Stereocontrolled Self-Assembly of a Helicate-Bridged CuI12L4 Cage That Emits Circularly Polarized Light

Control over the stereochemistry of metal–organic cages can give rise to useful functions that are entwined with chirality, such as stereoselective guest binding and chiroptical applications. Here, we report a chiral CuI12L4 pseudo-octahedral cage that self-assembled from condensation of triaminotriptycene, aminoquinaldine, and diformylpyridine subcomponents around CuI templates. The corners of this cage consist of six head-to-tail dicopper(I) helicates whose helical chirality can be controlled by the addition of enantiopure 1,1′-bi-2-naphthol (BINOL) during the assembly process. Chiroptical and nuclear magnetic resonance (NMR) studies elucidated the process and mechanism of stereochemical information transfer from BINOL to the cage during the assembly process. Initially formed CuI(BINOL)2 thus underwent stereoselective ligand exchange during the formation of the chiral helicate corners of the cage, which determined the overall cage stereochemistry. The resulting dicopper(I) helicate corners of the cage were also shown to generate circularly polarized luminescence.


Materials and methods
All starting materials were purchased from commercial sources and used as supplied.2,7,14trinitrotriptycene was prepared according to a published procedure. 1(S)-and (R)-BINOL were purchased from Sigma Aldrich and used as received.Solvents were used as supplied.A CEM Discover microwave reactor was used for the subcomponent self-assembly of 1.
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.Coupling constants (J) were reported in Hz to 1 decimal place. 1 H DOSY NMR experiments were conducted on a Bruker 400 MHz Avance III HD Smart Probe spectrometer.Maximum gradient strength was 5.35 G/cm A. The standard Bruker pulse program, ledbpgp2s,3 employing a stimulated echo and longitudinal eddy-current delay (LED) using bipolar gradient pulses for diffusion using 2 spoil gradients, was utilized.A gradient ramp of 5% to 90% was used.High-resolution electrospray ionisation mass spectra were recorded on a Waters Synapt G2-Si instrument.A (1.00 mg, 4 equiv, 3.34 μmol), B (1.35 mg, 12 equiv, 10.0 μmol), C (1.58 mg, 12 equiv, 10.0 μmol) tetrakis(acetonitrile)copper(I) triflate (3.76 mg, 12 equiv, 10.0 μmol) and 0.50 mL of acetonitrile were added into a small vial that was sealed in a glove box.The vial was kept at 393 K in a microwave reactor for 1h, affording a dark brown suspension.The solvent was reduced by nitrogen flow, followed by the addition of ethyl acetate.The precipitate was collected by centrifugation and washed with ethyl acetate and diethyl ether.After drying in vacuum, 1 was obtained as a black solid in approximately quantitative yield.

Host-guest studies of 1
The cavity size of 1 was determined to be 344 Å 3 by MoloVol (probe size 1.8 Å). 2 However large windows and the lack of aromatic walls impede the binding of neutral guests.NMR titrations indicated sodium 1-hexylsulfonate, potassium perfluoro-1-hexanesulfonate and tetrabutylammonium tetraphenylborate were complexed by 1, respectively. 1 was prepared in 0.5 mL CD3CN with a concentration of 0.0005 M; sodium 1-hexylsulfonate, potassium perfluoro-1-hexanesulfonate and tetrabutylammonium tetraphenylborate stock solutions were prepared in CD3CN with a concentration of 0.05 M. Known volumes of the guest solution were added into the host solution in an NMR tube. 1 H NMR spectra were collected after shaking and sonicating the host-guest mixture for 1-5 minutes.
Bindfit was used to determine the binding constant. 3Higher binding stoichiometry was excluded due to the small cavity size and repulsion between anions.
The Hill equation was also employed to determine the apparent binding constant and the Hill coefficient.(b) Hill function fitting.The apparent binding constant was determined to be 1616 M -1 .n was determined to be 1.39, which is close to 1.     K in an oil bath overnight, affording a dark brown suspension.For chiral induction with BINOL, oil bath heating at lower temperatures gave better stereocontrol.We infer that at lower temperatures, the chiral intermediate Cu I (BINOL)2 complex, is more stable, producing a higher ee.The solvent was reduced by nitrogen flow, followed by the addition of ethyl acetate.The precipitate was collected by centrifugation and washed with ethyl acetate and diethyl ether to remove BINOL completely, as confirmed by 1 H NMR. After drying under vacuum, P6-or M6biased 1 was obtained as a black solid in approximately quantitative yield.The maximum amount of BINOL tested was 100 equiv.Higher amounts of BINOL caused precipitation and decreased the yield of the cage.

Titration of potassium perfluoro-1-hexanesulfonate
Identical NMR spectra were obtained as those of racemic 1. CD spectra were recorded by diluting the solution of 1 to 1 × 10 -5 M (calibrated by UV absorbance) in acetonitrile.Pure acetonitrile was used to determine the baselines for UV-vis and CD spectra.Δ-TRISPHAT (2.00 equivalents relative to 1) was added to NMR samples of P6-or M6-biased 1 to enable chiral discrimination.The proton signal of H7 on the quinaldine imine splits into two peaks.
After deconvolution, the areas of the two peaks were calculated to determine the ee value of 1.

S25
The M6-biased 1 used in the experiments described below was prepared in the presence of (R)-BINOL.Equivalents of (R)-BINOL used are shown in each caption.(R)-BINOL was completely removed prior to any tests.Figure S44.CD spectra of M6-biased 1 prepared by oil bath heating at 323 K (black line) and microwave heating at 393 K (red line).50 equiv of (R)-BINOL was used for each reaction.The concentration was 1 × 10 -5 M. The lower CD intensity of 1 prepared at higher temperature also provides evidence for the formation of the Cu I -BINOL complex that determines the chirality of 1, because such a complex is expected to be less stable at higher temperature.

S33
Figure S45. 1 H NMR spectra (400MHz, 298 K, CD3CN) of M6-biased 1 prepared with 50 equiv of (R)-BINOL at different temperatures.2.00 equiv of Δ-TRISPHAT was added for chiral discrimination.A higher ee was obtained at lower temperature, consistent with the importance of the proposed Cu I -BINOL chiral intermediate.All compounds in Figure S45 were screened for their ability to induce the chirality of the helicates in 1 by adding 50 equiv of them before assembly.Racemic 1 was obtained in the presence of all monohydric alcohols and diols.The failure of the chiral induction could be ascribed to extremely weak binding between these alcohols and Cu I .Adding 50 equiv of tetramethylcyclopentanediamine, MeBINOL and BINAP prevented the cage formation, only insoluble precipitates were obtained.We infer that the binding between these three compounds and Cu I is too strong, whereby subsequent ligand exchange doesn't take place.This S34 phenomenon also indicates our proposed Cu I (BINOL)2 complex is crucial and BINOL has the most suitable binding affinity to Cu I .

Generation of the triptycenes geometries
The set of possible triptycene geometries was generated by setting four molecules as they could keep the T symmetry.These geometries were scanned varying along two independent modes at the same time.The modes consisted in the distance of the triptycene from the center of the assembly (Figure S47

Shape investigation
In order to investigate the possible cage structures that could be assembled given the components, we inquired the possible match that could be observed between the four preoriented tripticenes and the metal-binding domains.This match was checked by means of distance, angles and dihedrals formed by the neighbor components.In the cluster binging domain this was expressed as N-N distance (H-H in the pre-oriented tripticenes), CNN angle  Representing on a single plot (Figure S49) the distance angle and dihedral that can be observed by the generated structures (inner binding blue colored and outer bonding red colored) along with the values related to the cluster binding domain varying the distance between the two copper atoms, we can observe the possible assemblies of the cage.From this plot we can observe the match between of HH metal binding to form cages with outer bonding and with inner bonding, while for HT it is only possible to form a cage only with inner bonding.The two inner bonding cages lie on the surface of the possible conformations explored in the geometrical study (HT Figure 2a and HH Figure 2b, main text), while the only outer bonding cage (Figure 2c, main text) has some distance from the red-colored region.In the latter case the distance can be justified by the deformations endured by the triptycenes.

Figure S9. 1 H
Figure S9. 1 H NMR spectra (400 MHz, 298 K, CD3CN) of assembly reactions carried out in different solvents.Only the dicopper(I) helicate was obtained in acetone, methanol and 1,2dichloroethane.No trace of 1 was observed.We infer that as acetonitrile is a good ligand for copper(I), the self-assembly process in acetonitrile is more reversible, enabling the annealing processes that result in self-sorting into the cage structure instead of random oligomers.The reaction in other solvents results in kinetically-trapped insoluble polymers incorporating triaminotriptycene.Thus we only observed the soluble dicopper(I) helicate in solution.

Figure S17. 1 H
Figure S17. 1 H NMR spectra (400 MHz, 298 K, CD3CN) of Cage S4.The absence of the triplet between the two singlets of the alkyl bridges is consistent with the 4-position of pyridine being substituted by bromide.

Figure S18 .
Figure S18.Aromatic region of the 1 H NMR spectra (400 MHz, 298 K, CD3CN) of Cage S1 and S4.The absence of the triplet (marked with a blue circle) between the two singlets of the alkyl bridges is consistent with the the 4-position of pyridine being substituted by bromide.Orange circles represent two singlets on the substituted pyridine ring.

Figure
Figure S21.(a) Self-assemblies with planar tritopic subcomponents.(b) 1 H NMR spectra (400 MHz, 298 K, CD3CN) of the self-assemblies.No discrete structure was observed in the selfassembly reaction 1.Only simple dicopper(I) helicates were obtained from reactions 2 and 3, where rigid tritopic ligands formed insoluble polymers that precipitated out.The result suggested the curvature of triptycene was required for the preparation of 1.
where i is the fraction of host bound by the guest which is determined by observed chemical shifts (Δδ) against the maximum chemical shift during titrations (Δδmax), [G] is the guest concentration, n is the Hill coefficient describing cooperativity, and Ka is the apparent association constant.The coefficient n > 1 indicates positively cooperative binding, n = 1 indicates non-cooperative binding, and n < 1 indicates negatively cooperative binding.

Figure S22 .
Figure S22.Cavity map (green mesh) of 1. Carbon atoms are grey, nitrogen atoms are blue, hydrogen atoms are white and copper atoms are ruby.

Figure
Figure S26.NMR titration (left) of potassium perfluoro-1-hexanesulfonate and the corresponding non-linear curve fitting (right) for the associate constant.

Figure S27 .
Figure S27.Binding isotherms and residual plots of binding potassium perfluoro-1hexanesulfonate by 1 using BindFit with the 1:1 binding mode, Ka = 667 M -1 , error = ± 9%.The cavity size of the cage and electronic repulsion may exclude the binding of more than one anionic guest.

Figure
Figure S29.NMR titration (left) of tetrabutylammonium tetraphenylborate and the corresponding non-linear curve fitting (right) for the associate constant.The chemical shift changes occur from the protons that locate on the cage surface, rather than those pointing into the cavity.We therefore infer a peripheral binding mode.

Figure
Figure S33.(a) CD spectra of enantioenriched samples of 1, immediately after synthesis and after 10 days at either room temperature or 80 °C.(b) 1 H NMR spectra (400 MHz, 298 K, CD3CN) of 1 prepared in the presence of 50 equiv of BINOL, containing 2 equiv of Δ-TRISPHAT.BINOL was only used in the preparation of 1 and was not present during the 30day periods noted.The same CD intensities in (a) and integrated peak areas of M6-and P6-1 in (b) indicated the stereochemistry of 1 is stable, with a strong chiral memory effect.Although the Δ-TRISPHAT used in the NMR tests might have a stereochemical effect that stabilizes one enantiomer, the consistency between CD and NMR results confirmed that the chiral memory effect was not due to the presence of Δ-TRISPHAT.Note that different time periods and temperatures were used to test the stereochemical stability.

Figure S35 .
Figure S35.Plot of the relationship between CD intensity and equivalents of BINOL used in the preparation of 1.

Figure S37 .
Figure S37.Partial 1 H NMR (400MHz, 298 K, CD3CN) of M6-biased 1 prepared with different equivalents of (R)-BINOL in the presence of 2.00 equiv of Δ-TRISPHAT.Each group contains parallel experiments that are marked by brackets.

Figure S38 .
Figure S38.Plot of the relationship between ee and equivalents of BINOL used in the synthesis of 1.

Figure S40 .
Figure S40.Linear curve fitting of the relationship between CD signal intensity and ee in each sample prepared using different equivalents of (R)-BINOL.The linear correlation indicates CD signal intensity and ee match well with each other.

Figure S41 .
Figure S41.Aromatic region of the 1 H NMR spectra (400MHz, 298 K, CD3CN) of a mixture of 1, 50 equiv of (R)-BINOL and 2 equiv of Δ-TRISPHAT after heating at 70 o C for 16 and 36 h.The ratio of the two enantiomers remained unchanged, implying the chirality of the system does not change once the cage forms.This provides evidence for the proposed mechanism that the chirality is determined at the initial stage of assembly.

Figure S46 .
Figure S46.Compounds tested for their ability to induce the chirality of the helicates in 1.

Figure S47. 1 H
Figure S47. 1 H NMR spectra (400MHz, 298 K, CD3CN) of self-assembly reactions for the attempted preparation of cage 1 in the presence of 50 equiv of different chiral additives.Chiral additives were removed by washing with ethyl acetate or toluene and then diethyl ether before NMR spectra were recorded.

Figure S48. 1 H
Figure S48. 1 H NMR spectra (400MHz, 298 K, CD3CN) of mixtures of 1 and different equivalents of BINAP upon heating for 16 h.Cage 1 decomposed after heating, indicating the binding affinity between BINAP and Cu I is high.

Figure
Figure S51.Time-resolved PL spectra of 1 in acetonitrile (left) and methanol (right).All emissions faded in a nanosecond level, indicating singlet-involved radiative decay.

Figure
Figure S52.CPPL spectra of P6-and M6-1 in acetonitrile measured by a homemade spectrometer equipped with a rotating quarter waveplate.Photons are captured and thus all signals are positive only.Negative signals can be seen on the glum spectrum in the manuscript.

Figure S53 .
Figure S53.Isolated metal binding domain complexes.The two complexes (HT on the left and HH on the right), investigated at DFT level.The optimal conformations were inquired varying the distance between the two copper atoms (shown in violet).
left) and the rotation the triptycenes around their principal axis (Figure S47 right).Other modes were excluded to maintain the symmetry of the assembly.The distances were sampled every 0.055 Å, while the rotation was sampled every 0.09 deg.A total of 10000 structures were generated.The structures showing overlap or close contact between the molecules were excluded.

Figure S54 .
Figure S54.Triptycenes displacements modes.The two modes were used to span the possible conformations.Each color refers to the generated arrangement.

Figure S55 .
Figure S55.Inner and outer binding of a generic 4 triptycenes assembly.

Figure
Figure S56.3D representation of the areas explored by the geometrical investigation (inner binding, blue surface) and inner binding (red surface), along with the scanned configurations of HT (grey colored) and HT (green colored) metal binding observed.The three DFT optimized structures that could be generated.The three coordinates correspond to the distance, angle and dihedral, that should match in order to generate a cage assembly.