Effect of MAF-6 Crystal Size on Its Physicochemical and Catalytic Properties in the Cycloaddition of CO 2 to Propylene Oxide

: Zeolitic imidazolate frameworks MAF-5 and MAF-6 based on Zn 2+ and 2-ethylimidazole were demonstrated to be efﬁcient heterogeneous catalysts in solvent-free coupling of CO 2 and propylene oxide (PO) to produce propylene carbonate (PC) at 0.8 MPa of CO 2 and 80 ◦ C. Activity of MAF-5 was lower in comparison with MAF-6 due to the difference in their structural and textural characteristics. MAF-6 samples with particle size of 190 ± 20, 360 ± 30, and 810 ± 30 nm were prepared at room temperature from [Zn(NH 3 ) 4 ](OH) 2 and 2-ethylimidazole. Control of particle size was achieved by variation of type of alcohol in alcohol/cyclohexane media for the preparation of MAF-6. According to this comprehensive study, the yield of PC was found to decrease with increasing crystal size of the MAF-6 material, which was related to the change in textural properties and the number and localization of active sites. The combination of MAF-6 with particle size of with particle size of 190 ± 20 nm and tetrabutylammonium bromide ([n-Bu 4 N]Br) as co-catalyst led to an approximately 4-fold enhancement in the yield of PC (80.5%). Compared with reported ZIFs catalysts, the efﬁciencies of MAF-5/[n-Bu 4 N]Br and MAF-6/[n-Bu 4 N]Br binary systems were comparable and higher under similar reaction conditions.


Introduction
Nowadays, it nearly impossible to imagine modern life without the products of chemistry. The chemicals industry and other related industries supply us with a huge variety of essential products, from plastics to pharmaceuticals. However, these industries are based on the utilization of fossil fuels as a source of energy and feedstocks, and often release pollutants into the environment that provokes both economic and environmental challenges. Moreover, due to the technological progress and rapid expansion of the population in the World since the second half of the 20th century immediate availability would no longer be able to fulfill the increasing global demand for these raw materials. To use renewable carbon resources, i.e., biomass and CO 2 , in the chemical and energy industries, it is a key to finding lasting solutions to both economic and environmental challenges.
Carbon dioxide is a waste product from all combustion processes and many chemical industries. CO 2 represents a potential and alternative carbon feedstock for the preparation of a variety of useful chemicals [1][2][3][4]. Several routes having a significant potential for both the chemical utilization of large amounts of CO 2 and becoming economically profitable were pointed in the position paper «Utilization and storage of CO 2 » [5]. The application There have been many different types of homogeneous and heterogeneous catalysts suggested for the synthesis of PC via the coupling reaction of carbon dioxide with PO. Yano et al. [16] demonstrated that a commercial MgO could give PC with 41% yield at 135 °C and 2 MPa of CO2 for 12 h in a solvent of dimethylformamide. According to the work in [17], the catalytic efficiency of MgO was higher in comparison with ZnO, ZrO2, La2O3, CeO2, CaO, and Al2O3. In the presence of MgO a yield of PC was 32% with a 92% selectivity at 150 °C and 8 MPa of CO2 for 15 h. At the same time, a higher PC yield of 54% and lower selectivity (75%) were observed in the presence of La2O3. These differences were explained by a large amount of moderately basic sites in MgO.
The coupling reaction of carbon dioxide with PO to produce PC was also efficiently catalyzed by ionic liquids (ILs) based on large organic cations like imidazolium, ammonium, pyridinium ones, and inorganic anions [18][19][20] and systems based on immobilized ILs [21]. Type of cation and anion affect the reaction rate and selectivity towards PC [22]. In the case of 1-n-butyl-3-methylimidazolium (BMIM) derivatives, high conversion and selectivity were found to be observed in anions with Brand I -, while relatively low selectivity was observed in the cases of BMIM with Cl -, AcO -, HCO3 -, malonate, and proline. The high activity of systems based on ILs was explained by the interaction between the imidazolium ring proton at the C2-position and the epoxide oxygen atom that make easy epoxide ring-opening in the course of nucleophilic attack of Bron the C atom of PO. There have been many different types of homogeneous and heterogeneous catalysts suggested for the synthesis of PC via the coupling reaction of carbon dioxide with PO. Yano et al. [16] demonstrated that a commercial MgO could give PC with 41% yield at 135 • C and 2 MPa of CO 2 for 12 h in a solvent of dimethylformamide. According to the work in [17], the catalytic efficiency of MgO was higher in comparison with ZnO, ZrO 2 , La 2 O 3 , CeO 2 , CaO, and Al 2 O 3 . In the presence of MgO a yield of PC was 32% with a 92% selectivity at 150 • C and 8 MPa of CO 2 for 15 h. At the same time, a higher PC yield of 54% and lower selectivity (75%) were observed in the presence of La 2 O 3 . These differences were explained by a large amount of moderately basic sites in MgO.
The coupling reaction of carbon dioxide with PO to produce PC was also efficiently catalyzed by ionic liquids (ILs) based on large organic cations like imidazolium, ammonium, pyridinium ones, and inorganic anions [18][19][20] and systems based on immobilized ILs [21]. Type of cation and anion affect the reaction rate and selectivity towards PC [22]. In the case of 1-n-butyl-3-methylimidazolium (BMIM) derivatives, high conversion and selectivity were found to be observed in anions with Brand I -, while relatively low selectivity was observed in the cases of BMIM with Cl -, AcO -, HCO 3 -, malonate, and proline. The high activity of systems based on ILs was explained by the interaction between the imidazolium ring proton at the C2-position and the epoxide oxygen atom that make easy epoxide ring-opening in the course of nucleophilic attack of Bron the C atom of PO.
Metal-organic frameworks (MOFs)-materials fabricated by metal ions/clusters and organic ligands-were also found to catalyze the reaction between CO 2 and PO [23,24]. The catalytic activity of MOFs systems resulted in the high CO 2 adsorption capacity on the active sites, which are due to the defect sites within or on the surface of the crystalline mate- rials. According to the work in [25], the addition of the quaternary ammonium salts (QAS) to MOF-5 leads to an increase in yield of PC in the presence of MOF-5/[QAS]Br binary system, and the yield of PC increases with the alkyl chain length in the following order: In optimal experimental conditions (6 MPa, 50 C, 4 h), the yield of PC was 95.5% in the presence of MOF-5/[n-Bu 4 N]Br. Notable that this binary system could be used for three cycles without loss in activity.
Zeolitic imidazolate frameworks (ZIFs), as a sub-class of MOFs, were also reported as efficient catalysts in the cycloaddition of CO 2 to PO [23,24]. ZIFs are porous crystals with zeolite-type tetrahedral topologies, which are constructed from tetrahedral metal ions (Zn 2+ , Co 2+ , Fe 2+ , etc.) bridged by imidazolate ligands. ZIFs combine the benefits of zeolites and MOF materials: high chemical and thermal stabilities, uniform pore distribution, and high specific surface area and microporosity. ZIFs draw attention as catalysts due to the availability of Zn 2+ (as Lewis acidic sites, LAS) and imidazolate linkers (as basic sites, BS) in the structure. Some especially noticeable features of ZIFs are high hydrophobicity and excellent capability for CO 2 adsorption and capture. Nowadays, a series of ZIFs and ZIFs/[n-Bu 4 N]Br binary systems were reported (Tables 1 and 2) [26][27][28][29][30][31][32][33]. The yield of cyclic carbonate was 49-81% with 95-99% selectivity in the presence of ZIF-23, ZIF-90, ZIF-90, F-ZIF-90, and ZIF-95 at 0.8-1.2 MPa of CO 2 and 100-140 • C ( Table 1). The yield of PO was significantly enhanced in the presence of mixed CoZn-ZIF composite due to the synergistic effect of Lewis acid-base by comparing the catalytic performance of ZIF-8(Zn) and ZIF-67(Co) ( Table 1). Park et al. [32] suggested that this phenomenon was resulted in exploiting the advantageous properties of Zn (high activity) and Co (low leaching and high selectivity). The addition of co-catalyst favors the increasing catalytic efficiency of ZIFs that allows reducing reaction rate and pressure of CO 2 . Thus, high activity of ZIF-23/[n-Bu 4 N]Br [27] and ZIF-95/[n-Bu 4 N]Br [31] binary systems was observed at 80-100 • C and 1.2 MPa (Table 2).  Recently, Chen et al. [34] demonstrated that the reaction rate of cycloaddition between CO 2 and styrene oxide could be tuned easily by the morphology control synthesis of ZIF-8. Morphology of ZIF-8 was found to affect the porosity and the number of acidic and basic sites and, therefore, catalytic performance. The activity of ZIF-8 decreased in the following order: solid spheres > hollow spheres > hierarchical particles > stackable particle > nano-dodecahedrons.
The high activity of ZIF-8 solid spheres was due to its highest specific surface areas, most abundant acid-basic sites, and large mesoporous volume. The effect of the morphology of ZIF-8 particles was also demonstrated for the condensation reaction of 4-bromobenzaldehyde (BA-Br) with malononitrile [35]. The conversion of BA-Br for 30 min of the reaction decreased in the following order: truncated rhombic dodecahedron (141 ± 48 nm) > rhombic dodecahedron (500 nm) > cubes and a rhombic dodecahedron (~1050 nm).
This order was explained by the decrease in the number of available active sites due to a decrease in the area of the outer surface of ZIF-8 crystals. A similar effect of morphology and particle size of ZIF-8 was also demonstrated by Zhang et al. [36] for [3+3] cycloaddition reactions between dions and aldehydes.
We have recently studied the effect of ZIF-8 particle size (50-800 nm) in a hexagonal shape on their catalytic efficiency in the synthesis of propylene glycol methyl ether (PGME) from PO and methanol [37,38]. A pronounced size effect on the reaction rate and selectivity toward PGME was found to decrease along with the growth of the ZIF-8 crystal size. According to the physicochemical investigations, this phenomenon was explained by the change in textural characteristics and the surface-to-volume ratio of the number of the basic sites. Note that reaction rate and selectivity towards PGME depended also on the type of ZIFs (Table S1, Supporting Information) and decreased in the following order: MAF-5 (ZIF-71) > MAF-6 (ZIF-14) > ZIF-8. This correlation was caused by the increasing channel size and decreasing surface basicity of ZIFs [38,39].
Herein, we present an investigation of the catalytic properties of MAF-5 and MAF-6 in the coupling reaction of carbon dioxide with PO to produce PC. The main attention was focused on the examination of the effect of the MAF-6 catalyst crystallite size on its catalytic performance, because the reduction of the particle size from the micrometer to the nanometer scale could involve changes in its physicochemical and catalytic properties. A combination of catalytic and physicochemical methods was used for solving such a challenge. Thus, the nature of the active sites of the MAF-6 samples was analyzed by DRIFT spectroscopy using CDCl 3 as a probe molecule and DR-UV-Vis spectroscopy. Morphology and particle size were analyzed by SEM and XRD methods. We assume that investigation of size effect could open the way for the tuning and control catalytic properties of MAF-6, in particular, and ZIFs, in general, in this reaction.

Results and Discussion
2.1. Synthesis and Characterization of MAF-6 and MAF-5 MAF-5 was synthesized at room temperature with 2-EIM/Zn 2+ molar ratio of 2.0 in methanol solution. A combination of physicochemical methods points to the formation of the MAF-5 structure with high crystallinity [39]. SEM images, XRD pattern and textural characteristics are shown in Figures S1-S3 (Supporting Information) and Table 3. The zeolitic imidazolate framework MAF-6s with different particle sizes were synthesized from [Zn(NH 3 ) 4 ](OH) 2 and 2-EIM. The particle size estimated by SEM was tuned by the changing of solvent type (Table 3). Microscopy images of MAF-6 samples are shown in Figure 1. One can see that all samples have uniform particle size distributions. The morphology of crystals is a rhombic dodecahedron structure exposing 12{110} faces. The experimental data (Table 3) show the particle size of MAF-6 decreases in the following order, because of solvent effect: Earlier, the effect of solvent on the crystallization rate and size of crystal was demonstrated for ZIF-8 [40]. Similar to our results, the crystal size of ZIF-8 was demonstrated to decrease in the following order: MeOH > EtOH > PrOH. Table 3. Synthesis, abbreviation, textural properties, and crystal sizes of ZIFs.

Experimental Conditions
Crystal Size b (nm)

Textural Properties
Solvent α a S BET (m 2 /g)

MAF-6(S)
PrOH/cyclohexane 28 The impact of solvent was explained by its role as a structure-directing agent via its included in the framework by non-covalent interactions that can be confirmed by the correlation between crystal size of ZIF-8 and value of the hydrogen bond donation of solvents (α) [41]. Park et al. [42] also suggested the interaction between solvent and framework of ZIF-8 that followed from the presence of solvent molecules near the hydrogen atoms of 2-methylimidazolate groups.
Ideally, Zn 2+ ion bounds to the four ligands via nitrogen atoms in the structure of ZIFs (ZnIV atoms). However, DFT calculations [57,58] pointed that ZIF-8 can have lowcoordinated Zn atoms (Zn atoms without one or two linkers, ZnIII and ZnII atoms), and free N-moieties belonging to linkers can exist on the external surface of solid. For the identification of λ300 and λ360 bands, the following points should be taken into account: (1) The energy of ligand-to-metal charge transfer depends on the coordination of zinc atoms in the samples.
(2) The increasing integral intensity of the band at 300 nm in DR-UV-Vis spectra of MAF-6 samples after heating under vacuum at 150 °C for 6 h and the next increasing after wet air addition ( Figure S6, Supporting Information). Where D XRD is particle size (nm), β is the full-width at half maximum of the peak (radian), Θ is the Bragg angle of diffraction peak (grad), and λ is X-ray wavelength of CuK α (0.1542 nm). The D XRD value for MAF-6 was found to change in the following order (Table S2, Supporting Information): MAF-6(L) (36 ± 2 nm) > MAF-6(M) (29 ± 2 nm) > MAF-6(S) (24 ± 1 nm).
If we assume that MAF-6 particles consist of domains and the size of each domain is equal to that of the region of X-ray coherent scattering (D XRD ), the D SEM /D XRD ratio can be considered as a parameter indicating the relationship between regular and imperfect structural regions in the bulk of the MAF-6 particle (Table S2, Supporting Information). The data suggest that the size of imperfect structural regions rises with decreasing particle size (D SEM /D XRD ): MAF-6(L) (23) < MAF-6(M) (12) < MAF-6(S) (10) Textural properties of MAF-6 samples are shown in Table 3. It was found that the specific surface area and microporosity (V µ /V Σ ) of MAF-6 increased with increasing crystal size. This phenomenon can be explained by the impact of the nature of alcohols used in MAF-6 synthesis on the formation of pore structure. First, both basicity of solvent and size of molecule of solvent are important for the hydrogen bonding interaction between the alcohol and 2-MIM, i.e., the linker deprotonation, because this interaction affects the rate of nucleation and crystal growth. Moreover, molecules of alcohol can be incorporated in the as-synthesized MAF-6 structure and act as a space-filling molecule as kinetic diameters of alcohols (MeOH 3.8 Å, EtOH 4.30 Å and n-PrOH 4.7 Å [48] are smaller in comparison with the pore size of MAF-6 (theoretical pore size: 7.6 Å [49]). Note that the effect of particle size on the textural properties was demonstrated for ZIF-8 [37,50,51]. It was also found that the external surface decreases, while microporosity increases with increasing crystal size.

Nature of Active Sites of MAF-6
The nature and structure of active sites were investigated by different spectral methods, such as DRIFT spectroscopy, DR-UV-Vis spectroscopy, and DRIFT spectroscopy using CDCl 3 as a probe molecule.
DRIFT spectroscopy. The DRIFT spectra of MAF-6 samples with different particle sizes are shown in Figures 3 and S5 (Supporting Information). In spectra of samples, the bands observed in the regions of 3650-3300 cm −1 and 1550-1650 cm −1 may be attributed to the ν (OH) stretching mode of water and alcohol molecules [52][53][54]. The 2-ethylimidazole has bands in the region of 2800-3150 cm −1 and at 1152 cm −1 which are assigned to the ν (CH3) and ν (CH2) asymmetric and symmetric stretching frequencies and ρ (CH2) in-plane rocking vibration frequency, respectively. Moreover, the δ (CH3) in-plane and ω (CH3) out of plane bending modes are observed at 1172 and 1143 cm −1 in the DRIFT spectrum, respectively. The observed bands at 1470 and 1454 cm −1 are assigned to ν (C-N) stretching vibration in 2-ethylimidazole ligand. The band in the region of 410 to 440 cm −1 is particularly worth discussing. This band is assigned to the Zn-N stretching mode due to the interaction between a Zn atom and an N atom of 2-ethylimidazole. In a spectrum of MAF-6(L), a symmetric band at 423 cm −1 is observed. At the same time, this band splits to 418 and 434 cm −1 in spectra of MAF-6(M) and MAF-6(S), respectively. We can suggest that band at 434 cm −1 is attributed to the Zn-O stretching mode due to the interaction between the Zn 2+ ions of MAF-6 and the oxygen atom of alcohol [37,39]. The change in the character of bands in the region of 770-790 cm −1 (OH out-of-plane bend) ( Figure 3) and 1300-1330 cm −1 (OH in-plane bend) ( Figure S5, Supporting Information) point also to this interaction.
Ideally, Zn 2+ ion bounds to the four ligands via nitrogen atoms in the structure of ZIFs (ZnIV atoms). However, DFT calculations [57,58] pointed that ZIF-8 can have lowcoordinated Zn atoms (Zn atoms without one or two linkers, ZnIII and ZnII atoms), and free N-moieties belonging to linkers can exist on the external surface of solid. For the identification of λ300 and λ360 bands, the following points should be taken into account: (1) The energy of ligand-to-metal charge transfer depends on the coordination of zinc atoms in the samples.
Ideally, Zn 2+ ion bounds to the four ligands via nitrogen atoms in the structure of ZIFs (Zn IV atoms). However, DFT calculations [57,58] pointed that ZIF-8 can have lowcoordinated Zn atoms (Zn atoms without one or two linkers, Zn III and Zn II atoms), and free N-moieties belonging to linkers can exist on the external surface of solid. For the identification of λ 300 and λ 360 bands, the following points should be taken into account: (1) The energy of ligand-to-metal charge transfer depends on the coordination of zinc atoms in the samples.
(2) The increasing integral intensity of the band at 300 nm in DR-UV-Vis spectra of MAF-6 samples after heating under vacuum at 150 • C for 6 h and the next increasing after wet air addition ( Figure S6, Supporting Information).
Therefore, λ 300 and λ 360 bands can be ascribed to the metal-to-ligand charge transfer from low-coordinated Zn atoms (Zn III and Zn II ) to ligand and high-coordinated Zn atoms (Zn IV ) to ligand, respectively. This suggestion agrees with the correlation between D SEM /D XRD ratio and the ratio of intensities of λ 300 and λ 360 (I 300 /I 360 ) ( Figure 4B). The I 300 /I 360 ratio decreases with increasing D SEM /D XRD , i.e., with the increasing size of regular structural regions. Therefore, this correlation points that Lewis acidic sites are located in the defect area and formed by low-coordinated Zn atoms (Zn III and Zn II ) Basic properties of ZIFs. Basicity of MAF-6 samples was investigated by DRIFT spectroscopy using deuterated chloroform (CDCl 3 ) as a C-H acid probe (DRIFT-CDCl 3 ). This method was successfully used to analyze the basicity of metal-organic frameworks, such as MIL-100(Al), Cu(BTC) and UiO-66 [59,60], NH 2 -UiO-66 [60], ZIF-8, and MAF-5 [37,39,60].
The spectra of CDCl 3 adsorbed on MAF-6 with different particle sizes are shown in Figures 5A and S7 (Supporting Information). Two bands at 2242 and 2250 cm −1 assigned to the interaction of CDCl 3 with basic sites (BS) can be distinguished. The strength of these bands as the proton affinity (PA) is 870 and 839 kJ/mol, respectively (Table 4). These values indicate that MAF-6 possesses lower basicity than that of 2-ethylimidazole (978 kJ/mol) [61]. The decreasing basicity of linker in ZIFs is due to the formation of a Zn-N bond via the interaction between Zn 2+ and the nitrogen atom of the organic linker. The particle size of MAF-6 affects the number of basic sites that following from the increasing integral intensity of the band at 2242 cm −1 (I2242) with decreasing particle size of MAF-6 ( Figure 4B). According to DFT calculations and different studies [58,62], basic sites are formed by Zn-OH groups and N-moieties of nonbridging linkers due to the existence of low-coordinated Zn atoms (ZnIII and ZnII). This assertion agrees with the linear correlation between the intensity of band at 2242 cm −1 (I2242) and the ratio of intensities of λ300 and λ360 (I300/I360) ( Figure 5B). Note that decreasing the number of basic sites with increasing particle size is consistent with the changing of DSEM/DXRD ratio ( Figure 4B). The increasing size of regular structural regions leads to a decreasing number of basic sites.
It was interesting to compare the strength of basic sites of MAF-6 with that of ZIFs and MOFs reported in the literature. Data shown in Table 4 point that the basicity of ZIF-8 (858 kJ/mol) formed by 2-MIM is lower than that of MAF-6 (870 kJ/mol) formed by 2-   [60] 2253 839 According to experimental data, particle size does not affect the strength of MAF-6 samples. Rather, the amount and accessibility of basic sites depend more on this parameter. Comparing the rate of CDCl 3 adsorption onto MAF-6, one can easily see the effect of particle size on the accessibility of basic sites ( Figure S7, Supporting Information). When particle size decreases, the rate of CDCl 3 adsorption rises, i.e., sample with smaller particle faster uptakes CDCl 3 . Adsorption rate is dependent not only on particle size, but also on the microporosity of samples (V µ /V Σ ) ( Table 3) which affects the diffusion of CDCl 3 molecules inside the porous MAF-6 particle. Note that specific surface area and pore volume of MAF-6 samples play a minor role in the adsorption kinetics (Table 3). MAF-6(S) with 190 nm particle size has the highest adsorption rate despite of the lowest S BET and total pore volume (V Σ ).
The particle size of MAF-6 affects the number of basic sites that following from the increasing integral intensity of the band at 2242 cm −1 (I 2242 ) with decreasing particle size of MAF-6 ( Figure 4B). According to DFT calculations and different studies [58,62], basic sites are formed by Zn-OH groups and N-moieties of nonbridging linkers due to the existence of low-coordinated Zn atoms (Zn III and Zn II ). This assertion agrees with the linear correlation between the intensity of band at 2242 cm −1 (I 2242 ) and the ratio of intensities of λ 300 and λ 360 (I 300 /I 360 ) ( Figure 5B). Note that decreasing the number of basic sites with increasing particle size is consistent with the changing of D SEM /D XRD ratio ( Figure 4B). The increasing size of regular structural regions leads to a decreasing number of basic sites.
It was interesting to compare the strength of basic sites of MAF-6 with that of ZIFs and MOFs reported in the literature. Data shown in Table 4 point that the basicity of ZIF-8 (858 kJ/mol) formed by 2-MIM is lower than that of MAF-6 (870 kJ/mol) formed by 2-EIM that is related to the lower basicity of 2-MIM than 2-EIM and the difference in pK a of the conjugate acid for 2-MIM (7.85 [63,64]) and 2-EIM (8.0 [63,64]). At the same time, the basicity of MAF-5 also formed by 2-EIM is higher than that of MAF-6. We can suggest that this difference may be related to the high distortion of the framework of MAF-5 (Table S1, Supporting Information), which probably leads to changes in electron density in the basic sites. In general, basicity of ZIFs is higher than that of NH 2 -UiO-66(Zr) [60]. We can suggest that the low basicity of NH 2 -UiO-66(Zr) is due to the interaction strength between -NH 2 groups and carboxylic groups of the linker (2-amino-terephthalic acid).

Catalytic Properties of MAF-6
Catalytic properties of MAF-5 and MAF-6 were investigated in the cycloaddition of CO 2 to PO at 80 C under 0.8 MPa of CO 2 . The main results are shown in Tables 1 and 2. According to GC-MS analysis, PC was the major product with > 99% selectivity in all cases. A blank experiment pointed that the conversion of PO was << 1% after 5 h. In the presence of MAF-5, the PO conversion increased to 8.2% (Table 1). After the addition of [n-Bu 4 N]Br as co-catalyst, conversion of PO and yield of PC strongly rose ( Table 2). The activity of MAF-5/[n-Bu 4 N]Br binary system depended on the weight ratio of [n-Bu 4 N]Br/MAF-5. The activity of binary system rose with increasing this ratio ( Figure S8, Supporting Information). The maximal yield of PC was 66.6% at a ratio of 1/1, 80 C for 5 h. The yield of PC rose to 79.9% with increasing reaction temperature from 80 to 100 C. A kinetic profile of the coupling reaction of CO 2 with PO in the presence of MAF-5/[n-Bu 4 N]Br (1/1) binary system is shown in Figure S8 (Supporting Information). It can be clearly seen that no induction period is observed in the kinetic curve for PC accumulation.
The addition of [n-Bu 4 N]Br as co-catalyst had a similar effect on the reactivity of MAF-6 samples (Tables 1 and 2). Activities of [n-Bu 4 N]Br/MAF-6 binary systems were higher in comparison with pristine MAF-6 samples. It is interesting that the catalytic performance of MAF-6 was higher in comparison with MAF-5. The lower yield of PC is not in agreement with the highest strength of basic sites of MAF-5 in comparison with MAF-6 ( Table 4) which are important in reaction mechanism. According to a reaction mechanism (Scheme S1, Supporting Information), the Lewis acidic center interacts with the oxygen atom of the epoxide, while the Lewis basic site activates a CO 2 molecule [65,66]. Investigation of cycloaddition of CO 2 to PO by DRIFT spectroscopy pointed that the Zn-N couples (LAS-BS) are formed by the defect sites where proceeds ring-opening step and CO 2 activation [66]. The addition of [n-Bu 4 N]Br to catalyst accelerates the ring-opening step due to the nucleophilic attack to C1 atom of epoxide and producing an intermediate (Scheme S1, Supporting Information).
The difference in activities of MAF-5 and MAF-6 can be caused by a few reasons. First, this phenomenon can be related to the difference in their structures (Table S1, Supporting Information). The MAF-5 with a zeolite ANA (Analcime) framework topology is formed by cells with the D3 (6 2 8 3 ) symmetry, which have aperture (4.0 × 5.8 Å) and an elongated ellipsoidal cavity (7.0 × 10.0 Å) with windows of Zn 4 (2-EIM) 4 , Zn 6 (2-EIM) 6 and Zn 8 (2-EIM) 8 rings [49]. At the same time, the MAF-6 with a zeolite RHO framework topology has a large cavity (~18.7 Å) and aperture (~7.6 Å). Structure of MAF-6 is formed by cells with windows containing four-, six-, and eight-membered rings containing 48 Zn atoms per unit cell. Each cell is connected to six adjacent cells by means of polyhedral units, consisting of double eight-membered rings (known as D8R) [67]. The higher disorder of the framework and lower aperture size of MAF-5 in comparison with MAF-6 can lead to the decreasing accessibility of active sites for the reagents and increasing diffusion limitations. The other reason can be the difference in their textural properties (Table 3). Specific surface area and total pore volume of MAF-5 are lower in comparison with MAF-6 that can affect the number of active sites and therefore, activity.
Effect of particle size of MAF-6 samples in term of the PC yield was investigated in the presence and without [n-Bu 4 N]Br at optimal experimental conditions (0.8 MPa of CO 2 , 80 • C and 5 h) (Tables 1 and 2). Correlation between the crystal size of the MAF-6 material and the yield of PC is shown in Figure 6A. One can see the yield of PC decreased with increasing crystallite size. This phenomenon is caused by several reasons. First, the change in textural properties could affect the activity of MAF-6. The simultaneous increase in microporosity (Vμ/VΣ) and particle size of the MAF-6 material ( Table  3) gives rise to the reduction of activity due to the rising diffusion limitations. This trend can point that mesopores are beneficial for the high catalytic activity of MAF-6 and ZIFs materials in general.
The varying amount and localization of active sites can be another reason of the dependence of activity MAF-6 on crystal size. Correlation between the yield of PC and the number of Lewis acidic sites (I300/I360) is shown in Figure 6B. One can see that the yield of PC rises with increasing amount active sites formed by low-coordinated Zn atoms (ZnIII and ZnII). The yield of PC also increases with increasing the number of basic sites (I2242) of the MAF-6 material ( Figure S9, Supporting Information). Both correlations indicate that both LAS and BS play an important role in the reaction mechanism. The high yield of PC could be obtained in the presence of MAF-6 sample containing a large number of active centers. These trends are in agreement with the structure and textural properties of MAF-6 samples. The low activity of MAF-6(L) with large particle size is due to the low quantity of active sites that is derived from the high structural order of particles. Such interdependence indicates that active sites are formed by defects and located in the regions close to the surface of the crystallites. This suggestion is confirmed by linear correlation between yield of PC and DSEM/DXRD ratio ( Figure S10, Supporting Information). The interdependence between defects and active sites, and impact of active sites on the catalytic performance were demonstrated for ZIF-8 [37,68].

Stability and Efficiency of MAF-6
The stability of MAF-6 in the course of the reaction was investigated by SEM. SEM images of MAF-6(L) and MAF-6(S) samples, after application in catalytic reaction, are shown in Figure 7. Mechanical degradation of the MAF-6 (L) crystals was observed due to a result of stirring. At the same time, similar mechanical degradation of the MAF-6 (S) crystals was found. The damage of the MAF-6(S) structure was confirmed by elemental analysis which pointed to the Zn 2+ leaching due to the polar reagents and intermediates. Amount Zn 2+ leached from MAF-6(S) was ~1.43 wt%. Note that this amount of Zn 2+ leached did not affect the catalytic activity of MAF-6(S). The size effect on the stability of ZIF-8 samples towards etching and Zn 2+ leaching was also observed in the Knöevenagel condensation reaction and the transesterification of hexanol and vinyl acetate [68]. First, the change in textural properties could affect the activity of MAF-6. The simultaneous increase in microporosity (V µ /V Σ ) and particle size of the MAF-6 material (Table 3) gives rise to the reduction of activity due to the rising diffusion limitations. This trend can point that mesopores are beneficial for the high catalytic activity of MAF-6 and ZIFs materials in general.
The varying amount and localization of active sites can be another reason of the dependence of activity MAF-6 on crystal size. Correlation between the yield of PC and the number of Lewis acidic sites (I 300 /I 360 ) is shown in Figure 6B. One can see that the yield of PC rises with increasing amount active sites formed by low-coordinated Zn atoms (Zn III and Zn II ). The yield of PC also increases with increasing the number of basic sites (I 2242 ) of the MAF-6 material ( Figure S9, Supporting Information). Both correlations indicate that both LAS and BS play an important role in the reaction mechanism. The high yield of PC could be obtained in the presence of MAF-6 sample containing a large number of active centers. These trends are in agreement with the structure and textural properties of MAF-6 samples. The low activity of MAF-6(L) with large particle size is due to the low quantity of active sites that is derived from the high structural order of particles. Such interdependence indicates that active sites are formed by defects and located in the regions close to the surface of the crystallites. This suggestion is confirmed by linear correlation between yield of PC and D SEM /D XRD ratio ( Figure S10, Supporting Information). The interdependence between defects and active sites, and impact of active sites on the catalytic performance were demonstrated for ZIF-8 [37,68].

Stability and Efficiency of MAF-6
The stability of MAF-6 in the course of the reaction was investigated by SEM. SEM images of MAF-6(L) and MAF-6(S) samples, after application in catalytic reaction, are shown in Figure 7. Mechanical degradation of the MAF-6 (L) crystals was observed due to a result of stirring. At the same time, similar mechanical degradation of the MAF-6 (S) crystals was found. The damage of the MAF-6(S) structure was confirmed by elemental analysis which pointed to the Zn 2+ leaching due to the polar reagents and intermediates. Amount Zn 2+ leached from MAF-6(S) was~1.43 wt%. Note that this amount of Zn 2+ leached did not affect the catalytic activity of MAF-6(S). The size effect on the stability of ZIF-8 samples towards etching and Zn 2+ leaching was also observed in the Knöevenagel condensation reaction and the transesterification of hexanol and vinyl acetate [68].
MAF-6 samples were synthesized from 2-EIM and [Zn(NH3)4](OH)2 according to the method described in [70] after a small modification. 2-EIM (4 mmol) was dissolved in mixed solvent systems (alcohol/cyclohexane). Zn(OH)2 (2 mmol), which was dissolved in aqueous NH3 (25%) solution (30 ml), was added once to the above described 2-EIM solution. MAF-6 samples were obtained after vigorous stirring of the reaction mixture for 20 min under room temperature. The main experimental conditions for synthesis of MAF-6 samples, properties of solvents, particle size, and textural properties of materials are shown in Table 3
MAF-5 was synthesized from 2-EIM and Zn(NO 3 ) 2 ·6H 2 O according to the method described in Ref. [69]. The main textural and physicochemical properties are show in Figures S1-S3 (Supporting Information).
MAF-6 samples were synthesized from 2-EIM and [Zn(NH 3 ) 4 ](OH) 2 according to the method described in [70] after a small modification. 2-EIM (4 mmol) was dissolved in mixed solvent systems (alcohol/cyclohexane). Zn(OH) 2 (2 mmol), which was dissolved in aqueous NH 3 (25%) solution (30 ml), was added once to the above described 2-EIM solution. MAF-6 samples were obtained after vigorous stirring of the reaction mixture for 20 min under room temperature. The main experimental conditions for synthesis of MAF-6 samples, properties of solvents, particle size, and textural properties of materials are shown in Table 3. The main textural and physicochemical properties are show in Figures S4-S8 (Supporting Information).

Instrumental Measurements
X-ray diffraction (XRD) patterns were obtained with the use of a Thermo ARL X'tra device, with the CuKα-radiation (λ = 1.5418 Å), focusing geometry θ-2θ in the scanning mode within the range of angles of 2θ from 3 to 75 • with the step of 0.05 • .
The textural characteristics of the MAF-5 and MAF-6 samples were determined from the adsorption isotherm of N 2 at −196 • C using a Micromeritics ASAP 2400. Chemical analyses were done by an inductively coupled plasma-atomic emission spectrometry (ICP-AES) using a PERKIN-ELMER instrument OPTIMA 4300.
Scanning electron microscopy (SEM) images were obtained with a Hitachi Regulus microscope operated at an accelerating voltage of 20 kV in secondary electron mode. The size distribution of the nanoparticles was calculated based on a representative set of SEM images taken at different areas of the sample.
The basicity of ZIFs was determined by DRIFT spectroscopy with CDCl 3 as probe molecule to follow the technique that was explained in [71].
DRIFT spectra were recorded on a Shimadzu FTIR-8300S spectrometer with a DRS-8000 diffusion reflectance cell in the range between 400 and 6000 cm -1 with a resolution of 4 cm −1 . All spectra are presented in F(R) Kubelka-Munk scale: F(R) = (1−R) 2 2R , where R is reflection coefficient. The DR-UV-Vis spectra were recorded on a UV-2501 PC Shimadzu spectrometer with an IRS-250A accessory in the 190-900 nm range with a resolution of 2 nm, on samples in powder form placed into a special cell for DR-UV-Vis measurement. BaSO 4 was used as standard for measurements.

Conclusions
In this work, zeolitic imidazolate frameworks MAF-5 and MAF-6 were demonstrated to be effective catalysts in the cycloaddition of CO 2 to propylene oxide to propylene carbonate (without any solvent) at 0.8 MPa of CO 2  Size effect on the reaction rate and yield of propylene carbonate was investigated in the presence of MAF-6. MAF-6 samples with particle size of 190 ± 20, 360 ± 30, and 810 ± 30 nm (MAF-6(S), MAF-6(M), and MAF-6(L), respectively) were prepared at room temperature from [Zn(NH 3 ) 4 ](OH) 2 and 2-ethylimidazole. Particle size was controlled by variation of type of alcohol in alcohol/cyclohexane media. Particle size reduced in the order: MeOH > EtOH > PrOH. According to comprehensive study, textural properties, and the type and content of active sites, depend on the particle sizes of MAF-6 samples. The catalytic activity of MAF-6 rose with decreasing particle size due to the increasing number and location of active sites in surface of solid. Maximal yield of PC was observed in the presence of MAF-6(S). Results of this investigation indicated that mesopores are beneficial to the high catalytic activity of MAFs and ZIFs materials in general.
Particle size also affects the stability of MAF-6 crystals. SEM data pointed that in contrast to MAF-6(L), MAF-6(S) sample is exposed to the indiscriminate etching of surface of crystal due to the leaching of Zn 2+ by polar reagents. The efficiencies of MAFs/[n-Bu 4 N]Br binary systems were compared with that of ZIFs/[n-Bu 4 N]Br reported in the literature. It was found that activity of MAFs were comparable and higher under similar reaction conditions. Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/catal11091061/s1, Figure S1: SEM image of MAF-5, Figure S2: Isotherms of N 2 adsorptiondesorption onto MAF-5, Figure S3: XRD pattern of MAF-5, Figure S4: Isotherms of N 2 adsorptiondesorption onto MAF-5 and MAF-6 samples, Figure S5: DRIFT spectra of MAF-6 with different particle size, Figure S6: DR-UV-vis spectra of MAF-6(L) and MAF-6(S) after calcination at 150 • C for 2 h in vacuum and the next addition of wet air, Figure S7 Figure S9: Dependence of PO conversion on the number of basic sites (I 2242 ), Figure S10: Dependence of PO conversion on the ratio of D SEM /D X-ray , Table S1: Structural properties of ZIF-8, MAF-5 and MAF-6, Table S2: MAF-6 particle size estimated from XRD and SEM data, Scheme S1: Mechanisms of reaction between propylene oxide and CO 2 .