Exploring the Scope of Macrocyclic “Shoe-last” Templates in the Mechanochemical Synthesis of RHO Topology Zeolitic Imidazolate Frameworks (ZIFs)

The macrocyclic cavitand MeMeCH2 is used as a template for the mechanochemical synthesis of 0.2MeMeCH2@RHO-Zn16(Cl2Im)32 (0.2MeMeCH2@ZIF-71) and RHO-ZnBIm2 (ZIF-11) zeolitic imidazolate frameworks (ZIFs). It is shown that MeMeCH2 significantly accelerates the mechanochemical synthesis, providing high porosity products (BET surface areas of 1140 m2/g and 869 m2/g, respectively). Templation of RHO-topology ZIF frameworks constructed of linkers larger than benzimidazole (HBIm) was unsuccessful. It is also shown that cavitands other than MeMeCH2—namely MeHCH2, MeiBuCH2, HPhCH2, MePhCH2, BrPhCH2, BrC5CH2—can serve as effective templates for the synthesis of x(cavitand)@RHO-ZnIm2 products. The limitations on cavitand size and shape are explored in terms of their effectiveness as templates.


S1
Analysis of PXRD patterns was conducted using Panalytical X'Pert Highscore Plus [9] software, and raw data was converted into a suitable format using the PowDLL [10] program. Experimental patterns S3 were compared to simulated patterns calculated from single crystal structures using Mercury [11] crystal structure viewing software and/or the Lazy Pulverix [12] package implemented in X-Seed.
[13] Crystallographic Information Files containing published crystal structures were obtained from the Cambridge Structural Database (CSD) or the Crystallography Open Database (COD).
Gas adsorption analyses were conducted on a Quantachrome Instruments Autosorb-1 sorption analyzer. All samples were first activated by heating at 150 °C for a minimum of 18 hours, then analyzed in a 6 mm bulb cell, at 77K, with N2 as the analysis gas. The samples were weighed 5 times, liberally applying a static gun to the cells before each weigh, in order to reduce interferences from static electricity. The average masses of the empty cell, full cell before and after outgas were used. After milling, the resulting paste was transferred into a capped glass vial and left to age at room temperature. After the reaction was complete (based on PXRD), the reaction mixture was washed extensively with chloroform (according to a procedure previously shown to remove all excess template) and dried at room temperature under ambient conditions. 1.53 mL) were milled together in a PMMA jar for 2 min at 30 Hz. The resulting dense paste was transferred to a 5 mL glass vial and left to age. Reaction progress was monitored by PXRD ( Figure   S5). When reaction has reached completion, as evidenced by the disappearance of ZnO X-ray reflections from the powder pattern (approx. 4 days), the sample was washed extensively with chloroform to remove excess MeMeCH2, yielding essentially phase-pure 0.2MeMeCH2@RHO-Zn16(Cl2Im)32, with nearly quantitative conversion based on PXRD ( Figure S2). The excess template was recovered by evaporating the chloroform wash under low pressure and drying. It was reused in further experiments.
The BET plot is linear and has a positive intercept, giving the surface area of 1140.5 m 2 /g. The material maintains crystallinity and the RHO topology after activation, as evidenced by the postsorption PXRD pattern ( Figure S4). As a control reaction, the procedure from section 2.1. was repeated without addition of the cavitand template. Nanoparticulate zinc oxide (50 mg, 0.6 mmol), 4,5-dichloroimidazole (170 mg, 1.2 mmol), and 60 L of DEF were milled together in a PMMA jar for 2 min at 30 Hz. The resulting dense paste was transferred to a 5 mL glass vial and left to age. Reaction progress was monitored by PXRD ( Figure S5).
Similarly to the templated reaction, the non-templated control reaction also provides an RHO topology zinc 4,5-dichloroimidazolate, namely RHO-Zn(Cl2Im)2 (ZIF-71). However, monitoring of the templated and non-templated reaction mixtures by PXRD over time shows that the templated reaction is significantly faster. In fact, the templated reaction after 1 day shows approximately the same conversion (based on relative amount of ZnO) as the non-templated reaction does after 35 days, and while the non-templated reaction shows a large amount of residual ZnO after 3 days, the templated reaction undergoes full conversion in that time, both on a small and large scale ( Figure S5).

S7
This shows that not only does templation using MeMeCH2 help direct the topological outcome of a mechanochemical synthesis, it also allows the reaction to proceed faster.  The 1 H NMR spectrum shows that there is no enclathrated MeMeCH2 in the final product ( Figure   S7).

Control reaction and time monitoring of the LAG Procedure synthesis of RHO-ZnBIm2
As a control reaction, the procedure from section 2.3. was repeated without addition of the cavitand template. Nanoparticulate zinc oxide (50 mg, 0.6 mmol), benzimidazole (145 mg, 1.2 mmol), and 60 L of DEF were milled together in a PMMA jar for 2 min at 30 Hz. The resulting dense paste was transferred to a 5 mL glass vial and left to age. Reaction progress was monitored by PXRD ( Figure   S10).
The non-templated control reaction provides very small amounts of RHO-ZnBIm2 after a much longer period of time, as evidenced by PXRD ( Figure S10). Namely, the non-templated reaction shows no sign of an RHO topology product after 3 days, and only a very small amount after 12 days, with a very large amount of residual ZnO. Meanwhile, the templated reaction shows more conversion after 3 days than the non-templated reaction after 12 days, and is nearly quantitative after 12 days ( Figure S10). This again demonstrates that MeMeCH2 significantly accelerates the synthesis of the RHO topology product. S11 Figure S10. Comparison of PXRD patterns for the time monitoring and control reactions in the synthesis of RHO-ZnBIm2: a) -b) the reaction mixture from non-templated LAG Procedure synthesis using HBIm after 3, and 12 days, respectively, c) -d) reaction mixture from templated LAG Procedure synthesis using HBIm after 3, and 12 days, respectively, e) calculated pattern of DEF@MeMeCH2, and f) zinc oxide. Tick marks denote predicted peak positions of the RHO topology ZnBIm2 (CSD code VEJZOA, ZIF-11). Hz. The resulting dense pastes were transferred to a 5 mL glass vial and left to age. Reaction progress was monitored by PXRD ( Figure S12). None of the reactions appeared to show any trace of an RHO topology product, based on PXRD. The reaction using theophylline provided the known zinc theophylline dihydrate after a year, while the reactions using 5-nitrobenzimidazole and 5,6dimethylbenzimidazole showed almost no conversion after 37 days, as evidenced by PXRD. Figure S12. Comparison of PXRD patterns for the templated solid-state reactions using theophylline, 5,6dimethylbenzimidazole, and 5-nitrobenzimidazole: a) predicted peak positions of zinc theophylline dihydrate (CSD code CIBLIK), b)c) the reaction mixture from the MeMeCH2 templated LAG Procedure synthesis using theophylline after 1 year (washed), and 22 days, respectively, d) reaction mixture from the MeMeCH2 templated LAG Procedure synthesis using 5,6-dimethylbenzimidazole after 37 days, e) reaction mixture from the MeMeCH2 templated LAG Procedure synthesis using 5-nitrobenzimidazole after 37 days, f) calculated pattern of DEF@MeMeCH2, and g) zinc oxide. Tick marks denote predicted peak positions of the RHO topology ZnBIm2 (CSD code VEJZOA, ZIF-11). S13

SYNTHESIS AND CHARACTERIZATION OF RHO TOPOLOGY ZINC IMIDAZOLATES USING OTHER CAVITAND TEMPLATES
To further explore the scope of the templation effect, we employed a variety of cavitands other than MeMeCH2 in the mechanochemical synthesis of zinc imidazolates. We varied the top rim substituent, the bottom rim substituent, the top rim linker, as well as the number of repeating units in the cavitand.
In a typical LAG-aging experiment, nanoparticulate zinc oxide, imidazole, the template, and DEF were mixed in a 1 : 2 : 0.5 : 4 molar ratio, and milled in a PMMA jar for 2 min at 30 Hz. The resulting dense paste was transferred to a 5 mL glass vial and left to age. Reaction progress was monitored by PXRD. When reaction has reached completion, as evidenced by the disappearance of ZnO peaks from the powder pattern, the sample was washed extensively with chloroform to remove excess cavitand.
3.1. LAG-aging reaction using HPhCH2 as template LAG-aging experiment was performed using nanoparticulate zinc oxide (48 mg, 0.6 mmol), imidazole (82 mg, 1.2 mmol), HPhCH2 (246 mg, 0.3 mmol), and DEF (200 L). The mixture was washed after 12 days, giving 0.8HPhCH2@RHO-Zn16Im32, as confirmed by NMR and PXRD ( Figure   S14, S13). Interestingly, the PXRD pattern of 0.8HPhCH2@RHO-Zn16Im32 shows very different peak intensities compared to the other RHO materials ( Figure S14). Namely, the (100) peak (at ~3° 2 ) that is normally very small in PXRD patterns of RHO topology zeolitic imidazolate frameworks is very intense in this pattern, while the (110) peakwhich is normally the dominant oneis much smaller. In addition, there is significant broadening of several peaks, while the (100) peak is still sharp. Both effects could be due to the large quantity HPhCH2 trapped inside the double-8-rings of the material, causing the phenyl feet to protrude into the LTA cage. The electron density in the cage will affect peak intensities, while steric hindrance could cause twisting and breathing of the framework itself, resulting in peak broadening.  washed after 11 days, resulting in a mixture of products (major product is cag-ZnIm2) that does not contain any RHO topology product, as confirmed by PXRD ( Figure S15).       and nog-ZnIm2, as confirmed by PXRD ( Figure S24).
NMR analysis of the starting MePhCH2 cavitand shows that it is a mixture of the rctt and rccc isomers, in the ratio of 87:13. Similarly, the NMR analysis of the washed product ( Figure S24) shows that the encapsulated cavitand exactly mimics that stereochemical ratio, with the total cavitand : imidazole ratio being 1:17.3. Unfortunately, calculating the incorporation of cavitand into the RHO topology zinc imidazolate is impossible, as the relative quantities of the nog and RHO topology products cannot be determined. Figure S23. PXRD patterns of the product of the LAG-aging procedure using MePhCH2, washed after 5 days (top), and zinc oxide (bottom), and the predicted PXRD pattern of nog-ZnIm2 (middle, CSD code: HIFWAV). Tick marks denote predicted peak positions of a cubic unit cell, Pm-3m, a = 28.907(2) Å, corresponding to xMeMeCH2@RHO-Zn16Im32.