Ethanol, methanol, cobalt nitrate hexahydrate (Co(NO₃)₂·6H₂O), nickelous nitrate (Ni(NO₃)₂·6H₂O), and 2-methylimidazole were obtained from Aladin Ltd. (Shanghai, China). All reagents were of analytical grade and were used directly without further purification.

The morphology and microstructure of the products were examined using scanning electron microscopy (SEM, Hitachi S4800) and transmission electron microscopy (TEM) coupled with energy-dispersive X-ray spectroscopy (EDX). X-ray diffraction (XRD) patterns were collected on a D8 Focus (Bruker) diffractometer with Cu Kα radiation. The X-ray photoelectron spectrum (XPS) was recorded using an AXIS Ultra instrument (Kratos Analytical). Nitrogen adsorption-desorption was performed at liquid nitrogen temperature with a Micromeritics ASAP 2020 M automated sorption analyzer. Fourier transforms infrared (FT-IR) spectra were obtained on Thermo Fisher Nicolet-6700. The thermogravimetric analyses (TGA) were carried out using a TG/SDTQ600 instrument in nitrogen atmosphere within temperature range of 25~600 ℃. The ultrviolet-visible (UV-Vis) spectroscopic measurements were recorded using a JASCO V-550 UV-Vis spectrometer.

In a typical experiment^[30], two solutions were first prepared by separately dissolving 291 mg of Co(NO₃)₂·6H₂O in 15 mL of methanol and 492 mg of 2-methylimidazole in 15 mL of methanol. The solution of Co(NO₃)₂·6H₂O was added into the solution of 2-methylimidazole drop by drop under stirring. The mixed solution was stirred for 30 min and aged for 12 h at room temperature. The product was collected by centrifugation and washed with methanol 3 times and then dried at 60 ℃ overnight^[28].

The as-synthesized crystals of ZIF-67 (100 mg) were dispersed in 15 mL of ethanol containing 300 mg of Ni(NO₃)₂·6H₂O. After ultrasounding for 5 min, the mixture was transferred into a 20 mL Teflon reactor. The reactor was heated to 80 ℃ and maintained at 80 ℃ for 3 h. The product was collected by centrifugation, washed with anhydrous ethanol and dried at 60 ℃ overnight^[31].

A 5 mg of NiCo-LDH@ZIF-67 sample each was added into aqueous methyl orange solutions (5 mL each) with different initial concentrations and pH. The adsorbent-solution mixture was carefully sonicated for different periods of time, and then the NiCo-LDH@ZIF-67 adsorbent was separated from the solution by filtration using a 0.22 μm membrane. The supernatant was analyzed using a UV-Vis spectrometer. The concentrations of each dye before and after adsorption were obtained using a standard calibration curve. The adsorption percentage ( r) and adsorption capacity ( Q_t) of dye were calculated through the following equations:

$r=[(ρ_0-ρ_t)/ρ_0]×100$ (1)

$Q_t=(ρ_0-ρ_t)V/m$ (2)

where ρ₀ and ρ_t (mg/L) represent the initial and equilibrium mass concentrations of dye, respectively, Q_t (mg/g) is the adsorption capacity (mg/g), and V and m represent the volume (L) of the solution and the mass (g) of adsorbent used, respectively^[21].

To evaluate the adsorption kinetics of MO on NiCo-LDH@ZIF-67, the experimental data was fitted with the pseudo-first-order model and pseudo-second-order mode represented separately as follows:

The pseudo-first-order:

logqe-qt=logqe-k1t(3)

The pseudo-second-order:

t/qt=1/(k2qe2)+t/qe(4)

where q_e and q_t were the amounts of methyl orange adsorbed (mg/g) at equilibrium and at time t (min) respectively, and k₁ was the pseudo-first-order rate constant (min); k₂ was the pseudo-second-order rate constant (g·mg/min)^[32].

The Langmuir model, Freundlich model were commonly used to investigate the adsorption behavior of MO.

The Langmuir equation:

ρe/qe=1/(qmKL)+ρe/qm(5)

The Freundlich equation:

lgqe=1/nlgρe+lgKF(6)

where ρ_e was the equilibrium concentration of adsorbates in the solution (mg/L), q_e denoted the equilibrium adsorption capacity of adsorbates (mg/g) of the adsorbent, q_m was the monolayer adsorption capacity of adsorbent (mg/g), K_L was the Langmuir adsorption equilibrium constant (L/mg); K_F and n were Freundlich constant and intensity factor, respectively^[33].

The SEM image (Fig.1A) indicates that the as-prepared ZIF-67 particles have a typical rhombic dodecahedral shape with a smooth surface and that the dodecahedral particles have a narrow size distribution of around 300 nm. The particles of NiCo-LDH@ZIF-67 (Fig.1B) basically remain unchanged compared with ZIF-67 except for surface becoming rougher, indicating well inheritance of morphology and dimension of the ZIF-67. When Ni(NO₃)₂ is added to the solution containing ZIF-67, protons which was generated from the hydrolysis of Ni²⁺ ions will etch the ZIF-67 and the ZIF-67 nanoparticles become hollow. The Co²⁺ ions which was released from ZIF-67 are oxidized to Co³⁺ by NO₃^- ions and O₂. The amount of OH^- ions in the solution increases along with continuous consumption of protons. As a consequence, and the LDH was formed by the co-precipitation of Co, Ni ions around the ZIF-67 particles^[34]. The so-formed composite material has more adsorption sites for MO. As the nanosheet grows on the facets of the polyhedrons of the ZIF-67 particles, the hollow inner of the ZIF-67 particle can be clearly visualized in the TEM images (Fig.1D). The NiCo-LDH@ZIF-67 particles so obtained have thin nanosheet cover of NiCo-LDH and the thickness of NiCo-LDH nanosheet cover is about 60 nm.

Fig.1

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.1 SEM images of ZIF-6(A) and NiCo-LDH@ZIF-67(B); TEM images of ZIF-67(C) and NiCo-LDH@ZIF-67(D)

As shown inFig.2A, the diffraction peaks of the synthesized ZIF-67 are consistent with the single crystal data of ZIF-67 (CCDC 671073). After the reaction of ZIF-67 with nickelous nitrate, the diffraction peaks of ZIF-67 totally disappear because of the outward migration of the Co²⁺ ions. The broad diffuse background between 15° and 35° reveals the existence of some amorphous products. The diffraction peaks of NiCo-LDH are consistent with the standard card (JCPDS No.14-0191). The diffraction peaks of the (003), (006) and (009) crystal planes indicate the 2D lamellar structure^[35]. The peak at around 60° is attributed to (110) crystal plane^[36]. The existence of these peaks suggests that the synthesized NiCo-LDH@ZIF-67 materials contain components having typical hydrotalcite-like structures.

The FT-IR spectrum of NiCo-LDH@ZIF-67 is shown inFig.2B, all of the characteristic peaks for ZIF-67 disappear after reacting with Ni(NO₃)₂. The broad band at 3700~3200 cm^-1 and the peak at 1632 cm^-1 belong to O—H stretching of interlayer water molecules. The peaks at 1492 cm^-1 is attributed to C=O. The sharp peak at 1383 cm^-1 is ascribed to N—O stretching vibration of NO₃^- and the peak at 1281 cm^-1 is related to C—O stretching vibration^[16].

Fig.2

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.2 (A)XRD patterns of NiCo-LDH@ZIF-67, ZIF-67 and NiCo-LDH; (B)FT-IR spectra of NiCo-LDH@ZIF-67

Typical XPS spectra of NiCo-LDH@ZIF-67 are also conducted and shown inFig.3A-3C. The XPS survey spectrum (Fig.3A) confirms the presence of C, N, O, Co, and Ni. As shown inFig.3B, the binding energy located at 797.4 and 781.2 eV are indexed to Co³⁺ and Co²⁺ oxidation states, respectively. The intensities of the two satellite peaks at 803.0 and 786.9 eV (indicated as “Sat”) are very low, revealing the coexistence of Co²⁺ and Co³⁺ in the NiCo-LDH@ZIF-67 particles. In Ni2 p XPS spectrum of NiCo-LDH@ZIF-67 (Fig.3C), the two peaks centered at 873.3 and 856.0 eV can be attributed to Ni2 p_1/2 and Ni2 p_3/2, respectively. The other higher binding energies at 880.2 and 862.0 eV are assigned to the shake-up satellite^[31,35].

Fig.3

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.3 (A)The survey XPS spectra of NiCo-LDH@ZIF-67; (B)Co2 p and (C)Co2 p peak fitting of NiCo-LDH@ZIF-67

To characterize the porosity of the NiCo-LDH@ZIF-67, nitrogen adsorption-desorption was performed. The adsorption-desorption isotherm and pore size distribution of the NiCo-LDH@ZIF-67 are presented inFig.4A and 4B. It is clear that the adsorption-desorption isotherms of the material are of type IV according to the IUPAC classification of adsorption isotherms, suggesting the presence of mesopores^[14]. Further, the isotherms exhibit type H3 hysteresis loop over the relative pressure range of 0.45~0.99, indicating the coexistence of non-structural mesopores^[37]. The mesopores provide easy transportation for MO, conducive to the adsorption process. Based on the nitrogen quantity adsorbed at different relative pressures, the BET surface areas and pore volume of the NiCo-LDH@ZIF-67 are calculated to be 170.77 m²/g and 0.39 m³/g, respectively. The pore-size distribution of NiCo-LDH@ZIF-67 shows that the pore size is centered around 10.63 nm. The smaller mesopores reflected the presence of pores within nanosheets, while larger mesopores are associated with the pores formed between stacked nanosheets.

Fig.4

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.4 N2 sorption isotherms(A) and pore size distribution(B) of NiCo-LDH@ZIF-67

The TGA curve of NiCo-LDH@ZIF-67 sample is shown in Supporting Information Fig.S1. The first mass loss of NiCo-LDH@ZIF-67 could be resulted from the desorption of water from the interlayers and surface with temperature lower than 100 ℃^[38]. The second weight loss (about 30%) ranging from 200 to 300 ℃ should be attributed to the decomposition of NO₃^- resulting in collapse of nanosheets^[39]. The third mass loss ranging from 300 to 500 ℃ indicated the collapse of the organic carbon skeleton. It is worth noting that pure ZIF-67 usually loses mass above 500 ℃^[38]. Instead, no extra weight loss was observed for composite NiCo-LDH@ZIF-67 at temperature higher than 500 ℃^[37]. The differences in stability means the successful synthesis of NiCo-LDH@ZIF-67.

Generally, the removal of dye from waste water was dependent on the initial concentration of the dye. For determining the effect of initial concentration of MO in solution, the adsorption experiments were performed in several initial concentrations varying from 300 to 2400 mg/L at 30 ℃ for 20 minutes. The results were shown inFig.5A. The adsorption capacity of MO adsorption increases with the increase of initial dye concentration. The adsorption capacity increases rapidly with the initial MO concentration from 0 to 1200 mg/L, above which it increases slowly. This can be attributed to the high concentration difference and the presence of unoccupied adsorption sites^[17]. The maximum adsorption capacity of NiCo-LDH@ZIF-67 for methyl orange is 1274 mg/g, which is much better than that of ZIF-67 (168 mg/g) for methyl orange ( see Supporting Information Fig.S2). There were two models, Langmuir and Freundlich, to describe the relationship between the MO adsorption amount by the adsorbent and the MO equilibrium concentration in aqueous solution, and the results are shown inFig.5B, 5C and Table 2. Obviously, the equilibrium data for the adsorption of MO on adsorbents can be well fitted by the Langmuir model with correlation coefficients values ( R²) being higher than 0.999, and the q_e values for the adsorption of MO by NiCo-LDH@ZIF-67 were 1301 mg/g, which is the same as the experiment data 1274 mg/g. These results demonstrate that NiCo-LDH@ZIF-67 is homogeneous with the adsorption mechanism of monolayer uptake.

Fig.5

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.5 (A)Concentration effects on MO adsorption of NiCo-LDH@ZIF-67; (B)Adsorption isotherms fitted by Langmuir and (C)Freundlich model

Table 2

Table 2

Table 2 Isotherm parameters for the Langmuir, Freundlich and Tempkin models Langmuir Freundlich Qmax/(mg·g-1) KL/(L·mg-1) R2 KF n R2 1301.165 0.030902 0.99961 206.8719 3.057 0.78014

Table 2 Isotherm parameters for the Langmuir, Freundlich and Tempkin models

The effect of adsorption time on the adsorption capacities of MO is shown inFig.6A. The adsorption rate is fast during the first 3 minutes and reaches equilibrium after 15 minutes. Different concentrations of 600, 1200 and 1800 mg/L were selected as the initial MO mass concentrations to study the adsorption kinetics. Reasons accounting for the change in adsorption rate can be described as follow:At the beginning, adsorption rate of MO on NiCo-LDH@ZIF-67 was high due to more vacant adsorption sites available, whereas after a period of time, adsorption rate was getting slower. The remaining vacant adsorption sites were getting difficult to be occupied because of the repulsion between the solute molecules on the solid and bulk phases^[2,21]. To analyze the adsorption kinetics, the pseudo-first-order and pseudo-second-order kinetic models were tried to fit the adsorption data.Fig.6B and 6C shows the fitting results, and the parameters are given inTable 3. The R² obtained from the pseudo-second-order kinetics model were closer to unity ( R²>0.999), compared with that from the pseudo-first-order kinetic model. The calculated value of q_e was in good agreement with the actual operation data. This result indicates that adsorption of MO onto NiCo-LDH@ZIF-67 follows the pseudo-second-order kinetic model. As shown in the XRD data of NiCo-LDH@ZIF-67 before and after the dye adsorption ( see Supporting Information Fig.S3), after adsorption with MO, the intensity of XRD peaks obviously reduces. Still, it is obvious that the d₀₀₃ peak shifted to a smaller degree compared with the original sample. This is attributed to the reason that a part of MO ions are into the interlayer of LDH, which to increase of the interlayer distance by chemisorption. The possible adsorption mechanism is that MO molecules are arranged parallelly or tipsily in the interlayer space through ion exchange^[10,15,40]. Compared with ZIF-67, when MO is adsorbed by ZIF-67, it almost does not rely on ion exchange, which may be the reason why the adsorption capacity of ZIF-67 is much smaller than that of NiCo-LDH@ZIF-67^[41]. In addition, the composite NiCo-LDH@ZIF-67 prepared with ZIF-67 as template has higher BET surface areas and more mesopores than 3-dimensional NiCo-LDH, which will be beneficial to the exposure of more adsorption sites and the transport of MO molecules. Therefore, NiCo-LDH@ZIF-67 has a larger adsorption capacity than 3-dimensional NiCo-LDH^[9].

Fig.6

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.6 (A)The effect of contact time on the adsorption of MO on NiCo-LDH@ZIF-67; (B)the kinetics models for adsorption of MO on NiCo-LDH@ZIF-67 pseudo-first-order kinetic mode and (C)pseudo-second-order kinetic mode

Table 3

Table 3

Table 3 Kinetic parameters for the pseudo-first-order kinetic model, pseudo-second order kinetic model MO mass concentration/ (mg·g-1) qm/ (mg·g-1) Pseudo-First order model Pseudo-Second order model qe k1 R2 qe k2 R2 600 569.26 14.43 0.0591 0.82181 568.18 0.0401 0.99997 1200 1053.60 568.7 0.1605 0.93495 1069.04 0.0030 0.99909 1800 1274.79 412.9 0.0852 0.98806 1292.29 0.0016 0.99779

Table 3 Kinetic parameters for the pseudo-first-order kinetic model, pseudo-second order kinetic model

In order to know the pH impact on the adsorption capacity of MO on NiCo-LDH@ZIF-67, different pH values ranging from 3 to 11 were examined. The pH of the MO solution was adjusted with 0.1 mol/L NaOH or 0.1 mol/L HCl aqueous solutions at temperature 30 ℃ and fixed contact time, 20 min. The data show that when the pH increased from 2 to 4, the adsorption of MO increased from 1724 to 1766 mg/g on NiCo-LDH@ZIF-67, and the adsorption reached the highest value at pH=4.0(Fig.7A). At pH<4, on the one hand, MO is positively charged while the surfaces of NiCo-LDH@ZIF-67 are also positively charged. With the increasing pH from 2 to 4, it is obvious that MO carries fewer and fewer positive charges, the increasing adsorption of MO on NiCo-LDH@ZIF-67 are attributed to the electrostatic attraction^[42]. At pH>4.0, the adsorption amount was decreased with the increasing pH value ( see Supporting Information Fig.S4). This decrease in adsorption at higher pH is attributed to repulsion between the dye molecules (MO) and the negatively charged adsorbent surface^[15].

Selective adsorption ability is a crucial factor affecting the adsorption performance of NiCo-LDH@ZIF-67. A 5 mg NiCo-LDH@ZIF-67 were added into 5 mL of the aqueous mixture of MB and MO with a concentration of 50 mg/L. After adsorption at 30 ℃ for 20 min, the solid was separated out from the solution by filtration using a 0.22 μm membrane. A mixture of MB and MO was chosen because of their opposite charges but similar size. As showed inFig.7B, the color of MB/MO mixture solution turned from green to blue after separation. Besides, the color of NiCo-LDH@ZIF-67 became yellow, indicating a good separation phenomenon. The adsorption percentage of MO is more than 92% while the adsorption percentage of MB is 10%, which can be contributed to the electronic interaction. Generally, NiCo-LDH@ZIF-67 has a positive surface charge, so that MO as an anionic dye can be adsorbed over NiCo-LDH@ZIF-67 by electrostatic attraction. However, MB is a cationic dye, the interaction between MB and NiCo-LDH@ZIF-67 is electrostatically repulsive. Another possible adsorption mechanism is that the adsorption of MO on NiCo-LDH@ZIF-67 is by electrostatic attraction. However, there is a small amount of MB adsorbed, which indicates that other interaction mechanisms must be involved in the adsorption process^[15,43].

Fig.7

	Figure Option ViewDownloadNew WindowDownload As Powerpoint Slide
	Fig.7 (A)Effect of pH on adsorption of MO on NiCo-LDH@ZIF-67; (B)The selective adsorption capacity of NiCo-LDH@ZIF-67 toward the mixed dyes of MO/MB; (C)The recyclability of NiCo-LDH@ZIF-67

The regeneration of an adsorbent is an important step for practical applications. Therefore, the regenerative experiments were carried out. After adsorption, the desorption of MO from the solid particles was performed by washing with ethylene glycol and ethanol at 80 ℃ several times. Then the regenerated NiCo-LDH@ZIF-67 nanoparticles were separated out by centrifugation, and followed by drying. The adsorption result of the regenerated samples is shown inFig.7C. The adsorption percentage of NiCo-LDH@ZIF-67 is lowered by about 9% after four regeneration cycles, indicating that the regeneration might not be completed under this condition and MO might still exist in the NiCo-LDH@ZIF-67. The large loss may also be due to the loss of the adsorbent during washing and the progressively decreasing crystallinity of the NiCo-LDH@ZIF-67 in the heated regeneration. XRD pattern of the NiCo-LDH@ZIF-67 after regeneration ( see Supporting Information Fig.S5) showed that the NiCo-LDH@ZIF-67 can retain its structure during the regeneration process.