Unveiling the nature of supramolecular crown ether–C60 interactions

Preparation of exTTF-(crown ether)2 receptors, which host C60, to understand the nature of the fullerene–crown ether interaction. A combination of experimental and in silico studies suggest that it results from the interplay of donor–acceptor, ð–ð , n–ð and CH•••ð interactions.


General Methods
Reagents were used as purchased from commercial sources without further purification. Solvents were dried and distilled using standard techniques prior to use. [1] Compounds 2, [2] 9, [3], [4], [5] and 15 [6] were prepared according to previously reported procedures. All reactions were performed in standard glassware under an inert Ar atmosphere. Analytical thin-layer chromatography was performed using aluminum-coated Merck Kieselgel 60 F254. Visualization was made by UV light or I 2 vapor. Purification of crude reaction mixture was achieved by flash chromatography (FC) using neutral Al 2 O 3 gel (Panreac) or SiO 2 gel (Scharlau, Kieselgel 60, 0.04-0.06 mm). NMR spectra were recorded on a Bruker DPX-300 spectrometer at 298 K using partially deuterated solvents as internal standards. Multiplicities are denoted as follows: s = singlet, d = doublet, t = triplet, m = multiplet, br = broad, dd = double doublet. IR spectra were determined on a Bruker Tensor 27 (ATR device) spectrometer. Only neat picks are reported. UV/Vis spectra were recorded with a Shimadzu Spectrophotometer UV-3600. MALDI-TOF experiments were taken on a Brucker Ultraflex III using DCTB + NaI as matrix. Femtosecond transient absorption studies were performed with 150 fs laser pulses (1 kHz) from amplified Ti:Sapphire laser systems (CPA-2101 and CPA-2110 from Clark-MXR, Inc.), the laser energy was 200 nJ.
After 30 min, the solvent was removed under reduced pressure, the resulting residue was diluted with anhydrous DCM (30 mL) and Et 3 N (0.1 mL, 0.66 mmol) and 2,6-dihydroxy-exTTF 13 (62 mg, 0.15 mmol) were added. The reaction mixture was stirred at rt until no precipitate was observed.

Theoretical Calculations
A first exploration of the supramolecular potential energy surface was carried out by performing geometry optimizations of the different 1-6·C 60 host-guest associates at the semiempirical PM7 level of theory [10] using the MOPAC2012 program package. [11] The geometry optimization termination criterion (gradient norm) in both gradient minimization and energy minimization was set at 0.01 kcal/mol/Å. Figure S7 shows the minimum-energy structures for the 1-6·C 60 complexes obtained after PM7 optimization. Several conformers were designed (and subsequently optimized) by internal rotation around the single bonds of the ester groups but only the most stable rotamers are discussed. Non-embraced host-guest arrangements, in which the crown ethers fold themselves away from C 60 , and intermediate one-arm embraced conformations, in which the C 60 ball is embraced by only one arm of the exTTF-(crown ether) 2 receptor, were also optimized for complexes 1-3·C 60 .

S23
Accurate geometry optimizations of the supramolecular associates 1-6·C 60 were performed within the density functional theory (DFT) framework [12] using the B97-D Grimme's functional, [13] which includes an additional dispersion energy term, and the correlation-consistent cc-pVDZ basis set. [14] The B97-D functional is consolidated as an efficient and accurate quantum chemical approach to deal with large systems where dispersion forces are of general importance at a relative low-cost of computation. [15] Previously optimized structures at the PM7 level were used as starting geometries for the more accurate DFT optimizations. The different structural disposition adopted by the crown and aza-crown ether moieties in 2·C 60 and 5·C 60 , respectively, at the DFT minimum-energy geometries ( Figure 6 in the main text) were optimized by means of the Gaussian 09 (Rev. C01) suite of programs. [16] On the B97-D/cc-pVDZ optimized structures, the association binding energy of the complexes was estimated by single-point energy calculations using the revPBE0 correlation-exchange functional in combination with the -D3 Grimme's dispersion correction (revPBE0-D3) 12,13 and the correlationconsistent cc-pVTZ basis set. [14] The choice of the exchange-correlation functional revPBE0 is justified by its excellent performance when studying the very popular S22 [17] and S66 [18] non-covalent interaction databases [19] as well as when applied to other related supramolecular systems. [20] The basis set superposition error (BSSE) is expected to be negligible at the large correlation-consistent triple-ζ basis set employed and, therefore, the interaction energies are not counterpoise corrected.
Moreover, note that the counterpoise method is believed to overestimate the BSSE, for which some authors propose to scale it down by half of its value. [21] The original damping function in the -D3 approach has been replaced by the Becke-Johnson damping function to provide a better performance. [22] The "resolution of identity" (RI) [23] and "chain of spheres" (COSX) [24] techniques, for the Coulomb and exchange integrals, respectively, were used to alleviate the computational cost of the more demanding steps. Note that the three-body contribution to the dispersion energy has been included because it can be significant for medium and large supramolecular systems. [25] The association energy in each associate was computed as the difference between the energy of the associate and the sum of the energies for the two constituting fragments at the geometry of the complex [E bind = E(complex) -E(exTTF-tweezer) -E(C 60 )]. Geometry optimizations and single-point energy calculations at the revPBE0-D3/cc-pVTZ level were all performed using the ORCA program package (version 2.9.0). [26] Molecular orbitals ( Figure S9) were plotted using the Chemcraft 1.6 software with isovalue contours of ±0.03 au. [27] S24 Figure S9. Isovalue contours (±0.03 au) calculated for the frontier molecular orbitals (HOMO and LUMO) of the supramolecular associates at the revPBE0-D3/cc-pVTZ level.