Magnetic γ-Fe2O3/ZIF-7 Composite Particles and Their Application for Oily Water Treatment.

Crude oil spills are about global challenges because of their destructive effects on aquatic life and the environment. The conventional technologies for cleaning crude oil spills need to study the selective separation of pollutants. The combination of magnetic materials and porous structures has been of considerable interest in separation studies. Here, γ-Fe2O3/ZIF-7 structures were prepared by growing a ZIF-7 layer onto supermagnetic γ-Fe2O3 nanoparticles with an average size of 18 ± 0.9 nm in situ without surface modification at low temperatures. The product composite particles were characterized using X-ray diffraction, scanning electron microscopy, energy-dispersive X-ray spectroscopy, vibrating sample magnetometry, and N2 adsorption/desorption isotherms. The analyses revealed a time growth-dependent ZIF-7 rod thickness with abundant nanocavities. The γ-Fe2O3/ZIF-7 surface area available for sorption (647 m2/g) is ∼12-fold higher than that of the γ-Fe2O3 nanoparticles. Moreover, the crystal structure of γ-Fe2O3 remained essentially unchanged following ZIF-7 coating, whereas the superparamagnetism declined depending on the coating time. The γ-Fe2O3/ZIF-7 particles were highly hydrophobic and selectively and rapidly (<5 min) sorbed crude oil and other hydrocarbon pollutants from water. As high as 6 g/g of the hydrocarbon was sorbed by the γ-Fe2O3/ZIF-7 particles immersed into the hydrocarbon. A coefficient of determination, R 2 2, consistently >0.96 at all pollutant concentrations suggested a pseudo-second-order sorption kinetics. The thermal stability and 15 cycles of use and reuse confirmed a robust γ-Fe2O3/ZIF-7 sorbent.

Introduction

The global demand of oil prompts continued production of both light and heavy crude oils. Accidental oil spill into water bodies, on the other hand, poses catastrophic environmental problems, in addition to immeasurable financial challenges.[1−3] For example, the biggest recorded crude oil spill released 4.9 million barrels of crude oil into the Gulf of Mexico over a short period of time and cost $30 billion in reclamation effort.[4] Subsequently, continuous effort should be invested to develop feasible materials and protocols for oily water purification.[5]

Nanomaterials have been widely used in the advancement of enhanced oil recovery processes.[6] They were also explored as potential materials for advancing separation processes. Among other technologies, such as biological methods,[7] sedimentation–coagulation,[8] photocatalysts,[9] and filtration,[10] that have been used to remove organic/oily pollutants from water, adsorption is considered an effective technology owing to its simplicity, ease of handling, and favorable economy.[11] Effective sorbents such as hybrid foams,[12−14] sponges,[15,16] activated carbon,[17,18] fibers,[19,20] aerogels,[21] porous materials,[22,23] and magnetic materials[24] have been proposed over the past few decades. General attributes of ideal sorbents include superhydrophobicity, excellent recyclability, and environmentally friendly chemical and physical nature. Accordingly, the high specific surface area, surface roughness, low density, and nontoxicity of adsorbents have been studied.[25] Achieving these specifications for materials with effective magnetic properties provides the advantage of easily recoverable sorbents. Magnetic nanoparticles (NPs) such as Fe3O4/PS,[25] Fe3O4/SiO2,[26] and yeast magnetic bionanocomposites[27] have been investigated. Rapid removal of oil-loaded sorbents through separation with permanent magnets has been reported. On the other hand, porous materials with a large specific surface area have been fabricated to serve oil/water separation.[14,28,29] Metal–organic framework (MOF) porous materials in different shapes have shown good oily pollutant cleanup attributes due to their high surface area, hierarchical porous structure, and highly hydrophobic nature.[30] MOFs, specially zeolitic imidazole frameworks (ZIFs), with a much wider range of pore sizes and well-defined shapes and sizes have drawn considerable attention and have been tested for hydrogen storage,[31] gas storage/separation,[32] sensing,[33] drug delivery,[34] electrocatalysts,[35] electrochemical supercapacitors,[36] low carbon natural gas recovery from hydrate reservoirs,[37] catalysis,[38] dye degradation[39] wastewater treatment,[40,41] and oil absorption.[42] ZIFs are a group of hierarchical porous materials in which Zn is connected to organic ligands such as 2-methylimidazolite, benzimidazole, and so forth through organometallic bonds.[43−45]

Coupling magnetic properties and porous structures in a composite structure has recently led to novel sorbents with high sorption capacity and rapid separation attributes.[46,47] For example, magnetic iron oxides, owing to their nontoxic nature, low cost, and strong magnetic response are widely applied in the synthesis of porous magnetic materials using techniques such as the template method[48] and coprecipitation method.[49,50] Furthermore, synthesis protocols to produce composite structures such as the hydrothermal and solvothermal processes,[51] sol–gel method,[52] chemical vapor deposition,[53] and solution-immersion process[54] have been reported. However, these techniques suffer from complications pertaining to operating conditions, the use of many precursors/reagents, and the need for meticulous surface premodification steps. An effective process for preparing hydrophobic porous shell structures over a magnetic core ideal for oil spill separation would involve a simple experimental protocol with mild operating conditions and a minimum number of precursors/reagents. Precipitation coupled with in situ synthesis and coating on substrates has been recently explored.[55−57] Magnetic NP/MOF composites with a core–shell structure onto presynthesized magnetic NPs surrounded by in situ-grown MOFs have been explored. Nevertheless, meticulous surface premodification and/or stabilization by surfactants was still required.[58,59]

In this study, we report the synthesis of a novel γ-Fe2O3/ZIF-7 composite with a huge specific surface area using a simple precipitation route without surface premodification or stabilization. The resultant composite is evaluated for the selective removal of crude oil and other organic solvents from water. Moreover, we study the separation of the sorbent using a permanent magnet and its recyclability attributes. The γ-Fe2O3/ZIF-7 composite with powerful magnetic properties can meet the formidable demand of oil spills. In addition, the superparamagnetic properties and superhydrophobicity of the sorbent are also reported.

Materials and Methods

Materials

Ferric chloride hexahydrate (FeCl3·6H2O, 99%), FeCl2·4H2O (98%), NH4OH (ammonium hydroxide, 25–30% of ammonia), N,N-dimethylformamide [DMF, HCON(CH3)2, 99.8%], methanol (CH3OH, 99.8%), and ethanol (CH3CH2OH, 99.93%) were provided by Merck (Darmstadt, Germany). Zinc nitrate hexahydrate (Zn(NO3)2·6H2O, 98% pure), benzimidazole (C7H6N2, 98%), and HCl (hydrochloric acid, 37%) were purchased from Sigma-Aldrich (Darmstadt, Germany). Heavy crude oil was prepared from a well in the south of Iran with viscosity of 415.4 cP at 60 °C and an oil API gravity of 19. Toluene and acetone were obtained from Merck (99% pure, Darmstadt, Germany). All above materials were utilized as received without additional purification. Deionized (DI) water was used throughout the tests.

Preparation of the Absorbent

Synthesis of the γ-Fe2O3 NPs

The γ-Fe2O3 NPs were synthesized by a wet chemical method at room temperature.[60] Solutions of 2 M FeCl2 and 1 M FeCl3 in 2 M HCl were prepared under continuous magnetic stirring. Drops of a 2 M NH4OH solution were added, and the pH of the aqueous solution was adjusted to 9.73. After stirring for 2 h, a brown precipitate was obtained, which was filtered and washed three times with ethanol and DI water. The gathered precipitate was dried in an oven at 70 °C for 12 h.

Synthesis of the γ-Fe2O3/ZIF-7 Composite

The in situ synthesis of the ZIF-7 nanostructure onto the γ-Fe2O3 NP utilized the precipitation method reported in the literature[61,62] with slight modifications, as illustrated in Figure . In brief, 0.0500 g of the γ-Fe2O3 NPs and 0.2347 g benzimidazole were dispersed in 30 mL of DMF using ultrasonication for 30 min. A mass of 0.8025 g of Zn(NO3)2·6H2O was dissolved into 30 mL of DMF and slowly added to the sonicated suspension over 5 min at room temperature. The mixture was transformed into a water bath maintained at 60 °C. The in situ synthesis was carried out for different times of 1, 2, and 3 h in order to investigate the effect of the precipitation time on the sorption properties. During this time, ZIF-7 nanostructures emerged on the surface of the magnetic γ-Fe2O3 NPs. The resultant composite was collected by filtration and washed several times with ethanol and finally dried at 60 °C for 12 h.

Figure 1

Schematic preparation procedure of the γ-Fe2O3/ZIF-7 composite.

Characterization

The X-ray diffraction pattern of the synthesized products was collected on an X′Pert diffractometer (model, Vendor, Country) with Cu Kα radiation and 2θ ranging from 0 to 60° at 40 kV and 20 mA. The XRD pattern was compared with the standard patterns in X-pert HighScore software in order to identify the particles. The crystal size was determined using Scherrer eq .[9]where D is the crystal size (nm), λ is the X-ray radiation wavelength (nm), θ is the Bragg angle, and β is the full width at half-maximum peak (radians). The Bragg angle, 2θ = 21.211°, was selected for these calculations since it corresponded to the most pronounced peak. The morphology of the product particles was investigated by field-emission scanning electron microscopy (FE-SEM, Hitachi S4160, Tokyo, Japan) equipped with an energy-dispersive X-ray (EDX) probe. The magnetic properties of the NPs and the composite at different preparation times were measured on a vibrating sample magnetometer (model of LBKF, Meghnatis Daghigh Kavir, Iran). The N2 adsorption–desorption isotherms of the NPs and the nanostructural composite were determined using an automated surface area analyzer (model of Sorptometer Kelvin 1042). The surface area and pore volume were calculated using Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) equations, respectively.

Thermogravimetric analysis (TGA) of the samples was carried out on a thermoanalyzer (TG NETZSCH 209F3, Germany) under 50 cm3/min of flowing argon from room temperature to 1000 °C at a 10 °C/min of heating rate. The different functional groups and anchoring sites were assigned with the aid of a Fourier transform infrared (FT-IR) spectrometer (Tensor 27 model, Bruker Co, Germany) under ambient conditions for wave numbers ranging from 400 to 4000 cm–1. The powder was compressed into a tablet by applying pressure. A sessile droplet was placed on top of the tablet, and the image of the droplet was analyzed by fitting the Laplace equation to the shape with the aid of a contact-angle measuring system (G10 model, Kruss, Germany).

Oil and Hydrocarbon Removal

Oil/hydrocarbon sorption by the γ-Fe2O3/ZIF-7 particles was estimated based on a UV–vis spectrophotometer (CARY-100, Varian Australia Pty Ltd., Australia) measurements of the oil/hydrocarbon content. The samples of oily water models were prepared with different contents of crude oil, acetone, toluene, and a 50/50 mixture of toluene/acetone in water, including 250, 500, 1000, 2000, and 3000 mg/L. The samples were prepared under sonication for 30 min. Then, 0.5 g of γ-Fe2O3 or γ-Fe2O3/ZIF-7 particles was added to the 1000 mL mixture containing the oil. Five samples were collected every 5 min based on the equilibrium time (25 min) over 30 min under sonication. The maximum wavelength was obtained under an intermediate concentration of the oily water sample (1000 mg/L), and a linear calibration curve was constructed. The particles, together with the sorbed oil, were collected by a permanent magnet at the end of every time interval (5 min). The concentration of the remaining oil was measured using UV–vis spectroscopy. A plot of C/C0 versus sorption time was used to study the sorption process of oily pollutants using γ-Fe2O3 NPs and the γ-Fe2O3/ZIF-7 composite. The percent removal of the oil/hydrocarbon was determined as follows (eq ).where C0 is the initial concentration of the oil/hydrocarbon in the mixture and C is the residual concentration after sorption. The γ-Fe2O3/ZIF-7 particles were washed three times with ethanol before they could be reused. The recyclability of the particles was assessed by reusing the particles to separate the oil from water 15 times.

For the regeneration process, the γ-Fe2O3/ZIF-7 particles together with the sorbed oil were mixed with ethanol and sonicated for 30 min. Then, the γ-Fe2O3/ZIF-7 particles were removed using a permanent magnet and dried at 60 °C for 6 h before being reused. The reduction in the sorption efficiency is expressed aswhere %R1 is the removal of oil/hydrocarbon by the γ-Fe2O3/ZIF-7 particles prepared freshly and %R2 is the removal by the recycled particles, as defined by eq . It is noted that oily water mixtures were prepared at the same C0 concentration and 30 min of sorption was used.

Kinetic Study

The kinetics of oil/hydrocarbon sorption was studied by collecting samples every 5 min from the oily water mixtures over 25 min at 25 °C. First- and second-order kinetic models were fitted to the results.[63] First, the mass of the pollutant sorbed per mass of the sorbent was defined as follows.where q is the mass of the sorbed pollutant per mass of the sorbent at any instant in time (mg/g), C0 presents the initial concentration of the pollutant solution (mg/L), C is the concentration of the pollutant at any time (mg/L), M is the amount of the sorbent added (g), and V presents the volume of the solution (L). The linear form of the first-order kinetics is given in eq (63)where qe denotes the mass of the sorbed pollutant per mass of the sorbent at equilibrium, and k1 is the rate constant of first-order sorption (1/min). A plot of Ln(qe – q) versus t should give a straight line with a slope of −k1 and intercept of Ln(qe). Accordingly, k1 and qe can be evaluated from the slope and the intercept. The linear form of the pseudo-second-order kinetics is given in eq .[63]where k2 denotes the pseudo-second-order rate constant of sorption (g/mg min). A plot of t/q versus t gives a straight line with a slope of 1/k2qe2 and intercept of 1/qe. Accordingly, k2 and qe can be evaluated from the slope and the intercept.

Results and Discussion

Particle Characterization

Figure shows the XRD pattern of the γ-Fe2O3 NPs and the γ-Fe2O3/ZIF-7 composite. The XRD fingerprint of the γ-Fe2O3 matches a cubic structure of pure magnetic γ-Fe2O3 indexed with JCPDS no 00-039-1346, as shown in Figure a. According to Scherrer eq , the average crystallite size of the γ-Fe2O3 is 13 nm. The XRD pattern of γ-Fe2O3/ZIF-7 shows the peak positions of the porous ZIF-7 nanostructure with crystallite planes (011), (002), (112), (022), (013), (222), (114), (233), (134), (044), and (244).[64] The ZIF-7 nanostructures do not have an effect on the crystallinity of the γ-Fe2O3 NPs, as depicted in Figure b. In addition, a simple comparison between XRD patterns after 15 recycle runs (Figure c) confirms a similar composite structure as the freshly prepared particles (Figure b). The XRD pattern of recycled γ-Fe2O3/ZIF-7 displays sharper peaks likely due to heating during recycling, which increases the crystallinity. According to Figure , no leftover impurity due to the preparation procedure appears in both the γ-Fe2O3 NPs or the γ-Fe2O3/ZIF-7 particles.

Figure 2

XRD patterns for (a) γ-Fe2O3, (b) freshly prepared γ-Fe2O3/ZIF-7 composite particles, and (c) γ-Fe2O3/ZIF-7 composite particles after 15 cycles of oil separation. Blue circles highlight the ZIF-7 structure.

The mechanism of nucleation and deposition of ZIF-7 rods onto the magnetic NPs is schematically described in Figure . It can be noted that the coating process was carried out without surface modification of the γ-Fe2O3 NPs. The chemical structure shows the binding of Zn agents and benzimidazole ligands with a central porosity and crystal structure illustrated by the unit cell of ZIF-7 with sodalite topology.[65] The schematic of the hierarchical porous structure of ZIF-7 is shown in Figure b,c.

Figure 3

(a) Schematic representation of the buildup stages of the porous γ-Fe2O3/ZIF-7 composite nanostructure, (b) chemical structure, and (c) unit cell of the crystalline structure of ZIF-7.

The morphology of the γ-Fe2O3 NPs and the γ-Fe2O3/ZIF-7 particles is shown in Figure a,b,c–g, respectively. The γ-Fe2O3 NPs display a spherical morphology with an average diameter of 18 ± 0.9 nm, as estimated by Image J software (Figure b). This particle size is close the XRD estimate. Figure c–f shows the hexagonal rod morphology of ZIF-7 structures with a uniform thickness (Figure c–d) that, on a closer look (Figure e–f), roughly ranges in length from 0.5 to 1 μm. These hexagonal rods most likely coat agglomerated γ-Fe2O3 NPs. As evident with increasing magnification, the ZIF-7 rods having hexagonal surface morphology with ranging of 10–100 nm diameter grew onto the γ-Fe2O3 NPs in the C direction. It is likely that the γ-Fe2O3 NPs were agglomerated during the synthesis of the composite structure due to the absence of any capping agents in our synthesis protocol.

Figure 4

FE-SEM micrographs of (a,b) γ-Fe2O3 NPs and (c–f) γ-Fe2O3/ZIF-7 composite nanostructures with a growth time of 1 h at different magnifications.

The effect of the growth time of the ZIF-7 coating is depicted in Figure . Upon increasing the deposition time from 1 to 3 h, an elongated ZIF-7 hexagonal nanorod structure appeared under SEM imaging. Figure shows that the ZIF-7 nanoroads grow in both lengths (0.5–1 and 1–2 to 2–3 μm) and diameters (10–100 and 30–120 to 50–200 nm) with increased coating time (1 and 2–3 h), respectively. Furthermore, EDX analysis provides the elemental composition of the γ-Fe2O3 NPs (Figure a) and the γ-Fe2O3/ZIF-7 composite particles (Figure b), which reveals Zn and N together with Fe within the composite nanostructure. The quantitative analysis of γ-Fe2O3 gives the mass ratios of Fe (71.5%) and O (28.5%), which correspond to Fe2O3, while the mass ratios obtained from the γ-Fe2O3/ZIF-7 were Fe (24.4%), O (12.6%), Zn (20.1%), N (15.6%), and C (27.3%).

Figure 5

FE-SEM micrographs of the γ-Fe2O3/ZIF-7 composite at different coating times of (a) 1, (b) 2, and (c) 3 h.

Figure 6

EDX elemental analysis of (a) γ-Fe2O3 NPs and (b) γ-Fe2O3/ZIF-7 composite particles.

The magnetic properties of the γ-Fe2O3 NPs and the γ-Fe2O3/ZIF-7 composite particles collected on the vibrating sample magnetometer at room temperature are given in Figure . The saturation magnetization of γ-Fe2O3 and γ-Fe2O3/ZIF-7 at an external field of 5 KOe is 63 and 45 emu/g, respectively, without magnetic hysteresis, suggesting superparamagnetic behavior.[66] Generally, for coated particles, the magnetic properties depend on the particle size, specific surface area coverage, the degree of interaction of ZIFs with the γ-Fe2O3 NP, and the extent of the order of the ZIF coat.[66] Hence, with more loading of porous ZIF-7 nanorods, the magnetic properties of the γ-Fe2O3 weakened from 45 to 33 emu g–1 due to increased length and agglomeration. Therefore, the above results suggest that γ-Fe2O3/ZIF-7 prepared after 1 h precipitation time exhibits superparamagnetic properties and can be effectively removed by magnetic separation. Accordingly, this composite structure which can be easily removed by a magnet and display proper coverage of absorbent material is selected to perform further separation.

Figure 7

Magnetization curves (vibrating sample magnetometry) of γ-Fe2O3 NPs and the γ-Fe2O3/ZIF-7 composite with different coating times of 1, 2, and 3 h.

The surface area analysis of the γ-Fe2O3/ZIF-7 with 1 h coating time was carried out. The N2 adsorption/desorption isotherms of the γ-Fe2O3 NPs and the γ-Fe2O3/ZIF-7 particles are shown in Figure . Figure a reveals a H3-type hysteresis,[67,68] which suggests that the γ-Fe2O3 NPs have negligible pore volume and almost all the 55 m2/g surface area, as calculated based on the BET equation, is externally exposed. Assuming spherical particles, as suggested by SEM results in Figure a,b, this surface area corresponds to an average particles size of 20 nm, which is not very different from the SEM-obtained amounts. The γ-Fe2O3/ZIF-7 particles, on the other hand, display a H4-type hysteresis, which suggests type I and II isotherms with noticeable adsorption at low P/P0 and a large specific surface area of 647 m2/g, as calculated using the BET equation. The difference in surface area estimated between γ-Fe2O3 and γ-Fe2O3/ZIF-7 can be explained in light of the SEM results. The huge increase of the surface area corresponds to the ZIF-7 nanorods loaded onto the magnetic particles with microporous and mesoporous structures having average sizes of 1.8 and 4.5 nm, respectively, as confirmed by the pore volume curve in Figure b.

Figure 8

N2 adsorption–desorption isotherms and BJH of (a) γ-Fe2O3 NPs and (b) γ-Fe2O3/ZIF-7 composite measured at 77 K.

The thermal stability of γ-Fe2O3, ZIF-7, and γ-Fe2O3/ZIF-7 was studied using their TGA profiles under an inert atmosphere. Figure confirms that γ-Fe2O3 particles display excellent thermal stability with no weight loss up to 1000 °C, as also reported in the literature.[69,70] Accordingly, γ-Fe2O3 NPs make a reliable support material. The TGA profile for ZIF-7, on the other hand, displays a three-step mass loss of 6 wt % (50–300 °C), 10 wt % (590–700 °C), and 40 wt % (700–900 °C). The first two steps are attributed to the removal of free and terminal H2O and DMF molecules, while the last step relates to the decomposition of the ZIF-7 framework.[71] Some literature pointed to the formation of ZnO as a decomposition product of ZIF-7,[28,71,72] which is doubtful under an inert atmosphere, especially since no oxygen atoms exist in the structure of ZIF-7.

Figure 9

TGA profiles under an inert atmosphere for γ-Fe2O3 NPs, ZIF-7, and γ-Fe2O3/ZIF-7.

The TGA results of γ-Fe2O3/ZIF-7 display a shift toward higher temperatures relative of the ZIF-7 decomposition profile with the two major mass loss regions of 35 wt % shifting to 610–710 °C and 25 wt % shifting to 710–930 °C. This shift suggests a more stable ZIF-7 structure onto the γ-Fe2O3 particles than a standalone structure. The TGA profile of γ-Fe2O3/ZIF-7 helps determining the ZIF-7 content attached to the surface of the γ-Fe2O3 particles, especially since the mass loss profile displays the three step loss encountered for the standalone ZIF-7 structure. For the γ-Fe2O3/ZIF-7, ∼45% of the mass remains, which could be attributed to the ZIF-7 pyrolysis product and the original γ-Fe2O3 particles. It is noted that in the presence of the ZIF-7 coating, γ-Fe2O3 particles may act as an oxidant and contribute to the formation of CO2.[73]

According to the empirical formula, only 15% of this weight loss can be attributed to the removal of DMF solvent molecules.[72] The mass losses indicate that the weight changed, mainly as a result of the removal of DMF (<200 °C) and thermal decomposition of the frameworks (>600 °C).

Figure shows the FT-IR spectra of benzimidazole, ZIF-7 crystals, γ-Fe2O3 NPs, and γ-Fe2O3/ZIF-7 composite particles. In the FTIR spectra of benzimidazole, the peaks at 1500–1600 and 2500–3500 cm–1 are identified as the C=C stretch band of the aromatic ring of benzimidazole and the stretch band of the N–H group, respectively.[74] The C–H stretch vibration of benzimidazole occurs in the wavelength range of 3000–3100 cm–1, while the peak in the range of 1580–1650 cm–1 is attributed to the bending vibration of the N–H group.[74] For the ZIF-7 spectrum, the disappearance of the strong and broad N–H band between 2500 and 3250 and 1580–1650 cm–1 compared to benzimidazole suggests that benzimidazole has been fully deprotonated during the crystallization and leads to the formation of ZIF-7.[74] The ZIF-7 peaks in the range of 700–1700 cm–1 are attributed as follows. The peak at 777 cm–1 corresponds to the C=H band, and the peak at 1455 cm–1 is attributed to the C–C band arising from the benzene functional group of benzimidazole of the ZIF-7 crystals.[28] For the γ-Fe2O3 NPs, the Fe–O vibration corresponds to the peak at 587 cm–1.[60,75] The other peaks at 632, 795, 892, and 1629 cm–1 are attributed to the other peaks at pure maghemite.[75,76] Absorption peaks at 3170 and 3408 cm–1 are attributed to the hydroxyl group (−OH).[75] The γ-Fe2O3/ZIF-7 peaks display a slightly different pattern from γ-Fe2O3 NPs, which depicts most of the major ZIF-7 peaks.

Figure 10

FTIR spectra of benzimidazole, γ-Fe2O3 NPs, ZIF-7, and γ-Fe2O3/ZIF-7.

Oil/Water Separation

The γ-Fe2O3/ZIF-7 particles were dispersed into water by vigorous shaking but mostly separated at the surface over 24 h of sitting time (Figure ), suggesting a highly hydrophobic material[25] coupled with low particle density. The superhydrophobicity of a solid surface depends on its surface roughness and surface energy.[77,78] In addition, the results of the oil contact angle (OCA) of γ-Fe2O3 NPs and γ-Fe2O3/ZIF-7 in Figure a,b, respectively, confirmed the oleophilicity (hydrophobicity) of γ-Fe2O3/ZIF-7. As shown, the OCA of γ-Fe2O3 NPs is higher than 150° with superhydrophilic properties and the γ-Fe2O3/ZIF-7 composite is oleophilic because the OCA is close to zero (Table ). These results depict that the magnetic γ-Fe2O3 NP have been successfully coated with ZIF-7 structures as a hydrophobic layer. Consequently, the γ-Fe2O3/ZIF-7 composite can selectively sorb oil and repel water.

Figure 11

Images of the γ-Fe2O3/ZIF-7 particles dispersed into DI water after (a) 0, (b) 12, and (c) 24 h.

Figure 12

OCA of (a) γ-Fe2O3 NPs and (b) γ-Fe2O3/ZIF-7 composite.

Table 1

OCA of the γ-Fe2O3 NPs and γ-Fe2O3/ZIF-7 Composite

sorbent	OCA (deg)
γ-Fe2O3	3 ± 0.5
γ-Fe2O3/ZIF-7	155 ± 0.5

Oil and Hydrocarbon Removal

The removal of crude oil and different model pollutants, including toluene, acetone, and a 50/50 mixture of toluene/acetone, using the γ-Fe2O3 NPs and γ-Fe2O3/ZIF-7 particles was evaluated. Emulsified droplets of the organic pollutants in DI water under 30 min sonication appeared milky (Figure a). The sorption of the pollutants was accomplished by adding 0.5 g of the γ-Fe2O3 NPs or γ-Fe2O3/ZIF-7 particles under the same sonication, followed by extracting the particles using a permanent magnet, as detailed in the experimental methodology. Oil or hydrocarbon content was measured before particle addition and after particle removal using UV–vis spectroscopy every 5 min. A plot of C/C0 versus time for the γ-Fe2O3 NP and γ-Fe2O3/ZIF-7 particle sorption is given in Figures and 15, respectively. As shown in Figure , the γ-Fe2O3 NPs show very little sorption affinity for the different pollutants, whereas the γ-Fe2O3/ZIF-7 particles showed the highest affinity to toluene, acetone, 50/50 toluene/acetone mixture, and then oil (Figure ).

Figure 13

Images of (a) milky crude oil emulsion and (b–d) sorption and separation of the γ-Fe2O3/ZIF-7 particles.

Figure 14

Removal of (a) crude oil, (b) acetone, (c) toluene, and (d) 50/50 acetone/toluene mixture by the γ-Fe2O3 NPs.

Figure 15

Removal of (a) crude oil, (b) acetone, (c) toluene, and (d) 50/50 acetone/toluene mixture by the γ-Fe2O3/ZIF-7 particles.

The rate of sorption of the different pollutants can be determined from Figure . It is noted that sorption is rapid during the first 5 min, but then, it slows down. It is also evident that sorption is rapid for the solution with high pollutant content at 30 min. The higher rate of sorption in the first 5 min can be attributed to the strong interactions between the aromatic frameworks and the pollutant molecules. For times >15 min, more than 90% of the equilibrium pollutant uptake was sorbed. Hence, 15 min of sorption presents a practical sorption time limit. The equilibrium pollutant uptake was determined theoretically using eqs and E6, as detailed earlier. The hierarchical porous structure of ZIF-7 onto the γ-Fe2O3 as is shown in Figure . Decorating the NPs with ZIF-7 enhanced the hydrophobicity of the sorbent and hence the sorption capacity. Also, the sorption rate of the pollutant also increased owing to the hierarchical porous structure. Hierarchical-structured pores improve pore diffusion, especially into larger pores.

The sorption rate by γ-Fe2O3/ZIF-7 can be compared with other literature reports. The sorption rate by γ-Fe2O3/ZIF-7 is compared with other adsorbents in the literature. A magnetic demulsifier prepared from the Fe3O4 NP covered with silica showed an equilibrium time of at least 4 h for oil removal.[79] The magnetic nanocomposite Fe3O4@ZIF-8 was used as an adsorbent for the fast adsorption of UO22+ ions from aqueous solutions within 2 h.[80] The adsorption kinetics of As(III) by Fe3O4@ZIF-8 increased the adsorption rate in the first 60 min, and equilibrium was attained in 4 h.[61]

The main interactions responsible for the sorption of oily pollutants and hydrocarbons are schematically represented in Figure . The evaluation of the OCA of the sorbent in Figure and the difference in the sorption capacity in Figures and 15 demonstrate that hydrophobic interactions are the main reason behind the increased capacity/uptake of the sorbent.

Figure 16

Hydrophobic mechanism for the sorption of the oil/hydrocarbon pollutant using the γ-Fe2O3/ZIF-7 composite.

Table includes various magnetic and porous materials with similar structures to the γ-Fe2O3/ZIF-7 composite used for oil/water separation. As suggested by Table , the γ-Fe2O3/ZIF-7 composite portrays exceptional capacity for oil sorption. The surface wettability, that is, hydrophobicity–oleophilicity, of the sorbent, which is typically endowed by functionalizing the surface with surfactants plays a crucial role. The sorption capacities obtained in this study are comparable to those obtained in the literature for ZIF-8 particles with the separation capacity between 70 and 250 wt %.[81] The Fe3O4/PS magnetic NPs displayed a removal of as high as 240 wt % for different organic solvents having a water contact angle of 131.2°.[25] The separation capacities of organic solvents and oils ranging between 36–58 and 150–600 wt %, respectively, were reported for ZIF-8/carbon nitride foam[82] and HFGO@ZIF-8.[83] In the present work, surfactant-free synthesis was adopted and the resultant hierarchical cavities of the ZIF-7 oleophilic coating contributed to unique oil sorption capacity. In addition, the cavities created by the interconnected ZIF-7 fibers provide a large volume for the storage of oil molecules.

Table 2

Various Porous Materials as Oil/Organic Sorbents

sorbents	water contact angle (deg)	surface area (m2/g)	pollutant type	separation (wt %)	reference (year)
γ-Fe2O3/ZIF-7 particles		690	crude oil and hydrocarbons/water	600	present work
Fe3O4/PS magnetic NPs	131.2		organic solvents	240	(25)
ZIF-8/carbon nitride foam	138	211	different oils/water	36–58	(84)
HFGO@ZIF-8	125	590	nonpolar and polar organic solvents and oils	150–600	(83)
ZIF-8 particles	142	1408	different oils	70–250	(81)

The results from fitting the γ-Fe2O3/ZIF-7 particle sorption data to the first-order and pseudo-second-order kinetic models are illustrated in Figure . As shown in Table , the coefficient of determination, R22, consistently >0.96 for pollutants at all concentrations suggests that a pseudo-second-order model better fits the sorption data. Moreover, the equilibrium sorption capacity for the pseudo-second-order model is slightly closer to the experimental data. It is noted that ultrasonic mixing is essential to attain a sufficiently uniform mixture for accurate and reproducible UV–vis measurements of the oil content. Ultrasonic mixing does not promote pore diffusion and, hence, adsorption kinetics onto the γ-Fe2O3/ZIF-7 sorbent.

Figure 17

Matching experimental sorption data of the removal of crude oil: acetone; toluene; and 50/50 acetone/toluene mixture by the γ-Fe2O3/ZIF-7 particles with (a,c,e,g) first- and (b,d,f,h) pseudo-second-order kinetic models.

Table 3

Kinetic Parameters of Pseudo-First-Order (R12) and Pseudo-Second-Order (R22) and k1 (1/min) and k2 (g/mg min) for the Sorption of Different Pollutants from Oily Water Samples Using γ-Fe2O3/ZIF-7 Particles

	crude oil				acetone				toluene				50/50 acetone/toluene mixture
C0 (mg/L)	R12	R22	k1	k2	R12	R22	k1	k2	R12	R22	k1	k2	R12	R22	k1	k2
250	0.924	0.992	0.131	0.020	0.936	0.993	0.097	0.020	0.966	0.990	0.056	0.020	0.979	0.993	0.066	0.020
500	0.959	0.999	0.101	0.013	0.939	0.989	0.074	0.010	0.936	0.989	0.050	0.010	0.942	0.990	0.082	0.010
1000	0.931	0.964	0.094	0.005	0.963	0.983	0.088	0.005	0.945	0.989	0.072	0.005	0.982	0.993	0.072	0.005
2000	0.937	0.982	0.085	0.003	0.971	0.978	0.095	0.002	0.966	0.987	0.104	0.002	0.951	0.988	0.084	0.002
3000	0.953	0.961	0.084	0.002	0.957	0.963	0.112	0.001	0.966	0.971	0.108	0.001	0.988	0.984	0.104	0.001

Regeneration Process

Following the sorption step, the regeneration of γ-Fe2O3/ZIF-7 particles was evaluated. The results of sorption efficiency as calculated from eq are summarized in Table . 15 use and reuse cycles corresponded to only 9% reduction in the performance. This observation agrees well with the SEM photographs of recycled γ-Fe2O3/ZIF-7 particles in Figure . Rough particles of the same ranging sizes appeared similar to the freshly prepared particles (Figure c–f). It is the evidence that micrographs show only a little loss of ZIF-7 rods due to frequent interactions. The ZIF-7 particles remained intact, indicating good adhesion to the surface of the γ-Fe2O3 NPs.

Table 4

Drop in the Sorption Efficiency, η, after 15 Cycles of γ-Fe2O3/ZIF-7 Particle Reuse

pollutant	η for C0 = 200 mg/L	η for C0 = 3000 mg/L
crude oil	3.1	9
acetone	2	7.5
toluene	2.2	7.5
toluene/acetone (50/50)	2.5	7.8

Figure 18

FE-SEM micrographs of the γ-Fe2O3/ZIF-7 composite after 15 cycles of use and reuse during oil sorption.

Conclusions

This study reported the fabrication of superparamagnetic γ-Fe2O3/ZIF-7 composite particles with controllable structures. The ZIF-7 nanorods were in situ grown onto dispersed γ-Fe2O3 NPs by the addition of the precursors and controlling the time and the temperature without physically or chemically modifying the surface. The resultant γ-Fe2O3/ZIF-7 nanostructures maintained porous and highly crystalline morphology, which effectively accommodated the guest molecules. The combination of ZIF-7 nanorods and the magnetic γ-Fe2O3 enabled selective sorption of the organic contaminants (vs water) and ease of separation by means of magnetic force, especially for nanorods grown for no more than 1 h. The results showed that the γ-Fe2O3/ZIF-7 composite is an effective and selective sorbent for different organic pollutants from water, including crude oil, acetone, toluene, and 50/50 acetone/toluene. Rapid sorption at an equilibrium time of 25 min, which could be fit by the pseudo-second-order kinetic model, was reported in the first 5 min for different concentrations of pollutants. 15 cycles of use and reuse without loss of the ZIF-7 coating and with <9% reduction in the sorption efficiency confirmed the successful regeneration of the γ-Fe2O3/ZIF-7 sorbent.