Selective Dye Adsorption by Zeolitic Imidazolate Framework-8 Loaded UiO-66-NH2

In this study, Zeolitic Imidazolate Framework-8 (ZIF-8)-loaded UiO-66-NH2 was synthesized, characterized, and analyzed for its potential to efficiently remove dyes. The selective adsorption on ZIF-8-loaded UiO-66-NH2 or its parent MOFs (UiO-66-NH2 and ZIF-8) in the mixed dyes solution was explored, including anionic dye (methyl orange (MO)) and cationic dyes (methylene blue (MB) and rhodamine B (RhB)). ZIF-8-loaded UiO-66-NH2 displayed much better selectivity to MB than its parent MOFs. Adsorption capacity of ZIF-8-loaded UiO-66-NH2 (173 mg/g) toward MB was found to be 215% higher than UiO-66-NH2 (55 mg/g). A kinetics study based on adsorption data demonstrated that the adsorption process most closely matched with the model of pseudo-second-order kinetic and Langmuir isotherm. The adsorption was an exothermic and spontaneous physical process as revealed by the values of thermodynamic parameters. Furthermore, reusability of ZIF-8-loaded UiO-66-NH2 was investigated and revealed the significant regeneration efficiency in adsorption capacity for MB even after four adsorption cycles. Experimental results proved that the interaction between ZIF-8-loaded UiO-66-NH2 and MB was mainly affected by the mechanism, for instance, electrostatic interaction as well as π–π stacking interactions.


Introduction
Metal organic frameworks (MOFs) are three-dimensional porous crystal materials consisting of metal clusters or ions coordinated with organic ligands [1]. MOFs are the emerging highly ordered crystalline porous materials with advantages such as, for instance, large specific surface area, unconventional porosity, good chemical stability, and tunable structural and functional properties [2].
UiO-66-based composites, including polymer (polymer@UiO-66) [3], nanoparticles (NPs@UiO-66) [4], and MOF (MOF@UiO-66) [5] have become a focus of research due to the high stability of UiO-66. Owing to the interface between single MOF, the MOF@MOF structures can provide new capabilities [6] and have been widely employed for gas storage and catalysis. For instance, Ren et al. [7] synthesized MIL-101@UiO-66 nanocrystals, which were used for hydrogen storage and displayed significantly improved performance of hydrogen storage compared with parent MOF (MIL-101 or UiO-66). Zhuang et al. [8] synthesized core-shell UiO-66-NH 2 @ZIF-8 and further transformed it into Pd-UiO-66-NH 2 @ZIF-8 to achieve enhanced selective catalysis for alkene hydrogenations. Although MOF@MOF showed appreciably better performances than the parent MOFs, so far MOF@MOF has not been widely used for the removal of dye.
To date, many approaches have been developed for the elimination of dyestuff, such as adsorption [9], biodegradation [10], photocatalytic oxidation [11], and membrane separation [12]. electrons on the N-atom of amino groups can form complexation with Zn 2+ ions, which can chelate with 2-MI to produce ZIF-8 seeded layers in surface of UiO-66-NH 2 . After that, through the layer-by-layer solution deposition method, the continuous and compact ZIF-8 can be formed on the surface of UiO-66-NH 2 . The task consists of the following steps: 0.5 g UiO-66-NH 2 was dispersed in 0.1 M Zn (NO 3 ) 2 /methanol solution (20 mL) and stirred at 1200 rpm for 2 h under ambient conditions (25 • C), followed by ultrasonic washing with methanol for 20 min and centrifugation to the obtained product. To this, 0.8 M 2-methylimidazole/methanol solution (20 mL) was added and stirred at 1200 rpm for 2 h under ambient condition (25 • C), and followed by ultrasonic washing with methanol for 20 min, and finally vacuum dried at 75 • C for 12 h to get the ZIF-8-loaded UiO-66-NH 2 . The whole process represents one loop and this whole process was repeated for n number of cycles by changing the ZIF-8 content and represented as n-ZIF-8@UiO-66-NH 2 .

Adsorption Experiments of Dyes
Typical representative dyes with different charges were employed to measure the ability of the adsorbent for selectively adsorbing dyes, for example, MB (664 nm), RhB (554 nm), and MO (464 nm). Typically, 10 mg sample was dispersed into 100 mL dye solution (0-50 mg/L) and stirred continuously for 24 h under ambient condition (25 • C). 0.05 M HCl and NaOH were used to regulate pH of the MB solution. The adsorption capacities were calculated using the following Formula (1): where q e is the adsorption capacity of equilibrium (mg/g) and C 0 and C e are the initial and final concentrations (mg/L), respectively. m is the mass of adsorbent for the adsorption experiments (g), and V is the volume of dye solutions (L). Figure 1 depicts the XRD graphs of the synthesized UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 . All the characteristic diffraction peaks of UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 were sharp, thereby showing excellent crystallinity, which was consistent with the previous reports [21,24]. However, the intensities of diffraction for ZIF-8-loaded UiO-66-NH 2 were lower than UiO-66-NH 2 , suggesting that the crystallinity of UiO-66-NH 2 reduced after ZIF-8 loading. It could be due to the interaction between Zn 2+ and amino groups of ZIF-8 and UiO-66-NH 2 , which might slightly change the frame structure of UiO-66-NH 2 [25,26]. The diffraction peak of ZIF-8 did not emerge in the XRD pattern of ZIF-8-loaded UiO-66-NH 2 , which is due to the fact that the thickness of ZIF-8 shell was not good enough to form strong scattering crystalline peaks [21]. The FTIR spectra of ZIF-8-loaded UiO-66-NH2 and parent MOFs (UiO-66-NH2 and ZIF-8) were analyzed (Figure 2a), aiming to probe the loading amount of ZIF-8. The spectra of UiO-66-NH2 exhibited a peak at 1500 cm −1 due to C=C of the aromatic group. The peaks at 3368 and 3487 cm −1 represent symmetric and asymmetric vibrations of N-H groups, respectively [27]. However, all the characteristic peaks of UiO-66-NH2 remained in ZIF-8-loaded UiO-66-NH2 and both showed a similar pattern of FTIR spectrum. In addition, two characteristic absorption bands of ZIF-8 at 3135 and 2929 cm −1 corresponding to stretching vibration of aromatic and aliphatic C-H, respectively, also appeared, indicating the successful loading of ZIF-8 in the UiO-66-NH2 [28]. To determine the loading amount of ZIF-8, semi-quantitative analysis of FTIR with 3135 and 1500 cm −1 peaks of ZIF-8 and UiO-66-NH2, respectively, was performed ( Figure 2b) and found that the ZIF-8 content increased in the UiO-66-NH2 on increasing the number of loading cycles. The FTIR spectra of ZIF-8-loaded UiO-66-NH 2 and parent MOFs (UiO-66-NH 2 and ZIF-8) were analyzed (Figure 2a), aiming to probe the loading amount of ZIF-8. The spectra of UiO-66-NH 2 exhibited a peak at 1500 cm −1 due to C=C of the aromatic group. The peaks at 3368 and 3487 cm −1 represent symmetric and asymmetric vibrations of N-H groups, respectively [27]. However, all the characteristic peaks of UiO-66-NH 2 remained in ZIF-8-loaded UiO-66-NH 2 and both showed a similar pattern of FTIR spectrum. In addition, two characteristic absorption bands of ZIF-8 at 3135 and 2929 cm −1 corresponding to stretching vibration of aromatic and aliphatic C-H, respectively, also appeared, indicating the successful loading of ZIF-8 in the UiO-66-NH 2 [28]. To determine the loading amount of ZIF-8, semi-quantitative analysis of FTIR with 3135 and 1500 cm −1 peaks of ZIF-8 and UiO-66-NH 2 , respectively, was performed ( Figure 2b) and found that the ZIF-8 content increased in the UiO-66-NH 2 on increasing the number of loading cycles. Figure 3A,B displays the images of SEM and TEM for (a) UiO-66-NH 2 , and (b) ZIF-8-loaded UiO-66-NH 2 (1-ZIF-8@UiO-66-NH 2 ), showing that all the samples were agglomerated granules. The shape and size of the UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 did not change significantly, and the particle sizes were about 100 nm. Figure 3Ac,Bc displays the EDS mappings of the ZIF-8-loaded UiO-66-NH 2 and found that the Zn and Zr mappings were consistent with the C, N, and O mappings, indicating that the Zn was uniformly distributed into the UiO-66-NH 2 framework. TEM was used for a better comparison of difference in morphological structure between UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 . UiO-66-NH 2 framework showed smooth surface and no obvious lattice fringes (Figure 3Ba). After loading with ZIF-8, highly dispersed tiny black nanoparticles (about 10-20 nm) were observed on the UiO-66-NH 2 surface (Figure 3Bb). pattern of FTIR spectrum. In addition, two characteristic absorption bands of ZIF-8 at 3135 and 2929 cm −1 corresponding to stretching vibration of aromatic and aliphatic C-H, respectively, also appeared, indicating the successful loading of ZIF-8 in the UiO-66-NH2 [28]. To determine the loading amount of ZIF-8, semi-quantitative analysis of FTIR with 3135 and 1500 cm −1 peaks of ZIF-8 and UiO-66-NH2, respectively, was performed ( Figure 2b) and found that the ZIF-8 content increased in the UiO-66-NH2 on increasing the number of loading cycles.  Figure 3A,B displays the images of SEM and TEM for (a) UiO-66-NH2, and (b) ZIF-8-loaded UiO-66-NH2 (1-ZIF-8@UiO-66-NH2), showing that all the samples were agglomerated granules. The shape and size of the UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2 did not change significantly, and the particle sizes were about 100 nm. Figure 3Ac,Bc displays the EDS mappings of the ZIF-8-loaded UiO-66-NH2 and found that the Zn and Zr mappings were consistent with the C, N, and O mappings, indicating that the Zn was uniformly distributed into the UiO-66-NH2 framework. TEM was used for a better comparison of difference in morphological structure between UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2. UiO-66-NH2 framework showed smooth surface and no obvious lattice fringes ( Figure  3Ba). After loading with ZIF-8, highly dispersed tiny black nanoparticles (about 10-20 nm) were observed on the UiO-66-NH2 surface (Figure 3Bb).  Figure 4a. ZIF-8, UiO-66-NH2, and ZIF-8-loaded UiO-66-NH2 exhibited strictly Type I isotherm with a predominantly microporous structure. ZIF-8-loaded UiO-66-NH2 showed a BET (Brunauer-Emmet-Teller) specific surface area and pore volume of 651 m 2 /g and 0.233 cm 3 /g, respectively, which is much less than ZIF-8 and UiO-66-NH2. It might be attributed to the fact that ZIF-8 loaded on the surface of UiO-66-NH2 blocked the channels of UiO-66-NH2, which further confirmed that ZIF-8 were immobilized on the UiO-66-NH2 framework. The parameters of various samples are listed in Table 1. The average pore diameter of as-synthesized UiO-66 was 1.2 nm. However, the pore diameter of UiO-66 decreased to 0.5 nm after loading with ZIF-8. The weight loss of UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2 (1-ZIF-8@UiO-66-NH2) in N2 atmosphere is shown in Figure 4b. The first phase of weight loss is attributed to the loss of solvent molecules (H2O, DMF) in the temperature range 25-250 °C. The next weight loss is mainly because of destruction of organic framework. Finally, at high temperature, it stabilized, due to the generated metallic oxides. The trend of curves of ZIF-8-loaded UiO-66-NH2 were similar to that of UiO-66-NH2, which shows that the thermal stability of UiO-66-NH2 did not change after loading with ZIF-8. N 2 adsorption/desorption isotherms of UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 (1-ZIF-8@UiO-66-NH 2 ) are shown in Figure 4a. ZIF-8, UiO-66-NH 2 , and ZIF-8-loaded UiO-66-NH 2 exhibited strictly Type I isotherm with a predominantly microporous structure. ZIF-8-loaded UiO-66-NH 2 showed a BET (Brunauer-Emmet-Teller) specific surface area and pore volume of 651 m 2 /g and 0.233 cm 3 /g, respectively, which is much less than ZIF-8 and UiO-66-NH 2 . It might be attributed to the fact that ZIF-8 loaded on the surface of UiO-66-NH 2 blocked the channels of UiO-66-NH 2 , which further confirmed that ZIF-8 were immobilized on the UiO-66-NH 2 framework. The parameters of various samples are listed in Table 1. The average pore diameter of as-synthesized UiO-66 was 1.2 nm. However, the pore diameter of UiO-66 decreased to 0.5 nm after loading with ZIF-8. The weight loss of UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 (1-ZIF-8@UiO-66-NH 2 ) in N 2 atmosphere is shown in Figure 4b. The first phase of weight loss is attributed to the loss of solvent molecules (H 2 O, DMF) in the temperature range 25-250 • C. The next weight loss is mainly because of destruction of organic framework. Finally, at high temperature, it stabilized, due to the generated metallic oxides. The trend of curves of ZIF-8-loaded UiO-66-NH 2 were similar to that of UiO-66-NH 2 , which shows that the thermal stability of UiO-66-NH 2 did not change after loading with ZIF-8. Nanomaterials 2019, 9, x FOR PEER REVIEW 6 of 18

Separation Selectivity in the Mixed Dye Solution
The ideal adsorption selectivity of MB over MO and RhB on ZIF-8-loaded UiO-66-NH2 was calculated according to the following expression: where S is the dye selective adsorption index, qMB, qMO, and qRhB are the adsorption capacity of dye (MB, MO, and RhB, respectively) on adsorbents when in equilibrium (mg/g), CMB, CMO, and CRhB correspond to initial concentration of different dyes in mixed solution (mg/L). As shown in Figure 5, the adsorptive capacity of ZIF-8-loaded UiO-66-NH2 for MB was significantly higher than its parent MOFs (UiO-66-NH2 and ZIF-8). At the same time, the adsorption selectivity of MB over MO or RhB first increased and then decreased with increasing ZIF-8 loading. Pure UiO-66-NH2 exhibited the best adsorption effect on MO in mixed dyes due to the addition of HCl as a regulator during preparation. The dye preferential sorption was due to the hydrogen ions exposure on to the acid-promoted UiO-66-NH2 surface and resulted in zeta potential raise [29]. Pure ZIF-8 also showed the best adsorption effect on MO in mixed dyes due to the positive charge of its surface [30]. However, ZIF-8-loaded UiO-66-NH2 exhibited the best adsorption effect on MB in mixed dyes. Among them, 1-ZIF-8@UiO-66-NH2 showed the largest adsorptive capacity of MB. Owing to the opposite electrical properties of MB and MO, the zeta potential of ZIF-8-loaded UiO-66-NH2 could be lower than UiO-66-NH2. For the better insight of adsorption interactions between ZIF-8-loaded UiO-66-NH2 and MB, the adsorption of MB by 1-ZIF-8@UiO-66-NH2 has been discussed in detail in the proceeding sections.

Separation Selectivity in the Mixed Dye Solution
The ideal adsorption selectivity of MB over MO and RhB on ZIF-8-loaded UiO-66-NH 2 was calculated according to the following expression: where S is the dye selective adsorption index, q MB , q MO , and q RhB are the adsorption capacity of dye (MB, MO, and RhB, respectively) on adsorbents when in equilibrium (mg/g), C MB , C MO , and C RhB correspond to initial concentration of different dyes in mixed solution (mg/L).
As shown in Figure 5, the adsorptive capacity of ZIF-8-loaded UiO-66-NH 2 for MB was significantly higher than its parent MOFs (UiO-66-NH 2 and ZIF-8). At the same time, the adsorption selectivity of MB over MO or RhB first increased and then decreased with increasing ZIF-8 loading. Pure UiO-66-NH 2 exhibited the best adsorption effect on MO in mixed dyes due to the addition of HCl as a regulator during preparation. The dye preferential sorption was due to the hydrogen ions exposure on to the acid-promoted UiO-66-NH 2 surface and resulted in zeta potential raise [29]. Pure ZIF-8 also showed the best adsorption effect on MO in mixed dyes due to the positive charge of its surface [30]. However, ZIF-8-loaded UiO-66-NH 2 exhibited the best adsorption effect on MB in mixed dyes. Among them, 1-ZIF-8@UiO-66-NH 2 showed the largest adsorptive capacity of MB. Owing to the opposite electrical properties of MB and MO, the zeta potential of ZIF-8-loaded UiO-66-NH 2 could be lower than UiO-66-NH 2 . For the better insight of adsorption interactions between ZIF-8-loaded UiO-66-NH 2 and MB, the adsorption of MB by 1-ZIF-8@UiO-66-NH 2 has been discussed in detail in the proceeding sections.
where, qmax is maximum adsorption capacity (mg/g), KF and KL are Freundlich constant and Langmuir constant, respectively, and 1/n is Freundlich constant related to the adsorption. Figure 6 shows that the adsorption of MB by the UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2 conforms to Langmuir isotherm model, while the predicted result of Freundlich adsorption model was significantly different from experimental data for MB. The model parameters for adsorption are presented in the Table 2. Correlation coefficient (R 2 ) of Langmuir isotherm models were bigger than that of Freundlich isotherm models and R 2 value of Langmuir isotherm models was almost equal to unity, which suggested that the adsorption of MB on these two adsorbents was heterogeneous and a monolayer adsorption process [31]. The theoretical maximum adsorption capacity of the ZIF-8loaded UiO-66-NH2 (173 mg/g) was higher than pure UiO-66-NH2 (55 mg/g), indicating 215% improvement in the adsorption capacity, which could be due to the loading of ZIF-8 nanoparticles.
To date, several researches have been carried out for the removal of MB by MOFs [32]. Table 3 compares the adsorption capacities of MB over similar MOFs under the same experimental conditions [29,[33][34][35][36][37]. Yang [36] demonstrated the use of UiO-66-P for the adsorptive removal of MB, MO, Congo Red (CR), and Acid Chrome Blue K (AC). UiO-66-P could adsorb MB but not the other dyes. The adsorption capability of UiO-66-P (91.1 mg/g) for MB was improved by 272% compared to that of pristine UiO-66 (24.5 mg/g), which shows that the negative charge of UiO-66 was due to phosphate immobilization approach. Yang et al. [37] studied the adsorption of MB using Ce(III)doped UiO-66 and showed decrease in surface potential of UiO-66, thereby enhancing adsorption capacity for MB. However, the surface potential of UiO-66-P (−12.6 mv) was lower than Ce(III)-doped UiO-66 (+6.48 mv) but the adsorption capacity of MB on Ce(III)-doped UiO-66 (145.1 mg/g) higher than UiO-66-P. This behavior could be due to higher specific surface area of synthesized UiO-66 composite, which indicates that the adsorption process was controlled by many factors. These results
where, q max is maximum adsorption capacity (mg/g), K F and K L are Freundlich constant and Langmuir constant, respectively, and 1/n is Freundlich constant related to the adsorption. Figure 6 shows that the adsorption of MB by the UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 conforms to Langmuir isotherm model, while the predicted result of Freundlich adsorption model was significantly different from experimental data for MB. The model parameters for adsorption are presented in the Table 2. Correlation coefficient (R 2 ) of Langmuir isotherm models were bigger than that of Freundlich isotherm models and R 2 value of Langmuir isotherm models was almost equal to unity, which suggested that the adsorption of MB on these two adsorbents was heterogeneous and a monolayer adsorption process [31]. The theoretical maximum adsorption capacity of the ZIF-8-loaded UiO-66-NH 2 (173 mg/g) was higher than pure UiO-66-NH 2 (55 mg/g), indicating 215% improvement in the adsorption capacity, which could be due to the loading of ZIF-8 nanoparticles. To date, several researches have been carried out for the removal of MB by MOFs [32]. Table 3 compares the adsorption capacities of MB over similar MOFs under the same experimental conditions [29,[33][34][35][36][37]. Yang [36] demonstrated the use of UiO-66-P for the adsorptive removal of MB, MO, Congo Red (CR), and Acid Chrome Blue K (AC). UiO-66-P could adsorb MB but not the other dyes. The adsorption capability of UiO-66-P (91.1 mg/g) for MB was improved by 272% compared to that of pristine UiO-66 (24.5 mg/g), which shows that the negative charge of UiO-66 was due to phosphate immobilization approach. Yang et al. [37] studied the adsorption of MB using Ce(III)-doped UiO-66 and showed decrease in surface potential of UiO-66, thereby enhancing adsorption capacity for MB. However, the surface potential of UiO-66-P (−12.6 mv) was lower than Ce(III)-doped UiO-66 (+6.48 mv) but the adsorption capacity of MB on Ce(III)-doped UiO-66 (145.1 mg/g) higher than UiO-66-P. This behavior could be due to higher specific surface area of synthesized UiO-66 composite, which indicates that the adsorption Nanomaterials 2019, 9, 1283 8 of 18 process was controlled by many factors. These results further confirmed that ZIF-8-loaded UiO-66-NH 2 demonstrates enormous advantage for MB in adsorption capacity and selective adsorption.

Kinetic Studies
In order to better explore the adsorption rate, kinetic models were studied. The kinetic equation shows the relationship between concentration and other parameters with time. The kinetics models of pseudo-first-order (5) and pseudo-second-order (6) are as follows [38][39][40]: where K1 and K2 are the rate constant of pseudo-first-order kinetics model (min −1 ) and pseudo-secondorder kinetics model (g/mg min), respectively. t is the adsorption time (min). The values of the kinetic parameters are set out in the Table 4. Figure 7 shows the pseudo-first-order and pseudo-second-order kinetic models for the adsorption of MB onto UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2, and displays that the latter fitted

Kinetic Studies
In order to better explore the adsorption rate, kinetic models were studied. The kinetic equation shows the relationship between concentration and other parameters with time. The kinetics models of pseudo-first-order (5) and pseudo-second-order (6) are as follows [38][39][40]: where K 1 and K 2 are the rate constant of pseudo-first-order kinetics model (min −1 ) and pseudo-second-order kinetics model (g/mg min), respectively. t is the adsorption time (min). The values of the kinetic parameters are set out in the Table 4.  Figure 7 shows the pseudo-first-order and pseudo-second-order kinetic models for the adsorption of MB onto UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 , and displays that the latter fitted with better linearity compared to former. Moreover, the calculated data of the kinetics model for pseudo-second-order was very close to the adsorption data, which showed that the process of adsorption in our work conforms to the kinetics model of pseudo-second-order. with better linearity compared to former. Moreover, the calculated data of the kinetics model for pseudo-second-order was very close to the adsorption data, which showed that the process of adsorption in our work conforms to the kinetics model of pseudo-second-order.  Liquid phase adsorption generally consists of three parts [41]: Step one represents film transport involving the diffusion of pollutants through a boundary layer of hypothetical film for the adsorbent. Liquid phase adsorption generally consists of three parts [41]: Step one represents film transport involving the diffusion of pollutants through a boundary layer of hypothetical film for the adsorbent. The second step represents intra-particle diffusion, and the last step represents adsorption and desorption to achieve balance. Film and intra-particle transport are the main constraints that control the absorption rates by porous adsorbents. In our experiments, using the model of intra-particle diffusion, we simulated the adsorption process in order to understand the diffusion mechanism between the adsorbent and MB [42]. The related model is as following (7): where k i is the rate constant of intra-particle diffusion during i stage (mg/g min) 1/2 and C i is constant of boundary layer theory. The parameters of the model are listed in Table 5.  Figure 8 shows that the curves were divided into two regions, which suggested that the adsorption process contains two or more steps. The first half of the curve represents the adsorbed molecules diffusion in the fluid film on the surface of the particles. The slope of the second part of the curve was significantly smaller, indicating the lower adsorption rate of this part. This behavior represents the process of intra-particle diffusion. Using a model of intra-particle diffusion, all R 2 values were > 0.9, which indicates that adsorption of MB on UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 could be described by intra-particle diffusion. Therefore, the adsorption of MB by the ZIF-8-loaded UiO-66-NH 2 or UiO-66-NH 2 was based on both the diffusion of external and intra-particle. Moreover, the coordinate origin was not passed by the curve, which signified that the diffusion of intra-particle was not the only dominant factor of adsorption rate [43]. Nanomaterials 2019, 9, x FOR PEER REVIEW 10 of 18 Figure 8 shows that the curves were divided into two regions, which suggested that the adsorption process contains two or more steps. The first half of the curve represents the adsorbed molecules diffusion in the fluid film on the surface of the particles. The slope of the second part of the curve was significantly smaller, indicating the lower adsorption rate of this part. This behavior represents the process of intra-particle diffusion. Using a model of intra-particle diffusion, all R 2 values were > 0.9, which indicates that adsorption of MB on UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2 could be described by intra-particle diffusion. Therefore, the adsorption of MB by the ZIF-8loaded UiO-66-NH2 or UiO-66-NH2 was based on both the diffusion of external and intra-particle. Moreover, the coordinate origin was not passed by the curve, which signified that the diffusion of intra-particle was not the only dominant factor of adsorption rate [43].

Effect of pH on Adsorption
The pH of the dye solution has a huge influence on the dye adsorption capacity of absorbent. Figure 9 represents

Effect of pH on Adsorption
The pH of the dye solution has a huge influence on the dye adsorption capacity of absorbent. Figure 9 represents the impact of various pH on MB adsorption onto UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 under ambient conditions (25 • C). Due to the fact that MB aqueous solution can result in the blue shift in absorption wavelength under strongly alkaline conditions, the adsorption experiments were carried out only under acidic and weakly basic conditions (pH [3][4][5][6][7][8][9][10][11]. For UiO-66-NH 2 , the adsorptive capacity of MB increased rapidly on raising the pH to 10. Nevertheless, the adsorption capacity decreased with pH >10. At low pH, the UiO-66-NH 2 with positive surface charge hinders the adsorption of cationic dye, MB. The highest adsorptive capacity was observed at pH = 10, which indicates that the hydroxyl ion promoted the adsorption of MB. However, the adsorption capacity of MB was lowered with pH > 10, which may be due to excessive hydroxide ions that might compete with the adsorbent, leading to the decrease in adsorption capacity [44]. For ZIF-8-loaded UiO-66-NH 2 , the adsorption capacity of MB increased rapidly as the pH rose to 5 and the adsorption capacity of MB remained stable at pH = 5-10. Therefore, it can be concluded that the stability of MB on ZIF-8-loaded UiO-66-NH 2 was higher compared with UiO-66-NH 2 . From 5 to 3, the adsorption capacity of the ZIF-8-loaded UiO-66-NH 2 rapidly decreased, which may be due to the fact that ZIF-8 might dissolve in acidic conditions [45]. Nanomaterials 2019, 9, x FOR PEER REVIEW 11 of 18 capacity of the ZIF-8-loaded UiO-66-NH2 rapidly decreased, which may be due to the fact that ZIF-8 might dissolve in acidic conditions [45].

Adsorption Thermodynamics
To understand the adsorption mechanism of MB with ZIF-8-loaded UiO-66-NH2, the change in adsorption capacity of MB onto UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2 under different temperatures (25,35, and 45 °C) was studied. In general, the parameters were calculated by the following Equations [46]: where ΔS°, ΔH°, and ΔG° are the standard entropy change (J/mol K), enthalpy changes (kJ/mol), and free energy (kJ/mol), respectively. T and R are the adsorption temperature (K) and gas constant (8.3145 J/mol K), respectively. Table 6 shows the thermodynamic parameter values.
The negative values of enthalpy change show that the adsorption processes were exothermic, which was in accordance with the results in Figure 10, in which the adsorption capacity of MB decreased with the rise of temperature. Moreover, the value of adsorption heat on UiO-66-NH2 was about −20 kJ/mol, which shows that the interaction between UiO-66-NH2 and MB could be because of the dipole-dipole interaction and/or hydrogen-bond [47]. By comparison, the value of adsorption heat for ZIF-8-loaded UiO-66-NH2 slightly declined, which suggested that it may have other interaction. Based on the correlative references, the ΔG° of chemisorption and physisorption were −400 to −50 kJ/mol and −40 to 0 kJ/mol, respectively [48]. The free energy in our case was found to be negative, which indicates that the adsorption course was an irreversible and spontaneous processes. The values of free energy of UiO-66-NH2 and ZIF-8-loaded UiO-66-NH2 were in range of −11 to −5 (kJ/mol) and −8 to −1 (kJ/mol), respectively, which suggests that both the adsorption course of MB were primarily physisorption. Furthermore, the negative value of entropy change shows that MB mobility decreased after adsorption with UiO-66-NH2 or ZIF-8-loaded UiO-66-NH2. Since both entropy change and enthalpy change were negative, it can be concluded that the enthalpy change is an important driving force with such kind of adsorption processes.

Adsorption Thermodynamics
To understand the adsorption mechanism of MB with ZIF-8-loaded UiO-66-NH 2 , the change in adsorption capacity of MB onto UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 under different temperatures (25,35, and 45 • C) was studied. In general, the parameters were calculated by the following Equations [46]: where ∆S • , ∆H • , and ∆G • are the standard entropy change (J/mol K), enthalpy changes (kJ/mol), and free energy (kJ/mol), respectively. T and R are the adsorption temperature (K) and gas constant (8.3145 J/mol K), respectively. Table 6 shows the thermodynamic parameter values. The negative values of enthalpy change show that the adsorption processes were exothermic, which was in accordance with the results in Figure 10, in which the adsorption capacity of MB decreased with the rise of temperature. Moreover, the value of adsorption heat on UiO-66-NH 2 was about −20 kJ/mol, which shows that the interaction between UiO-66-NH 2 and MB could be because of the dipole-dipole interaction and/or hydrogen-bond [47]. By comparison, the value of adsorption heat for ZIF-8-loaded UiO-66-NH 2 slightly declined, which suggested that it may have other interaction. Based on the correlative references, the ∆G • of chemisorption and physisorption were −400 to −50 kJ/mol and −40 to 0 kJ/mol, respectively [48]. The free energy in our case was found to be negative, which indicates that the adsorption course was an irreversible and spontaneous processes. The values of free energy of UiO-66-NH 2 and ZIF-8-loaded UiO-66-NH 2 were in range of −11 to −5 (kJ/mol) and −8 to −1 (kJ/mol), respectively, which suggests that both the adsorption course of MB were primarily physisorption. Furthermore, the negative value of entropy change shows that MB mobility decreased after adsorption with UiO-66-NH 2 or ZIF-8-loaded UiO-66-NH 2 . Since both entropy change and enthalpy change were negative, it can be concluded that the enthalpy change is an important driving force with such kind of adsorption processes.

Reusability of ZIF-8-Loaded UiO-66-NH 2
The recycling adsorption experiments were performed on 1-ZIF-8@UiO-66-NH 2 in order to analyze the reusability of ZIF-8-loaded UiO-66-NH 2 . The used adsorbent recovered by centrifugation was immersed into 100 mL ethanol for desorption by ultrasonication for 60 min and washed with methanol several times, followed by drying at 373 K for the next adsorption cycle. The recycling performance of ZIF-8-loaded UiO-66-NH 2 ( Figure 11) indicates the significant regeneration efficiency in adsorption capacity for MB after four adsorption cycles (less than 25% decrease in adsorption capacity), which provides its applicability to regenerate and reuse in actual dye removal.

Adsorption Mechanism
To further reveal the MB adsorption mechanism, XRD and FT-IR analysis of ZIF-8-loaded UiO-66-NH2 before and after adsorption of MB were examined. XRD spectra (Figure 12a) showed all the peaks of ZIF-8-loaded UiO-66-NH2 at the same position after the adsorption of MB, demonstrating the unchanged crystalline structure of ZIF-8-loaded UiO-66-NH2. Nevertheless, the characteristic peak intensities of ZIF-8-loaded UiO-66-NH2 recovered after adsorption was found to be weaker than ZIF-8-loaded UiO-66-NH2 before used, which may be due to binding of dye molecules with the active binding sites of adsorbents [49]. In the FTIR spectra, after adsorption of MB on ZIF-8-loaded UiO-66-NH2 (Figure 12b), an intense and broad band centered at 3368 cm −1 remain unchanged (the broad band could be due to peak overlap of N-H and O-H [50]), which showed that the hydrogen bonding might not play a major role in the adsorption process [51]. However, the new absorption peaks for the vibration of aromatic ring of MB at 890 cm −1 and 1600 cm −1 could be clearly observed [52], which showed that the MB was immobilized on surface of ZIF-8-loaded UiO-66-NH2. The peaks at 1571 cm −1 (C=O), 1257 cm −1 (C-N), and 426 cm −1 (Zn-N) of ZIF-8-loaded UiO-66-NH2 shifted and showed a significant decrease in intensity after MB adsorption [28,53,54]. This could be explained by the following assumptions: (i) MB, a cationic dye, is more inclined to be absorbed on negatively charged surfaces by electrostatic interaction. The surface potential of the UiO-66-NH2 changed from positive (+13) to negative (−2) after ZIF-8 loading as revealed by zeta potential (Figure 12c) and the cause of the change in surface potential might be because the basic sites (OHgroups and Nmoieties mainly) adsorbed on the surface of ZIF-8 in the solution [55], which could provide the adsorption sites for electrostatic interaction with MB. (ii) Both MB and ZIF-8-loaded UiO-66-NH2 contain aromatic rings and π-π stacking interactions could occur between MB and ZIF-8-loaded UiO-66-NH2, resulting in the shift of the peak (1257 to 1253 cm −1 ; 1571 to 1567 cm −1 ) [56]. Moreover, the characteristics of good water stability and relatively high specific surface area make ZIF-8-loaded UiO-66-NH2 contact and adsorb MB efficiently, which provides an advantage for the ZIF-8-loaded UiO-66-NH2 to adsorb the cationic dyes. However, RhB is also cationic dye, but the bad performance of ZIF-8-loaded UiO-66-NH2 in adsorption of RhB might be due to the steric hindrance because of large molecular size. The complex structure of RhB (as shown in Table 7) with bigger molecular size compared with MB might hinder it from entering the pores of ZIF-8-loaded UiO-66-NH2 (RhB: 1.41 × 0.98 nm > pore width: 0.5 nm). In conclusion, the ZIF-8-loaded UiO-66-NH2 was suitable for selective adsorption of small-size cationic dye and their interaction was mainly affected by electrostatic and π-π stacking interactions. The possible adsorption mechanism is schematically illustrated in Scheme 1.

Adsorption Mechanism
To further reveal the MB adsorption mechanism, XRD and FT-IR analysis of ZIF-8-loaded UiO-66-NH 2 before and after adsorption of MB were examined. XRD spectra (Figure 12a) showed all the peaks of ZIF-8-loaded UiO-66-NH 2 at the same position after the adsorption of MB, demonstrating the unchanged crystalline structure of ZIF-8-loaded UiO-66-NH 2 . Nevertheless, the characteristic peak intensities of ZIF-8-loaded UiO-66-NH 2 recovered after adsorption was found to be weaker than ZIF-8-loaded UiO-66-NH 2 before used, which may be due to binding of dye molecules with the active binding sites of adsorbents [49]. In the FTIR spectra, after adsorption of MB on ZIF-8-loaded UiO-66-NH 2 (Figure 12b), an intense and broad band centered at 3368 cm −1 remain unchanged (the broad band could be due to peak overlap of N-H and O-H [50]), which showed that the hydrogen bonding might not play a major role in the adsorption process [51]. However, the new absorption peaks for the vibration of aromatic ring of MB at 890 cm −1 and 1600 cm −1 could be clearly observed [52], which showed that the MB was immobilized on surface of ZIF-8-loaded UiO-66-NH 2 . The peaks at 1571 cm −1 (C=O), 1257 cm −1 (C-N), and 426 cm −1 (Zn-N) of ZIF-8-loaded UiO-66-NH 2 shifted and showed a significant decrease in intensity after MB adsorption [28,53,54]. This could be explained by the following assumptions: (i) MB, a cationic dye, is more inclined to be absorbed on negatively charged surfaces by electrostatic interaction. The surface potential of the UiO-66-NH 2 changed from positive (+13) to negative (−2) after ZIF-8 loading as revealed by zeta potential (Figure 12c) and the cause of the change in surface potential might be because the basic sites (OH − groups and N − moieties mainly) adsorbed on the surface of ZIF-8 in the solution [55], which could provide the adsorption sites for electrostatic interaction with MB. (ii) Both MB and ZIF-8-loaded UiO-66-NH 2 contain aromatic rings and π-π stacking interactions could occur between MB and ZIF-8-loaded UiO-66-NH 2 , resulting in the shift of the peak (1257 to 1253 cm −1 ; 1571 to 1567 cm −1 ) [56]. Moreover, the characteristics of good water stability and relatively high specific surface area make ZIF-8-loaded UiO-66-NH 2 contact and adsorb MB efficiently, which provides an advantage for the ZIF-8-loaded UiO-66-NH 2 to adsorb the cationic dyes. However, RhB is also cationic dye, but the bad performance of ZIF-8-loaded UiO-66-NH 2 in adsorption of RhB might be due to the steric hindrance because of large molecular size. The complex structure of RhB (as shown in Table 7) with bigger molecular size compared with MB might hinder it from entering the pores of ZIF-8-loaded UiO-66-NH 2 (RhB: 1.41 × 0.98 nm > pore width: 0.5 nm). In conclusion, the ZIF-8-loaded UiO-66-NH 2 was suitable for selective adsorption of small-size cationic dye and their interaction was mainly affected by electrostatic and π-π stacking interactions. The possible adsorption mechanism is schematically illustrated in Scheme 1.

Conclusions
In summary, the positively charged UiO-66-NH2 was synthesized successfully using the solvothermal method and ZIF-8 was successfully loaded into UiO-66-NH2, as revealed by FTIR and

Conclusions
In summary, the positively charged UiO-66-NH2 was synthesized successfully using the solvothermal method and ZIF-8 was successfully loaded into UiO-66-NH2, as revealed by FTIR and

Conclusions
In summary, the positively charged UiO-66-NH2 was synthesized successfully using the solvothermal method and ZIF-8 was successfully loaded into UiO-66-NH2, as revealed by FTIR and

Conclusions
In summary, the positively charged UiO-66-NH2 was synthesized successfully using the solvothermal method and ZIF-8 was successfully loaded into UiO-66-NH2, as revealed by FTIR and

Conclusions
In summary, the positively charged UiO-66-NH2 was synthesized successfully using the solvothermal method and ZIF-8 was successfully loaded into UiO-66-NH2, as revealed by FTIR and

Conclusions
In summary, the positively charged UiO-66-NH 2 was synthesized successfully using the solvothermal method and ZIF-8 was successfully loaded into UiO-66-NH 2 , as revealed by FTIR and XRD. The content of ZIF-8 present after several loading cycles was determined by semi-quantitative analysis of FTIR spectrum. The adsorption capacity of ZIF-8-loaded UiO-66-NH 2 (173 mg/g) toward MB was found to be 215% higher than UiO-66-NH 2 (55 mg/g), which was due to the improved electrostatic attraction between MB and UiO-66-NH 2 after loading with ZIF-8. Kinetic studies indicated that the adsorption process in this work conforms to the pseudo-second-order kinetics model and the adsorption of MB by the ZIF-8-loaded UiO-66-NH 2 or UiO-66-NH 2 was based on the joint control of the diffusion of external and intra-particle. The adsorption capacity of MB by ZIF-8-loaded UiO-66-NH 2 increased rapidly as the pH rose to 5 and remained stable at pH = 5-10. The recycling adsorption experiments revealed the significant regeneration efficiency in adsorption capacity for MB even after four adsorption cycles and displayed no change in the crystalline structure of the reused ZIF-8-loaded UiO-66-NH 2 as revealed by XRD and IR. Thus, the prepared ZIF-8-loaded UiO-66-NH 2 showed that high selectivity and good water stability can be of great potential in the water treatment.