Xenon binding by a tight yet adaptive chiral soft capsule

Xenon binding has attracted interest due to the potential for xenon separation and emerging applications in magnetic resonance imaging. Compared to their covalent counterparts, assembled hosts that are able to effectively bind xenon are rare. Here, we report a tight yet soft chiral macrocycle dimeric capsule for efficient and adaptive xenon binding in both crystal form and solution. The chiral bisurea-bisthiourea macrocycle can be easily synthesized in multi-gram scale. Through assembly, the flexible macrocycles are locked in a bowl-shaped conformation and buckled to each other, wrapping up a tight, completely sealed yet adjustable cavity suitable for xenon, with a very high affinity for an assembled host. A slow-exchange process and drastic spectral changes are observed in both 1H and 129Xe NMR. With the easy synthesis, modification and reversible characteristics, we believe the robust yet adaptive assembly system may find applications in xenon sequestration and magnetic resonance imaging-based biosensing.

1 H NMR and 13 C NMR spectra were recorded on Bruker 400 and 500 MHz NMR spectrometer. 1 H-1 H COSY, HSQC, HMBC, NOESY and 2D-EXSY spectra were recorded on Bruker 500 MHz NMR spectrometer. Chemical shifts are reported in ppm and referenced to tetramethylsilane or the residual solvent resonance. 129 Xe NMR spectra were recorded on a Bruker Avance III 500WB spectrometer (nominal frequency for 129 Xe =138.3 MHz) using a 5 mm BBFO probe (temperature = 298 K). The chemical shift of free xenon in (CDCl2)2 was referenced to the resonance frequency of pure xenon gas extrapolated to zero pressure according to literature. [1] Mass spectra were obtained on LCMS-2010, Shimadzu Co. (for CSI). Infrared spectra were recorded on JASCO-4800 (for compounds characterization) and Nicolet-6700 FT-IR spectrometer.
Elemental analysis was recorded on Carlo Erba 1106. Melting points are uncorrected.

(3aS,7aS)-1,3-bis(3-amino-5-(trifluoromethyl)benzyl)octahydro-2H-benzo[d]imid
azol-2-one (1): To a solution of 6 (546 mg, 1 mmol) in THF (50 mL) was added a solution of SnCl2·2H2O (1.50 g, 6.4 mmol) in conc. hydrochloric acid (2.5 mL). The mixture was stirred at room temperature for 12 h. Then a solution of 40% aq. NaOH was added to reach pH > 10, and the organic layer was separated. The aqueous layer was extracted with ethyl acetate (20 mL× 3). The combined organic layers were dried over Na2SO4, and then evaporated under reduced pressure. The residue was subjected S5 to column chromatography on silica gel (dichloromethane/methanol = 10:1) to give 1 as a white solid (481 mg, yield: 98%).  Phasing and refined with the ShelXL (Sheldrick, 2015) refinement package using Least Squares minimisation. All non-hydrogen atoms were refined anisotropically, and hydrogen atoms attached to carbon atoms were fixed at their ideal positions. For the included xenon atom, the refining mode of the occupancy factor was set as free, so that an exact occupancy factor of xenon can be obtained. See Supplementary Tables 1-5 for crystal data, structure refinement and related structural parameters.    Figs. 7-14). Based on 2D EXSY spectrum, two-site position exchange signals were observed and the two divided sets of signals can thus be determined (in blue and purple, respectively). The assignments of all the signals in each set can be obtained by a combination of the above spectra. The identifying of which set is which was realized by the correlation signal observed between NHa and C=S in HMBC spectrum (no correlation peak was observed between NHa and C=S; this was compared to that the similar correlation peak was also observed in Xe  M2, vide infra). Supplementary Fig. 7 1 H NMR (298 K, 500 MHz) of M2 in (CDCl2)2 ([M]initial = 10 mM). "*" denotes the minor N2 inclusion peaks. These minor peaks disappeared through bubbling of O2 to the solution, and re-appeared after bubbling of N2.  Figs. 15-22). The assignments of the signals in each set (in blue or purple) can be obtained by a combination of the spectra.
In the 2D EXSY spectrum, two-site position exchange signals were no longer observed (the xenon binding suppressed the exchange). The identifying of which set is which was realized by the NOE correlation signal observed between NHa and Hc, Hc, and the correlation signal between NHa and Hb in the NOESY spectrum. In the HMBC spectrum, a correlation signal between NHa and C=S was observed, but no correlation peak was observed between NHa and C=S (this also gave an indication for identifying the two sets of signals for M2, vide supra).

Determination of xenon binding constants. The binding constant of xenon with
M2 was determined from 1 H NMR integration of the related species, including free M2, and Xe  M2. The concentration of the free xenon was estimated to be the xenon solubility in 1,1,2,2-tetrachloroethane (0.101 M). [2] The calculation was carried out The activation energy ΔG ‡ for the exchange between the two equally populated sites can be determined according to the below equation, whereas Tc is the coalescence temperature and Δν corresponds to the difference of the chemical shifts in hertz between the two signals in the absence of exchange. [3] Here the difference of the chemical shifts at 233 K was used as Δν for calculation as the value reached almost constant at these low temperatures (Supplementary Table 6).

Supplementary Note 1
In order to roughly assess the sequestration ability of the crystal capsule toward xenon gas, we have built the below thermodynamics cycle and obtained the dissociation pressure for xenon bound within the crystal capsule (Pdissociation = 0.1 atm).
- competent for sequestration of xenon from atmosphere directly. What we suppose is that development of xenon host materials (like our system) could have the potential for the separation of xenon from the already-enriched sample (e.g. a mixture of xenon and krypton) after using the traditional cryogenic methods. [4][5]