An S10-Symmetric 5-Fold Interlocked [2]Catenane

The reaction of sym-pentakis(4-aminothiophenyl)corannulene with 2-formyl-6-methylpyridine and CuI or 2-formyl-1,10-phenanthroline and MII (M = Co, Zn) yields an S10-symmetric 5-fold interlocked [2]catenane of two interpenetrating [CuI5L2]5+ cages or D5-symmetric [MII5L2]10+ cages, respectively. The new structures were characterized by X-ray crystallography, NMR spectroscopy, and mass spectrometry. Density functional theory computations point to dispersive energies on par with traditional covalent bond energies. Subcomponent exchange reactions favored formation of the [CoII5L2]10+ cage over the [CuI10L4]10+ catenane. The single cage and catenane each cocrystallized with a corannulene guest to form a bowl-in-bowl substructure.


. General
Unless otherwise specified, all starting materials were purchased from commercial sources and used as supplied. Cobalt(II) bis(trifluoromethanesulfonyl)imide 1 and sym-pentachlorocorannulene 2 were prepared following literature procedures. NMR spectra were recorded using 400 MHz Avance III HD Smart Probe (routine 1 H NMR, DOSY) and DCH 500 MHz dual cryoprobe (high-resolution 13 C and 2D experiments) NMR spectrometers. Chemical shifts (δ) for 1 H NMR spectra are reported in parts per million (ppm) and are reported relative to the solvent residual peak. DOSY experiments were performed on a Bruker DRX-400 spectrometer. Maximum gradient strength was 6.57 G/cmA. The standard Bruker pulse program, ledbpgp2s, employing a stimulated echo and longitudinal eddy-current delay (LED) using bipolar gradient pulses for diffusion using 2 spoil gradients was utilized. Rectangular gradients were used with a total duration of 1.5 ms. Gradient recovery delays were 1200 μs. Diffusion times were 50 ms. Individual rows of the S4 quasi-2D diffusion databases were phased and baseline corrected. Low-resolution electrospray ionization mass spectra (ESI-MS) were obtained on a Micromass Quattro LC infused from a Harvard Syringe Pump at a rate of 10 µL per minute. Highresolution ESI mass spectra were obtained by the EPSRC UK National Mass Spectrometry Facility at Swansea University using a Thermofisher LTQ Orbitrap XL.
To a flame-dried and degassed round bottom flask, 4-aminothiophenol (225 mg, 1.8 mmol), sodium hydride (58 mg, 1.45 mmol) and 1,3-dimethyl-2-imidazolidinone (DMI) (10 mL) were added. The S3 mixture was stirred at room temperature for 15 min. The yellow suspension changed to a clear yellow solution. sym-Pentachlorocorannulene (100 mg, 0.24 mmol) was added and the mixture was warmed up to 60 °C for 24 h. After cooling down to room temperature, the mixture was extracted three times with EtOAc and water, dried over MgSO4 and the solvent was evaporated under reduced pressure. The crude orange-red oil was redissolved in DCM (ca. 10 mL) and hexane was added (ca. 300 mL) to slowly precipitate out the product. After filtration the desired compound could be obtained as an orange solid (95 mg, 0.11 mmol) in 46 % yield. 1

Synthesis and characterization of [Cu10L4](BF4)10 (1)
To a Teflon-capped J-Young NMR tube was added sym-pentakis(4-aminothiophenyl)corannulene (3.0 mg, 3.5 μmol, 4 equiv.), 2-formyl-6-methylpyridine (2.5 mg, 21 μmol, 24 equiv.) and tetrakis(acetonitrile)copper(I) tetrafluoroborate (3.3 mg, 10.5 μmol, 12 equiv.) as stock solutions of known concentration in CD3CN (total volume 0.6 mL). The solution was degassed by three evacuation/nitrogen fill cycles and then sonicated for 5 min. The NMR tube was then rotated at room temperature for 24 hr and the formation of 1·10BF4 was confirmed by 1 Figure S11: High-resolution ESI-mass spectrometry analysis of 1·10BF4 showing the (a) +3 peak (b) +4 peak (c) +5 peak and (d) +6 peak. The fragmentation patterns observed for the +4 and the +6 peaks are attributed to homolytic fragmentation of the [Cu10L4] 10+ assembly under the MS conditions. The ionisation conditions of the instrument used to record the high-resolution ESI-MS are harsher than those used to record the low-resolution ESI-MS in Figure S10, as evidenced by much higher levels of fragmentation overall and the observation of strong peaks for low charged +1 and +2 fragments. No other spectroscopic evidence suggested the formation of [Cu5L2] 5+ single cages in solution.

Structural transformation of 1 into 2
A sample of 1·10BF4 was initially prepared according to the method described in Section 1.3, starting from 2.5 mg (2.9 µmol) of subcomponent A. Assembly 1·10BF4 was combined with 2-formyl-1,10phenanthroline (3.0 mg, 14 µmol, 20 equiv. per assembly, assuming quantitative formation of 1·10BF4 in the first step) and Co(BF4)2·6H2O (2.5 mg, 7.3 µmol, 10.2 equiv. per assembly) in CD3CN (0.65 mL) and the mixture degassed by bubbling with nitrogen for 15 min. The mixture was stirred at 333 K for 24 h and then at 353 K for 72 h. The 1 H NMR ( Figure S31) and ESI mass ( Figure S32) spectra of the resulting mixture were consistent with the formation of 2·10BF4 as the major product in solution.

General procedure
The host-guest chemistry of the diamagnetic assemblies 1 and 3 was investigated with the prospective guests shown in Figure S33. Host-guest complexes were prepared on an NMR scale and characterized by 1 H NMR spectroscopy. A solution of 1 or 3 in CD3CN (1.0-3.0 mM) was transferred to an NMR tube and the prospective guest molecule (typically 5-10 equiv) was added as a solid. The mixtures were sonicated for 10 minutes and allowed to equilibrate for at least 24 hours at 298 K prior to measurement of the 1 H NMR spectrum. In the case of 3, the mixtures were subsequently heated at 353 K for 24 hr and the 1 H NMR spectra measured again.
In both cases a fast-exchange binding interaction was inferred to have occurred with corannulene as indicated by the observation of clear shifts in both the host and guest signals. Characterization of the resulting host-guest complexes is given below.
In all other cases no binding was inferred to have taken place, as the signals for the host appeared at the same chemical shifts as in the absence of the guest and the signals for the guest were identical to those in the absence of host for guests that were soluble in CD3CN. In general small shifts of <0.05 ppm were attributed to weak non-specific π-stacking interactions while guest binding was inferred by more significant shifts of >0.05 ppm. Attempts were also made to synthesise 1 and 3 in the presence of C60 and coronene resulting in 1 H NMR spectra identical to those obtained in the absence of guest.

X-ray crystallography
Data were collected at Beamline I19 of Diamond Light Source employing silicon double crystal monochromated synchrotron radiation (0.6889 Å) with ω and ψ scans at 100(2) K. 3 Data integration and reduction were undertaken with Xia2. 4 Subsequent computations were carried out using the WinGX-32 graphical user interface. 5 Multi-scan empirical absorption corrections were applied to the data using the AIMLESS 6 tool in the CCP4 suite. 7 The structures were solved by direct methods using SHELXT 8 then refined and extended with SHELXL. 9 In general, non-hydrogen atoms with occupancies greater than 0.5 were refined anisotropically. Carbon-bound hydrogen atoms were included in idealised positions and refined using a riding model. Disorder was modelled using standard crystallographic methods including constraints, restraints and rigid bodies where necessary. Crystallographic data along with specific details pertaining to the refinement follow.
In all cases no restraints were applied to the corannulene moieties within the structures. Distances

Specific refinement details:
Crystals of 1·10BF4·16.25MeCN·2.5C4H10O were grown by diffusion of diethyl ether into an acetonitrile solution of 1·10BF4. The crystals employed immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data. Despite these measures and the use of synchrotron radiation few reflections at greater than 0.9 Å resolution were observed. The asymmetric unit was found to contain two crystallographically unique Consequently, the molecular weight and density given above are likely to be slightly underestimated.
CheckCIF gives 12 B level alerts. These alerts all result from the limited data resolution and thermal motion and/or unresolved disorder of some anions and solvent molecules as described above.   immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data. Even so data were obtained to 0.84 Å resolution by employing synchrotron radiation. The asymmetric unit was found to contain one Cu5L2 assembly, one 3.69(1) Å 6.93(1) Å

S36
corannulene molecule and associated counterions and solvent molecules. The Cu5L2 assembly forms a Cu10L4 [2]catenane with its enantiomer. Thermal parameter restraints (SIMU, RIGU) were applied to all atoms except for copper in order to facilitate anisotropic refinement.
The structure shows evidence of substantial disorder. The main Cu5L2 residue was found to exhibit whole molecule disorder with a refined occupancy of 0.8085(10) for the major occupancy part. Bond lengths and angles within the two chemically identical ligands (excluding the corannulene portion of the organic ligands to which no restraints were applied) of the major occupancy assembly were restrained to be similar to each other. In order to obtain a reasonable model for the minor occupancy part the GRADE program 12 was employed using the GRADE Web Server 13 to generate a full set of bond distance and angle restraints (DFIX, DANG, FLAT) for the organic ligands. The minor occupancy part was modelled with isotropic thermal parameters. The stacked corannulene molecule shows evidence of thermal motion or dynamic disorder which could not be adequately modelled with discrete atom positions. Consequently the thermal parameters of this residue are larger than the main Cu5L2 residue. The occupancy of this corannulene was initially refined but a value close to unity was obtained hence it was fixed at full occupancy.
The diethyl ether solvent molecules also show evidence of disorder. Two of these solvent molecules were modelled as disordered over two locations. The disordered atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement. The anions within the structure also show evidence of substantial disorder. Four of the BF4 − anions were modelled as disordered over two locations. Some disordered atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement.
Further reflecting the solvent loss and poor diffraction properties there is a small amount of void volume in the lattice containing smeared electron density from 0.5 unresolved anions per Cu5L2 assembly (assigned as BF4 − in the formula) and further highly disordered solvent. Consequently the SQUEEZE 10 function of PLATON 11 was employed to remove the contribution of the electron density associated with the highly disordered anion and solvent molecules which gave a potential solvent accessible void of 7354 Å 3 per unit cell (a total of approximately 2152 electrons). Since the diffuse solvent molecules could not be assigned conclusively to acetonitrile or diethyl ether they were not included in the formula.
Consequently, the molecular weight and density given above are likely to be slightly underestimated.
CheckCIF gives 2 B level alerts. These alerts both result from thermal motion and/or unresolved disorder of some anions and solvent molecules as described above.     Specific refinement details: The crystals of 2·coronene·10BF4·20.5MeCN·iPr2O [+ solvent] were grown by diffusion of diisopropyl ether into an acetonitrile solution of 2·10BF4 containing excess coronene. The crystals employed immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data. Even so data were obtained to 0.84 Å resolution by employing synchrotron radiation. The asymmetric unit was found to contain one complete Co5L2 assembly, half of two separate coronene molecules (located over special positions) and associated counterions and solvent molecules.
Bond lengths and angles within pairs of chemically identical organic ligand arms were restrained to be similar to each other as were the two coronene molecules. No restraints were applied to the corannulene portion of the organic ligands. Thermal parameter restraints (SIMU, RIGU) were applied to all atoms except for cobalt to facilitate anisotropic stable refinement. One of the coronene molecules shows evidence of thermal motion or minor unresolved disorder resulting in a large average Ueq value for that residue.
The anions within the structure show evidence of substantial disorder. The 10 BF4 − anions were modelled as disordered over 12 lattice sites; eight of these lattice sites were further modelled as disordered over two locations. Some disordered atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement.
Many acetonitrile solvent molecules were also modelled as disordered over multiple locations and/or S40 with partial occupancy. The hydrogen atoms of some of the disordered solvent molecules could not be located in the electron density map and were not included in the model.
Further reflecting the solvent loss there is a small amount of void volume in the lattice containing smeared electron density from disordered solvent. Consequently the SQUEEZE 10 function of PLATON 11 was employed to remove the contribution of the electron density associated with this highly disordered solvent which gave a potential solvent accessible void of 874 Å 3 per unit cell (a total of approximately 210 electrons). Since the diffuse solvent molecules could not be assigned conclusively to acetonitrile or diisopropyl ether they were not included in the formula. Consequently, the molecular weight and density given above are likely to be slightly underestimated.
CheckCIF gives 3 A and 47 B level alerts. These alerts (both A and B level) all result from solvent molecules for which the hydrogen atoms were not modelled (singly bonded carbons) and thermal motion and/or unresolved disorder of some anions, solvent molecules and one of the coronenes as described above. Figure S43: Side-on views of 2 from the crystal structure of 2·coronene·10BF4·20.5MeCN·iPr2O in stick and space-filling representations. Counterions, solvents and disorder are omitted for clarity. The structure is similar to that of 2 in the structures of 2·10ClO4·3MeCN ( Figure S49) and 2·10BF4·11MeCN ( Figure S48). Figure S44: Top views of 2 from the crystal structure of 2·coronene·10BF4·20.5MeCN·iPr2O in stick and space-filling representations. Counterions, solvents and disorder are omitted for clarity. The structure is similar to that of 2 in the structures of 2·10ClO4·3MeCN ( Figure S49) and 2·10BF4·11MeCN ( Figure S48). Figure S45: The structure of 2 from the crystal structure of 2·coronene·10BF4·20.5MeCN·iPr2O highlighting the arrangement of corannulenes. Counterions, solvents and disorder are omitted for clarity. The distance between the corannulenes is marked, highlighting the absence of a cavity in the solid state. Similar distances were obtained for the structures 2·10ClO4·3MeCN (3.23(3) Å) and 2·10BF4·11MeCN (3.28(3) Å). Figure S46: Crystal packing within the crystal structure of 2·coronene·10BF4·20. 5MeCN·iPr2O showing the arrangement of the co-crystallized coronene (shown in red) which intercalates between molecules of 2 forming a 1D tape. Counterions, solvents and disorder are omitted for clarity. Figure S47: Crystal packing within the crystal structure of 2·coronene·10BF4·20. 5MeCN·iPr2O showing the packing of the co-crystallized coronene (shown in red) and 2 into 2D sheets. The counterions and solvents which occupy the remaining voids in the structure are omitted for clarity.

Specific refinement details:
The crystals of 2·10BF4·11MeCN [+ solvent] were grown by diffusion of diethyl ether into an acetonitrile solution of the complex. The crystals employed immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data. Despite these measures and the use of synchrotron radiation few reflections at greater than 1.1 Å resolution were observed and the data were trimmed accordingly. Nevertheless, the quality of the data is far more than sufficient to establish the connectivity of the structure. The asymmetric unit was found to contain one half of a Co5L2 assembly and associated counterions and solvent molecules.
Due to the limited resolution, bond lengths and angles within pairs of chemically identical organic ligand arms were restrained to be similar to each other. No restraints were applied to the corannulene portion of the organic ligands. Two of the five ligand arms show evidence of substantial thermal motion or dynamic disorder which could not be adequately modelled with discrete atom positions. In order to obtain a reasonable model for these ligand arms the GRADE program 12 was employed using the GRADE Web Server 13 to generate a full set of bond distance and angle restraints (DFIX, DANG, FLAT) for these residues. Thermal parameter restraints (SIMU, RIGU) were applied to all atoms except for cobalt to facilitate anisotropic stable refinement. Even with these restraints some thermal parameters remain larger than ideal as a consequence of the high level of thermal motion or minor unresolved disorder present throughout the structure resulting in a large average Ueq value.
The anions within the structure also show evidence of disorder. One BF4 − anion was modelled as disordered over two locations and several more were modelled with partial occupancy. Some disordered S44 atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement. Some acetonitrile solvent molecules were also modelled as disordered over multiple locations and/or with partial occupancy. The hydrogen atoms of some disordered acetonitrile molecules could not be located in the electron density map and were not included in the model.
A further two anions per Co5L2 assembly remain unaccounted for and no satisfactory model for these anions could be obtained despite numerous attempts at modelling, including with rigid bodies.
Therefore the SQUEEZE 10 function of PLATON 11 was employed to account for the highly disordered anions and further disordered solvent molecules, which gave a potential solvent accessible void of 4517 Å 3 per unit cell (a total of approximately 1008 electrons). These anions are included as BF4 − in the formula given above. Since the diffuse solvent molecules could not be assigned conclusively to acetonitrile or diethyl ether they were not included in the formula. Consequently, the molecular weight and density given above are likely to be underestimated.
CheckCIF gives one A and sixteen B level alerts. These alerts (both A and B level) all result from the limited data resolution, solvent molecules for which the hydrogen atoms were not modelled, thermal motion and/or unresolved disorder of some anions and solvent molecules and the generally high level of thermal motion present throughout the structure as described above. Figure S48: Two views of the structure of 2 from the crystal structure of 2·10BF4·11MeCN.

2·10ClO4·3MeCN [+ solvent]
The crystals of 2·10ClO4·3MeCN [+ solvent] were grown by diffusion of diethyl ether into an acetonitrile solution of 2·10NTf2 containing excess TBAClO4. The crystals employed immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data. Even so data were obtained to 0.84 Å resolution by employing synchrotron radiation. The asymmetric unit was found to contain one half of a Co5L2 assembly and associated counterions and solvent molecules.
Bond lengths and angles within pairs of chemically identical organic ligand arms were restrained to be similar to each other. No restraints were applied to the corannulene portion of the organic ligands. One of the five ligand arms was modelled as disordered over two locations with a refined occupancy of 0.723(5) for the major part. In order to obtain a reasonable model for the lower occupancy part of the disordered ligand arm the GRADE program 12 was employed using the GRADE Web Server 13 to generate a full set of bond distance and angle restraints (DFIX, DANG, FLAT) for these residues.
Thermal parameter restraints (SIMU, RIGU) were applied to all atoms except for cobalt to facilitate anisotropic stable refinement. Even with these restraints some thermal parameters remain larger than ideal as a consequence of the high level of thermal motion or minor unresolved disorder present throughout the structure resulting in a large average Ueq value.
The anions within the structure also show evidence of disorder. One ClO4 − anion was modelled as disordered over two locations and several more were modelled with partial occupancy. Some disordered atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement. Some acetonitrile solvent molecules were also modelled as disordered over multiple locations and/or with partial occupancy. The hydrogen atoms of some disordered acetonitrile molecules could not be located in the electron density map and were not included in the model.
A further three anions per Co5L2 assembly remain unaccounted for and no satisfactory model for these anions could be obtained despite numerous attempts at modelling, including with rigid bodies.
Therefore the SQUEEZE 10 function of PLATON 11 was employed to account for the highly disordered anions and further disordered solvent molecules, which gave a potential solvent accessible void of 8373 formula given above. Since the diffuse solvent molecules could not be assigned conclusively to acetonitrile or diethyl ether they were not included in the formula. Consequently, the molecular weight and density given above are likely to be underestimated.
CheckCIF gives one A and seven B level alerts. These alerts (both A and B level) all result from solvent molecules for which the hydrogen atoms were not modelled, thermal motion and/or unresolved disorder of some anions and solvent molecules and the generally high level of thermal motion present throughout the structure as described above. Figure S49: Two views of the structure of 2 from the crystal structure of 2·10ClO4·3MeCN.
Counterions, solvents, hydrogen atoms and disorder are omitted for clarity. The structure is similar to that of 2 in the structures of 2·10BF4·11MeCN ( Figure S48) and 2·coronene·10BF4·20.5MeCN·iPr2O ( Figure 1 in the main text, Figures S43-S45). Specific refinement details: The crystals of (corannulene)0.6•2·4ReO4·6NTf2·3.5MeCN·2.5Et2O·H2O [+ solvent] were grown by diffusion of diethyl ether into an acetonitrile solution of 2·10NTf2 containing excess TBAReO4 and a large excess of corannulene. The crystals employed immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data.
Even so data were obtained to 0.7 Å resolution by employing synchrotron radiation. The asymmetric unit was found to contain one complete Co5L2 assembly, one partial occupancy corannulene molecule and associated counterions and solvent molecules.
Bond lengths and angles within the two chemically identical organic ligands were restrained to be similar to each other (excluding the corannulene portion of the organic ligands to which no restraints were applied). Thermal parameter restraints (SIMU, RIGU) were applied to all atoms except for cobalt and rhenium to facilitate anisotropic stable refinement. The occupancy of the stacked corannulene was initially refined and then fixed at the obtained value of ca. 0.6 (rounded to the nearest 0.1).
The anions within the structure show evidence of very substantial disorder. Almost all located anions were modelled as disordered over two or three locations with some lattice sites modelled as a mixture of perrhenate and triflimide. The occupancies of the disordered anions were allowed to refine freely and then fixed at the obtained values. Some additional minor occupancy positions of the anions could not be located in the electron density map and were not included in the model resulting in a discrepancy of 1.6 counterions per Co5L2 assembly (included as triflimide in the formula since no large electron density S49 peaks corresponding to rhenium remained). Some disordered atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement. Many acetonitrile solvent molecules were also modelled as disordered over multiple locations and/or with partial occupancy. The hydrogen atoms of one disordered solvent molecule could not be located in the electron density map and were not included in the model.
Further reflecting the solvent loss there is a small amount of void volume in the lattice containing smeared electron density from disordered solvent. Consequently the SQUEEZE 10 function of PLATON 11 was employed to remove the contribution of the electron density associated with this highly disordered solvent which gave a potential solvent accessible void of 978 Å 3 per unit cell (a total of approximately 314 electrons). Since the diffuse solvent molecules could not be assigned conclusively to acetonitrile or diethyl ether they were not included in the formula. Consequently, the molecular weight and density given above are likely to be slightly underestimated.
CheckCIF gives two B level alerts, both resulting from solvent molecules for which the hydrogen atoms were not modelled (singly bonded carbons).

Specific refinement details:
The crystals of 3·2CB11H12·8NTf2·2.5MeCN [+ solvent] were grown by diffusion of diethyl ether into an acetonitrile solution of 3·10NTf2 containing excess CsCB11H12. The crystals employed immediately lost solvent after removal from the mother liquor and rapid handling prior to flash cooling in the cryostream was required to collect data. Despite these measures and the use of synchrotron radiation few reflections at greater than 1.15 Å resolution were observed and the data were trimmed accordingly.

3.57(4) Å
structure. The asymmetric unit was found to contain one complete Zn5L2 assembly and associated counterions and solvent molecules.
Due to the less than ideal resolution, extensive restraints were required to facilitate realistic modeling for the organic parts of the structure. The GRADE program 12 was employed using the GRADE Web Server 13 to generate a full set of bond distance and angle restraints (DFIX, DANG, FLAT) for the organic ligand arms (excluding the corannulene portion of the organic ligands to which no restraints were applied). Thermal parameter restraints (SIMU, RIGU) were applied to all atoms except for zinc to facilitate anisotropic stable refinement. Even with these restraints some thermal parameters remain larger than ideal as a consequence of the high level of thermal motion or minor unresolved disorder present throughout the structure resulting in a large average Ueq value for the main residue.
The anions within the structure also show evidence of disorder. Three of the triflimide anions were modelled as disordered over two locations and the other two were modelled with partial occupancy.
The occupancies of the disordered anions were allowed to refine freely and then fixed at the obtained values. Some disordered atoms were modelled with isotropic thermal parameters and bond length and thermal parameter restraints were applied to facilitate a reasonable refinement. The CB11H12 − anions were modelled as rigid groups; for these anions the carbon site could not be clearly discerned from the electron density map so all atoms of the CB11H12 − anions were modelled as boron. Some acetonitrile solvent molecules were also modelled with partial occupancy. The hydrogen atoms of one of these acetonitrile molecules could not be located in the electron density map and were not included in the model.
A further 4.4 anions per Zn5L2 assembly remain unaccounted for and no satisfactory model for these anions could be obtained despite numerous attempts at modelling, including with rigid bodies.
Therefore the SQUEEZE 10 function of PLATON 11 was employed to account for the highly disordered anions and further disordered solvent molecules, which gave a potential solvent accessible void of 7641 Å 3 per unit cell (a total of approximately 2217 electrons). These anions are included as triflimide in the formula given above. Since the diffuse solvent molecules could not be assigned conclusively to acetonitrile or diethyl ether they were not included in the formula. Consequently, the molecular weight and density given above are likely to be underestimated.
CheckCIF gives 8 A and 9 B level alerts. These alerts (both A and B level) all result from the limited data resolution, the solvent molecule for which the hydrogen atoms were not modelled, thermal motion and/or unresolved disorder of some anions and solvent molecules and the generally high level of thermal motion present throughout the structure as described above. Figure S52: Side-on views of 3 from the crystal structure of 3·2CB11H12·8NTf2·2.5MeCN in stick and space-filling representations. Counterions, solvents and disorder are omitted for clarity. The structure is similar to that of 2 in the structures of 2·coronene·10BF4·20.5MeCN·iPr2O (Figure 1 in main text, Figures S43-S45), 2·10ClO4·3MeCN ( Figure S49) and 2·10BF4·11MeCN ( Figure Figure S48). Figure S53: Top views of 3 from the crystal structure of 3·2CB11H12·8NTf2·2.5MeCN in stick and space-filling representations. Counterions, solvents and disorder are omitted for clarity. The structure is similar to that of 2 in the structures of 2·coronene·10BF4·20.5MeCN·iPr2O (Figure 1 in main text, Figures S43-S45), 2·10ClO4·3MeCN ( Figure S49) and 2·10BF4·11MeCN ( Figure Figure S48). Figure S54: The structure of 3 from the crystal structure of 3·2CB11H12·8NTf2·2.5MeCN highlighting the arrangement of corannulenes. Counterions, solvents and disorder are omitted for clarity. The distance between the corannulenes is marked, highlighting the absence of a cavity in the solid state. Figure S55: Crystal packing within the crystal structure of 3·2CB11H12·8NTf2·2.5MeCN showing the 1D stacking of 3 with the CB11H12 − counterions along the b axis. Other counterions, solvents and disorder are omitted for clarity. No significant interaction was observed between 3 and CB11H12 − in solution by 1H NMR.

Computational methods
Full geometry optimizations were performed and uniquely characterized via second derivatives (Hessian) analysis to establish stationary points. Several Density Functional types were investigated in this work, including, B3LYP, 14 B3LYP-D3, 15 and B97-D 16