Core–Shell Crystals of Porous Organic Cages

Abstract The first examples of core–shell porous molecular crystals are described. The physical properties of the core–shell crystals, such as surface hydrophobicity, CO2 /CH4 selectivity, are controlled by the chemical composition of the shell. This shows that porous core–shell molecular crystals can exhibit synergistic properties that out‐perform materials built from the individual, constituent molecules.

structure prediction. [15b] Because the methyl groups in CC15 partially block the cage windows,t he (CC3-S, CC15-R) cocrystal becomes selectively porous to H 2 but not N 2 at 77 K, 1bar. [15b] Another cage molecule with an analogous tetrahedral architecture,first reported by Petryk et al., [19] can be prepared by 2-hydroxy-1, 3, 5-benzenetricarbaldehyde with R, R-CHDA. We will refer to this covalent cage here as CC19 (Figure 1a,r ight). Thed isordered hydroxyl groups occupy the four cage windows. CC19-R crystallizes to form aw indow-to-window packing with 3D diamondoid pores, isostructural with CC3a (Figure 1c right). CC19-R shows permanent porosity to ar ange of gases and exhibits at ype I N 2 sorption isotherm with aSA BET of 514 m 2 g À1 ( Figure S2 in the Supporting Information).
Three different heterochiral cage compositions were used in this study:r acemic CC3-RS,r acemic CC19-RS,a nd quasiracemic CC3-R, CC15-S. In each case,c age particles were fabricated by simple mixing of the corresponding R and S solutions,t aking advantage of the lower solubility product of the racemic or quasiracemic materials. [16] All heterochiral cage particles were crystalline and each had the same basic packing mode,a sd emonstrated by powder X-ray diffraction (PXRD) ( Figure S3,4). Thesimilar lattice parameters for the three different compositions suggested the potential for epitaxial growth to create core-shell structures.A ll cage particles showed uniform, octahedral crystal morphologies (e.g.,F igure 2b). Thep article size could be controlled systematically in the range 250 nm to 2 mmb yv arying the mixing temperature ( Figure S5). To probe the potential for core-shell structure generation, we first investigated the sequential addition of CC3-R and CC3-S solutions to see whether this would make larger particles by seeded, epitaxial growth, or whether new particles would be nucleated. The particle sizes measured by dynamic light scattering (DLS) and by scanning electron microscopy (SEM) for each addition confirmed that progressively larger particles were formed ( Figure S6, Table S1), suggesting epitaxial growth and the possibility of core-shell structure generation by sequential addition of solutions of distinct cages.
Next, we prepared core-shell structures using CC3 and CC19 cage molecules.T he schematic structure is shown in Figure 2a;t he core molecules are colored purple.T wo coreshell crystal systems were prepared: CC3-RS core /CC19-RS shell and its inverse structure, CC19-RS core /CC3-RS shell ,b oth using the sequential addition method described above using DCM solutions at 30 8 8C. Theaverage DLS particle diameters for the core-shell cocrystals, CC3-RS core /CC19-RS shell and CC19-RS core /CC3-RS shell ,w ere 744 nm and 721 nm, respectively,a s compared to 212 nm and 474 nm for the CC3-RS and CC19-RS core seed particles ( Figure S7, Table S2). This would suggest a CC19-RS shell thickness of 266 nm in CC3-RS core / CC19-RS shell and a CC3-RS shell thickness of 124 nm in CC19-RS core /CC3-RS shell .T he particle size was further verified by SEM, as shown in Figure S8. There was ag ood agreement between the DLS and SEM measurements.L arger crystals were required to confirm the core-shell morphology by microscopy.W et herefore mixed the solutions in CHCl 3 at ahigher temperature (50 8 8C), whereupon the average particle size of the core-shell crystals was increased to 3-4 mm, as shown in Figure 2b-d: CC3-RS ( % 2 mm) and CC19-RS (1-2 mm) prepared under the same conditions ( Figure S9,10). A terraced surface structure was observed by SEM ( Figure 2b, Figure S11) indicating the epitaxial growth of the shell. The core-shell samples showed uniform octahedral shape morphologies without any apparent particle aggregation during the formation of the shell.
Since no contrast could be seen between the chemicallysimilar core and shell by TEM (Figure 2c), the morphologies  of the CC3-RS core /CC19-RS shell and CC19-RS core /CC3-RS shell cocrystals were explored by confocal laser scanning microscopy (CLSM). This was possible because CC19-RS,u nlike CC3-RS,isstrongly fluorescent. To visualize the layered coreshell structure,w eu sed % 5m icrometer-sized core-shell cocrystals prepared in CHCl 3 at 60 8 8C. Theh orizontally sliced confocal image of CC3-RS core /CC19-RS shell revealed an on-fluorescent inner core (CC3-RS)e ncapsulated by af luorescent outer shell layer (CC19-RS), as shown in Figure 3c and the corresponding 3D structural model (Movie S1). By contrast, the CC19-RS core /CC3-RS shell crystals comprise an on-fluorescent CC3-RS shell encapsulating af luorescent core (CC19-RS) ( Figure 3d). Thei ntensity profiles are presented in Figure 3e,f,which correspond to the core-shell crystals shown in the horizontally sliced images (Figure 3c,d). Thedistance across the crystal is approximately 6 mmf or CC3-RS core /CC19-RS shell ,a nd this representative crystal has an on-fluorescent core of approximately 3 mmi n diameter and a1 .5 mm-thick shell, as estimated from the fluorescence intensity profiles.T he diameter of the CC19-RS core /CC3-RS shell crystal was 4 mmw ith a3mmf luorescent core and a5 00 nm non-fluorescent shell. Z-stack of CLSM images of CC3-RS core /CC19-RS shell and CC19-RS core /CC3-RS shell are shown in Figure S12,13. A3 Ds tructural model for CC3-RS core /CC19-RS shell was constructed based on the zstack of CLSM analysis (SI, Movie S2).
Thestructural relationship between the core seed crystals, separate crystals of the shell components,a nd the core-shell cocrystals was further explored by synchrotron X-ray diffraction. Both CC3-RS and CC19-RS crystallized in the cubic space group F4 1 32 with unit cell parameters of a = 24.7069-(1) for CC3-RS and a = 24.6914(3) for CC19-RS. Lattice parameter matching is important in allowing the growth of the core-shell morphology.T he PXRD patterns for CC3-RS, CC19-RS,a nd CC3-RS core /CC19-RS shell ( Figure S14) indicate that the core-shell particles retain as imilar crystal packing: the core-shell cage crystals also crystallize with cubic symmetry and window-to-window packing motifs,a nalogous to CC3-RS and CC19-RS,w ith as mall expansion in the unit cell parameters compared to the individual racemic crystals (Table S3).
CC3-RS core /CC19-RS shell demonstrates as ignificantly higher oxygen content as measured by X-ray photoelectron spectroscopy (XPS) due to an outer layer containing hydroxyl groups (oxygen elements), while CC19-RS core /CC3-RS shell does not (Table S4). Also,t he solution UV absorption spectrum for CC19-RS shows absorption peaks at 300 and 375 nm. By contrast, a CC3-RS solution exhibits no UV adsorption in this region. Theabsorption peaks for the coreshell, CC3-RS core /CC19-RS shell ,asm easured by dispersing the cage particles in the hexane suspension, showed aslight blue shift relative to the CC19-RS solution spectrum, while ar ed shift was observed for the CC19-RS core /CC3-RS shell material ( Figure S16). Thei ntensity of the fluorescence excitation/ emission spectra for CC19-RS core /CC3-RS shell was significantly decreased as compared to CC19-RS,i nk eeping with af luorescent core of CC19-RS that is encapsulated by an onfluorescent CC3-RS layer ( Figure S17).
This synthetic method can also be applied to other cage molecules:for example,acore-shell crystal with racemic CC3 as the core and quasi-racemic CC3-R/CC15-S as the shell was also prepared. The CC3-RS core crystals had an average particle size of 1-2 mm, as measured by SEM. Subsequent addition of solutions of CC3-R and CC15-S formed as hell, creating a CC3-RS core /CC15S-CC3R shell cocrystals with an average diameter of 3 mm( Figure S18, S19).
Core-shell structures can be exploited to control particle surface properties,which are important in applications such as gas storage and separation. [20] Contact angles with water for cage crystals (1-3 mmd iameter) gradually increased from 55.68 AE 2.58 8 (CC19-RS)t o7 8.71 AE 0.808 8 (CC3-RS)t o8 3.06 AE 3.048 8 (CC3-R/CC15-S)a st he constituent cage materials become more hydrophobic ( Figure S20). CC3-RS core /CC19-RS shell shows acontact angle of 59.71 AE 6.58 8:that is,very close to the pure,relatively hydrophilic CC19 material (Figure 4a), showing that the shell dominates the surface properties. Likewise,t he inverse CC19-RS core /CC3-RS shell cocrystal showed ac ontact angle of 79.01 AE 3.18 8,c lose to pure CC3-RS. Thec ontact angle of CC3-RS core /CC15S-CC3R shell is 83.40 AE 0.878 8;t his material is slightly more hydrophobic due to the methyl groups in CC15.
( Figure S21). We found that CO 2 /CH 4 selectivity was defined by the crystal shell. CC19-RS core /CC3-RS shell was porous to both CO 2 and CH 4 at 273 K, 1bar and had rather poor selectivity for these two gases (Figure 4d). By contrast, CC3-RS core /CC19-RS shell was selectively porous to CO 2 under the same conditions (Figure 4e). Thei deal adsorbed solution theory (IAST) selectivity of CC3-RS core /CC19-RS shell was 33, as calculated using experimental single-component isotherms at 273 Kw ith CO 2 /CH 4 mixtures (50/50 molar ratio;s ee Figure S22b). This core-shell material combines ah igh capacity for CO 2 (2.5 mmol g À1 )with good CO 2 /CH 4 selectivity.T he high CO 2 sorption capacity is attributed to the CC3-RS core while the selectivity results from the CC19-RS shell, which inhibits CH 4 diffusion into the core.T he CC3-RS core / CC19-RS shell material therefore has synergistic properties that are not exhibited by the individual cage components,n or by the inverse CC19-RS core /CC3-RS shell morphology,i llustrating the power of this approach. Asummary of gas sorption data is given in Table S4.
Forp ractical application, it is preferable for core-shell crystals to be defect and crack free,s ince cracks in the shell layer could allow direct access to the core,r educing selectivity.N either SEM nor TEM images revealed any obvious cracks on the cage particle surfaces ( Figure S23). Moreover, core-shell crystals were immersed into as olution of af luorescent organic dye (Rose Bengal) that would be size excluded from the cage pores but not from larger cracks or defects.F or most crystals (approx. 90 %), horizontally sliced confocal images showed that most of the dye was coated onto the surface of the core-shell cage crystal ( Figure S24), indicating that there were no significant cracks or defects in the shell layer. However, around 10 %ofthe crystals that we measured appeared to show some sort of mechanical damage, which might affect the adsorption properties ( Figure S25).
In conclusion, we have successfully prepared core-shell cage crystals.T he surface chemistry is controlled by the functionality in the shell layer, thus allowing control over surface hydrophobicity.H ence, CC3,w hich was shown previously to have multiple practical applications, [21] can be rendered either more hydrophobic or more hydrophilic, depending on the choice of shell. A CC3-RS core /CC19-RS shell material was shown to have asynergistic combination of CO 2 sorption capacity and CO 2 /CH 4 selectivity that surpassed either of individual constituent cages.T his approach has the potential to open up new applications for porous organic cages.T ogive one example, CC3 crystals have been incorporated into polymers of intrinsic microporosity to form organic mixed matrix membranes (MMMs) for molecular sieving. [22] In MMMs,agood interaction between the polymer and filler components is essential, and this core-shell approach offers an ew strategy for optimizing the polymer-cage particle interface.Itisalso possible that cage shells could be deposited from solution onto porous crystals of other materials such as MOFs,C OFs and zeolites,p roviding that conditions can be identified to promote epitaxial growth.