Mixed-Ligand Strategy for the Creation of Hierarchical Porous ZIF-8 for Enhanced Adsorption of Copper Ions.

The adsorption of heavy metals using metal-organic framework-based adsorption technology has been pointed out as a promising technique for the removal of these toxic elements from water. However, their adsorption capacity needs to be enhanced. Thus, the current work reports the effect of using a mixed-ligand strategy on the MOF framework and its effect on the removal of copper ions from water by adding terephthalic acid (BDC) linker to the ZIF-8precursors (2-methylimidazole (mI) and Zn2+) under solvothermal synthesis, leading to the formation of a hierarchical microporous mesoporous MOF, named Zn-mI-BDC, which was characterized by SEM, EDX, XRD, TGA, BET, and FTIR. As a result, all of these techniques revealed that the addition of a controlled amount of BDC did not alter the crystallinity of ZIF-8, resulting in the creation of a pore size of 4.2 nm. The new hierarchical porous MOF was tested for the adsorption of copper and exhibited an enhanced adsorption capacity compared to pristine ZIF-8 and many other standard adsorbents. The adsorption isotherm matched well with the Langmuir isotherm model, suggesting that the adsorption process chemisorption had a dominant role in the adsorption of Cu2+ species. Therefore, the current work is considered as an important step toward the use of a mixed-ligand strategy in enhancing the adsorption capacity of heavy metals using MOF materials.

Introduction

During the past two decades, extensive work has been performed on the synthesis and study of materials known as metal–organic frameworks (MOFs), which have emerged as a new family of crystalline hybrid organic/inorganic porous materials with outstanding performance in various potential applications including catalysis,[1,2] adsorption,[3−5] sensing,[6] and so on,[7−10] due to their fascinating properties such as ultrahigh surface area, tunable porosity, and excellent thermal and chemical stability.[11−13] However, in some cases, their microsized pores and narrow channels inherently restrict their adsorption capacity.[11,14,15]

For example, Yaghi et al. reported zeolitic imidazolate frameworks (ZIFs) with a high surface area and high thermal and water stability with small pores since the organic linker is imidazole.[16] For this reason, strategies to enhance the adsorption performance of MOFs are needed. In this context, different approaches have been investigated to create defects in MOFs with accessible active sites to boost their adsorption capacity for heavy metals, particularly copper ions.[17] Building MOFs with different ligands or different metals for the synthesis of mixed-ligand or mixed-metal MOFs, respectively, could generate more defects and larger pore sizes.[18−21]

Water is a vital element in life which is used for several human needs and activities such as drinking, domestic use, and food production.[22] Despite intensive efforts, access to safe drinking water is still at risk owing to the intolerable amounts of heavy metals found in numerous water streams and food webs.[23] However, heavy metals released from industrial activities and daily use are a major concern for the environmental ecosystem which can engender adverse health effects on humans, animals, plants, and microorganisms.[23] For instance, copper is one of the most widely used heavy metals in many industrial sectors, such as electric and electronic applications as well as industries of alloys, paints, and agriculture, etc.[23] This chemical element is toxic even at low doses for many organisms, especially in aquatic environments.[23] Accordingly, effective solutions for the efficient remediation of trace copper from water are key alternatives to promote environmental and human well being. Indeed, various adsorbents have been used in the removal of copper such as activated carbon,[24] zeolite,[25] and biochar.[26] However, these conventional adsorbents suffer from low adsorption capacity and modest removal efficiency due to their low surface area and limited adsorptive sites. Therefore, designing and synthesizing a new adsorbent is quite important for the development of a significant adsorption technology for the removal of toxic metals from aquatic systems.

Mixed-ligand MOFs offer many opportunities for the creation of functional active sites, which can effectively enhance the adsorption efficiency of these materials. Herein, a ZIF-8 material (Zn-mI) was chosen as a model for our study, as it was employed previously in the removal of copper species,[27] and a mixed-ligand MOF strategy was chosen to demonstrate the proof of this concept in accessing more active sites and large cavities to improve the diffusion and mass transfer and thus enhance the adsorption efficiency of ZIF-8 for copper removal from water. Dicarboxylic benzene acid (H2BDC) was added as a second ligand to form a new mixed-ligand MOF material (Zn-mI-BDC), which was used for the adsorption of Cu2+, and subsequently compared to pristine ZIF-8 under the same working conditions. The Langmuir model, Freundlich model, and pseudo-first order kinetics, and pseudo-second order kinetics were used to describe the adsorption behaviors.

Results and Discussion

Characterization of Zn-mI-BDC and ZIF-8

The XRD patterns of pristine ZIF-8 and mixed-ligand Zn-mI-BDC materials are displayed in Figure . While the pristine ZIF-8 shows eight main characteristics peaks at 2 tetha of 7° and 3°, 10° and 4°, 12° and 7°, 14° and 7°, 16° and 4°, 18° and 1°, 24° and 5°, and 26° and 7° typical for the ZIF-8 material,[27] Zn-mI-BDC exhibits the same peaks as pristine ZIF-8, confirming that the addition of a controlled amount of H2BDC did not alter the crystallinity of our synthesized ZIF-8. The small additional peaks at 8° and 9°, 11° and 1°, 17° and 1°, and 17° and 5° can be attributed to the coordination framework between terephthalic acid (BDC) and Zn (Zn–BDC).[28] This interference of the coordination between BDC and mI on the Zn cluster can create some defects within the new material, leading to the creation more active sites.

Figure 1

X-ray diffraction patterns of ZIF-8 and Zn-mI-BDC.

It is well known that the pore size distribution and specific surface area of the adsorbent have an important effect on its adsorption performance. N2 adsorption–desorption isotherms were carried out, and the results are recorded in Figure . The curve of ZIF-8 shows that the N2 adsorption–desorption isotherm is type I. Although Zn-mI-BDC exhibited a N2 adsorption–desorption isotherm of type I, a hysteresis loop was observed which is typical for a hierarchical microporous and mesoporous material (Figure a).[29]

Figure 2

(a) N2 adsorption–desorption isotherms and (b) pore-size distributions of ZIF-8 and Zn-mI-BDC materials.

The isotherm profile indicates that Zn-mI-BDC is a hierarchical microporous mesoporous material with a specific surface area of 1106 m2/g and a narrow pore size of 4.2 nm (calculated by the BJH desorption), while ZIF-8 is a microporous material (<2 nm) with a specific surface area of 1477 m2/g (Table and Figure b), which confirms the creation of mesopores after adding the second ligand.

Table 1

Textural Properties of ZIF-8 and Zn-mI-BDC

samples	pore size (nm)	SBET (m2/g)	Vmicro (cm3/g)	Vtotal (cm3/g)
ZIF-8	<2	1477.2	0.52	0.63
Zn-mI-BDC	4.2	1106.9	0.44	0.803

The morphology of the synthesized materials was characterized with scanning electron microscopy (SEM) as shown in Figure a and 3b. The synthesized ZIF-8 has a typical rhombic dodecahedral shape with a particle diameter of ∼122 μm, similar to that reported in the literature.[30,31] However, a decrease in the crystal size from 100–122 to 30–60 μm was observed after adding H2BDC with the appearance of more broken pieces, proving that the addition of a second ligand constrains the normal growth of ZIF-8 crystals.

Figure 3

SEM images of normal ZIF-8 (a, b) and Zn-mI-BDC (c, d).

In order to identify the elemental composition of ZIF-8 and Zn-mI-BDC, energy-dispersive X-ray analysis (EDX) was performed in parallel with SEM characterization (Figure a and 4b). As it can be seen, zinc, azote, carbon, and oxygen are present in both materials. In addition, the elemental composition of the mixed-ligand Zn-mI-BDC was almost identical with the pristine ZIF-8 since the added ligand (BDC) was just 10%; a slight increase of oxygen atoms was observed due to the presence of carboxylate groups in the BDC ligand, while a minor decrease in azote was noted due to the decrease of mI ligands (Figure b).

Figure 4

Energy-dispersive X-ray analysis for ZIF-8 (a) and Zn-mI-BDC (b).

To determine the quantity of BDC incorporated in the framework and reveal the thermal stability of these porous compounds. Both MOFs were subjected to thermogravimetric analysis (TGA), and the results are presented in Figure . The TGA curve of the as-synthesized sample of ZIF-8 exhibited two gradual weight-loss steps: the first one corresponding to the removal of guest molecules (mainly H2O) from the cavities, and the second one is associated with a long plateau in the temperature range of 400–600 °C, indicating the high thermal stability of ZIF-8. On the other hand, the TGA curve of Zn-mI-BDC was comprised of three stages, and the thermal decomposition of Zn-mI-BDC began at about 200 °C. While the first stage was due to the removal of water molecules at temperatures under 200 °C, the second stage between 200 and 370 °C resulted from the decomposition of BDC groups since it corresponds to a weight loss of ∼10% as introduced in the experimental part. The third stage was because of the decomposition of 2-methyl imidazole of the MOF, and there is a loss of 64% by weight. The thermal stability of Zn-mI-BDC was slightly lower than that of ZIF-8, which is attributed to the creation of defects.[32]

Figure 5

TGA curves of ZIF-8 and Zn-mI-BDC.

Further evidence to characterize the functional groups of ZIF-8 and mixed-ligand MOF Zn-mI-BDC was provided by Fourier-transform infrared (FTIR) spectroscopy. The FTIR spectra of ZIF-8, coordinated ligands, terephthalic acid (H2BDC), and 2-methyl imidazole (2-mI) are shown in Figure . As observed, 2-methylimidazole exhibits a broad and strong absorption in the range of 2200–3250 cm–1 due to vibrations of the hydrogen bonds established between the pyrrole group and the pyridinic nitrogen (N–H···N). Also, N–H stretching vibrations can be observed as well at 1849 cm–1,[33] which completely disappeared in the ZIF-8 and Zn-mI-BDC spectra, indicating that the 2-mI ligands are fully reacted during formation of both MOFs. Besides, the band at 1584 cm–1 represents the stretching vibration of C=N. In addition, all vibrational modes of ZIF-8 are present in the infrared spectra of the Zn-mI-BDC material, which indicates that the addition of BDC did not change the functional groups of ZIF-8. However, the appearance of vibrational modes in the infrared spectra of Zn-mI-BDC such as those at 1669 and 1396 cm–1, which are related to the asymmetric and symmetric stretching modes of coordinated carboxylic acid, respectively, indicates the attachment of a carboxylate group of BDC in the mixed-ligand MOF.

Figure 6

FTIR spectra of 2-mI, BDC, ZIF-8, and Zn-mI-BDC.

In an attempt to further investigate the charge state of ZIF-8 and Zn-mI-BDC particles, zeta potential measurements across a wide pH range were conducted (Figure ). Both ZIF-8 and Zn-mI-BDC demonstrated a positive zeta potential in acidic pH and then flipped to a negative zeta potential at pH values of 3.7 and of 4.2 for Zn-mI-BDC and ZIF-8, respectively. The value of the point of zero charge implied that the surface charge of Zn-mI-BDC particles was positive when the pH of the solution is lower than 3.7; then, it is reversed into negative charge at a pH greater than 3.7.

Figure 7

Zeta potential measurements for Zn-mI-BDC and ZIF-8 at different pH values.

Heavy Metal Adsorption Study

MOFs are typically porous materials; their excellent properties for the removal of metal ions are due to the presence of pores and active adsorptive sites in their frameworks.[34] In this work, we synthesized a mixed-ligand Zn-mI-BDC adsorbent with more accessible sites and investigated its adsorption ability for the removal of copper from aqueous solution.

Effect of Different Parameters Affecting the Adsorption

The adsorption capacity for Cu2+ on Zn-mI-BDC is shown in Figure . It can be seen that the adsorption of copper is obviously improved after adding the BDC ligand, presumably due to increased adsorption sites and the presence of mesopores that promoted access to these active sites, which all together play an important role in improving the Zn-mI-BDC adsorption efficiency.

Figure 8

(a) Effect of contact time on the removal of copper by Zn-mI-BDC and ZIF-8 ([Cu2+] = 200 mg/L; adsorbents = 5 mg; V = 10 mL; T° = 298 K). (b) Effect of initial concentration on the removal of copper on Zn-mI-BDC and its adsorption capacity ([Cu2+] = 200 mg/L; adsorbents = 5 mg; V = 10 mL; contact time: 60 min; pH = 5; T° = 298 K).

The effect of contact time on Cu2+ ion removal was also studied. As expected, Figure a shows that the rate of adsorption increases as the contact time increases. Also, the Cu2+ removal rate was faster using Zn-mI-BDC and higher when compared to that of ZIF-8. Thereby, in the first 5 min, Zn-mI-BDC removed 47% of Cu2+ while ZIF-8 adsorbed only 29%. Besides, the removal rate for both MOFs was rapid at the beginning, which could be explained by the availability of pores and the number of active sites at the beginning of the adsorption process.[27] Thereafter, the number of active sites was reduced, thus decreasing the removal rate.

The effect of Cu2+ concentration on the mixed-ligand MOF Zn-mI-BDC was also investigated. Figure b shows the effect of the initial concentration on the removal of copper ions on Zn-mI-BDC and its adsorption capacity. The results demonstrated that the removal performance decreased gradually with increasing copper concentration. This was attributed to saturation of available active sites, indicating that Zn-mI-BDC did not provide enough binding sites to adsorb all of the Cu2+ ions in the highly concentrated solution.[17] Furthermore, the adsorption efficiency was positively related to temperature (Figure S1), as higher temperatures are expected to provide enough energy to accelerate the diffusion of copper ions into the pores of the adsorbent.[27] In addition, the increase of copper removal at higher temperatures indicates that the adsorption of Cu2+ ions on Zn-mI-BDC is an endothermic process.[35]

Adsorption Isotherm and Adsorption Kinetic Models

The correlation between the amount of Cu(II) ions adsorbed onto the mixed-ligand MOF Zn–Cu–BDC and the equilibrium concentration of Cu(II) ions was estimated using Langmuir and Freundlich isotherm models (Text S1).[27,36] Batch experiments were carried out with different initial copper concentrations (100–800 mg·L–1) at room temperature over 1 h.

The linear plots of the Langmuir (Figure S2a) and Freundlich (Figure S2b) models were studied, and their related relative parameters were determined (Table S1). As a result, the Langmuir isotherm model was found to be linear over the entire concentration range with a good linear correlation coefficient (R2 = 0.9980), which is higher than that of the Freundlich model (R2 = 0.9726), indicating that the Langmuir equation represents the best fit of the experimental data. Also, the process occurs on a heterogeneous surface, and the calculated RL values (Text S1) for the studied concentrations are all between 0 and 1 (0.15 and 0.58), which revealed an efficient adsorption process. In addition, n was greater than 1, demonstrating the high affinity of Zn-mI-BDC for copper. The nonlinear plots of the Langmuir and Freundlich models (Figure c) and their related parameters (Table S1) exhibited primarily the same results of that of the linear fit.

The adsorption kinetics of copper on Zn-mI-BDC were explored as well, not only to study the rate of adsorption but also to evaluate the adsorption efficiency. In this regard, pseudo-first-order and pseudo-second-order kinetics models were used to confirm the relation between adsorption and time (Text S2).[37] Accordingly, pseudo-second-order kinetics was more appropriate to describe the adsorption process since the correlation coefficient R22 was higher (R22 = 0.9977 > R12 = 0.9610) and the calculated Qe (Qe = .197) was closer to that of the measured data (Qeexp = 0.189) (Text S2, Figure S3, and Table S2). Therefore, these results suggest that the adsorption process is subjected to a pseudo-second-order kinetic model which assumes a chemical adsorption process. Thus, the adsorption may not be fully reversible and requires more effort for material regeneration.[37]

In addition, a comparison of the adsorption capacity between our newly developed hierarchical porous Zn-Im-BDC adsorbent with pristine ZIF-8 and other reported adsorbents is given in Table . The adsorption capacity of Zn-Im-BDC (200 mg·g–1) is superior to or identical with most of the reported adsorbents, indicating that the designed adsorbent holds great promise in copper treatment from water and that more hierarchical porous MOFs can be prepared and used for the removal of Cu2+ and other heavy metals.

Table 2

Copper Adsorption Data Using Various Adsorbents

adsorbent	adsorption capacity (mg·g–1)	pH	equilibrium time (min)	ref
Zn-mI-BDC	200	6.8	100	this work
ZIF-8	116.98	5	30	(27)
Cu-ZIF-8	135.12	4–6		(38)
MOF-5	290	5.2	60	(39)
Ag–Fe MOF	213	5	65	(17)
carbon foam	246.66	6–7	1440	(40)
Activated carbon	43.47	5	180	(24)
PGCB	100	5	180	(41)
PNIPAM-Co-AA	67.25	5		(42)
keratin/PA6 nanofibers	103.5	5.8	300	(43)

Conclusions

In this study, a hierarchical microporous mesoporous MOF, Zn-mI-BDC, was successfully synthesized using a mixed-ligand MOF strategy by adding terephthalic acid and 2-methylimidazole (mI) linkers. It was revealed that the addition of a controlled amount of a second ligand did not alter the crystallinity of ZIF-8 and interestingly resulted in the creation of larger pores and shrunken size in morphology, leading to an enhanced adsorption capacity of Cu2+ compared to pristine ZIF-8. Moreover, the maximum adsorption capacity was better than many existing adsorbents. The adsorption of Cu2+ using Zn-mI-BDC fitted well with both linear and nonlinear pseudo-second-order models, and the adsorption isotherm matched well with the Langmuir isotherm model, suggesting that the adsorption process was controlled by chemical reactions. Therefore, this work will inspire other researchers who further attempt to improve heavy metal adsorption using metal–organic frameworks.

Experimental Section

Chemicals

The chemicals used in the study were of analytical grade and used without any further purification. Zinc nitrate hexahydrate (Zn(NO3)2·6H2O), 2-methylimidazole (mI), terephthalic acid (H2BDC), copper sulfate anhydrous (CuSO4), sodium hydroxide (NaOH), N,N-dimethylformamide (DMF), ethanol, and acetone were received from Sigma-Aldrich.

ZIF-8 Synthesis

ZIF-8 was prepared according to the literature procedure with some modifications.[44] Briefly, zinc nitrate hexahydrate (1.78 g) and 2-methylimidazole (0.49 g) were completely dissolved in 60 mL of DMF and stirred vigorously until a clear solution was obtained. Then, the mixture was introduced to a 100 mL bottle with a Teflon-taped screw cap and heated at 140 °C for 24 h in a convection oven. Eventually, the product was centrifuged, washed with DMF several times, kept in MeOH for 3 days to remove the remaining DMF in the pores, and dried at room temperature.

Zn-mI-BDC Mixed-Ligand MOF Synthesis

Zn-mI-BDC was synthesized using H2BDC and mI linkers; the molar ratio of H2BDC/mI was 0.9:0.1. Typically, 2.009 g of zinc nitrate hexahydrate (6.756 mmol), 0.501 g of 2-methylimidazole (6.102 mmol), and 0.101 g of H2BDC (0.609 mmol) were completely dissolved in 60 mL of DMF and stirred vigorously until a clear solution was obtained. Then, the mixture was introduced to a 100 mL bottle with a Teflon-taped screw cap and heated at 140 °C for 24 h in a convection oven. Eventually, the product was centrifuged, washed with DMF several times, kept in MeOH for 3 days, and dried at room temperature.

Characterization

The physicochemical properties of both MOFs (Zn-mI-BDC and ZIF-8) were characterized with a series of analytical methods. X-ray diffraction (XRD) measurements were performed using a D8 Advance X-ray diffractometer. The crystal size and particle morphology of the materials were examined by scanning electron microscopy (SEM) (operated at 20 kV, QUATTRO S-FEG-ThermoFisher Scientific). Energy-dispersive X-ray analysis (EDX) was investigated in parallel with SEM analysis. Thermal analysis was carried out under an air atmosphere at a rate of 10 °C/min over the range of 25–700 °C (TA DSC Q20). A JASCO FTIR-4600 spectrometer was used to record the FTIR spectra from 400 to 4000 cm–1. The Brunauer–Emmett–Teller (BET) surface area, pore size, and the pore volume were measured with a gas sorption analyzer (Micromeritics 3-Flex Surface Characterization Analyzer). The zeta potential of the materials was measured using a Zetameter (Zetasizer Nano ZS90, Malvern) at room temperature. A simultaneous inductively coupled plasma atomic emission spectrometer ICP-AES (PerkinElmer Avio 500) was used to determine the copper concentration in aqueous solutions.

Batch Adsorption Experiments

Multiple batch adsorption tests were managed to investigate the adsorption performance of the developed MOF adsorbents. For this reason, an aqueous solution of copper with different concentrations was used to carry out the adsorption experiments. In brief, 20 mg of MOF as the adsorbent was introduced to each tube containing 5 mL of copper ions, which was well mixed and kept at room temperature. Subsequently, the solution was separated from the adsorbent via centrifugation and filtration. The content of copper in the solution was measured by ICP-AES. Batch adsorption results were the average values of three replicate experiments.

In general, the removal efficiency (R%) and adsorption capacity Qe (mg/g) of the copper ions were determined according to eqs and 2, respectivelywhere C0 (mg·L–1) and Ce (mg·L–1) are the initial and equilibrium concentrations of the adsorbate in solution, respectively, V (mL) is the volume of the solution, and m (g) is the mass of the adsorbent.