attraction and the ␲–␲ stacking interaction between MRGO and ZIF. The resultant MRGO/ZIF exhibited an ultra-high adsorption capacity for MG (∼3000 mg g−1 ). The adsorption kinetics, isotherm, activation and thermodynamics were also determined. Other factors affecting the adsorption were examined including temperature, pH and co-existing ions/compound. To demonstrate that MRGO/ZIF can be recovered and reused, a multiplecycle of MG adsorption using the regenerated MRGO/ZIF was revealed and the recyclability remained highly efficient and stable. The highly-effective, recoverable and re-usable features enable MRGO/ZIF a promising adsorbent to remove MG from water. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Malachite green (MG) is a common dye in textile industry [1]. MG can be also used as a disinfectant and an antifungal agent, particularly, for aquaculture [2,3]. The exposure to MG, however, is considered to cause carcinogenesis, mutagenesis, teratogenesis and respiratory diseases [3]; thus it is essential to remove MG from water in order to avoid the exposure to MG. Up to date, MG can be removed from water via several methods including coagulation/flocculation [4], biological decolorization [5,6] and adsorption [7–11]. Coagulation/flocculation, nevertheless, is ineffective to treat MG-containing wastewater because of high solubility of MG in water [7]. On the other hand, the biological decolorization typically requires long operation time [12] and dyes can be resistant to biodegradation [13]. Considering these issues, adsorption appears to be a more feasible method to remove

∗ Corresponding author. Tel.: +886 4 22854709. E-mail address: [email protected] (K.-Y.A. Lin). http://dx.doi.org/10.1016/j.apsusc.2015.11.108 0169-4332/© 2015 Elsevier B.V. All rights reserved.

MG because adsorption process is easy to implement and scale up with low initial cost [14,15]. Therefore, many adsorbents have been reported to remove MG from water and most of them are wastederived activated carbon [7–11], chemically modified biomass [16], ordered mesoporous carbon [17] and graphene oxide [18]. These conventional materials generally exhibited quite limited adsorption capacities (i.e., ∼500 mg g−1 ). Recently, a novel inorganic-organic hybrid material, zeolitic imidazolate framework (ZIF)-67, has been proposed as an adsorbent to remove MG from water and ZIF showed significantly high adsorption capacity for MG (i.e., >2000 mg g−1 ) [19]. Although ZIF seems to be a promising adsorbent to remove MG, such a nanoscale material has to be recovered after the adsorption process to avoid its release to the environment. However, to our best knowledge, almost no ZIF with recoverability has been developed for the MG adsorption. Thus, this present study proposes to prepare a highly efficient and recoverable adsorbent incorporating with ZIF. To load ZIF on a controllable substrate, we particularly adopted a magnetic reduced graphene oxide (MRGO) as a support. MRGO was selected because it can be prepared from a one-pot synthesis

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

hydrothermally [20] and its sheet-like structure can provide large planar surfaces to load ZIF. Via a simple mixing of MRGO and ZIF, a self-assembled MRGO/ZIF nanocomposite can be obtained, which can be magnetically controllable, allowing MRGO/ZIF to be recovered conveniently. Characteristics of the as-prepared MRGO/ZIF were determined, including morphology, surface charge, crystalline structure, surface chemistry and conjugated carbon bonding. To examine the magnetic property of MRGO/ZIF, a magnetometer was also used, while thermogravimetric analyzer (TGA) was adopted to reveal its thermal decomposition behaviors. The MG adsorption kinetics and isotherm using MRGO/ZIF were measured and analyzed by theoretical models. Adsorption activation energy and thermodynamics were also determined to probe in the adsorption mechanism. Factors affecting the MG adsorption were also investigated, such as temperature, pH and co-existing ions/compounds. Recyclability of MRGO/ZIF for the MG adsorption was examined and a multiplecycle of MG adsorption using the regenerated MRGO/ZIF was demonstrated.

2. Experimental 2.1. Materials All chemicals used in this study were purchased from commercial chemical suppliers. Graphite was from Showa Chemicals (Japan). 2methylimidazole (2-MIM) and cetyltrimethylammonium bromide (CTAB) were from Acros Organics (USA). Malachite green dye and cobalt nitrate were obtained from Choneye Pure Chemicals (Taiwan). Iron sulfate (Fe(SO4 )) and glucose were from

115

Sigma-Aldrich (USA). Deionized (DI) water was prepared to exhibit less than 18 MOhm-cm. 2.2. Synthesis and characterization of MRGO/ZIF Preparation of MRGO/ZIF can be illustrated as Fig. 1a. First, MRGO was synthesized based on the reported protocol [20]. In brief, GO (100 mg), prepared according to Hummers’ method [21], was added to 50 mL of DI water and sonicated for 1 h, followed by addition of 100 mg of glucose and 1 mL of NH4 OH solution. On the other hand, 100 mg of Fe(SO4 ) was added in 5 mL of DI water, to which 20 mg of NaOH in 5 mL of ethanol was subsequently added. The resultant GO suspension was mixed with the ferrous solution for 30 min and then the mixture was transferred to a Teflon-lined autoclave and heated at 180 ◦ C for 12 h. The precipitate was collected, washed with DI water and dried at 65 ◦ C for 12 h to obtain MRGO. To prepare MRGO/ZIF, 0.36 g of Co(NO3 )2 and 1.32 g of 2MIM were added to 150 mL of DI water and stirred at ambient temperature for 2 h to obtain a ZIF (i.e., ZIF-67) suspension. Next, 1.6 g of MRGO was sonicated in 200 mL of DI water for 30 min and the resultant MRGO suspension was added to the ZIF suspension and the resultant mixture was stirred for 12 h at ambient temperature. The self-assembled nanocomposite was collected using a permanent magnet, and then washed and dried to yield the final product, MRGO/ZIF. MRGO/ZIF was first characterized using a TEM (JEOL JEM-2010, Japan) to reveal its morphology. Surface charges of MRGO/ZIF and precursors were measured in water by a zetasizer (NanoZS, Malvern Instruments Ltd, Malvern, UK). Crystalline structures of MRGO/ZIF and precursors were determined using an X-ray diffractometer (BRUKER D8 DISCOVER). IR spectra were obtained by an infrared spectrometer (Jasco 4100, Japan). Conjugated

Fig. 1. MRGO/ZIF: (a) synthesis scheme, (b) a TEM image of the precursor, ZIF-67 and (c) a TEM image of MRGO/ZIF (scale bar = 500 nm).

116

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

carbon bonding of MRGO/ZIF was analyzed by a Raman spectrometer (Tokyo Instruments Inc. Nano-finder, Japan). Thermogravimetric analyses (TGA) of MRGO/ZIF and precursors were determined using a thermogravimetric analyzer (i1000, ISI, USA) at a heating rate of 20 ◦ C min−1 from 25 to 800 ◦ C in nitrogen. Saturation magnetization of MRGO/ZIF was obtained by a superconducting quantum interference device (SQUID) vibrating sample magnetometer (Quantum Design MPMS SQUID VSM, USA) at 27 ◦ C. 2.3. Removal of MG from water using MRGO/ZIF The removal of MG by MRGO/ZIF was evaluated by batchtype adsorption experiments. In a typical experiment, 5 mg of MRGO/ZIF was added to a glass vial containing 20 mL of MG solution with a concentration of 600 mg L−1 . The vial was then placed on a temperature-controllable orbital shaker at 300 rpm to increase contact between MG and MRGO/ZIF. After a certain mixing time t, the vial was withdrawn and MRGO/ZIF was immediately separated from the solution by a permanent magnet. The remaining MG concentration (Ct ) of the solution was analyzed by a UV–vis photospectrometer (Chrom-Tech CT-2000, Taiwan) at 633 nm. The adsorption capacity, qt (mg g−1 ), of MRGO/ZIF at time t was calculated by the following equation: qt =

V (C0 − Ct ) M

(1)

where M (g) is the amount of MRGO/ZIF added in the MG solution and V (L) means the volume of the MG solution. At adsorption equilibrium, the adsorption capacity was represented as qe (mg g−1 ). To obtain adsorption isotherm data, a fixed amount of MRGO/ZIF (5 mg) was added to MG solutions (20 mL) with various concentrations, ranging from 200 to 600 mg L−1 at a constant temperature. Experiments of adsorption isotherm were conducted for 3 h to achieve the adsorption equilibrium. In this study, adsorption kinetics and adsorption were performed at three different temperatures (i.e., 20, 30 and 40 ◦ C) to investigate effect of temperature and to obtain thermodynamic parameters. Considering that color of MG changes spontaneously under alkaline condition, we particularly examined effect of low pH on the removal efficiency by adjusting initial pH of MG solutions using 0.1 M of HCl solution. Effect of co-existing compounds was also investigated by adding NaCl, NH4 Cl and a cationic surfactant, CTAB, to MG solutions. Recyclability of MRGO/ZIF for the removal of MG was also evaluated. The used MRGO/ZIF was regenerated by washing with ethanol until no green color was detected in the ethanol solution and the regenerated MRGO/ZIF was then used to adsorb MG. Adsorption experiments were performed in duplicate and repeated at least twice.

3. Results and discussion 3.1. Characterization of MRGO/ZIF MRGO/ZIF consists of ZIF nanocrystals and MRGO sheets. Fig. 1b shows the morphology of ZIF nanocrystals with a size range of 100–500 nm. It has been indicated that when cobalt nitrate salt was used to synthesize ZIFs, the resulting size should be smaller than 1000 nm as observed in the present study [22]. In addition, the shape of ZIF nanocrystals was more close to chamfered-cube, which is commonly found in ZIF-67 comprising 2-MIM [23]. Once ZIF was mixed with MRGO, the resultant composite can be seen in Fig. 1c. One can find that MRGO, although contains many wrinkles, still exhibits sheet-like structures with a number of iron oxide nanoparticles (IONs) embedded. A few ZIF nanocrystals were immobilized on the surface of MRGO sheets, showing that ZIF was assembled with MRGO via the mixing in the aqueous solution. To probe in possibilities leading to such a self-assembled nanocomposite, surface charges of ZIF, MRGO and MRGO/ZIF were measured and shown in Fig. 2a. One can see that ZIF nanocrystals exhibited positive surface charges at most of pH, especially under acidic and neutral conditions. In contrast, MRGO displayed highly negative surface charges because of oxygenic functional groups (e.g., carboxylates) remained on the GO surface. The electrostatic attraction between ZIF and MRGO might cause the attachment of ZIF to MRGO to form such a self-assembled nanocomposite. Thus, the surface charge of the resultant MRGO/ZIF was found to be close to neutral as revealed in Fig. 2a. In addition, 2-MIM in ZIF, which was considered as aromatic, could also interact with MRGO via the ␲–␲ stacking effect to facilitate the self-assembling process of MRGO/ZIF. The crystalline structures of MRGO/ZIF and precursors can be seen in Fig. 2b. The XRD pattern of ZIF is almost identical to the reported and simulated patterns [24], indicating the welldevelopment of ZIF nanocrystals. The XRD pattern of MRGO can be also seen in Fig. 2b, in which a broad peak at 2 = 23–26◦ is derived from chemically reduced GO (CRGO) [20]. Other peaks at 18.3◦ , 30.2◦ , 35.6◦ , 37.2◦ , 43.3◦ , 53.7◦ , 57.3◦ and 62.8◦ can be readily indexed according to the standard pattern of Fe3 O4 (JCPDS card #75-0033), validating that IONs had been immobilized on the surface of reduce graphene oxide. Furthermore, the XRD pattern of MRGO/ZIF shows combined features from the XRD patterns of MRGO and ZIF, suggesting that ZIF nanocrystals were attached to MRGO and the crystalline structures of ZIF and MRGO were well preserved even after the self-assembling process. Surface chemistry of MRGO/ZIF and precursors were also determined and shown in Fig. 3a. Compared to GO (Fig. S1, see ESI†), peak intensities of the oxygenic groups in MRGO (e.g., peaks at 1040 cm−1

Fig. 2. Characteristics of MRGO/ZIF and the precursors: (a) surface charges at ambient temperature and (b) XRD patterns.

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

117

the D and G bands, respectively. Besides, additional short peaks appear at 182, 460 and 663 Raman Shift (cm−1 ), which are derived from cobalt ions of ZIF [30,31], validating the existence of ZIF in MRGO/ZIF. Furthermore, to investigate thermal stability of MRGO/ZIF, TGA of MRGO/ZIF and precursors were determined and shown in Fig. 3c. MRGO exhibited gradual weight losses from 200 ◦ C to 450 ◦ C, while ZIF started losing weight at a relatively low temperature (i.e., 100 ◦ C) due to the escape of moisture or gas molecules [32]. The weight loss occurred from 200 to 500 ◦ C can be ascribed to the decomposition of the organic ligand [32]. In the case of MRGO/ZIF, the weight loss profile was between the profiles of MRGO and ZIF. This suggests that the combination of MRGO and ZIF can exhibit a higher thermal stability compared to the pristine ZIF. Since MRGO/ZIF was proposed to be a recoverable sorbent, it is necessary to examine its magnetic property. Fig. 4a displays a plot of the magnetization of MRGO/ZIF versus magnetic field. The saturation magnetization approached to 20 emu g−1 and the remanence of magnetization was found to be insignificant (i.e., <100 Oe) as shown in Fig. S2. Such a magnetic property allows MRGO/ZIF to be readily separated and recovered from a solution as shown in Fig. 4b. 3.2. Adsorption of MG to MRGO/ZIF: Comparison with precursors and kinetics

Fig. 3. Chemical and physical properties of MRGO/ZIF and the precursors: (a) surface chemistry by IR analysis, (b) Raman spectroscopic analysis and (d) TGA results.

(C O C) and 1220 cm−1 (C O)) were significantly decreased, indicating the chemical reduction of GO [25,26]. Another peak is noticed at around 600 cm−1 , which can be attributed to the IONs attached to CRGO surfaces [27]. On the other hand, the spectrum of ZIF was found to consist of peaks mainly from 2-MIM [24,28,29]. A series of peaks at 600–1500 cm−1 can be assigned to the stretching and bending modes of imidazole ring, whereas the peak at 1584 cm−1 is arisen from the stretching mode of C N bonding in 2-MIM. The peaks at 2929 and 3135 cm−1 can be attributed to the stretching mode of C H bonding in 2-MIM. After ZIF was assembled with MRGO, the surface chemistry of MRGO/ZIF was found to comprise characteristic peaks both from MRGO and ZIF. This suggests that surface functional groups of ZIF and MRGO were remained even after the self-assembling process. Raman spectroscopic analysis of MRGO/ZIF can be seen in Fig. 3b. Since MRGO/ZIF consisted of graphene, two notable peaks at 1350 and 1590 Raman Shift (cm−1 ) can be detected and assigned to

Adsorption capacity of MRGO/ZIF for MG is shown in Fig. 5a, in which individual adsorption capacities of MRGO and ZIF for MG are also included for comparison. It can be seen that at 30 ◦ C, adsorption of MG to MRGO/ZIF gradually increased and approached to equilibrium after 120 min. The adsorption capacity at t = 120 min had reached ∼1850 mg g−1 , revealing a very high adsorption capacity of MRGO/ZIF for MG. On the other hand, ZIF alone also exhibited a quite high adsorption capacity of 1550 mg g−1 under the same condition, whereas MRGO alone showed almost negligible adsorption capacity for MG (i.e., 30 mg g−1 ). This result suggests that the high adsorption capacity of MRGO/ZIF can be attributed to ZIF and the combination of MRGO and ZIF appeared to provide a synergic effect to enhance MG adsorption. Such an enhancement could be owing to inclusion of MG in inter-space existing between MRGO and ZIF. Furthermore, adsorption kinetics at different temperature was investigated as shown in Fig. 5b. When temperature was increased from 30 to 40 ◦ C, the adsorption capacity obviously increased from 1850 to 2200 mg g−1 at t = 150 min. In contrast, the adsorption capacity was decreased to 1500 mg g−1 at 20 ◦ C after the same mixing time. This indicates that the effect of temperature was significant and the elevated temperature could facilitate the MG adsorption. While the effect of temperature on the adsorption capacity can be readily distinguished, the adsorption kinetics had to be analyzed using rate laws to reveal the effect of temperature. In this study, two common rate laws are adopted including the pseudo first and second equations. The pseudo first order equation is employed to model adsorption behaviors in a solid–liquid system where adsorbate (i.e., MG) is adsorbed to single adsorption site of adsorbent (i.e., MRGO/ZIF). Thus, the pseudo first order equation typically can be expressed in the following equation: ln(qe − qt ) = ln qe −

k1 t 2.303

(2)

where k1 (min−1 ) denotes the pseudo first order rate constant. The fitting results are represented by the dotted lines shown in Fig. 5b and the correlation coefficients (R1 2 ) of fitting are listed in Table 1. The k1 was found to increase noticeably from 0.0021 to 0.0054 min−1 as the temperature changed from 20 to 40 ◦ C, revealing that the elevated temperature also significantly improved the adsorption kinetics. Furthermore, the pseudo second order

118

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

Fig. 4. Magnetic properties of MRGO/ZIF: (a) saturation magnetization and (b) pictures showing the quick separation of MRGO/ZIF from water.

a.

b.

2000

2000

qt (mg g−1)

1500

qt (mg g−1)

2500

1000 MRGO/ZIF ZIF MRGO

500

1500 1000 ο

20 C ο 30 C 40 οC

500

0

0

0

50

100

150

200

0

50

t (min)

100

150

200

t (min)

Fig. 5. Adsorption of MG to MRGO/ZIF: (a) comparison between MRGO/ZIF and the precursors (T = 30 ◦ C) and (b) adsorption kinetics at different temperatures (MRGO/ZIF = 5 mg, C0 of MG = 600 mg L−1 ).

equation was also adopted considering that MG might interact with dual adsorption sites of MRGO/ZIF. The pseudo second order equation can be described as follows: t 1 t = + qt qe k2 q2e

(3)

where k2 (g mg−1 min−1 ) means the pseudo second order rate constant. The fitting results can be seen as the solid lines in Fig. 5b and the R2 2 of fitting and rate constants are also summarized in Table 1. The k2 value was also found to increase noticeably when the temperature rose up. This validates that the elevated temperature could facilitate the adsorption of MG to MRGO/ZIF. 3.3. Adsorption isotherm of MG to MRGO/ZIF Adsorption isotherm of MG to MRGO/ZIF at different temperatures can be seen in Fig. 6a. At a constant temperature, as the initial concentration increased, the qe gradually increased and then approached to saturation. Once the temperature was elevated, the saturation value was correspondingly raised up, validating the highly positive effect of higher temperature on adsorption capacity. Thus, the adsorption of MG to MRGO/ZIF may be considered as an

endothermic reaction. To further analyze the adsorption isotherm quantitatively, a number of adsorption isotherm models were adopted including Langmuir, Freundlich and Langmuir–Freundlich (also called Sips) isotherm models. First, the Langmuir isotherm model was employed, which assumes that adsorption occurs as a mono-layer on a homogenous surface. Since the number of adsorption sites is finite, a maximal adsorption capacity (i.e., qmax ) is expected. Typically the Langmuir isotherm can be expressed by the following equation: Ce 1 Ce = + qe KL qmax qm

(4)

where KL represents the Langmuir isotherm constant which is associated with the adsorption bonding energy. The fitting results using the Langmuir isotherm model can be seen as the dotted lines in Fig. 6a with correlation coefficients >0.95 at all testing temperatures. qmax and KL values are also listed in Table 2 and qmax was found to increase significantly as the temperature rose up, validating the positive effect of the elevated temperature. Next, the adsorption isotherm data was analyzed using the Freundlich isotherm model. In the Freundlich isotherm, adsorption involves both the chemisorption and physisorption into account because

Table 1 Kinetic parameters for the MG adsorption to MRGO/ZIF at various temperatures. Conditions

Pseudo-first-order

Temp. (◦ C)

k1 × 103 (min−1 )

qe , estimated (mg g−1 )

R1 2

Pseudo-second-order k2 × 106 (g mg−1 min−1 )

qe , estimated (mg g−1 )

R2 2

20 30 40

21 37 54

1475 1826 2171

0.987 0.991 0.995

11 20 28

1882 2137 2454

0.991 0.995 0.995

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

a.

b.

2500

2500 2000

1500 1000 ο

20 C ο 30 C 40 οC

500 0

qe (mg g−1)

qe (mg g−1)

2000

119

1500 1000 500 0

0

20

40

60

80 100

3

−1

Ce (mg L )

4

5

6

7

pH

Fig. 6. Adsorption characteristics: (a) adsorption isotherm and (b) the effect of initial pH on the adsorption capacity (T = 30 ◦ C).

Table 2 Modeling parameters of adsorption isotherm derived from Langmuir model, Freundlich model, and Langmuir–Freundlich (Sips) models at various temperatures. Langmuir

Freundlich

Langmuir–Freundlich

Temp. (◦ C)

qmax (mg g−1 )

KL (L mg−1 )

RL 2

KF (mg g−1 ) (L mg−1 )1/n

n

RF 2

qm (mg g−1 )

KLF (L mg−1 )1/n

N

RLF 2

20 30 40

1680 2045 2116

0.507 0.650 2.400

0.965 0.973 0.957

827 1062 1340

5.71 5.60 5.91

0.990 0.985 0.992

2328 2682 3165

0.163 0.292 0.456

2.34 2.05 2.68

0.999 0.999 0.999

the adsorption is assumed to occur as mono-layer as well as multiple-layer. Typically, the Freundlich isotherm can be expressed as follows: ln qe = ln KF +

1 ln Ce n

(5)

where KF denotes the Freundlich isotherm constant and n means the heterogeneity factor of MRGO/ZIF related to the surface heterogeneity. The fitting results can be seen as the dashed lines in Fig. 6a. It can be seen that the fitting curves correlate with the data points more properly than the fitting result using the Langmuir isotherm. The correlation coefficients (Table 2) also validate that the Freundlich isotherm seemed more satisfactory than the Langmuir isotherm. Considering that the Freundlich model was more properly to describe the adsorption isotherm data, we further employed the Langmuir–Freundlich (L–F) isotherm model in order to estimate the maximal adsorption capacity. The L–F model is developed based on the Freundlich isotherm with an asymptotic characteristic that adsorption eventually reaches a maximal value. In particular, when the heterogeneity factor (n) is unity, the L–F isotherm tends to become the Langmuir isotherm. Thus, the L–F isotherm model can be expressed as follows: qe =

qm KLF (Ce )1/n 1 + KLF (Ce )1/n mg−1 )1/n

reporting adsorption capacities for MG to be included in this study, only those adsorption capacities higher than 200 mg g−1 were listed here. It can be seen that most of adsorbents merely exhibited adsorption capacities of 500 mg g−1 or even lower, whereas ZIF-based materials (i.e., ZIF-67 and MRGO/ZIF) show significantly higher adsorption capacities than the reported materials. It has been indicated that adsorption of organic pollutants to MOFs can be attributed to the following interactions: electrostatic interactions, Lewis acid–base interactions, hydrogen bonding as well as the ␲–␲ stacking interactions [33,34]. The high adsorption capacity of MRGO/ZIF is essentially owing to ZIF- which consists of imidazole rings. The aromatic imidazole rings are considered to interact with the aromatic rings of MG via the ␲–␲ stacking interaction which enables MRGO/ZIF to attract MG and results in such a high adsorption capacity. Table 3 Comparison of the MG adsorption capacity of MRGO/ZIF with the reported sorbents. Material

T (◦ C)

Maximal adsorption capacity

Reference

MRGO/ZIF

20 30 40 40 60

2328 2682 3165 3020 214

In this study

30

263

30

222

Hameed et al. [9] Malik et al. [10]

60

280

Wang et al. [16]

30

384

60

418

Nuengmatcha et al. [18] Ahmad et al. [7]

25

354

Tian et al. [17]

30

648

Song et al. [11]

(6)

denotes the L–F isotherm constant and qm where KLF (L represents the estimated maximal adsorption capacity. The fitting results using the L–F model can be represented as the solid lines in Fig. 6a. Compared to the fitting curves obtained using the Langmuir and Freundlich separately, the fitting curve of the L–F model can fit the data points most properly. The correlation coefficients were 0.999 at all testing temperatures, indicating the excellent applicability of the L–F model for describing the isotherm data. Besides, at 20 ◦ C, qm had reached 2328 mg g−1 and even higher to 2682 and 3165 mg g−1 at 30 and 40 ◦ C, respectively, showing a very high adsorption capacity for MG. To compare with the reported adsorption capacities, we summarized adsorption capacities of various adsorbents in Table 3. Considering that there are too many studies

ZIF-67 Coconut-derived activated carbon Bamboo-derived activated carbon Groundnut-derived activated carbon Sawdust functionalized with oxalic acid Graphene oxide Rambutan peel-derived activated carbon Ordered mesoporous carbon Tyre-derived activated carbon

Lin et al. [19] Bello et al. [8]

120

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

Table 4 Activation energy and thermodynamic parameters of the MG adsorption to MRGO/ZIF. Ea (kJ mol−1 )

k (g mg-1 min− )

Temp. (K)

G◦ (kJ mol− )

H◦ (kJ mol−1 )

S◦ (kJ mol−1 K−1 )

35.7

0.026

293 303 313

−26.78 −29.16 −31.28

39.1

0.225

3.4. Activation energy and thermodynamic parameters of the MG adsorption As revealed in the earlier section, the elevated temperature noticeably facilitated the adsorption kinetics. The relationship between the effect of temperature and kinetics is further correlated using the Arrhenius equation by the following equation: ln k2 = ln k −

Ea 1 R T

(7)

k is the temperature-independent parameter where (g mg−1 min−1 ); Ea denotes the activation energy of adsorption (kJ mol−1 ); R represents the universal gas constant and T is temperature in Kelvin. Fig. S3a (see ESI†) shows a plot of ln k2 versus 1/T, in which the data points can fit to the linear regression line quite properly (R2 = 0.981), indicating that the effect of temperature can be correlated to the kinetics via the Arrhenius equation. Based on the slope and intercept of the linear regression line in Fig. S3a, Ea and the temperature-independent parameter, k were also determined to be 35.7 kJ mol−1 and 0.026 g mg−1 min−1 , respectively (Table 4). In addition, the effect of temperature can also be correlated to adsorption isotherm constant to determine the adsorption enthalpy (H◦ ) and entropy (S◦ ) using the following equations: G◦ = −RT ln KLF ◦

◦

G = H − TS

(8) ◦

(9)

where G◦ denotes the free energy of the adsorption at a given temperature and KLF is the Langmuir–Freundlich isotherm constant when the heterogeneity factor (n) becomes unity. G◦ values at the testing temperatures are summarized in Table 4. One can see that as the temperature increases, the G◦ value became even more negative. This suggests that the adsorption of MG is more likely to occur spontaneously at a higher temperature. A plot of G◦ versus T was further compiled and shown in Fig. S3b. The data points were almost lined up with the linear regression line with R2 of 0.999. The intercept and slope of the regression line are then used to calculate the enthalpy (H◦ ) and entropy (S◦ ) which are 39.1 kJ mol−1 and 0.225 kJ mol−1 K−1 , respectively. The positive value of H◦ validates that the MG adsorption to MRGO/ZIF was an endothermic reaction, whereas the low S◦ suggests that the entropy change derived from the MG adsorption to MRGO/ZIF was not significant.

Wastewater may also contain other compounds, particularly cationic ions and molecules, which could also compete with MG during the adsorption to MRGO/ZIF. Thus, we further examined the effect of co-existing compounds by adding salts (i.e., NaCl and NH4 Cl) and a cationic surfactant (i.e., CTAB) to MG solutions (C0 = 500 mg L−1 ). Fig. 7a shows removal efficiencies without and with co-existing compounds as a function of concentration of co-existing compounds varied from 100 to 500 mg L−1 . When relatively low concentration of NaCl (e.g., 100−300 mg L−1 ) was added to the MG solution, the removal efficiency remained almost the same. As C0 of NaCl was increased to 400 or 500 mg L−1 , a slight decrease in the removal efficiency can be observed from 100 to 98%, which, however, was negligible. The similar result can be obtained when NH4 Cl was added to MG solutions. In the case of CTAB, while the decrease due to the presence of CTAB in the removal efficiency became more notable, the effect of CTAB was still not significant (i.e., 100 → 93%) even though the equal-concentration of CTAB was added to the MG solution. These results suggest that MRGO/ZIF exhibited sufficient selectivity toward MG even in the presence of other common cationic compounds.

3.6. Recyclability of MRGO/ZIF for the MG adsorption Since MRGO/ZIF was proposed to be a recoverable adsorbent for MG, MRGO/ZIF can be more advantageous and competitive if it can be used for multiple cycles of adsorption after regeneration. In this study, a simple method was adopted to regenerate the used MRGO/ZIF by washing it with ethanol. Pure ethanol was continuously used to wash the used MRGO/ZIF until no green color was detected in the ethanol solution. The regenerated MRGO/ZIF then was used to adsorb MG. The recyclability of MRGO/ZIF, which is defined as the percentage of qe of the regenerated MRGO/ZIF over the pristine MRGO/ZIF, can be seen in Fig. 7b. The first-cycle recyclability was close to 100%, indicating no noticeable loss in adsorption capacity of the regenerated MRGO/ZIF. Even after 4 cycles of regeneration, the recyclability was still quite close to 100%. This suggests that MRGO/ZIF can be regenerated by desorbing MG from MRGO/ZIF and the regenerated MRGO/ZIF can exhibit stable and high recyclability over multiple cycles.

3.5. Effects of pH of solution and co-existing compounds pH of wastewater containing MG may be varied; therefore, the effect of pH has to be investigated. Nevertheless, the color of MG becomes unstable under alkaline conditions. Thus, in this study pH was changed only from 3 to 7 to probe into adsorption capacities under relatively acidic conditions. When pH was reduced from 7 to 5, the adsorption capacity seemed quite stable without notable capacity loss. However, the capacity decreased slightly at pH = 4 and further dropped from 2200 to 200 mg g−1 at pH = 3. Such a significant decrease could be attributed to proton accumulation on the surface of MRGO/ZIF under highly acidic conditions. The electrostatic repulsion between the proton and MG hindered the approach of MG to MRGO/ZIF, thereby decreasing the adsorption capacity.

Fig. 7. Adsorption behaviors: (a) effect of co-existing ions and (b) recyclability (T = 30 ◦ C).

K.-Y.A. Lin, W.-D. Lee / Applied Surface Science 361 (2016) 114–121

4. Conclusion In the present study, a nanocomposite consisting of MRGO sheet and ZIF nanocrystals was developed via the self-assembling process. The magnetic property of MRGO equipped MRGO/ZIF with a magnetic controllability which allows MRGO/ZIF to be easily recovered from water. ZIF, immobilized on the surface of MRGO, can exhibit significantly high adsorption capacity for MG. Considering these features, the self-assembled MRGO/ZIF appeared to be an ideal adsorbent for MG. The adsorption kinetics and isotherm of MG to MRGO/ZIF were determined and analyzed using the rate laws and isotherm models. In particular, the adsorption isotherm data could well fit to the Langmuir–Freundlich model with R2 = 0.999 at all testing temperatures. The effects of pH and co-existing ionic compounds were also examined. While the highly acidic condition was unfavorable for the MG adsorption to MRGO/ZIF, the presence of co-existing salts and cationic surfactant caused insignificant effect on the MG adsorption. In addition, MRGO/ZIF can be simply regenerated by washing it with ethanol and the regenerated MRGO/ZIF exhibited high and stable recyclability. Considering its extremely high adsorption capacity, convenient magnetic controllability and great recyclability, MRGO/ZIF can be a promising adsorbent for MG. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apsusc.2015.11. 108. References [1] J. Zhang, Y. Li, C. Zhang, Y. Jing, Adsorption of malachite green from aqueous solution onto carbon prepared from Arundo donax root, J. Hazard. Mater. 150 (2008) 774–782. [2] S. Nethaji, A. Sivasamy, G. Thennarasu, S. Saravanan, Adsorption of Malachite Green dye onto activated carbon derived from Borassus aethiopum flower biomass, J. Hazard. Mater. 181 (2010) 271–280. [3] C. Berberidou, I. Poulios, N.P. Xekoukoulotakis, D. Mantzavinos, Sonolytic, photocatalytic and sonophotocatalytic degradation of malachite green in aqueous solutions, Appl. Catal., B: Environ. 74 (2007) 63–72. [4] L.W. Man, P. Kumar, T.T. Teng, K.L. Wasewar, Design of experiments for Malachite Green dye removal from wastewater using thermolysis – coagulation – flocculation, Desalin. Water Treat. 40 (2012) 260–271. [5] N. Daneshvar, M. Ayazloo, A.R. Khataee, M. Pourhassan, Biological decolorization of dye solution containing Malachite Green by microalgae Cosmarium sp, Bioresour. Technol. 98 (2007) 1176–1182. [6] A.R. Khataee, M. Zarei, M. Pourhassan, Bioremediation of Malachite Green from contaminated water by three microalgae: neural network modeling, CLEAN—Soil, Air, Water 38 (2010) 96–103. [7] M.A. Ahmad, R. Alrozi, Removal of malachite green dye from aqueous solution using rambutan peel-based activated carbon: equilibrium, kinetic and thermodynamic studies, Chem. Eng. J. 171 (2011) 510–516. [8] O.S. Bello, M.A. Ahmad, Coconut (Cocos nucifera) shell based activated carbon for the removal of Malachite Green Dye from aqueous solutions, Sep. Sci. Technol. 47 (2011) 903–912. [9] B.H. Hameed, M.I. El-Khaiary, Equilibrium, kinetics and mechanism of malachite green adsorption on activated carbon prepared from bamboo by K2 CO3 activation and subsequent gasification with CO2 , J. Hazard. Mater. 157 (2008) 344–351. [10] R. Malik, D.S. Ramteke, S.R. Wate, Adsorption of malachite green on groundnut shell waste based powdered activated carbon, Waste Manage. 27 (2007) 1129–1138.

121

[11] X. Song, R. Xu, K. Wang, High capacity adsorption of malachite green in a mesoporous tyre-derived activated carbon, Asia-Pac. J. Chem. Eng. 8 (2013) 172–177. [12] T. Robinson, G. McMullan, R. Marchant, P. Nigam, Remediation of dyes in textile effluent: a critical review on current treatment technologies with a proposed alternative, Bioresour. Technol. 77 (2001) 247–255. [13] S.-H. Huo, X.-P. Yan, Metal–organic framework MIL-100(Fe) for the adsorption of malachite green from aqueous solution, J. Mater. Chem. 22 (2012) 7449–7455. [14] M.N. Rashed, Adsorption Technique for the Removal of Organic Pollutants from Water and Wastewater, in: M.N. Rashed (Ed.), Organic Pollutants Monitoring, Risk and Treatment, InTech, Rijeka, Croatia, 2013. [15] J.L. Sotelo, G. Ovejero, A. Rodríguez, S. Álvarez, J. Galán, J. García, Competitive adsorption studies of caffeine and diclofenac aqueous solutions by activated carbon, Chem. Eng. J. 240 (2014) 443–453. [16] H. Wang, X. Yuan, G. Zeng, L. Leng, X. Peng, K. Liao, L. Peng, Z. Xiao, Removal of malachite green dye from wastewater by different organic acid-modified natural adsorbent: kinetics, equilibriums, mechanisms, practical application, and disposal of dye-loaded adsorbent, Environ. Sci. Pollut. Res. 21 (19) (2014) 11552–11564. [17] Y. Tian, P. Liu, X. Wang, H. Lin, Adsorption of malachite green from aqueous solutions onto ordered mesoporous carbons, Chem. Eng. J. 171 (2011) 1263–1269. [18] P. Nuengmatcha, R. Mahachai, S. Chanthai, Thermodynamic and kinetic study of the intrinsic adsorption capacity of graphene oxide for Malachite green removal from aqueous solution, Orient. J. Chem. 30 (2014) 1463–1474. [19] K.-Y.A. Lin, H.-A. Chang, Ultra-high adsorption capacity of zeolitic imidazole framework-67 (ZIF-67) for removal of malachite green from water, Chemosphere 139 (2015) 624–631. [20] J. Shen, M. Shi, H. Ma, B. Yan, N. Li, M. Ye, Hydrothermal synthesis of magnetic reduced graphene oxide sheets, Mater. Res. Bull. 46 (2011) 2077–2083. [21] W.S. Hummers, R.E. Offeman, Preparation of graphitic oxide, J. Am. Chem. Soc. 80 (1958) 1339. [22] J. Shao, Z. Wan, H. Liu, H. Zheng, T. Gao, M. Shen, Q. Qu, H. Zheng, Metal organic frameworks-derived Co3 O4 hollow dodecahedrons with controllable interiors as outstanding anodes for Li storage, J. Mater. Chem., A 2 (2014) 12194–12200. [23] J. Cravillon, C.A. Schroder, H. Bux, A. Rothkirch, J. Caro, M. Wiebcke, Formate modulated solvothermal synthesis of ZIF-8 investigated using time-resolved in situ X-ray diffraction and scanning electron microscopy, CrystEngComm 14 (2012) 492–498. [24] A.F. Gross, E. Sherman, J.J. Vajo, Aqueous room temperature synthesis of cobalt and zinc sodalite zeolitic imidizolate frameworks, Dalton Trans. 41 (2012) 5458–5460. [25] S. Zhang, L. Zhu, H. Song, X. Chen, B. Wu, J. Zhou, F. Wang, How graphene is exfoliated from graphitic materials: synergistic effect of oxidation and intercalation processes in open, semi-closed, and closed carbon systems, J. Mater. Chem. 22 (2012) 22150–22154. [26] D.C. Marcano, D.V. Kosynkin, J.M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L.B. Alemany, W. Lu, J.M. Tour, Improved synthesis of graphene oxide, ACS Nano 4 (2010) 4806–4814. [27] J. Shen, Y. Hu, M. Shi, N. Li, H. Ma, M. Ye, One step synthesis of graphene oxide–magnetic nanoparticle composite, J. Phys. Chem., C 114 (2010) 1498–1503. [28] M. He, J. Yao, Q. Liu, K. Wang, F. Chen, H. Wang, Facile synthesis of zeolitic imidazolate framework-8 from a concentrated aqueous solution, Microporous Mesoporous Mater. 184 (2014) 55–60. [29] J. Yao, R. Chen, K. Wang, H. Wang, Direct synthesis of zeolitic imidazolate framework-8/chitosan composites in chitosan hydrogels, Microporous Mesoporous Mater. 165 (2013) 200–204. [30] X. Wang, L. Song, H. Yang, W. Xing, H. Lu, Y. Hu, Cobalt oxide/graphene composite for highly efficient CO oxidation and its application in reducing the fire hazards of aliphatic polyesters, J. Mater. Chem. 22 (2012) 3426–3431. [31] K.-Y. Andrew Lin, F.-K. Hsu, W.-D. Lee, Magnetic cobalt–graphene nanocomposite derived from self-assembly of MOFs with graphene oxide as an activator for peroxymonosulfate, J. Mater. Chem., A 3 (2015) 9480–9490. [32] A. Schejn, L. Balan, V. Falk, L. Aranda, G. Medjahdi, R. Schneider, Controlling ZIF-8 nano- and microcrystal formation and reactivity through zinc salt variations, CrystEngComm 16 (2014) 4493–4500. [33] N.A. Khan, Z. Hasan, S.H. Jhung, Adsorptive removal of hazardous materials using metal–organic frameworks (MOFs): a review, J. Hazard. Mater. 244–245 (2013) 444–456. [34] Z. Hasan, S.H. Jhung, Removal of hazardous organics from water using metal–organic frameworks (MOFs): plausible mechanisms for selective adsorptions, J. Hazard. Mater. 283 (2015) 329–339.