Synthesis and Thermophysical Properties of Ether‐Functionalized Sulfonium Ionic Liquids as Potential Electrolytes for Electrochemical Applications

Abstract During this work, a novel series of hydrophobic room temperature ionic liquids (ILs) based on five ether functionalized sulfonium cations bearing the bis{(trifluoromethyl)sulfonyl}imide, [NTf2]− anion were synthesized and characterized. Their physicochemical properties, such as density, viscosity and ionic conductivity, electrochemical window, along with thermal properties including phase transition behavior and decomposition temperature, have been measured. All of these ILs showed large liquid range temperature, low viscosity, and good conductivity. Additionally, by combining DFT calculations along with electrochemical characterization it appears that these novel ILs show good electrochemical stability windows, suitable for the potential application as electrolyte materials in electrochemical energy storage devices.


Introduction
Ionic liquids (ILs) have gained considerable interest over the past few decades due to their numerous attractive properties, such as extremelyl ow vapor pressure,l ow flammability,h igh thermals tability and large liquid range. This has led to an umber of groups researching IL applicationsp articularly in the fields of catalysis, [1] and separations. [2] In comparison with molecular solvents, ILs also possess high ionic conductivity coupled with good electrochemical stabilitya nd, therefore, have been touted as new electrolytes fore nergy devices such as solar cells, [3] fuel cells, [4] lithium batteries, [5] and supercapacitors. [6] Within the fields of lithiumb atteries and capacitors, an umber of variousI Ls has been developed with specific physiochemical applications.T od ate, however,m ost of these ILs have been based on imidazolium, [7] tetraalkylammonium, [8] pyridinium, [9] and quaternary phosphonium cations. [10] Despite having al ower viscosity,h igher conductivitya nd lower melting points;I Ls based on trialkylsulfonium cations have attracted limited attention compared to their ammonium and phosphonium analogues [11] with other research groups emphasizing the use of cyclic [12] and acyclics ulfonium ILs [13] as electrolytes for energy devices. Recently,w ea lso reported the use of protic and aprotics ulfonium ILs as potentiale lectrolytes for electric double-layer capacitors (EDLC) devices. [14] In addition, numerousp ublications have highlighted the influence of the presenceo fa ne ther functionalization on the physiochemical properties of acyclic ammonium-based [15] and cyclic ammonium-based ILs. [16] Research surrounding the influence of mono-etherf unctionalization on the properties of sulfonium-based ILs has only been studied by Han et al. [17] and Orita et al. [18] In this paper,w ee xtended this family of hydrophobic room temperature ILs based on mono-ether and diether containing alkyl chains appended to sulfonium cations. These ether containing chains are of particulari nterest for energy storages ystems such as supercapacitors and Li batteries due to the possible improved transport properties and good electrochemical stability. Furthermore, the presence of ethereal groups may also offer improved dissociation of the Li + cation and improved transportp roperties in the bulk solution in Li-ion batteries, for example. The investigation of viscosity,d ensity,t hermalp hase behavior,t hermals tability,e lectrochemicals tabilitya nd ionic conductivity were undertaken. Finally,I Ls electrochemical stabilityr esults were also analyzed in combination with DFT calculations to further understand the effect of thee ther functionalization in the cation on this property.
During this work, an ovel series of hydrophobic room temperature ionic liquids (ILs) based on five ether functionalized sulfonium cations bearing the bis{(trifluoromethyl)sulfonyl}imide, [NTf 2 ] À anion weres ynthesized and characterized. Their physicochemical properties, such as density,v iscosity and ionicc onductivity,e lectrochemical window,a long with thermal properties including phase transition behavior and decomposition temperature, have been measured. All of these ILs showed large liquid range temperature, lowv iscosity,a nd good conductivity. Additionally,b yc ombining DFTc alculations along with electrochemical characterization it appears that these novel ILs showg ood electrochemical stabilityw indows, suitable for the potential application as electrolyte materials in electrochemical energy storaged evices.

Synthesis of the SulfoniumI Ls
The general synthesis method of the selected sulfonium based [SR 3 ][NTf 2 ]I Ls is described in the experimental section. The various ether and alkyl substituents Ra ttached to the sulfonium center are listed in Table 1a nd the ILs and intermediates (1a, 1b, 2a and 2b)a re shown in Figure 1. Several synthetic strategies were undertaken to develop these ILs with the maximum yield and purity.I nitial attempts to synthesis these materials using conventionalp rocedures employed for imidazolium-or ammonium-based ILs proved problematic due to the lower alkylation yield observed between the sulfide and the ethers, as shown in Figure1,r oute 1. [19] Therein, low to moderate yields of deeplyc olored ILs were obtained. Therefore, the proposed strategy was to firstly synthesize the corresponding thioethers 1a and 1b and, if possible, alkylate at the final step. These thioethers 1a and 1b were easily synthesized using the conventional Williamsonp rocedureu nder aqueous conditions. [20] The corresponding alkylation of the thioethers using either iodomethane or diethylsulfate, via route 2i nF igure1also proceeded smoothly to yield the corresponding dialkylether-sulfonium ILs.
Alkylation of the thioethers 1a and 1b with other bromoethers to form the alkyldiether sulfoniumI Ls proved to be more difficult due decreases in nucleophilicity of the sulfur center resultingi nl ower yields of the desired ILs, see route 3 in Figure 1. Therefore, ad ifferent synthetic strategy was adopted to produce alkyldiether-sulfonium ILs, as shown in Figure 1, route 4. Therein, sodium sulfide was reactedw ith 2equivalents of the corresponding bromogylme to form thiodiethers 2a and 2b.A lkylation of these thioethers with iodomethane pro- For lab-scale synthesis of thesee ther-functionalised ILs, the commercial availability and good reactivity of 1-bromo-2-(2methoxyethoxy)ethane and 2-bromoethyl methyl ether,m ake them attractive reagents for relatively straightforward alkylation. However,f or larger scale preparation, the bromide containing groupsb ecome less feasible. In this context, alternative synthetic strategies wherein alkoxy functionalizationi sc ompleted using less toxic reagents, (e.g. sulfonate esters, tosylates) have been described in literature and, furthermore, reviewed in depth. [21]

Physical Characterization of the Sulfonium ILs
The thermalp roperties of selected ILs werec haracterized by differential scanning calorimetry (DSC) and thermalg ravimetric analysis( TGA). Furthermore, the density (1), dynamic viscosity (h)a nd ionic conductivity (s)w ere determined as the function of temperature at atmospheric pressure. Thisc haracterization data is summarized in Ta ble 2.

ThermalProperties
The thermalb ehavior of the studied ILs was investigated by DSC from 183.13 to 373.15 K. As shown in Figure S6 (left) of the SI, most of the ILs only exhibited glass transitions, T g ,w ith only the [S 1,1,G1 ][NTf 2 ]f unctionalized IL showing an observable meltingt ransition, T m ,a t2 67.54 Ka nd freezing transition, T f ,a t 223.43 K. Thiss upercooling phenomenon has been previously well documented fori midazolium-and ammonium-based ILs. [22] Interestingly,t he analogous non-functionalized [S 1,1,4 ][NTf 2 ] was observed to have a T m between2 69.55 [23] and 271.55 K [18] as measured by two independentr esearch groups. Although we are only comparing one set of ILs, generally,r eportsi nt he literatures uggest that functionalization of IL alkyl chainsw ith ether groups results in ad ecrease in meltinga nd increasei n liquid range. [24] Within the ILs studied, increased ether func-tionalization with either G 1 ( Figure S6 do ft he SI) or with G 2 (Figures S6 ba nd S6 eo ft he SI) did result in as ignificant increase in liquid range. Furthermore, by increasing the alkyl chain length from C 1 ( Figure S6 ao ft he SI) to C 2 in (Figure S6 c of the SI) also led to an increasei nt he liquid range temperature. Therein, the most plausible explanation fort he observed lack of T m is due to the increased rotational freedom and subsequent reduction in lattice energy as has been described elsewhere. [25] The T g valuess howed no correlation with regards to the IL functionalization. In addition, as evidenced from the transport studies, the low T g values observed do not always lead to low fluidities due to fragility. [26] To further investigate the thermalp roperties of the selected sulfonium-based ILs, TGA analysis of each sample has been carried out. As shown in Figure S6 (right) of the SI, the decomposition temperatures, T d ,o ft hese sulfonium ILs are in the range of 492.15-517.15 K, significantly lower (< 150 K) than that observed for analogous ammonium-based ILs. [22b] In addition, the short-chain-ether functionalized ILs tended to be less thermally stable than their analogous non-functionalized ILs primarily due to the weakened cation-anion electrostatic interaction freeingt he anionst oa ct as nucleophiles. Despite this lowering of the decomposition temperature, the T d values are still significantly higher than those temperatures (373-398K)e ncountered in high temperature energy storage devices. [27] Additionally,a sd epicted from the Figure S6 of the SI and Ta ble 2, all investigated ILs have al arge liquid range temperature, which is highert han 240 Ki nt he case of the [S 1,1,G1 ][NTf 2 ]a nd exceeds 300 Kf or the other ILs. In fact, each IL may be potentially used as electrolytes in low (e.g. 268 K) and/or high (e.g. 473 K) temperaturese nergy storaged evices.

Physicochemical Properties of Selected ILs
The fundamentalp roperties of the sulfonium ILs, including physicochemical quantities of density (1), viscosity (h), conductivity (s), and their ionicity have been investigated as detailed below. The densities (1/g cm À3 )h ave been measured from 293.15 to 363.15 Ka ta tmospheric pressure for all the investigated ILs (see tabulated data in Ta ble S1 of the SI). As shown in Figure 2, the density of selected sulfoniumI Ls are between 1.38 and 1.53 gcm À3 at 298.15 K, values whicha re in line with other sulfonium [NTf 2 ] À -based ILs previously reported. [23,28] Furthermore, the density decreases with temperature as expected. Generally, and as shown in Figure 2, over an arrow range of temperatures, the temperature dependence on the density can be expressed as follows: where a, b are the fitting parameters and T is the temperature in Kelvin. Withint he ILs studied as trong linear relationship (R 2 > 0.9999) with temperature was obtained and the best fitting parameters of Equation (1) are summarizedi nT able S2 of the SI. Furthermore, as shown in Figure 2, the density of the ILs decreasesb yi ncreasing the degree of functionalization within the alkyl chain and follows the trend, [S 1,1,G1 ] + > [S 1,2,G1 ] + > [S 1,1,G2 ] + > [S 1,G1,G1 ] + > [S 1,G2,G2 ] + .A st he density is governed by the cation-anion interaction and molecular packing, [29] which reduces with increasing ether functionalization and/or increase in alkyl chain, this trend mirrors those already reported in the case of the ether functionalized imidazolium-basedI Ls. [30] The viscosity (h/mPa s) is also an important property for assessing ILs with respect to their use as electrolyte mediai n energy devices, because of the strong association of the rate of mass transport. The electrolytesi nm any energy devices are required to operate at near ambient temperature ranges. In other words, any IL used as an electrolyte should have al ow viscosity particularly at room temperature. Currently,o ne of the largestb arriers for the application of the ILs as pure electrolytesi st heir high viscosity in comparison with classical electrolytesf or either supercapacitors (55 mS cm À1 at 0.6 mPa sf or 1mol dm À3 [Et 4 N][BF 4 ]i na cetonitrile at 298.15 K) [31] or Li batter-ies (9.7 mS cm À1 at 0.5 mPa sf or 1mol dm À3 Li[PF 6 ]i ne thylene carbonate:dimethylcarbonate blend at 298.15 K). [32] In this regard, the investigated sulfonium-basedI Ls show viscosities ranging from 28 to 53 mPa sa t2 98.15 K( see Ta ble S3 of the SI). As shown in Figure 3, the viscosity of the ILs decreases with increasingt he temperature, as expected. Furthermore, the viscosity seems to be strongly affected by the size and the symmetricity of the cation as well as by the degree of functionalization within the alkyl chain and followst he trend: To the best of our knowledge,t he viscosity of the [ [18] This is, however,s omewhat higher than the correspondingb is(fluorosulfonyl)imide [FSI] À anion which was reported to have av iscosity of 30.0 mPa sa t2 98.15 K. [17] This decrease in viscosity upon the addition of ether units into the alkyl chain is not surprising given the decrease in van der Waals interactions and has also been reported foro ther ether functionalized ILs. [33] In addition, all of the ILs showed comparable or significantly lower viscosities than the corresponding mono-ether containing pyrrolidinium [24b] (53 mPa sa t2 98.15 K), piperidinium [34] (55 mPa sa t2 98.15 K), linear guandinium [35] (58 mPa sa t 298.15 K) andc yclic guandinium [35] (46 mPa sa t2 98.15 K). More surprising, wast he similarv iscosities observed for the remaining ILs regardlesso ft he degree of ether functionalization. In this regard, negligible differences in viscosity have been reportedf or di-and triether based imidazoliuma cetate ILs [36] and the analogousd i-and trialkyl ammonium acetate based ILs. [37]   The conductivity of an IL is also of vital importance if it is to be considered as as upporting electrolyte in energy devices. Generally,t he ionic conductivity of ILs is mainly governed by their viscosity,f ormula weight, density and ions ize. [38] As shown in Figure 4, the conductivity of the ILs increases with increasingt emperature, as expected. Furthermore,t he ionic conductivity of the ILs decreases with the followingt rend, [S 1,2,G1 ] + > [S 1,1,G1 ] + > [S 1,G1,G1 ] + > [S 1,1,G2 ] + > [S 1,G2,G2 ] + and covers the range of 2.3-5.0 mS cm À1 and of 10.8-20.2mScm À1 at 298.15 and 353.15 K, respectively (see Table S4 of the SI).
The small difference in ionic conductivity data, even at lower temperatures, suggests that the transport of the ions in these sulfonium ILs is perhaps more dependento nt heir viscosity. Conversely,f or example at 298.15K,t he [S 1,1,G1 ][NTf 2 ]a nd [S 1,2,G1 ][NTf 2 ]I Ls show significantly lower ionic conductivities (3.94 and 4.93 mS cm À1 )c ompared to the analogous corresponding ILs with the [FSI] À anion (9.5 and 7.6 mS cm À1 ,r espectively), [17] although they have comparable viscosities. This also clearly indicates that an IL containing ac ommonc ation and an anion with al ower molecular weight ands maller size is also more favorable for the production of more conductive salts.
The ). In this regard, the distinction must be made between these two transport properties. The viscosity measured experimentally is macroscopic in nature,w hichi sc learly related to the cohesive energy of the solution;w hereas the conductivity of selected sulfonium-based ILs seems to be more strongly affected by the structure (volume), interaction (packing) and then the mobility of ions in the solution. One way to investigate the relationship between the viscosity (fluidity) and the conductivity( resistivity) of selected sulfonium-based ILs is driven by the determination of their ionicity thanks to the utilization of the Walden plot. [26] Prior to investigation of the effect of the temperature on the ionicity of the studied ILs, each property has to be correlated as the function of the temperature.
Herein, as shown in Ta ble S5 of the SI, all of these ILs deviated slightly from the Arrhenius behavior [Eq. (2)] but could be betterd escribed by the Vogel-Tamman-Fulcher (VTF) type equation[ Eq.
(3)].T hisi sn ot as urprising result as many other ILs are generally well described by the VTF equation. [39] h where h 0 (mPa s), s 0 (mS cm À1 ), E a (kJ mol À1 ), B h/s (K), and T 0 (K) are the Arrhenius and VTF fitting constants.The best fitting parameters for the viscosity and conductivity as af unction of temperature are reportedi nt he Table S5 of the SI, together with correlation coefficient R 2 for the fit. When comparing the ideal glass-transitiont emperatures derived from VTF-type fitting (i.e. T 0 values in the Table S5 of the SI) using the ionic conductivity and viscosity measurements, it can be seen that the corresponding values for each IL show good correlations (i.e. < 10 %) except in the case of the [S 1,G2,G2 ] [NTf 2 ]( i.e. % 22 %). Nevertheless, irrespectiveo ft he IL examined, no trend can be observed between T g and T 0 values from both VTF-type fittings.W ith the exception of [S 1,G2,G2 ][NTf 2 ], all of the ILs showtheoretical T 0 significantly lower than those obtained fort he DSC derived T g highlighting the fragile nature of these ILs. [40] The Arrhenius activation energies for both viscosity and conductivitya re also depicted in Table S5 of the SI, and are calculated between the temperature range 293.15 and 353.15K.A se xpected, for each IL the activation energy for conductivity is lower than that for viscosity due to the fractional Walden rule. [41] Unlike the Arrhenius equation, all VTF equations seem to correlate accuratelye ach property over aw ider temperature range allowing better predictions for both the limiting viscosity (h 0 )a nd limitingc onductivity (s 0 ), see Ta bleS5o ft he SI. This correlation allowsb etter determination of the Walden ionicity of the selected ILs as af unction structure and temperature. Prior to constructingt he Walden plot,t he molar conductivity (L m /S cm 2 mol À1 )d ata must be calculatedw ithin the same temperaturerange for each IL according to Equation (4): where, s (S cm À1 ), Mw (g mol À1 )a nd 1 (g cm À3 )a re the ionic conductivity,t he molecular weight and the density of the selected ILs.
The Walden plot (i.e.t he variation of log 10 (L m /S cm 2 mol À1 ) vs. log 10 (h À1 /P À1 )) for the five neat ILs within at emperature range of 293.15 to 353.15 Ki sp resented in Figure 5. The spe-  Figure 5r epresents as o-called ideal KCl line, that is, the ideal Walden behavior of a0 .01 mol dm À3 aqueous solution of KCl, as trong electrolyte, where the constituent ions are known to be fully dissociated and equallym obile in solution. [42] As shown in Figure 5, ad eviation below this ideal line is observed for all investigated sulfonium-based ILs, which may indicate, in each case, that not all ionic species are available for the conductiono fc harge and their apparent electrolytic conductivity is lower than may be expectedf or ag iven viscosity.I no ther words, this may be relatedt oas light tendency for the formation of ion pairs in each IL. This is not as urprising result as such behavior has been reported for several ILs. [43] Furthermore, as reported in Ta ble S6 of the SI, by applying the "classical"W alden rule, it appearsc learly that the corresponding Walden product, W = L m ·h is temperature dependent.A gain, this behaviori sa lready well described in the literature for al arge range of ILs, and according to Schreinere tal., [43] ab etter illustration of the conductivity-viscosity relationship could be achieved by using af ractional Walden rule as follows: where a and W' are an additional exponent fitting parameter and the "Walden product"oft he fractional Walden rule.

Electrochemical Window
One of the reasons fort he recent growing interesti nI Ls is the wide electrochemical windows,w hich may allow them to be used as solvent-free supporting electrolytes in high-energy density devices, including LiS batteries and electrochemical capacitors. [46] Therefore, of particular interest is the relationship between the cationic structures of the ILs andt heir electrochemicalw indows. Cyclic voltammograms (CVs) have been measured ( Figure 6) for all investigated ILs using a3 -electrode cell inside an Ar filled glove box, with a0 .3 cm diameter glassy carbon disk workinge lectrode, af resh lithium strip as reference electrode and ap latinum wire as counter electrode. A small quantity of Li[NTf 2 ]s alt (i.e. 20 mg of Li salt per mL of IL) has been added to the IL to be used as an internal reference.
The reduction (E red )a nd oxidation (E ox )l imiting potentials of the ILs were then determined by ag raphical methoda tthe potentialv alue where tangent lines to the main reduction/oxidation walls and the "flat" electroactivity domains cross. These values are presentedi nt he Table 3. The reduction limiting potentials suggest that both the number and the length of the ether chains attached on the sulfoniumcation have anegligible effect on the reduction potential of the ILs. However,s uch values seem to be lower than those reported for [S 1,1,4 ][NTf 2 ] (0.9 Vv s. Li + /Li). [28] This difference mayb ee xplained by comparingt he LUMO energy of selectedc ations (average close to À4.39 eV) vs. [S 1,1,4 ] + (À4.60 eV) as shown in Table S8 of   On the otherh and, the data reported in Ta ble 3s hows that the length of the ether chains has an important effect on the oxidation potential of the ILs even if they are all based on the same anion.F or example, ac omparison of these data between [ ]. As the reduction limiting potentials do not seem to be affected by the degree of functionalization,c ontrary to theo xidation limiting potentials,i t can be concluded that the degree of functionalization of the cation does not impact the reduction of the ILs (which is quite surprising) while it does influence the environment( interaction) of the [NTf 2 ] À anion and thusi ts oxidative stability.T he impact of the degree of functionalization on the oxidation potentialo ft he cations was then further examined by comparing the HOMO energy of selected cationsd eterminedf rom DFT calculations as shown in Table S8 of the SI. From these data, it can be observedt hat the HOMO energy is alwaysl ocalized on the oxygen atom furthest from the sulfonium centre and seemst od ecrease by increasing the degree of functionalization. For example, HOMO energies closet oÀ11. 20 ]. On the one hand, strong cation-anion interaction may indicatet he possibility for the cation to participate to the oxidationp rocess as claimed by some authors, [48] on the other hand, al ow degree of freedom may also influence the probability of finding various anion and cation conformations in solution. Interestingly, DFT calculations of the ion pair [S 1,G1,G1 ][NTf 2 ]s uggest that the conformation cis-[NTf 2 ] À surrounding the cation is the most stable, for example (see Table S8 of the SI). However, in the case of the single anion DFT calculations, this conformation is not the most stable and led to an increase of the anion HOMO energy close to + 0.22 eV versus the trans-conformer.I nb oth cases, the degree of functionalization of the cation seems to impact,s trongly,o nt he oxidation potentialo ft he ILs. This may be driven by achange on the cation-anion interactions inducingaradicalc hange of the anion conformer ratio (cis vs. trans)i nvolving ad ecrease of the oxidation potential.T his hypothesis could also explain why the electrochemical windows of the selected ILs, which range from 3.3 to 3.9V ,a re lower than those reportedf or the [S 1,1,x ][NTf 2 ]s eries with x = 2t o5 (i.e. from 3.8to4 .2 V). [28]

Conclusions
An ovel family of cation-functionalized sulfonium ILs has been synthesized and characterized. Different synthesis routesh ave  been investigated to be able to make targeted ILs in large scale within ag ood yield (63-87 %y ield). Firstly,t hioethers were made by following the conventional Williamsonp rocedure under aqueous conditions. Then, dialkyl-ether-sulfoniumbased ILs can be easily synthesized with agood yield by the alkylation of the thioethers using either iodomethane or diethylsulfate. Similarly,b ased on our investigations, alkyl-diether-sulfonium-based ILs can only be obtained with ag ood yield by the alkylation of the thiodiethersu sing iodomethane. These thiodiethers were obtained by reactings odium sulfide with 2equivalents of the corresponding bromogylme. In each case, anion exchange was achieved by as imple metathesis reaction with Li[NTf 2 ]s alt. These ILs exhibit good thermals tability,l ow meltingp oints, and good transport properties. As expected, the structure of the cation affects strongly all investigated properties. Both the density and conductivity decrease by increasing the degree of functionalization, whilem ore complexr elationships were observed in the case of the viscosity.F urthermore, according to the Walden rule, all investigated ILs can be classified as "good" ionic liquidsw ith ionicity ranging from 30 %t o5 0%.T he ionicity seems to be strongly affected by the degree of asymmetry and of functionalization on the cation structure. ]. Interestingly, as imilart rend was observedf or the electrochemical windows of these ILs with as trong impact of the degree of functionalization of the cation on this property.T his impact may be related to strongi nteractions between functionalized cation and the [NTf 2 ] À anion leadingt oaslightd ecrease of the oxidation limiting potential of these ILs. However,t hese functionalized sulfonium ILs show quite large electrochemical windows higher than 3.3 V. In other words, the resultso btainedd uring this work indicate that these ILs could be appliedi ne lectrochemicale nergy storaged evices. In particular,t he application of this class of ether-functionalized sulfonium [NTf 2 ] À -based ILs in electrochemical double layer capacitors is currently being investigated.

MethodsUsed During the ILs Characterization
Prior to any physiochemical, thermophysical and electrochemical measurement, all the ILs were dried under high vacuum (2 10 À3 mbar) for 48 ha t3 43.15 K. The water content in ILs was ana-lyzed by means of ac oulometric Karl-Fischer titration using an 899 Coulometer (Metrohm) with an accuracy better than 10 ppm.
Density measurements were performed using aD M40 (Mettler To ledo) oscillating tube densitometer in the range of 293.15-363.15 K( AE 0.01 K) within an accuracy close to AE 10 À4 gcm À3 .P rior to any measurements, the instrument was cleaned with acetone and dried with dehumidified air.
The viscosity of the ILs was measured using aB ohlin Gemini Rotonetic Drive 2c one and plate rheometer from 267 to 370 K( AE 0.01 K) at atmospheric pressure. The viscosity standard (ASTM Oil Standard S600 of CANNON, 1053 mPa sa t2 98.15 K) and ultra-pure water were used to calibrate the viscometer.B ased on these measurements, the accuracy of reported viscosity measurements is close to AE 1%.
Dynamic thermogravimetric analysis (TGA) of each IL was determined using aT GA Q5000 (TAI nstruments) under nitrogen flow with ah eating rate and terminal temperature set at 5Kmin À1 and 773.15 K, respectively.T he decomposition temperature onset, T d ,i s taken when the samples had lost 5% of their initial masses. Reported thermal properties are given with accuracy close to AE 1K.
Thermal phase transitions of each IL were recorded using differential scanning calorimetry (DSC) traces on aD SC Q2000 (TAI nstruments). Hermetically sealed aluminum pans containing the respective IL sample were prepared inside the Ar-filled glove box for DSC analysis. As ample of average weight of % 5mgw as hermetically sealed in an aluminum pan, and then heated and cooled at ar ate of 5Kmin À1 from 183.15 to 273.15 Ku nder af low of nitrogen. The glass transition temperature (T g ,o nset of the heat capacity change), crystallization temperature (T c ,o nset of the exothermic peak), and melting point (T m ,o nset of the endothermic peak) were recorded on the first or second heating scans with accuracy close to AE 0.25 K.
Conductivity measurements were performed using as ensION + EC71 benchtop meter with a3 -pole platinum sensION + 5070 conductivity probe with an in-built Pt1000 temperature probe (Hach Lange). The conductivity probe was calibrated using aqueous KCl standard conductivity solutions (147 mScm À1 ,1 413 mScm À1 ,a nd 12.88 mS cm À1 at 298.15 K). The immersion and sealing of the conductivity probe in the liquid sample was carried out in an Ar-filled glovebox. The conductivity probe (disconnected from the meter) was immersed in the liquid sample inside ag lass sample tube with as mall magnetic stirrer.T he sample was sealed using an O-ring seal and parafilm. The conductivity of the sample was then recorded with stirring as af unction of temperature (using the temperature reading built into the conductivity probe). The temperature of the sample was varied from 263.15 to 353.15 Ku sing as mall oil bath and ah ot-plate with at hermocouple control. The temperature and conductivity of the sample was recorded when the values were stable for about1 min with accuracies close to 0.05 Ka nd 1%, respectively.
All electrochemical measurements were performed using aV ersatile multichannel potentiostat (VMP 3, Biologic S.A.) inside an Ar-filled glovebox. The measurements have been conducted at as can rate of 2mVs À1 in a3 -electrode cell using a0 .3 cm diameter glassy carbon as the working electrode, af resh lithium metal strip as the reference, and ap latinum wire as the counter electrode. The electrolyte consisted of ca. 1mLo fp ure IL in which as mall quantity of Li[NTf 2 ]( approx. 20 mg) was dissolved to introduce aL i + /Li reference system. HOMO and LUMO energies of each species have been determined by using the Turbomole 7.0 program package. [49] Prior to visualization of these orbitals using TmoleX (version 4.1.1), the structure of each ion involved was optimized, with ac onvergence criterion of 10 À8 Hartree in the gas phase, by using DFT calculations combining the Resolution of Identity (RI) approximation [50] within the Turbomole 7.0 program package utilizing the B3LYP functional with the def-TZVP basis set. [51] Each resultant optimized structure was then used as an input for the generation of the conformers of each species using the COSMOconfX program (version 4.0). The orbitals of each conformer were then determined using single point energy calculations (DFT/B3LYP/def-TZVP + RI approximation) within Turbomole.

General Synthesis of the Sulfonium-based ILs
Synthesis of the Thioethers 1a and 1b: To af lask containing a2 1wt% aqueous solution of sodium thiomethoxide (182.942 g, 0.548 mol, 1.1 equiv.) was slowly added, over 2h,t he corresponding bromoether (0.498 mol, 1equiv.) The reaction mixture was covered by an aluminum foil and immersed in an ice bath. After complete addition of the bromoether,t he mixture was stirred and allowed to return slowly to room temperature for approximately 15 h. To this was added 50 mL of diethyl ether to dissolve and extract the methylthioether intermediates, 1a and 1b.T he aqueous phase was further extracted with portions of diethyl ether (2 25 mL) and all of the organic extracts were gathered in ar ound bottom flask covered by aluminum foil and immersed into an ice bath. The diethyl ether was removed under vacuum to leave the crude thioethers which were characterized by 1 H-NMR and found to be of ah igh enough purity to be used for the subsequent synthesis steps. : Iodomethane (107.215 g, 0.748 mol, 1.5 equiv.) was added slowly,d ropwise over ap eriod of 2h,t ot he thioether (0.49 mol, 1.0 equiv.) in ultrapure water (100 mL), and the biphasic mixture was stirred at room temperature for 15 h. After this, the resulting aqueous solution was washed successively with 100 mL fractions of ethyl acetate (three times) and diethyl ether (three times) to remove excess iodomethane and traces of thioether.T he resulting sulfonium iodide solution was then stirred with as olution of Li[NTf 2 ]( 153.805 g, 0.525 mol, 1.05 equiv.) in 150 mL of ultrapure water for 15 ha tr oom temperature in af lask covered by aluminum foil. Thereafter,1 00 mL of dichloromethane was added to dissolve the IL, which was then washed 15 times with 20 mL fractions of ultrapure water.T he absence of iodide traces was verified with as ilver nitrate test. Dichloromethane was then removed on ar otary evaporator and the IL was further dried under vacuum (10 À3 mbar) at 353.15 Kd uring three days.
Synthesis of Dialkylether sulfonium-based IL [S 1,2,G1 ][NTf 2 ]: Diethyl sulfate (77.09 g, 0.5 mol, 1.0 equiv.) was added slowly,d ropwise over ap eriod of 2h,t ot he thioether 1a (0.49 mol, 1.0 equiv.) in as olution of ethyl acetate (200 mL), and the mixture was stirred at room temperature for 15 h. After this, av iscous lower layer consisting of the sulfonium ethylsulfate, [S 1,2,G1 ][EtSO 4 ], was removed. The resulting ethyl sulfate IL was dissolved in 100 mL of ultrapure water,b efore being washed successively with 100 mL fractions of ethyl acetate (3 times) and diethyl ether (3 times) to remove traces of thioether.T he resulting ethyl sulfate IL was subjected to metathesis with Li[NTf 2 ]i na ni dentical procedure to that described above.
Synthesis of the Thiodiethers 2a and 2b: Sodium sulfide nonahydrate (82.459 g, 0.336 mol, 1equiv.) was dissolved in 100 mL ultrapure water.T he flask was covered by an aluminum foil and immersed into an ice bath, before 98.461 go fb romoether (0.673 mol, 2equiv.) was added dropwise over ap eriod of 2h.T he mixture was then stirred and allowed to return slowly to room temperature for 48 h, yielding two phases. The organic phase was removed and the aqueous phase was extracted with 50 mL fractions of diethyl ether (three times) and the combined organic phase and extracts were gathered in ar ound-bottom flask. The diethyl ether was removed under vacuum to leave the crude thioethers 2a and 2b.T hese were deemed pure enough by 1 H-NMR to continue with the synthesis. : Iodomethane (72.420 g, 0.505 mol, 1.5 equiv.) was added slowly,d ropwise over ap eriod of 2h,t ot he respective thioether (0.49 mol, 1.0 equiv.) in ultrapure water (100 mL), and the biphasic mixture was stirred at room temperature for 15 h. After this, the resulting aqueous solution was washed successively with 100 mL fractions of ethyl acetate (three times) and diethyl ether (three times) to remove excess iodomethane and traces of thioether.T he resulting sulfonium iodide solution was then stirred with as olution of Li[NTf 2 ]( 103.534 g, 0.353 mol, 1.05 equiv.) in 150 mL of ultrapure water for 15 ha tr oom temperature in af lask covered by aluminum foil. Thereafter,1 00 mL of dichloromethane was added to dissolve the IL, which is then washed 15 times with 20 mL fractions of ultrapure water.T he absence of iodide traces was verified with as ilver nitrate test. Dichloromethane was then removed on arotary evaporator.
Chemical Characterization of the Sulfonium-based ILs 1 H-and 13 C-NMR spectra were recorded at 293.15 Ko naBruker AvanceD PX spectrometer at 300 and 75 MHz, respectively,a nd are reported as Figures S1 to S5 of the SI. To avoid further contamination with moisture from the atmosphere, the ILs were stored in ag lovebox under an Ar atmosphere with am oisture content below 3ppm. Microanalysis and lithium content were performed by Analytical Services at Queen's University,B elfast. Ethylmethyl(2-methoxyethyl)sulfonium bis{(trifluoro-methyl)sul-fonyl}imide [S 1,2,G1 ][NTf 2 ]: Following the synthesis procedure, the title compound was obtained as ac olorless liquid, 87 %y ield.