Interface-Active Metal Organic Frameworks for Knoevenagel Condensations in Water

: It is desirable but challenging to locate solid catalysts at the oil-water interface to stabilize “Pickering emulsions”, which is one of the promising ways to developefﬁcient green chemical processes. Herein, water-stable metal organic framework ZIF-8 without any chemical modiﬁcation was demonstrated to be an interface-active catalyst for Knoevenagel condensation in a biphasic system. Pickering emulsion formed under the reaction conditions due to its amphiphilic property, which was beneﬁcial to the mass transfer and led to high catalytic performance. Moreover, it can be repeatedly applied for Knoevenagel condensation for at least six successive cycles without losing its catalytic activity and framework integrity.


Introduction
Using water as a solvent is one of the main ways toward the development of environmentally-friendly chemical processes. However, the low mass transfer efficiencies, ascribed to a modest interfacial contact due to the inhomogeneous mixing even under vigorous stirring, leads to poor catalytic activities in biphasic reaction systems. Although adding surfactants is a frequently-used way toincrease the liquid-liquid interfacial area, itresults in a negativeenvironmental impact. Alternatively, using an interfacial solid catalyst to stabilize a "Pickering emulsion" is an emerging strategy to develop greener chemical processes. "Pickering emulsion" is an emulsion that is stabilized by interfacial particles. The catalytic efficienciesin biphasic catalytic systems could be improved greatly by the location ofinterfacial catalysts at the interfaces. Likewise, it could not only decrease the phase transfer limitations, but alsocouldcatalyze reactions at the interface of two immiscible solvents. Furthermore, the interfacial solid catalystscan be easily recoveredfrom thePickering emulsions after the reaction [1]. In line with the aforementioned properties of Pickering emulsions, a couple of interface-active particles, such as silicas [2][3][4][5][6][7][8], zeolites [9], carbon [10] and metal organic frameworks [11,12], were attempted to be used to stabilizePickering emulsions.
Metal organic frameworks (MOFs) are a new class of highly-ordered porous coordination polymers that can be assembled from plenty of metal ions/clusters andmulti-topic organic struts. MOFs are more attractive materials than conventional porous materials in heterogeneous catalysis, due to their exceptional modular properties, imparting ultrahigh porosity, structural diversity, tunable surface properties and diverse functionalities [13,14]. Despite progressiveadvances made so far in the synthesis and catalytic applications of MOFs, the instability in waterhas considerably limited these MOFs'

Results and Discussion
ZIF-8 nanoparticles with narrow size distributions were prepared accordingly. Scanning electron microscopy (SEM) was used to characterize theirmorphology. Figure 2A shows representative SEM micrographs of ZIF-8. Regular nanoparticles with a narrow size distribution (about 100 nm) were observed. X-ray powder diffraction (XRD) was further used to determine thecrystal structure of ZIF-8. As shown in Figure 3a, the diffraction pattern for the prepared solid has several sharp and intense peaks that can be well indexed to ZIF-8 [21]. Its structural stability in water was tested, and the diffraction pattern for recovered ZIF-8 is shown in Figure 3b. No obvious change was observed before and after treatment, indicating that it waswater stable, which is a prerequisite for potential applications in an aqueous medium. Meanwhile, images of its transmission electron microscopies (TEM) further confirmedits water stability.

Results and Discussion
ZIF-8 nanoparticles with narrow size distributions were prepared accordingly. Scanning electron microscopy (SEM) was used to characterize theirmorphology. Figure 2A shows representative SEM micrographs of ZIF-8. Regular nanoparticles with a narrow size distribution (about 100 nm) were observed. X-ray powder diffraction (XRD) was further used to determine thecrystal structure of ZIF-8. As shown in Figure 3a, the diffraction pattern for the prepared solid has several sharp and intense peaks that can be well indexed to ZIF-8 [21]. Its structural stability in water was tested, and the diffraction pattern for recovered ZIF-8 is shown in Figure 3b. No obvious change was observed before and after treatment, indicating that it waswater stable, which is a prerequisite for potential applications in an aqueous medium. Meanwhile, images of its transmission electron microscopies (TEM) further confirmedits water stability. thecrystal structure of ZIF-8. As shown in Figure 3a, the diffraction pattern for the prepared solid has several sharp and intense peaks that can be well indexed to ZIF-8 [21]. Its structural stability in water was tested, and the diffraction pattern for recovered ZIF-8 is shown in Figure 3b. No obvious change was observed before and after treatment, indicating that it waswater stable, which is a prerequisite for potential applications in an aqueous medium. Meanwhile, images of its transmission electron microscopies (TEM) further confirmedits water stability. To the best of our knowledge, MOFs including ZIF-8 have been confirmed to be good catalysts for the Knoevenagel condensations in organic solvents [17,18]. The stability in aqueous medium of ZIF-8 has been proven by the XRD patterns and TEM images as shown in Figures 2 and 3 in our experiments. Therefore, the activities of ZIF-8 nanoparticles for the Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein various solvents were firstly investigated, and the results are summarized in Table 1. Similar to the results observed in previous reports, ZIF-8 gave high or medium yields in polar solvents and poor results in nonpolar solvents. As we know, catalytic reaction in water or biphasic systems is a greener route than in organic solvent. To our delight, ZIF-8 could also promote the reaction in water in an efficient way. Over an 80% yield was obtained in water after 30 min, a little bit lower than that in DMF, but higher than that in THF. Then, a water/ZIF-8/ethyl acetate system was used to examine the interfacial property of ZIF-8. After adding 2 mLof pure or 2-nitrobenzaldehydedissolved ethyl acetate to 3 mL of water containing 0.8 wt % of ZIF-8 (with respect to water), as shown in Figure 4A,B, ZIF-8 is predominantlydistributed at the interface between the oil and aqueous phase. After subsequentvigorous stirring or shaking, the appearances of its mixture ( Figure 4C) and its optical micrograph demonstrated the formation of Pickering emulsion. Optical micrographs revealed that it was an oil-in-water Pickeringemulsion, and the droplets sizes were about 2-10 μm. To the best of our knowledge, MOFs including ZIF-8 have been confirmed to be good catalysts for the Knoevenagel condensations in organic solvents [17,18]. The stability in aqueous medium of ZIF-8 has been proven by the XRD patterns and TEM images as shown in Figures 2 and 3 in our experiments. Therefore, the activities of ZIF-8 nanoparticles for the Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein various solvents were firstly investigated, and the results are summarized in Table 1. Similar to the results observed in previous reports, ZIF-8 gave high or medium yields in polar solvents and poor results in nonpolar solvents. As we know, catalytic reaction in water or biphasic systems is a greener route than in organic solvent. To our delight, ZIF-8 could also promote the reaction in water in an efficient way. Over an 80% yield was obtained in water after 30 min, a little bit lower than that in DMF, but higher than that in THF. Then, a water/ZIF-8/ethyl acetate system was used to examine the interfacial property of ZIF-8. After adding 2 mLof pure or 2-nitrobenzaldehydedissolved ethyl acetate to 3 mL of water containing 0.8 wt % of ZIF-8 (with respect to water), as shown in Figure 4A,B, ZIF-8 is predominantlydistributed at the interface between the oil and aqueous phase. After subsequentvigorous stirring or shaking, the appearances of its mixture ( Figure 4C) and its optical micrograph demonstrated the formation of Pickering emulsion. Optical micrographs revealed that it was an oil-in-water Pickeringemulsion, and the droplets sizes were about 2-10 µm.   Having established that ZIF-8 is an efficient catalyst for the Knoevenagel condensation between benzaldehyde and ethylcyanoacetate in an aqueous or biphasic system, we extended the study to various aromatic aldehydes with ethyl cyanoacetate or malononitrile. The results are listed in Table 2. To our delight, most of the aldehyde derivatives could be converted efficiently. Substitution and size effects were also observed. In our previous works, the effect of substitution on the reactivityof benzaldehyde catalyzed by UiO-66-NH2was also observed for the Knoevenagel condensation [22,23]. The electron-donating substitutions lowered the catalytic efficiency, but the electron-withdrawing groups enhanced its catalytic efficiency. ZIF-8 was also active for the Knoevenagel reactions of larger aromatic aldehydes (1-naphthaldehyde and 9-anthraldehyde) with cyanoacetate or malononitrile, though the yields were somewhat decreased when the aldehyde substrates become bulkier (comparing Entries 1, 7 and 8 in Table 2). The decrease of yields may be an indicator of size effects. As shown by Entries 1, 7 and 8, the yield of 9-anthraldehyde reacting with ethyl cyanoacetate (2 h, 23%) is much lower than that of 1-naphthaldehyde (1 h, 45%), which is much lower than that of benzaldehyde (1 h, >99%). The activities vary in the size order of benzaldehyde> 1-naphthaldehyde > 9-anthraldehyde. This trend holds true forthat of malononitrile. All results above reflect that the probability of the bulkier substrates forming transition-state complexes was significantly reduced due to the limited space in the porous catalyst.   Having established that ZIF-8 is an efficient catalyst for the Knoevenagel condensation between benzaldehyde and ethylcyanoacetate in an aqueous or biphasic system, we extended the study to various aromatic aldehydes with ethyl cyanoacetate or malononitrile. The results are listed in Table 2. To our delight, most of the aldehyde derivatives could be converted efficiently. Substitution and size effects were also observed. In our previous works, the effect of substitution on the reactivityof benzaldehyde catalyzed by UiO-66-NH2was also observed for the Knoevenagel condensation [22,23]. The electron-donating substitutions lowered the catalytic efficiency, but the electron-withdrawing groups enhanced its catalytic efficiency. ZIF-8 was also active for the Knoevenagel reactions of larger aromatic aldehydes (1-naphthaldehyde and 9-anthraldehyde) with cyanoacetate or malononitrile, though the yields were somewhat decreased when the aldehyde substrates become bulkier (comparing Entries 1, 7 and 8 in Table 2). The decrease of yields may be an indicator of size effects. As shown by Entries 1, 7 and 8, the yield of 9-anthraldehyde reacting with ethyl cyanoacetate (2 h, 23%) is much lower than that of 1-naphthaldehyde (1 h, 45%), which is much lower than that of benzaldehyde (1 h, >99%). The activities vary in the size order of benzaldehyde> 1-naphthaldehyde > 9-anthraldehyde. This trend holds true forthat of malononitrile. All results above reflect that the probability of the bulkier substrates forming transition-state complexes was significantly reduced due to the limited space in the porous catalyst. Having established that ZIF-8 is an efficient catalyst for the Knoevenagel condensation between benzaldehyde and ethylcyanoacetate in an aqueous or biphasic system, we extended the study to various aromatic aldehydes with ethyl cyanoacetate or malononitrile. The results are listed in Table 2. To our delight, most of the aldehyde derivatives could be converted efficiently. Substitution and size effects were also observed. In our previous works, the effect of substitution on the reactivityof benzaldehyde catalyzed by UiO-66-NH 2 was also observed for the Knoevenagel condensation [22,23]. The electron-donating substitutions lowered the catalytic efficiency, but the electron-withdrawing groups enhanced its catalytic efficiency. ZIF-8 was also active for the Knoevenagel reactions of larger aromatic aldehydes (1-naphthaldehyde and 9-anthraldehyde) with cyanoacetate or malononitrile, though the yields were somewhat decreased when the aldehyde substrates become bulkier (comparing Entries 1, 7 and 8 in Table 2). The decrease of yields may be an indicator of size effects. As shown by Entries 1, 7 and 8, the yield of 9-anthraldehyde reacting with ethyl cyanoacetate (2 h, 23%) is much lower than that of 1-naphthaldehyde (1 h, 45%), which is much lower than that of benzaldehyde (1 h, >99%). The activities vary in the size order of benzaldehyde> 1-naphthaldehyde > 9-anthraldehyde. This trend holds true forthat of malononitrile. All results above reflect that the probability of the bulkier substrates forming transition-state complexes was significantly reduced due to the limited space in the porous catalyst.  The recyclability and reusability of solid catalysts arethe most positiveaspectscompared to homogeneous catalysts for practical applications. Recovery and reuse of ZIF-8 was firstly studied using Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein a biphasic system at 80 °C. After the completion of the reaction, the liquidlayer was decanted, and the catalyst  The recyclability and reusability of solid catalysts arethe most positiveaspectscompared to homogeneous catalysts for practical applications. Recovery and reuse of ZIF-8 was firstly studied using Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein a biphasic system at 80 °C. After the completion of the reaction, the liquidlayer was decanted, and the catalyst            The recyclability and reusability of solid catalysts arethe most positiveaspectscompared to homogeneous catalysts for practical applications. Recovery and reuse of ZIF-8 was firstly studied using Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein a biphasic system at 80 °C. After the completion of the reaction, the liquidlayer was decanted, and the catalyst   The recyclability and reusability of solid catalysts arethe most positiveaspectscompared to homogeneous catalysts for practical applications. Recovery and reuse of ZIF-8 was firstly studied using Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein a biphasic system at 80 °C. After the completion of the reaction, the liquidlayer was decanted, and the catalyst     The recyclability and reusability of solid catalysts arethe most positiveaspectscompared to homogeneous catalysts for practical applications. Recovery and reuse of ZIF-8 was firstly studied using Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein a biphasic system at 80 °C. After the completion of the reaction, the liquidlayer was decanted, and the catalyst           The recyclability and reusability of solid catalysts arethe most positiveaspectscompared to homogeneous catalysts for practical applications. Recovery and reuse of ZIF-8 was firstly studied using Knoevenagel condensation between benzaldehyde and ethyl cyanoacetatein a biphasic system at 80 • C. After the completion of the reaction, the liquidlayer was decanted, and the catalyst was reused for the next run under the same conditions. Furthermore, phase structures of recovered ZIF-8 were determined by powder X-ray diffraction. It was observed thatthe phase structure of ZIF-8 was completely destroyed after the reaction (Figure 5b). The following experiments were carried out to find out the possible destroyers. Previous resultshaveproven that ZIF-8 is a water-stable MOF. Therefore, the stability of ZIF-8 in the aqueous solution of cyanoacetate and benzaldehyde wasfurther tested, respectively. As shown in Figure 5c,d, the phase structure of ZIF-8 was maintained in the aqueous solution of cyanoacetate, but changed in that of benzaldehyde. The same results were observed even under mild reaction conditions by replacingethyl cyanoacetate with malononitrile at 40 • C. The tested pH value of the aqueous solution of benzaldehyde wasbelow five, indicating that the acid impurity was the possible destroyer of ZIF-8. Many reports indicated that the stability of ZIF-8 in aqueous medium is pH dependent. It is stable in a basic system, but instable in acidic medium [24,25]. Therefore, 2-nitro-benzaldehyde with high purity was selected as a new substrate. To our delight, as shown in Figure 6c, the XRD pattern of the recovered ZIF-8 is same asthat of the fresh solid.
The recycle experiment was carried out using the Knoevenagel condensation between 2-nitrobenzaldehyde and malononitrileat 40 • C in a biphasic system as a test reaction. Although some solid was lost during the centrifugation, as shown in Figure 7, the yields in the consecutive cycleswerealmost the same asthoseof the first cycle. XRD patterns of ZIF-8 before and afterreaction ( Figure 8) revealedbetter integrity in the framework structure, but the TEM image (Figure 9c) demonstrated the presence of newly-formed mesoporous cages in the recovered solid. The retainedor improved specific activity canprobably be attributedto the more exposed active site and better mass transfer in the hierarchical ZIF-8. A hot filtration experiment was further performed to confirm theheterogeneous nature of the catalytic reaction ( Figure 10). The solid catalyst was removed from hot solution by filtration 10 min after initiating the catalytic test run. The reaction of the filtrate was then monitored for another 50 min. No significant further conversions were observed, indicating that most of the conversions were theresult ofthe heterogeneous catalysis. was reused for the next run under the same conditions. Furthermore, phase structures of recovered ZIF-8 were determined by powder X-ray diffraction. It was observed thatthe phase structure of ZIF-8 was completely destroyed after the reaction (Figure 5b). The following experiments were carried out to find out the possible destroyers. Previous resultshaveproven that ZIF-8 is a water-stable MOF. Therefore, the stability of ZIF-8 in the aqueous solution of cyanoacetate and benzaldehyde wasfurther tested, respectively. As shown in Figure 5c,d, the phase structure of ZIF-8 was maintained in the aqueous solution of cyanoacetate, but changed in that of benzaldehyde. The same results were observed even under mild reaction conditions by replacingethyl cyanoacetate with malononitrile at 40 °C. The tested pH value of the aqueous solution of benzaldehyde wasbelow five, indicating that the acid impurity was the possible destroyer of ZIF-8. Many reports indicated that the stability of ZIF-8 in aqueous medium is pH dependent. It is stable in a basic system, but instable in acidic medium [24,25]. Therefore, 2-nitro-benzaldehyde with high purity was selected as a new substrate. To our delight, as shown in Figure 6c, the XRD pattern of the recovered ZIF-8 is same asthat of the fresh solid.   The recycle experiment was carried out using the Knoevenagel condensation between 2-nitrobenzaldehyde and malononitrileat 40 °C in a biphasic system as a test reaction. Although some solid was lost during the centrifugation, as shown in Figure 7, the yields in the consecutive cycleswerealmost the same asthoseof the first cycle. XRD patterns of ZIF-8 before and afterreaction ( Figure 8) revealedbetter integrity in the framework structure, but the TEM image (Figure 9c) demonstrated the presence of newly-formed mesoporous cages in the recovered solid. The retainedor improved specific activity canprobably be attributedto the more exposed active site and better mass transfer in the hierarchical ZIF-8. A hot filtration experiment was further performed to confirm theheterogeneous nature of the catalytic reaction ( Figure 10). The solid catalyst was removed from hot solution by filtration 10 min after initiating the catalytic test run. The reaction of the filtrate was then monitored for another 50 min. No significant further conversions were observed, indicating that most of the conversions were theresult ofthe heterogeneous catalysis.  The recycle experiment was carried out using the Knoevenagel condensation between 2-nitrobenzaldehyde and malononitrileat 40 °C in a biphasic system as a test reaction. Although some solid was lost during the centrifugation, as shown in Figure 7, the yields in the consecutive cycleswerealmost the same asthoseof the first cycle. XRD patterns of ZIF-8 before and afterreaction ( Figure 8) revealedbetter integrity in the framework structure, but the TEM image (Figure 9c) demonstrated the presence of newly-formed mesoporous cages in the recovered solid. The retainedor improved specific activity canprobably be attributedto the more exposed active site and better mass transfer in the hierarchical ZIF-8. A hot filtration experiment was further performed to confirm theheterogeneous nature of the catalytic reaction ( Figure 10). The solid catalyst was removed from hot solution by filtration 10 min after initiating the catalytic test run. The reaction of the filtrate was then monitored for another 50 min. No significant further conversions were observed, indicating that most of the conversions were theresult ofthe heterogeneous catalysis.

Characterization
The crystal structures of materials were analyzed by a Bruker D4 Endeavour Powder X-ray diffractometer (Cu Kα, λ = 0.15405 nm, 40 kV and 40 mA). The morphology and size of the ZIF-8 were determined by scanning electronic microscopy (SEM, HitachiS4800, Tokyo, Japan, 5.0 kV) and transmission electron microscopy (TEM, JEM-2100, Tokyo, Japan, 200 kV). The Pickering emulsions were measured on an Olympus BX51 microscope.

Catalytic Activity Measurements
Typically, 25 mg of ZIF-8, 4 mmolof ethyl cyanoacetate or malononitrile and 5 mL of solvents were added into a 10-mL glass reactor. After stirring for 10 min at the desired temperature under an inert atmosphere, 3.5 mmol of benzaldehyde or derivatives were added to start the reaction. After the reaction, the catalyst was separated by centrifugation, successively rinsed withfresh ethyl acetate, ready for the next cycle. Yields of the reactions were estimated by gas chromatography (GC7900, Techcom, Shanghai, China) equipped with a flame ionization detector (FID) and an SE-30 capillary column (30 m × 0.25 mm × 0.25 µm).

Conclusions
In summary, ZIF-8 wasdemonstrated to be a water stable and an interface-active catalyst for Knoevenagel condensation in a biphasic system. It efficiently promoted the Knoevenagel condensations of various aromatic aldehydes in biphasic systems as an interfacial catalyst. Pickering emulsions formed during the reaction due to its amphiphilic property, which wasbeneficial to mass transfer and enhanced its catalytic performance in a biphasicsystem. It exhibited good structural stability during the Knoevenagel condensation reactions and could be recovered and recycled as a heterogeneous catalyst for more than sixtimes without obvious loss of activity in the condensation between 2-nitrobenzaldehyde and malononitrile at mild reaction conditions. The enhanced specific activity in the consecutive runs wasascribed toits newly-formed hierarchical structure observed by TEM images. Furtherdevelopment of catalytic reactions by using hierarchical ZIF-8is currently under investigation. The findings mentioned above will open an avenue for using metal-organic frameworks as an interfacial solid catalyst.