A Dynamic and Responsive Host in Action: Light‐Controlled Molecular Encapsulation

Abstract The rational design of a flexible molecular box, oAzoBox 4+, incoporating both photochromic and supramolecular recognition motifs is described. We exploit the E↔Z photoisomerization properties of azobenzenes to alter the shape of the cavity of the macrocycle upon absorption of light. Imidazolium motifs are used as hydrogen‐bonding donor components, allowing for sequestration of small molecule guests in acetonitrile. Upon E→Z photoisomerization of oAzoBox 4+ the guest is expelled from the macrocyclic cavity.


Computational Calculations
Gas phase geometry optimisations of the complex, host and guest molecules were performed using B3LYP functional in combination with TZVP and including the Grimme's D3 dispersion correction with Becke-Johnson damping [3] . Frequency calculations were performed at the same level of theory to obtain the thermostatistical corrections from energy to free energy in the rigid rotor/harmonic-oscillator approximation and including zero-pointvibrational energy in the gas phase at 298K and 1 atm (G T RRHO ). Solvation free energy was obtained at the same level of theory and using acetonitrile SMD continuum model (Gacet). Association free energy ∆G a is then calculated as the sum of those contributions to the gas phase association energy ∆E [4] .
∆G a = ∆E + ∆G T RRHO + ∆G acet (S1) The ∆ symbol represents that the supramolecular approach ∆X=X(complex)-X(host)-X(guest) have been used. For E,E -oAzoBox 4+ ⊂4DPDO the calculated Gibbs energy of association is -4.06 kcal mol −1 , which is in accordance with our experimental association constant.

Kinetic Data Fitting for Z→E Thermal Isomerisation
The evolution of E,Z -oAzoBox 4+ at a given temperature was determined from the integrated intensities of its corresponding Hα resonances in the 1 H NMR spectra. However, as the decay of these states depends on the decay of E,E -oAzoBox 4+ , the first step was to characterise its decay. This was done by fitting the integrated intensities of its Hα resonance as a function of time with the following equation S2.
A 0 is the initial concentration of E,E -oAzoBox 4+ , k 1 is its decay rate, t is time, and C is a constant representing the background noise of our detector.
The model fitting throughout this work was conducted using the Levenberg-Marquardt algorithm (LMA). LMA is a least squares fitting algorithm -whereby the squares of the residuals between the data and the model are minimised -and is more robust than some other least squares fitting algorithms, e.g. the Gauss-Newton algorithm. However, LMA still only finds a local minimum in the solution space and so requires a sufficiently good guesses for the model parameters as initial input. We therefore visually compared each fit to the data and found them all to approximate the data well. The uncertainties of the derived fit parameters quoted in this study represent one standard deviation.
Having determined the initial concentration and decay rate of E,E -oAzoBox 4+ , the evolution of E,Z -oAzoBox 4+ was then fit the following equation S3.
B ez 0 is the initial concentration of E,Z -oAzoBox 4+ (where the E -Hα resonance was monitored), k ez 2 is the decay rate of E,Z -oAzoBox 4+ , K 1 = k 1 /(2k 1 − k ez 2 ), and other variables have the same meaning as in Equation S2. When fitting, A 0 and k 1 were fixed as the best fit values obtained when fitting the evolution of E,E -oAzoBox 4+ with Equation S2. However, C was allowed to vary to account for any change in the noise of our detector between the E,E -oAzoBox 4+ and E,Z -oAzoBox 4+ measurements. Equation S3 was then fit for Z,E -oAzoBox 4+ (where the Z -Hα resonance was monitored), which hence allowed us to obtain values for B ze 0 and k ze 2 . Finally the combined evolution of E,Z -oAzoBox 4+ and Z,E -oAzoBox 4+ resonances was found by taking the average of the k ez 2 and k ze 2 , which we defined as k 2 .
This process described was repeated for the E,E, E,Z, and Z,E isomers at different temperatures. S5 2.

4,4'-[Bis
1 mmol) and 1,1'-carbonyldiimidazole (1.7 g, 10.5 mmol) in 1-methyl-2pyrrolidinone (20.0 mL) was heated at 170 • C for 1 h. After cooling to RT, the reaction mixture was diluted with ethyl acetate (60.0 mL) and washed with H 2 O (2 x 40.0 mL), brine (40.0 mL) and dried over MgSO 4 . The organic phase was filtered and the solvent was distilled from the filtrate. The residual solid was recrystallised from ethanol (1.0 g, 71%). O (80 mL) was added to the residue and sonicated for 10 minutes to give an orange suspension, which was subjected to centrifugation. AgBF 4 (317 mg, 1.63 mmol) in H 2 O (10 mL) was then added to the supernatant in the dark, resulting in the precipitation of AgBr. The suspension was subjected to centrifugation, after which the supernatant was set aside. The solid residue was then washed with H 2 O (30 mL) and centrifuged to obtain the supernatant, which was combined with the previous supernatant. This procedure was repeated three times.
The combined supernatant was stirred at room temperature in the light for 48 h, after which the H 2 O was removed via freeze drying. CH 3 CN (70 mL) was then added to the residue, which was then centrifuged and the supernatant filtered through a 13 mm 0.45 µm PTFE syringe filter. The filtrate was then heated to reflux for 10 minutes. Upon cooling to RT in the dark, the solvent was removed via rotary evaporation and the residue dried in vacuo. The crude product was then purified via recrystallisation (slow vapour diffusion of i-Pr 2 O into CH 3 CN, 10 mM) to yield oAzoBox 4+ ·4BF 4 − (74 mg, 24%).  Absorption (a.u.) Figure S1. Electronic absorption spectra of AzoBI 2+ (CH 3 CN) upon increased irradiation at 350 nm to achieve the Z -photostationary state (a, c and e) and at 420 nm to achieve the E -photostationary state (b, d and f ) (top) and the kinetic profiles of the E →Z photoisomerisation as monitored by the change in optical density at 320 nm (bottom).  Figure S2. Electronic absorption spectra of oAzoBox 4+ (CH 3 CN) upon increased irradiation at 350 nm to achieve the Z -photostationary state (a, c and e) and at 420 nm to achieve the E -photostationary state (b, d and f ) (top) and the kinetic profiles of the E →Z photoisomerisation as monitored by the change in optical density at 320 nm (bottom).    Figure S4. COSY 1 H NMR spectrum (CD 3 CN, 500 MHz, aromatic-aromatic resonance correlation) of oAzoBox 4+ upon partial Z -conversion via irradiation at 350 nm.    Figure S9. 1 H NMR spectrum (CD 3 CN, 500 MHz, aromatic region between 6.6-8.0 ppm) with resonance assignments of oAzoBox 4+ upon partial Z -conversion via irradiation at 350 nm.     5 Kinetics and Thermodynamics of E →Z Thermal Isomerisation Figure S15. Scheme illustrating the thermal Z →E isomerization of oAzoBox 4+ . As either of two azobenzene units may undergo the initial Z →E isomerization in Z,Z -oAzoBox 4+ , its the observed decay of is twice that of the actual rate of thermal isomerization of its two azobenzene components (κ 1 ). E,Z -oAzoBox 4+ may be equivalently written as Z,E -oAzoBox 4+ .

Differential Equations for the thermal Z →E isomerisation of oAzoBox 4+
A(t) represents the concentration of Z,Z -oAzoBox 4+ , B(t) represents the concentration of E,Z -oAzoBox 4+ (which is equivalent to Z,E -oAzoBox 4+ ). C a and C b are constants.  a Fit of the total integration area of the Z,Z -Hα proton resonance, such that the obtained rate constant is twice that of a single AB unit (2κ1).  a Fit of the total integration area of the Z,Z -Hα proton resonance, such that the obtained rate constant is twice that of a single AB unit (2κ1).  a Fit of the total integration area of the Z,Z -Hα proton resonance, such that the obtained rate constant is twice that of a single AB unit (2κ1).

S25
Erying Plot (κ 1 )      Figure S28. Temporal 1 H NMR thermal relaxation spectra (CD 3 CN, 500 MHz, 318 K) of Z -predominant AzoBI 2+ , obtained by irradiation at 350 nm.          − . Crystals formed under ambient conditions at room temperature over a period of days and were seen to be single and free of defects by use of an optical microscope fitted with a crossed-polarizer. Crystals were removed from the mother liquor and protected from desolvation by submersion in paratone oil before being mounted using an appropriate MiTeGen tip and flash frozen under a continuous stream of N 2 .
All crystallographic data are available free of charge from the Cambridge Crystallographic Data Centre (CCDC) via www.ccdc.cam.ac.uk/data_request/cif.

Crystallographic Data
The crystallographic information, structural parameters and additional refinement details for oAzoBox 4+ ·4BF 4 − is given below.    Figure S41. Structures, relative energies and thermal Z →E isomerisation activation energies of the three stereoisomers of oAzoBox 4+ and E,E -oAzoBox 4+ ⊂4DPDO. The energy is measured in kcal mol −1 and are compared to the lowest energy conformation of E,E -oAzoBox 4+ , which is set to 0 kcal mol −1 . The energy of E,E -oAzoBox 4+ ⊂4DPDO was calculated by taking the ground state energy of uncomplexed 4DPDO into account.