Peripheral Templation Generates an MII 6L4 Guest‐Binding Capsule

Abstract Pseudo‐octahedral MII 6L4 capsules result from the subcomponent self‐assembly of 2‐formylphenanthroline, threefold‐symmetric triamines, and octahedral metal ions. Whereas neutral tetrahedral guests and most of the anions investigated were observed to bind within the central cavity, tetraphenylborate anions bound on the outside, with one phenyl ring pointing into the cavity. This binding configuration is promoted by the complementary arrangement of the phenyl rings of the intercalated guest between the phenanthroline units of the host. The peripherally bound, rapidly exchanging tetraphenylborate anions were found to template an otherwise inaccessible capsular structure in a manner usually associated with slow‐exchanging, centrally bound agents. Once formed, this cage was able to bind guests in its central cavity.

All reagents were purchased from commercial sources and used as received. For electrochemical experiments, dry solvents were purchased from Sigma-Aldrich and purged with argon before use. n Bu4NPF6 was recrystallised three times from ethanol before use. [1] and Cd(OTf)2 [2] were prepared by literature procedures.

Nuclear Magnetic Resonance (NMR)
NMR spectra were recorded using a 400 MHz Avance III HD Smart Probe (routine 1 H NMR), DCH 500 MHz dual cryoprobe (high-resolution 13  and longitudinal eddy-current delay (LED) using bipolar gradient pulses for diffusion using 2 spoil gradients was utilised. Rectangular gradients were used with a total duration of 1.5 ms.
Gradient recovery delays were 875-1400 µs. Individual rows of the S4 quasi-2D diffusion databases were phased and baseline corrected.
Through-space 1 H-1 H NMR and variable temperature NMR experiments were performed on cages with 2-4 equivalents of guest.

Mass spectrometry (MS)
Low resolution electrospray ionisation (LR-ESI) mass spectrometry was undertaken on a Micromass Quattro LC mass spectrometer (cone voltage 10-30 eV; desolvation temp. 313 K; S4 ionization temp. 313 K) infused from a Harvard syringe pump at a rate of 10 μL min −1 . High resolution electrospray ionisation mass spectrometry (HRMS-ESI) was performed on a Waters LCT Premier Mass Spectrometer featuring a Z spray source with electrospray ionisation and modular LockSpray interface.

UV-Vis spectroscopy
UV-Visible absorption spectroscopy was performed using a Perkin Elmer Lambda 750 UV-Vis-NIR spectrophotometer fitted with a PTP-1 Peltier temperature controller accessory.
Spectra were obtained in double beam mode using only the (front) analyte beam to record spectra, with air in the (rear) reference path. A background spectrum of the neat solvent was recorded using the analyte beam prior to each experiment and baseline correction applied using the Perkin Elmer WinLab software suite. Samples were analysed using quartz cuvettes with optical path lengths of 10 mm. Upon each addition, the solution was manually stirred for 1 min before acquiring the spectrum, which allowed equilibrium to be reached between the host and guest.

Titrations and Job Plots
Procedure for Job Plots: [3] A series of solutions containing 1, 2 or 3 and guest were prepared such that the sum of the total guest and host concentration remained constant (1.6−2.3 × 10 −6 M). The mole fraction of the guest was varied from 0.1 to 1.0. The corrected absorbance (mole fraction * absorbance) at 387 nm was plotted against the molar fraction of the guest solution. S5

Cyclic voltammetry
Solution state cyclic voltammetry (CV) was performed using a BioLogic SP-150 potentiostat with ferrocene (Fc) as an internal reference. Measurements were conducted under an Ar atmosphere using a conventional three-electrode cell: a glassy carbon working electrode, a Pt wire auxiliary electrode, and a Ag/Ag + quasi-reference electrode. A 0.1 M n Bu4NPF6/CH3CN electrolyte was used, with scan rates in the range 25-1000 mV s -1 .

Crystallography
Data for 1 were collected using a Bruker D8 VENTURE diffractometer equipped with high-brilliance IμS Cu-Kα radiation (1.54178 Å), with ω and ψ scans at 180(2) K. Data for 2 were collected at Beamline I19 of Diamond Light Source employing silicon double crystal monochromated synchrotron radiation (0.6889 Å) with ω scans at 100(2) K. Data integration and reduction for 1 were undertaken with SAINT [4] in the APEX3 software suite; data integration and reduction for 2 were undertaken with Crysalis PRO. [5] In both instances, multi-scan empirical absorption corrections were applied to the data using SADABS. [6] Subsequent computations were carried out using the WinGX-32 graphical user interface. [7] Structures were solved by direct methods using SHELXT-2013 [8] then refined and extended with SHELXL-2013. [8] 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. Crystallographic data have been deposited with the CCDC (CCDC 1456932 and 1456933).

Specific refinement details
Crystals of 1·12BF4·5.25MeCN·0.5Et2O were grown by slow diffusion of diethyl ether into an acetonitrile solution of 1(BF4)12. The crystals employed rapidly 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 high intensity laboratory source, few reflections at greater than 1.35 Å resolution were observed. Nevertheless, the quality of the data is more than sufficient to establish the connectivity of the structure. Due to the less than ideal resolution bond lengths and angles within pairs of organic ligands were restrained to be similar to each other and thermal parameter restraints (SIMU, DELU) were applied to all non-metal atoms to facilitate anisotropic refinement. Most tetrafluoroborate anions showed a significant amount of thermal motion; bond length and thermal parameter restraints were required for the realistic modelling of these anions. Some anions displayed positional disorder and were modelled over two (sometimes three) locations. All anions and solvent molecules were refined with isotropic thermal parameters. Three of the twenty-four anions present in the asymmetric unit could not be successfully resolved despite numerous attempts at modelling, including the use of rigid bodies. Consequently, the SQUEEZE [9] function of PLATON [10] was employed to remove the contribution of the electron density associated with the remaining anions and further highly disordered solvent molecules.

Specific refinement details
Crystals of 2·12(CF3SO3)·0.5MeCN were grown by slow diffusion of diisopropylether into an acetonitrile solution of 2(OTf)12. The crystals employed rapidly 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 1.2 Å resolution were observed. In addition, the crystals appeared to decay further during data collection, resulting in lower than ideal completeness and high residuals. Nevertheless, the quality of the data is more than sufficient to establish the connectivity of the structure. Due to the less than ideal resolution bond lengths and angles within pairs of organic ligands were restrained to be similar to each other and thermal parameter restraints (SIMU, DELU) were applied to all non-metal atoms to facilitate anisotropic refinement. Bond length and thermal parameter restraints were required for the realistic modelling of all triflate anions and solvent molecules; these were refined with isotropic thermal parameters. Only 5 anions could be successfully resolved in the asymmetric unit, despite numerous attempts at modelling, including the use of rigid bodies.
Consequently, the SQUEEZE [9] function of PLATON [10] was employed to remove the contribution of the electron density associated with the remaining anions and further highly disordered solvent molecules. S21

VOIDOO calculations
In order to determine the available void space within 1-4, VOIDOO calculations [11] were performed using MM2 minimized CACHE models (based on the crystal structures of 1 and 2). A virtual probe with the minimum radius such that it would not exit the cavity of the largest structure (4) was employed for all cages. The following parameters were changed from their default values, following a previously published procedure. [12] Probe radius: 3.          The diffusion coefficient of cage 2 was measured to be 5.2 × 10 -6 cm 2 s -1 (blue). The occluded BPh4displayed a diffusion coefficient of 9.4 x 10 -6 cm 2 s -1 (red), slower than the literature value for free BPh4in CD3CN (1.7 x 10 -5 cm 2 s -1 [13] ). Figure S40. Job plot of n Bu4NBPh4 with 2 in CH3CN to determine the 1:1 binding stoichiometry, measured by UV-Vis spectroscopy.         Figure S50. (a) Binding isotherm (1:1 system) fit to the chemical shift of the imine proton vs. the concentration of BPh4added to determine binding affinity (Ka). (b) Binding isotherm (1:2 system) calculated using Bindfit. [14] High and sigmoidal residuals, along with large errors and a negative value of K12, indicate that a 1:2 model does not fit the data better than a 1:1 binding isotherm. [15] We hypothesise that, while the binding is approximately 1:1 at low concentrations of guest, the cage may begin to bind more than one guest when a large excess of guest is added (necessary to perform an adequate data fit). The resulting isotherm may be a combination of 1:1 and 1:2 (and potentially 1:3 and 1:4) binding events.

CBr4 (tetrabromomethane)
No significant changes in the spectrum of 2 were observed with up to 8 equivalents of CBr4. The addition of 15 and 25 equivalents of CBr4 produced the spectral changes noted in Figure  S53.

Other halogenated guests
No significant changes in the spectrum of 2 were observed upon the addition of up to 50 equivalents of CCl4, CHBr3, CH2Cl2, CHCl3 or CH2BrI. The addition of 1 mL of CCl4 led to slight spectral changes, which we attribute to solvent changes, rather than direct encapsulation of the guest.

K2B12F12
Significant broadening of the signals was observed past the addition of more than 2.05 equivalents of anion. Precipitation was observed past 3.29 equivalents, even at UV-Vis concentrations (10 -5 M), thus preventing determination of a binding constant.
Note: (a) While dilute samples were pure by 1 H NMR, some portion of subcomponent B was always found to persist in solution in more concentrated samples, despite numerous purification cycles. We attribute these signals to dissociation of the cage in solution; however, the cage was stable in its solid form. Thus, 2D NMR and 13 C NMR spectra all contain a portion of free B. Resonances attributed to the free subcomponents are indicated as such with an asterisk. (b) n Bu4N(OTf) was observed to be washed out of the reaction mixture during work-up.       . This is slower than the literature value for free BPh4 -(1.7 × 10 -5 cm 2 s -1 ) [13] , indicating binding to 4.