Implications of Anion Structure on Physicochemical Properties of DBU-Based Protic Ionic Liquids

Protic ionic liquids (PILs) are potential candidates as electrolyte components in energy storage devices. When replacing flammable and volatile organic solvents, PILs are expected to improve the safety and performance of electrochemical devices. Considering their technical application, a challenging task is the understanding of the key factors governing their intermolecular interactions and physicochemical properties. The present work intends to investigate the effects of the structural features on the properties of a promising PIL based on the 1,8-diazabicyclo[5.4.0]undec-7-ene (DBUH+) cation and the (trifluoromethanesulfonyl)(nonafluorobutanesulfonyl)imide (IM14–) anion, the latter being a remarkably large anion with an uneven distribution of the C–F pool between the two sides of the sulfonylimide moieties. For comparison purposes, the experimental investigations were extended to PILs composed of the same DBU-based cation and the trifluoromethanesulfonate (TFO–) or bis(trifluoromethanesulfonyl)imide (TFSI–) anion. The combined use of multiple NMR methods, thermal analyses, density, viscosity, and conductivity measurements provides a deep characterization of the PILs, unveiling peculiar behaviors in DBUH-IM14, which cannot be predicted solely on the basis of differences between aqueous pKa values of the protonated base and the acid (ΔpKa). Interestingly, the thermal and electrochemical properties of DBUH-IM14 turn out to be markedly governed by the size and asymmetric nature of the anion. This observation highlights that the structural features of the precursors are an important tool to tailor the PIL’s properties according to the specific application.


Sample preparation
Synthesis of the PIL sample S3 Purification of the PIL samples S3 NMR measurements 1 H, 13 C and 15 N NMR experimental details S4 1 H NMR spectra S8 13 C-{ 1 H} NMR spectra S9 1D 15 N NMR spectra S10 15 N chemical shift as a function of temperature S11 15 N INEPT spectra S12 Thermal properties DSC analysis S14 DSC traces S15 TGA analysis S15 TGA curves S16 DTG curves S17 Isothermal TGA S19

Density and viscosity measurements
Density data and fitting S20 Viscosity data and fitting S23  Table S2. Experimental details of the 15 N NMR coupled spectra acquired at 318 K for DBUH-TFSI, and 305 K for DBUH-IM14, and DBUH-TFO, and displayed in Figure 4 of the manuscript.  10.0 NS: number of scans; TD: time domain -number of raw data points; SW: spectral width; D1: relaxation delay between scans. Figure S1. 1 H NMR spectra and chemical shift assignment of the PILs investigated at 328 K with spectral width up to 20 ppm. Figure S2. Quantitative 13 C-{ 1 H} NMR spectra acquired at 308K for the PILs investigated and integrals of selected isolated signals. Figure S3. 1D 15 N NMR spectra acquired without proton decoupling at 318 K for DBUH-TFSI, and 305 K for DBUH-IM14, and DBUH-TFO.

DSC analysis
Thermal properties of the PILs samples were analyzed by differential scanning calorimetry (DSC), using the instrument DSC 1 (Mettler Toledo), equipped with a liquid nitrogen cooling system. About 10 mg of the PILs samples were weighted using the microanalytical balance MX5 (Mettler Toledo) and tightly sealed in standard aluminum crucible. The liquid samples, DBUH-IM14 and DBUH-TFO, were subjected to the same temperature program. First, the samples were cooled down to 153 K, with isothermal treatment at this temperature for 2 minutes. Then, they were heated up to 323 K. The heating and cooling scan rates were 10 K.min -1 . An additional evaluation was performed on both samples, fully crystallizing them before the heating scan. Thus, the PILs were subjected to a refrigeration cycle from 153 K until their cold-crystallization temperature in the DSC instrument 3 . Then, the thermograms were recorded during the heating scan from 153 K to 323 K at 10 K.min -1 .
For the DBUH-TFSI sample, solid at room temperature, the protocol was slightly different. The sample was initially heated from 298 K to 323 K, at 10 K.min -1 , with isothermal treatment at this temperature for 2 minutes. Then, the sample was cooled to 153 K at 2 K.min -1 and kept at 153 K for 5 minutes, for the crystallization process. Finally, the sample was heated to 323 K at a scan rate of 2 K.min -1 . The DSC thermograms were recorded during the last heating scan.
In the present work, the melting point (Tm) was determined as the extrapolated onset temperatures of an endothermic peak during the heating (or reheating) scan. The cold crystallization temperature (Tc-c) was defined as the onset of an exothermic peak upon heating from a subcooled liquid state to a crystalline solid phase. The glass transition temperature (Tg) was obtained as the midpoint of a small heat capacity change upon S15 heating (or reheating) from the amorphous glass state to a liquid phase. The procedure to establish the thermal events in ionic liquids is well described in the literature 4 . The decomposition temperature (Td) is frequently defined as the onset temperature 5 or the temperature where the mass loss is 3%. However, a better approach is based on to estimate the Td taking the minimum of the TGA derivative curve (DTG) 6 . In the present work, the derivative method (DTG) was applied to determine the decomposition temperatures of the PILs studied. Besides, the samples were evaluated by isothermal TGA S16 in synthetic air using different temperature steps to obtain more profound information on the stability at higher temperatures. Figure S9. Dynamic TGA curve of the PIL samples in synthetic air. Scan rate 10 K.min -1 S17 Figure S10. DTG curve of the PIL samples in nitrogen. Scan rate: 10 K.min -1 . S18 Figure S11. DTG curve of the PIL samples in synthetic air. Scan rate: 10 K.min -1 . S19 Figure S12. Isothermal TGA of the PIL samples in synthetic air.
There, densities (ρ) were recorded using the density meter DDM2910 from Rudolph Research Analytical in 5 K steps. The other two samples, DBUH-IM14 and DBUH-TFO, were externally manipulated without a controlled environment and the air exposure was minimized as much as possible. The densities of those two samples were measured using the density meter DM45 Delta Range from Mettler Toledo in 10 K steps. For all PILs, the samples were allowed to equilibrate for at least 15 minutes before the measurements.  The calculation of the isobaric thermal expansivity ( ), leads to useful information on the dependence of the volumetric properties on temperature. To calculate , first, a second-order polynomial function for the temperature dependence of the ln(ρ) was chosen (Eq. S1) 7,8 : where ρ is the experimental density at each pressure and temperature; ρ0 is assumed to be 1.0 kg.m³; T is the temperature; and A, B and C are constant parameters determined from the experimental data using a second-order polynomial The , and C parameters for each PIL are reported in Table S4 and the temperature dependent isobaric thermal expansion coefficient, (αP) at ambient pressure P is shown in Figure S14. Overall, αP increases with the increase of the anion size, similarly to other class of ionic liquids 9,10 . As shown in Fig. S13, αP is linearly dependent on temperature.
Comparing the slope of these linear trends (Table S5) shows that DBUH-IM14 is more affected by the temperature, followed by the DBUH-TFSI and finally DBUH-TFO.

Viscosity data and fitting
The viscosity measurements were carried out in a controlled environment (dry-room, relative humidity < 0.1% at 293 K) to prevent any contamination and air contact. The measurements were performed using a MCR 102 Rheometer (Anton Paar) in a plate-plate geometry (PP25/DI/Ti rotor) with a 100 µm gap. The temperature was controlled using a Peltier-heated P-PTD200/DI base plate and an H-PTD200 actively heated geometry housing. The viscosities values were measured from the loading temperature (i.e., 293 K for DBUH-IM14 and DBUH-TFO and 303 K for DBUH-TFSI, up 373 K with a 5 K temperature step. Each temperature was held for 10 minutes for thermal equilibration, and the viscosity was recorded with a constant shear rate of 10 s -1 . All the measurements were performed in triplicates. The temperature-dependence of the data can be described by the Vogel-Fulcher-Tammann: where (mPa.s), B (K), and T0 (K) are adjustable parameters. S24 Figure S15. Viscosity of the PILs studied. Solid lines correspond to the VFT fitting.

Specific conductivity
The specific conductivity was determined by impedance spectroscopy using the After that, the conductivity was measured in 5 K steps by running a heating scan at 1 K.h -1 from 233 K to 373 K. The specific conductivity (σ) was determined by resolving the Nyquist plots of the impedance data. The uncertainty for the measured specific conductivity is 10%.

Molar Conductivity
The molar conductivities (Λ) of the PILs were obtained from the specific conductivities via Eq. (S4): where is the molar mass and is the density of the PIL. Results are reported in Table   S8. The density values used here for the molar conductivity determination were obtained from the linear regression using the parameters reported in Table 3 of the manuscript. *Relative standard uncertainty of the specific conductivity is 0.12; standard uncertainty of the temperature is 0.05 K.
The VFT model can also describe the molar conductivity (Eq. S5) when the density shows a slightly linear dependence with temperature 11 .
with Λo (S.cm 2 .mol -1 ), B (K), and T0 (K) adjustable parameters. The best-fit VFT parameters for the molar conductivity of the studied DBUH-PILs are listed in Table S9 and the fit curves are plotted in Figure 6b of the manuscript.