Linking the structures, free volumes, and properties of ionic liquid mixtures

SAXS, 129Xe NMR and PALS were used to interrogate the relationship between the structure, free volume and physicochemical properties of ionic liquid mixtures.


1-Butyl-3-methylimidazolium trifluoromethanesulfonate ([C 4 C 1 im][OTf]).
[C 4 C 1 im]Cl (108.30 g, 0.620 mol) was dissolved in dichloromethane (300 mL). This solution was added to a stirred slurry of lithium trifluoromethanesulfonate (105.09 g, 0.674 mol) in dichloromethane (300 mL) and the resultant slurry stirred at room temperature for 60 h. The slurry was filtered and the dichloromethane phase washed successively with water until the aqueous phase tested negative for halide using a 0.1 M AgNO 3 solution (14 × 5 mL). The dichloromethane was then removed and the resultant liquid dried at 50 °C in vacuo to produce [C 4

Scattering Cross-Sections
The scattering cross-sections of the [C 4 C 1 im] + cation, including the separation of ring and alkyl chain components, and the anions employed are detailed in Table S1. These cross-sections correspond to values at 18.000 keV, which is the energy of the incident X-rays and assume no interference between neighbouring atoms and hence complete additivity of the atomic cross-sections for each ion. These values have been obtained from the tables compiled by McMaster et al. 1 Table S1. X-ray scattering cross-sections in barns/ion at 18.000 keV for the IL cation and anions used in this study.

Ion or Ion
[ [  Figs S1 and S2 show that the effect of temperature is similar for both sets of mixtures. Across both mixtures there are no significant differences in the correlation distances for peak I whereas peaks II and III show an increase of around 0.1 Å on increasing the temperature from 298 K to 353 K for all compositions. The greater correlation distance for these peaks at higher temperatures is consistent with the thermal expansion of the liquid increasing the inter-ion distance. Thermal expansion effects of this order of magnitude are too small to be able to be determined for the alkyl chain correlation due to the inherent errors involved in the fitting of this peak, which accounts for the lack of any significant trends for peak I.

IL or IL Mixture τ 3 (ns) v h (Å 3 ) v h E (Å 3 ) I 3 (%)
[       As can be observed from Fig S8, Fig. S9 and S10 it can be seen that this correlation is even stronger when examined separately for the mixtures rather than for the combined data. A direct linear relationship between PALS hole volumes and 129 Xe NMR chemical shift data is unusual and indicates that there is a strong link between the dynamic and static free volume within these ILs and their mixtures.

Quantum Chemical Volume Calculations
The volume of each ion was calculated by using the Volume keyword available in G09, which performs a molecular volume computation evaluated at the 0.001 electrons/Bohr 3 density surface around the molecule. The electron density was obtained from a B3LYP-D3BJ/6-311+G(d,p) calculation with improved convergence on the electronic structure (energy of 10 -7 and RMS density of 10 -9 ) and an improved integration grid. A Monte-Carlo integration was carried out to determine the molecular volume, the tight option was employed which increases the number of evaluation points per Bohr 3 to 100 and the computed accuracy to ≈10%. In addition, 5 assessments were performed of the molecular volume for the isolated lower energy optimised structures of each ion. The average and standard deviation of these values were reported. In the case of [NTf 2 ] -5 calculations were carried out for each of the cis and trans conformers and the results averaged. In the case of [C 4 C 1 im] + 5 calculations were carried out for the lowest energy conformer (with the alkyl chain wrapped over the ring) and for a slightly higher energy structure with the alkyl chain fully extended (trans-trans-trans) and the results averaged. The results of these calculations and the associated uncertainties are summarised in Table S15. The ionic volumes obtained from these crystal structures for each method are shown in Table S16 alongside the average volumes reported for the other ions. While the absolute value of these volumes differ from those obtained using quantum chemical methods (Table S15)   . The anion-anion distances were not originally determined from these calculations; however, as it was of interest to compare the computed values to the inter-anion distances determined by SAXS, these values were extracted from each of these calculations. For the purposes of these calculations, the anion-anion distances were defined as the distance between the central heavy atom of the anion for those containing multiple atoms (e.g. P-P distance for [Me 2 PO 4 ] − ).
Table S17 details the anion-anion distances observed for each of the low energy conformers determined for the ion-pair dimers from the DFT calculations. The conformations which display π + -π + stacking of the imidazolium rings are those which feature both anions in the middle, denoted by the letter M at the beginning of the abbreviation. Those which do not feature this interaction have anions on the diagonal of the structure, denoted by the letter D at the beginning of the abbreviation. The full details of the conformer abbreviations used have been described elsewhere. 6   Table S17. Anion-anion distances (Å) determined from DFT calculations for each of the low energy conformers for ion-pair dimers of [C 1 C 1 im]Cl, [ The radial distribution functions (RDFs) of the anion-anion distances obtained from the MD simulations of the ILs and IL mixtures are depicted in Figs. S11 and S12.