Synthesis and physical properties of tris(dialkylamino)cyclopropenium bistriflamide ionic liquids

Abstract

The synthesis and properties of 23 tris(dialkylamino)cyclopropenium (TDAC) cations with the bistriflamide anion, NTf₂⁻, are described. D_3h- and C_3h-symmetric cations ([C₃(NR₂)₃]NTf₂ (R = Me, Et, Pr, Bu, Pent, Hex, Dec) and [C₃(NRMe)₃]NTf₂ (R = Et, Bu, St), respectively) were synthesised by reaction of C₃Cl₅H with the corresponding amine. Reaction of alkoxydiaminocyclopropenium salts ([C₃(NMe₂)₂(OMe)]⁺ and [C₃(NEt₂)₂(OMe)]⁺) with amines led to two series of C_2v-symmetric salts ([C₃(NMe₂)₂(NR₂)]NTf₂ (R = Et, Pr, Bu, Hex) and [C₃(NEt₂)₂(NR₂)]NTf₂ (R = Me, Bu, Hex), respectively) and two series of C_s-symmetric salts ([C₃(NMe₂)₂(NRMe)]NTf₂ (R = Et, Pr, Bu, Hex) and [C₃(NEt₂)₂(NRMe)]NTf₂ (R = Bu, Hex), respectively). In addition to characterisation by NMR, mass spectrometry and microanalysis, the salts were characterised by DSC, TGA, density, viscosity, conductivity and miscibility/solubility studies. Along with molecular weight, symmetry plays a significant role in determining melting points and viscosity, whereas density was found to depend only on molecular weight. Methyl groups were found to significantly decrease thermal stability, while increasing the size of the other alkyl groups was found to increase stability; this increase in stability is contrary to observations with other classes of ionic liquids and indicates an associative decomposition mechanism. Walden plots indicated that these are “good ionic liquids” but that significant ion-pairing occurs when at least two alkyl chains of size C₆ or larger are present. Diffusion coefficients of [C₃(NBu₂)₃]NTf₂ revealed a relatively small loss of conductivity due to ion correlations. The chemical stability of [C₃(NEt₂)₃]NTf₂ to various reagents (acid, base, redox) was investigated at 25 °C and 60 °C. Cyclic voltammetry indicated a relatively small electrochemical window of 3.6 V (due to a relatively low oxidation potential of 1.2 V). The X-ray structures of [C₃(NMe₂)₃]NTf₂ and [C₃(NPr₂)₃]NTf₂ are reported.

---

1. Introduction

Ionic liquids (ILs) have undergone a rapid development over the past 15 years, particularly as many of these materials are now available and some are used on industrial scales.¹ The great attraction of these liquids arises from their general properties of almost zero vapour pressure, low flammability, ease of tunability, excellent solubilising properties and potential for efficient recycling. For these reasons, they are frequently considered as green alternatives to classical organic solvents.² Consequently, there are now a large number of applications, as detailed in several books on the subject.³ Currently, there are five major cation-based classes of ILs (pyrrolidinium, imidazolium, pyridinium, phosphonium and ammonium) with the guanidinium^4,5 salts also being under recent investigation; the imidazolium, pyridinium and guanidinium families are notable in that the positive charge is delocalised and this reduces cation–anion interactions. Cyclopropenium salts, [C₃R₃]X, although known since the synthesis of [C₃Ph₃]⁺ in 1957,⁶ have not been investigated as ILs. The main reason for this is likely to be an expected lack of stability due to the strained three-membered ring. However, this instability is off-set in the case of the triaminocyclopropenium (TAC) salts (R = NR′₂) due to the additional stability imparted by π donation from the three amino groups to the C₃ ring. TAC salts have been known for more than 40 years,⁷ and have been recently reviewed by Komatsu and Kitagawa as well as Bandar and Lambert.^8,9 However, there have been essentially no investigations of their properties as ILs: Gompper and Schönafinger¹⁰ reported the isolation of [C₃(NMe₂)(NⁱPr₂)₂]ClO₄ and [C₃(NⁱPr₂)₃]ClO₄ as red oils (although we isolated the former as a light brown solid¹¹ and the chloride hydrate of the latter as a colourless solid¹²) while Wilcox and Breslow reported a melting point of 37.5–38.5 °C for [C₃(NEt₂)₃]ClO₄.¹³ Some salts have reported melting points greater than 100 °C: Yoshida and Tawara reported melting points of 146 °C (dec) and 270 °C (dec) for the piperidine and morpholine perchlorate analogues, respectively,⁷ while some mono- and di-amino-substituted cyclopropenium systems have also been found to have melting points greater than 100 °C.¹⁴

Compared to ammonium, phosphonium, and guanidinium cations, TAC cations have a greater delocalisation of the positive charge. Compared to the imidazolium, pyridinium and triazonium cations, they have reduced hydrogen bond donor capabilities due to a lack of aromatic C–H groups. Furthermore, it has been shown that these cations have a high-lying non-bonding HOMO that gives particularly weak cation–anion interactions.¹⁵ A number of reports indicate that the TAC cation consequently has unusual properties as a result of these weak ionic interactions: Weiss and co-workers have prepared iodide–iodoacetylene and iodide–iodoarene adducts with “isolated” anions.¹⁶ Similarly, we were able to isolate a discrete dichloride hexahydrate cube, [Cl₂(H₂O)₆]²⁻, in which the solid state structure is essentially the same as the calculated gas-phase structure.¹²

This paper reports on the IL properties of tris(dialkylamino)cyclopropenium (TDAC) cations using the hydrophobic bistriflamide ([N(SO₂CF₃)₂]⁻, NTf₂⁻) counterion. Preliminary aspects of this work have been communicated.^17,18

2. Results and discussion

2.1 Synthesis

TDAC salts were first prepared by reaction of tetrachlorocyclopropene with secondary amines,⁷ and more recently by reaction with pentachlorocyclopropane.^17–19 These routes (Scheme 1) provide the D_3h- and C_3h-symmetric cations [C₃(NR₂)₃]⁺ (1) and [C₃(NRR′)₃]⁺ (2), respectively, as the chloride salts. In a significant improvement for the synthesis of 1a, we found that this salt can be prepared using 40% aqueous NMe₂H rather than the highly-volatile pure amine. Treatment of the chloride salts with aqueous LiNTf₂ readily provides the corresponding bistriflamide salts (3 and 4, respectively) which can be extracted into an organic solvent such as chloroform or dichloromethane. Earlier, we communicated the syntheses and some properties of [C₃(NR₂)₃]NTf₂ salts for R = Me, Et, Pr and Bu (3a–d) as well as the salt [C₃(NBuMe)₃]NTf₂ (4a).¹ Here we additionally include the related syntheses for R = Pent, Hex and Dec (3e–g) as well as [C₃(NEtMe)₃]NTf₂ (4a) and [C₃(NStMe)₃]NTf₂ (4c, St = C₁₈H₃₇). In some cases, the intermediate chloride salts were isolated and the properties of these materials will be described elsewhere.

Scheme 1 Synthesis of D_3h- and C_3h-symmetric TDAC salts.

It is notable, in the case of the small amines HNMe₂ and HNEtMe, that significant amounts of the corresponding ring-opened allyldiamidinium cations [HC₃(NMe₂)₄]⁺ (5a) and [HC₃(NEtMe)₄]⁺ (5b), respectively, are formed (Scheme 2). These can be separated from the TDAC salts by addition of acid to convert the allyldiamidinium cations to the corresponding diamidinium dications [H₂C₃(NMe₂)₄]²⁺ (6a) and [H₂C₃(NEtMe)₄]²⁺ (6b), respectively, which are much more water soluble and are, therefore, not extracted into the organic phase. [HC₃(NMe₂)₄]⁺ and [HC₃(N^tBuH)₄]⁺ have been reported previously^19,20 and we will be describing the bistriflamide salts of these interesting cations in due course.

Scheme 2 Formation of allyldiamidinium and diamidinium cations.

If the secondary amine is bulky, such as HNiPr₂ or HN(C₆H₁₁)₂, then its reaction with C₃Cl₅H or C₃Cl₄ gives the corresponding diaminochlorocyclopropenium cation [C₃(NR₂)₂Cl]⁺ which can then be treated with a smaller secondary amine to provide a limited range of cations with C_2v and C_s symmetry: [C₃(NR₂)₂(NR′₂)]⁺ and [C₃(NR₂)₂(NR′R′′)]⁺, respectively, in which R is bulky and NR′R′′ is reasonably small.^10,21 Due to the limited versatility of this route, we developed a route via reaction of secondary amines with the alkoxydiaminocyclopropenium cations [C₃(NMe₂)₂(OMe)]⁺ (7) and [C₃(NEt₂)₂(OMe)]⁺ (8) (Scheme 3). These are readily prepared in two steps: hydrolysis in hot aqueous base of the TDAC cation gives the diaminocyclopropenone which is then alkylated with dimethylsulfate to provide 7 or 8 as the methylsulfate salt. Reaction with secondary amines to generate the TDAC cation generally occurs quite readily, in a few hours or less, at ambient temperature. As with the chloride salts above, addition of aqueous LiNTf₂ followed by extraction with an organic solvent allows one to isolate the bistriflamide salts. Scheme 3 illustrates the four series of ILs that were prepared in this way: two C_2v-symmetric series, 9 and 11, via 7 and 8, respectively; and two C_s-symmetric series, 10 and 12, similarly via 7 and 8, respectively. When looking at trends within these four series, note that some higher-symmetry species will also belong to some these series, i.e., 1a can be included in series 9 and 10; 1b in series 11; and 11a in series 12. Similarly, 1a can be considered part of the C_3h 4 series of cations.

Scheme 3 Synthesis of C_2v- and C_s-symmetric TDAC salts.

In some cases, it proved to be more convenient, or significantly less expensive, to prepare class 10 compounds via the protic TAC IL [C₃(NMe₂)₂(NRH)]NTf₂ (13) (Scheme 4). Salts 13 were prepared by treatment of 7 with the appropriate primary amine. Deprotonation with n-BuLi gives the cyclopropenimine 14 which is readily alkylated, by reagents such as dimethylsulfate, to give salts 10 after anion exchange. Here, we provide details for the synthesis of 10a and 10b via 13a and 13b, respectively. In principle, this route can be used to generate a large range of TDAC ILs related to 9–12. The properties of 13 and other related protic TAC ILs will be discussed elsewhere.

Scheme 4 Synthesis of C_s-symmetric TDAC salts via protic TAC salts.

All new compounds were characterised by ¹H- and ¹³C{¹H}-NMR spectroscopy as well as ES-MS and microanalysis. Chloride and water contents were determined for ILs prior to measurement of their physical properties; namely, DSC, TGA, viscosity, conductivity and density.

With respect to the NMR spectra, it should firstly be borne in mind that rotation about the exocyclic C–N bonds is fast on the NMR timescale; thus, a C_s-symmetric cation such as 10a exhibits C_2v symmetry on the NMR timescale, and there is only one ¹H and ¹³C-NMR signal for the NMe₂ groups. ¹H- and ¹³C-NMR ranges are tabulated in the ESI.† These are much as expected. Perhaps most noticeable is that the ¹³C chemical shifts for the propyl C_β and C_ω atoms are ca. 10 and 3 ppm lower, respectively, than other alkyl groups (21–22 vs. 29–32 ppm and 10.5–11 vs. 13–15 ppm, respectively).

2.2 DSC data

DSC data were collected at 10 °C min⁻¹ and the results are given in Table 1. Not surprisingly, the only salt that is not an IL by definition is that with the smallest and most symmetric, D_3h, cation, 1a, with a mp of 105 °C. A large number of factors are known to influence melting points: the various intermolecular forces (Coulombic, van der Waals, hydrogen-bonding etc.), conformational flexibility and the shape or symmetry of the species. Generally, it is found that ILs have high melting points for small ions in which Coulombic attractions dominate. They also have high melting points for large ions in which van der Waals interactions dominate. Reducing the symmetry or increasing the conformational flexibility of the side chains tends to reduce melting points since they reduce the factors that favour efficient packing in the solid state. Due to the large variety of symmetry classes presented here, it is most instructive to consider each class in turn.

Table 1 DSC (10 °C min⁻¹) and TGA data for bistriflamide salts

Salt	Tg/°C	TS–S/°C	Tm/°C	Td at 1 °C min^−1/°C	Td at 10 °C min^−1/°C
[C3(NMe2)3]NTf2 (3a)	—	45, 65, 83	105	309	339
[C3(NEt2)3]NTf2 (3b)	−86	−34, 1	18	349	393
[C3(NPr2)3]NTf2 (3c)	−72	−2	34	364	409
[C3(NBu2)3]NTf2 (3d)	−75	—	7	351	403
[C3(NPe2)3]NTf2 (3e)	−73	—	4	343	395
[C3(NHex2)3]NTf2 (3f)	−71	—	3	346	406
[C3(NDec2)3]NTf2 (3g)	—	—	8	349	401
[C3(NEtMe)3]NTf2 (4a)	—	2	7	275	366
[C3(NBuMe)3]NTf2 (4b)	−81	—	—	349	398
[C3(NStMe)3]NTf2 (4c)	—	38, 48	65	338	384
[C3(NMe2)2(NEt2)]NTf2 (9a)	—	27	44	232	334
[C3(NMe2)2(NPr2)] NTf2 (9b)	—	—	52	243	317
[C3(NMe2)2(NBu2)]NTf2 (9c)	−42	—	36	250	314
[C3(NMe2)2(NHex2)]NTf2 (9d)	−48	—	21	295	348
[C3(NMe2)2(NEtMe)]NTf2 (10a)	—	6, 32	63	211	315
[C3(NMe2)2(NPrMe)]NTf2 (10b)	—	—	20	247	291
[C3(NMe2)2(NBuMe)]NTf2 (10c)	−83	—	6	274	321
[C3(NMe2)2(NHexMe)]NTf2 (10d)	—	—	−14	264	295
[C3(NEt2)2(NMe2)]NTf2 (11a)	—	−38, −31	17	340	379
[C3(NEt2)2(NBu2)]NTf2 (11b)	−86	—	−4	355	403
[C3(NEt2)2(NHex2)]NTf2 (11c)	−81	—	—	356	396
[C3(NEt2)2(NBuMe)]NTf2 (12a)	−89	—	—	341	384
[C3(NEt2)2(NHexMe)]NTf2 (12b)	−86	—	—	351	398

---

Fig. 1 shows a plot of mp versus cation MW and carbon number for the D_3h class of cations. This shows a rapid drop in mp from 1a as both size and conformational flexibility rapidly increase. Remarkably, the hexapropyl salt 3c has a higher mp than the hexaethyl salt 3b. It's not clear whether this is a result of chain flexibility issues or is related to the different ¹³C-NMR chemical shifts mentioned earlier. The melting point appears to start increasing after the hexahexyl salt 3f. However, with the exceptions of 3a and 3c, the changes are very small when you consider that on going from 3b to 3f there are an additional 24 CH₂ groups while on going from 3f to 3g there are a further 24 CH₂ groups.

Fig. 1 Melting points of the D_3h-symmetry classes (salts 3). The line is indicative of the trend.

Fig. 2 Illustrates the melting point trends of three series: the [C₃(NMe₂)₂(NR₂)]NTf₂ series which has C_2v symmetry, other than 1a which is D_3h; the [C₃(NEt₂)₂(NR₂)]NTf₂ series which has C_2v symmetry, other than 1b which is D_3h; and the [C₃(NMe₂)₂(NRMe)]NTf₂ series which has C_s symmetry, other than 1a which is D_3h. The D_3h-symmetric salts are indicated on the figure. The two C_2v series show a similar rate of decrease in mp with increasing cation size, but with the bis(diethylamino) (11) salts having lower melting points than the corresponding isomer of the bis(dimethylamino) (9) salt. This may be attributed to the 9 series having up to six conformationally-flexible alkyl groups whereas the 11 series has at the most two conformationally-flexible alkyl groups. Another way of looking at this would be to say that the 11 series has a greater degree of branching from the rigid tris(dimethylamino)cyclopropenium core. Notable exceptions to the decreasing mp with MW are that 9b has a higher mp than 9a, perhaps for the same reason that 3c has a higher mp than 3b, and that 3b has a higher mp than 11a, probably due to the increase in symmetry. Interestingly, the C_s series 10 shows a rapid decrease in the mp as the length of the alkyl chain increases. We attribute this to a more rapid decrease in symmetry from the disc-like D_3h-symmetric 1a to a shape in which one chain protrudes out of the disc's edge at an angle (approximately 60°) and increasingly disrupts the packing efficiency.

Fig. 2 Melting points of the C_2v (salts 9 and 11) and C_s (salts 10) symmetry classes. The lines are only indicative of the trends.

For the C_3h series of salts 4, 4a has a surprisingly low mp of 7 °C. This can be compared to its isomer 10c which has essentially the same mp (6 °C) despite having much lower symmetry. This may be due to greater “branching” in 4a or be related to the low mps observed for 3b and 9a relative to the salts next to them in their respective series; these ILs also have multiple ethyl groups. Salt 4b is a liquid at ambient temperature, unfortunately, we were unable to observe a mp. Salt 4c contains three long C₁₈ chains and consequently has a high mp of 65 °C. This salt can be compared to 3g with a mp of only 8 °C despite a slightly larger MW (884 versus 924 g mol⁻¹) and higher D_3h symmetry. In this case, the higher mp of 4c can be attributed to fewer branching.

As might be expected for cations with four ethyl groups and C_s symmetry, salts 12a and 12b are liquids at ambient temperature. However, we were unable to obtain mps for these ILs.

2.3 Thermal decomposition

Thermal decomposition data were collected at both 1 °C min⁻¹ and 10 °C min⁻¹ and are given in Table 1. The major factor in determining the thermal decomposition onset temperature, T_d, is the number of methyl groups, and especially the number of dimethylamino groups. When none of the six alkyl groups are methyl groups (3b–g, 11b,c), T_d is relatively invariant and there is no obvious trend with the size of the alkyl groups: at 1 °C min⁻¹, T_d ranges 343–364 °C, while at 10 °C min⁻¹, it ranges 393–409 °C. Fig. 3 illustrates the trends for the 1 °C data only and further discussion focuses on this data. The most stable IL is the hexapropyl salt 3c. The introduction of one methyl group, 12a and 12b, lowers T_d slightly to 341 and 351 °C, respectively. A second methyl group, in 11a, again lowers T_d slightly (340 °C). As the number of methyl groups increases to three, 4a–c, a dependence on the length of the non-methyl groups becomes apparent: the short ethyl groups in 4a gives very low T_d values, 275 °C, while the longer C₄ and C₁₈ chains in 4b and 4c, respectively, appear to provide enough steric protection to give relatively high T_d values, 349 °C and 338 °C, respectively.

Fig. 3 T_d values of the tetramethyl (9a–d) and pentamethyl (10a–d) TDAC bistriflamide salts at 1 °C min⁻¹.

The presence of four or five methyl groups, 9a–d and 10a–d, respectively, not only significantly lowers T_d, but also dramatically increases the dependency on the alkyl chain length, with longer chains leading to greater stabilities. There is approximately a 50–60 °C increase in stability from 9a to 9d and 10a to 10d. This trend is contrary to what is observed with imidazolium chloride and bistriflamide salts in which a small decrease (ca. 20 °C) in stability is found on going from [C₁mim]NTf₂ to [C₄mim]NTf₂, with even longer chains having no noticeable further effect.^22,23 This reduction in stability with increasing chain length is also seen in piperidinium bistriflamide salts and has been attributed to increased stability of carbocation and carbon radicals when the alkyl chain length increases.²⁴ IL thermal stabilities have recently been reviewed by Stevens and coworkers.²⁵ Given that longer alkyl chains afford more steric protection, an S_N2 type of mechanism is suggested for the decomposition of TDAC bistriflamide salts, possibly a reverse Menshutkin-type of reaction to generate a cyclopropenimine intermediate.

Somewhat curiously, the T_d value for the hexamethyl salt 3a of 308 °C is higher than is found for all of the tetramethyl and pentamethyl salts. We have no good explanation for this observation.

These T_ds are lower than most imidazolium and pyrrolidinium bistriflamide salts, but higher than pyridinium and phosphonium salts. For example, the T_d at 10 °C min⁻¹ for [bmim]NTf₂ is 422 °C. Perhaps a better comparison is with the 2-Me isomer 1-propyl-2,3-dimethylimidazoium which has a T_d of 462 °C, well above the TDAC bistriflamide salts.²⁶ N,N-butylmethylpyrrolidinium bistriflamide has a T_d of 435 °C.²⁷

For consideration of practical applications, isothermal decomposition profiles are desired. Workers have devised a number of methods to rapidly acquire this information. Measurement of the first-order rate constant for isothermal weight loss allows the determination of t_0.99 (the time for 1% of the sample to decompose). A plot of t_0.99 versus T generates an exponential decay curve from which t_0.99 can be calculated for any temperature. This is shown in Fig. 4 for 3d in both nitrogen and air atmospheres with t_0.99(N₂) = (4.5 × 10¹¹)exp(−0.044T) and t_0.99(air) = (7.1 × 10⁶)exp(−0.022T), respectively. This can be compared to butylmethylpyrrolidinium bistriflamide for which t_0.99(N₂) = (5.3 × 10¹⁶)exp(−0.059T).²⁷ The decomposition rates for TDAC salts appear to be less sensitive to temperature and this suggests a smaller contribution from entropy, perhaps indicating again an S_N2 mechanism. It is curious that the sensitivity of t_0.99 to temperature is less when in air (exponential factor of 0.022 compared to 0.044).

Fig. 4 t_0.99 versus isothermal temperature for 3d under nitrogen and air atmospheres.

2.4 Density

Density data were determined from 20–90 °C where possible; the results for the 20 °C and 50 °C data are given in Table 2 and molar volumes are provided in the ESI.† Unlike melting points, viscosities and conductivities, vide infra, densities for TDAC bistriflamide salts are independent of cation shape and depend only on the cation size (or mass). Fig. 5 shows a plot of densities at 20 °C versus the molecular weight (MW) of the cations for the bistriflamide salts as well as for simple alkanes. Although it is frequently stated that densities decrease linearly with the number of CH₂ groups, clearly this is only true for short ranges of alkyl chain lengths. At long chain lengths, densities must approach that of very low density polyethylene (VLDPE), 0.88–0.915 g mL⁻¹, as the cation core and anions are increasingly diluted. The equations on Fig. 5 describe the fitted curves. These are made up of a volume for the permethyl salt 3a of 528 ų, a factor of 27.7 ų per CH₂ group added (n) to this core and a volume of 27.3 ų per CH₂ group for alkanes (a volume of 52 ų is used for the CH₄ core). These CH₂ volumes are in agreement with that obtained by Ye and Shreeve for ILs of 28 ų.²⁸ They also estimated a volume of 248 ų for the bistriflamide anion in ILs which, therefore, gives us a volume for the [C₃(NMe₂)₃]⁺ cation in ILs of 280 ų.

Table 2 MW, cation hydrodynamic radius, and selected density, viscosity and conductivity data for TDAC bistriflamide ILs

	MW (g mol^−1)	r^+ at^a 20 °C (Å)	Density		Viscosity		Conductivity
			20 °C	50 °C	20 °C	50 °C	20 °C	50 °C
a Based on density data, see supplementary material for this calculation.
[C3(NEt2)3]NTf2 (3b)	532.61	4.74	1.277	1.251	94.7	27.5	1.387	3.681
[C3(NPr2)3]NTf2 (3c)	616.72	5.27	1.196	1.171	219.7	50.1	0.498	1.904
[C3(NBu2)3]NTf2 (3d)	700.89	5.71	1.134	1.111	230.4	55.2	0.428	1.535
[C3(NPe2)3]NTf2 (3e)	785.05	6.09	1.086	1.064	268.7	62.1	0.245	1.011
[C3(NHex2)3]NTf2 (3f)	869.21	6.42	1.059	1.037	273.0	65.3	0.141	0.508
[C3(NDec2)3]NTf2 (3g)	1205.86	7.51	0.991	0.970	407.6	84.6	0.027	0.131
[C3(NEtMe)3]NTf2 (4a)	490.48	4.43	1.333	1.306	72.5	22.0	2.42	7.05
[C3(NBuMe)3]NTf2 (4b)	574.64	5.02	1.224	1.199	101	26.8	0.939	2.825
[C3(NMe2)2(NEt2)]NTf2 (9a)	476.45	4.31	—	—	—	25.1	—	5.15
[C3(NMe2)2(NPr2)] NTf2 (9b)	504.51	4.53	—	1.287	—	28.9	—	5.09
[C3(NMe2)2(NBu2)]NTf2 (9c)	532.56	4.74	1.272	1.244	117.5	28.6	1.23	4.50
[C3(NMe2)2(NHex2)]NTf2 (9d)	588.67	5.10	1.209	1.184	—	42.3	0.29	1.41
[C3(NMe2)2(NPrMe)]NTf2 (10b)	476.45	4.31	1.356	1.328	72.5	20.0	2.71	8.30
[C3(NMe2)2(NBuMe)]NTf2 (10c)	490.48	4.43	1.331	1.303	76.1	21.9	2.36	7.23
[C3(NMe2)2(NHexMe)]NTf2 (10d)	518.53	4.64	1.292	1.266	94.0	24.7	1.31	4.83
[C3(NEt2)2(NMe2)]NTf2 (11a)	504.51	4.53	1.308	1.282	83.6	24.7	1.572	4.106
[C3(NEt2)2(NBu2)]NTf2 (11b)	588.67	5.10	1.216	1.191	125.7	32.7	0.850	2.532
[C3(NEt2)2(NHex2)]NTf2 (11c)	644.78	5.42	1.171	1.147	182.1	44.2	—	—
[C3(NEt2)2(NBuMe)]NTf2 (12a)	546.59	4.84	1.260	1.235	106.2	29.7	1.225	3.346
[C3(NEt2)2(NHexMe)]NTf2 (12b)	574.64	5.02	1.228	1.203	101.8	28.3	1.176	3.690

---

Fig. 5 Density at 20 °C versus the MW for TDAC bistriflamide salts and alkanes.

At 20 °C and n = ∞, a density of 0.842 g mL⁻¹ is calculated for TDAC bistriflamide salts and 0.853 g mL⁻¹ for alkanes, i.e. effectively “liquid polyethylene”. These are in quite good agreement and, interestingly, are only a little lower than is found in VLDPE, which presumably has some degree of crystallinity.

The densities of ILs have a linear dependency on temperature and can be well-fitted by the equation ρ = a − bT. Parameter a represents a theoretical density at 0 K. Pleasingly, a plot of a versus cation MW can also be fit (3e, 9b and 10c are possibly outliers, for reasons that will be apparent later) to a similar equation as the density data obtained at 20 °C (ESI, Fig. 1S†). In this case, we calculate a volume for each CH₂ group at 0 K of 23 ų while the volume of 3a at 0 K was found to be 440 ų. This gives a “free volume” of 88 ų for 3a at 20 °C and 4.7 ų per CH₂. Parameter a for “liquid polyethylene” (density at 0 K and n = ∞) is 1.012 g mL⁻¹.

Density parameter b represents the temperature dependency of the density. This parameter is rarely commented on, however, we find that this parameter is also well-fitted (again with the exceptions of 3e, 9b and 10c) by an equation of the type used for the density at 20 °C and parameter a (ESI, Fig. 2S†). This illustrates a decrease in temperature dependence with MW. Parameter b for “liquid polyethylene” is 5.82 × 10⁻⁴ g mL⁻¹ K⁻¹. The combination of the equations for a and b allow us to derive a temperature-dependent equation for all TDAC bistriflamide salts (eqn (1)).

The thermal expansion coefficient α_p can be obtained from the slope of a plot of ln(ρ) versus T, i.e. −[∂ln(ρ)/∂T]_P = −c.²⁹ Although α_p can vary with temperature,²⁷ and ln(ρ) versus T can be fit with quadratic or cubic functions, we have not done so due to the limited temperature range. The values of α_p vary from 0.653 × 10⁻³ K⁻¹ for 9b to 0.709 × 10⁻³ K⁻¹ for 10c, however, most values are in the range 0.680–0.705 × 10⁻³ K⁻¹. These are similar to, but on the high end of, values found for other ILs: phosphonium ILs (0.575–0.692 × 10⁻³ K⁻¹),^30–33 imidazolium ILs (0.579–0.705),^27,34 and pyridinium ILs (0.530–0.543).³⁴ Our high values can probably be attributed to relatively higher MWs. A plot of α_p versus MW (Fig. 6) shows some dependency on MW, with α_p generally greater at higher MWs (as also found in imidazolium and phosphonium ILs),^30,31 although it is not as obvious as in the plots of a and b (Fig. 1S and 2S,† respectively). This plot also strongly suggests that 3e, 9b and 10c are outliers (these ILs have also been indicated in Fig. 1S and 2S† where they are not so obviously outliers).

Fig. 6 Thermal expansion coefficient versus MW. ILs 3e, 9b and 10c are indicated by red squares.

2.5 Viscosity

Viscosity data were collected from 20–90 °C where possible; the results for the 20 °C and 50 °C data are given in Table 2. As is to be expected, viscosity generally increases with MW, as shown in Fig. 7 for viscosity at 20 °C. Also shown on this plot is the 2-Me isomer of [C₄mim]NTf₂, [C₃dmim]NTf₂. The lower viscosity of [C₄mim]NTf₂ relative to [C₃dmim]NTf₂ has been attributed to a lower density and greater free volume in [C₄mim]NTf₂.³⁵ It is interesting then that the TDAC salts have viscosities that are much lower than the 2-Me imidazolium salts of similar MWs, despite having very similar densities.

Fig. 7 Viscosity at 20 °C for TDAC bistriflamide salts and [C₃dmim]NTf₂ (green square). The trendline is indicative.

When the viscosity data is colour-coded by symmetry class, as shown in Fig. 8 for the data at 50 °C, some further trends become apparent. Relative to their MWs, the D_3h ILs 3b–g and the C_2v ILs 9a–d have similar viscosities (notably, 3b and 9c appear to be low and 3c possibly high). On the other hand, the C_2v salts 11a–c and the C_s salts 10b–d have similarly lower viscosities. Interestingly, 3b can also be considered part of the 11 series of salts [C₃(NEt₂)₂NR₂]NTf₂, in which R = Et, and its viscosity falls on this line rather than the upper line for the D_3h-symmetric salts 3.

Fig. 8 Viscosity at 50 °C for TDAC bistriflamide salts. Trend lines are indicative.

Of the C_3h-symmetric salts, 4a falls on the low viscosity line whereas 4b lies even lower. The C_s-symmetric salt 12a also lies on the low viscosity line, whereas 12b lies below this line and even below 12a.

Given the lack of symmetry effects on molar and, hence, free volumes. These differences in viscosities appear to due to a combination of conformational flexibility and symmetry: flexible chains facilitate the motion of molecules past each other and so lower viscosity, whereas higher symmetry increases viscosity, perhaps due to a greater degree of short range ordering. Of the “high” viscosity ILs, the D_3h salts 3 are highly symmetric whereas the slightly less-symmetric C_2v salts 9 have only two alkyl groups with conformational flexibility. Of the “low” viscosity IL series, The C_2v salts 11 have six conformationally-flexible alkyl groups while the C_s-symmetric salts 10 have only one flexible alkyl group but lowest symmetry. It appears, therefore, that low symmetry and conformationally-flexible alkyl groups lead to low viscosity. This possibly explains why 12b has such a low viscosity for its MW: it has lowest symmetry and six flexible alkyl chains.

The viscosity data was fit to both the Arrhenius (η = A exp(E_a/RT)) and Vogel–Fulcher–Tammann (VFT, eqn (2)) equations (also, D = B/T₀); these parameters are given in the ESI.† There are not many obvious trends: E_a tends to increase with MW; the values of D, a measure of the deviation from Arrhenius behaviour, lie in the range 2.6–13.9 which is typical for “fragile” liquids.³⁶ A “fragility plot” of log(viscosity) versus T_gT⁻¹ (ESI, Fig. 4S†) similarly shows that these materials are typical of fragile ILs, although the C_2v salts 9c and 9d appear to be more fragile than the others due to higher T_g values. However, it is not obvious why this is the case.

2.6 Conductivity

Conductivity data was collected from 20–90 °C where possible; results at 20 °C and 50 °C are given in Table 2. Fig. 9 illustrates the data at 50 °C for which a strong trend of decreasing conductivity with MW is observed. Unlike viscosity, there appears to be no obvious dependence of conductivity on the symmetry class of the cation. Given the normally expected strong relationship between viscosity and conductivity, this is quite surprising. A plot of log(conductivity) versus MW gives a reasonably straight line (ESI, Fig. 5S†).

Fig. 9 Conductivity at 50 °C for TDAC bistriflamide salts.

2.7 Ionicity

Conductivity and viscosity are linked through Walden's rule: Λη = k (in which k is a temperature-dependent constant, the Walden product). A Walden plot, log(Λ) versus log(1/η), can be used to investigate the ionicity of a conducting solution or liquid, with deviations from the ideal diagonal (represented by 1 M KCl(aq)) being ascribed to the formation of ion pairs or aggregates. The Walden plot for the ILs presented in this paper is given in Fig. 10. The majority of ILs fall in a narrow band with only a small deviation from ideal. These would thus be considered as “good ILs”. Two lie further off this ideal line, namely 3f and 9d, both of which have dihexylamino groups. Furthest off the ideal line is 3g which has three didecylamino groups. Salts 3e–g and 9d are shown separately in the ESI (Fig. 6S†) for clarity. MacFarlane and co-workers have observed similar deviations in phosphonium chloride salts as the alkyl chain lengths have increased, and that was attributed to encapsulation of the chloride anion by the alkyl chains leading to ion pairs.³⁷ We suggest a similar explanation here in which two hexyl or decyl chains are able to encapsulate (or at least strongly interact via dispersion forces) the bistriflamide anion and thus form transient ion pairs. Interestingly, 9d has the greatest slope in the Walden plot (1.12 versus 0.78–1.02 for the others), as well as noticeable deviation from linearity (the slope increases from 0.96 to 1.32 as T increases over the investigated temperature range), such that the ionicity is greatly increased at elevated temperatures and it becomes similar in ionicity to the other ILs. To confirm that the anomalous Walden plot of 9d is due to changes in ionicity, comparison of viscosity and conductivity with 3c, which has a similar MW, is useful: At 303 K, 9d has a lower viscosity than 3c (110.3 versus 127.1 cP, respectively) but also a lower conductivity (0.53 versus 0.86 mS cm⁻¹, respectively). At 363 K, 9d has a higher viscosity (13.2 versus 12.6 cP, respectively) and a correspondingly slightly lower conductivity (5.52 versus 5.64 mS/cm, respectively). This confirms that ion-pairing at lower temperatures has decreased both viscosity and conductivity in 9d and suggests that the dihexyl–bistriflamide interactions get broken at higher temperatures to reduce the number of ion pairs. For 3f and 3g, on the other hand, presumably the bistriflamide anions can get passed from one diamino group to another. It is not clear why 3e, with six pentyl groups, appears to have insufficient alkyl–bistriflamide interactions for ion pairing.

Fig. 10 Walden plot for TDAC bistriflamide salts.

MacFarlane developed an adjusted Walden plot to take into account the role of the ion sizes r⁺ and r⁻.³⁷ This plot and further discussion is presented in the ESI.† The plot (ESI, Fig. 7S†) brings most of the salts to within likely uncertainties of the KCl line. Salts 3f, 3g and 9d, however, are still distinctly below the line.

Ionicity was also investigated by measurement of the diffusion coefficients of 3d at 20 °C: D⁺ was found to be 3.17 × 10⁻¹² m² s⁻¹ and D⁻ is 5.62 × 10⁻¹² m² s⁻¹. From the Nernst–Einstein equation, Λ_NE = 0.336 S cm² mol⁻¹. This compares to the measured molar conductivity, Λ_M, of 0.264 S cm² mol⁻¹. The ionicity can be defined as the ratio Λ_M/Λ_NE, which for 3d is 0.79 (this can also be viewed as a 21% reduction in molar conductivity due to ion correlations). This is remarkably high compared to other bistriflamide-based ILs ([C₄mim]NTf₂ (Λ_M/Λ_NE = 0.61), [N₄₁₁₁]NTf₂ (0.65), [C₄py]NTf₂ (0.63) and [C₄mpyr]NTf₂ (0.70)),³⁸ indicating possibly one of the highest values of ionicity ever observed in ILs at around room temperature. This result is even more surprising considering that Watanabe and co-workers showed that ionicity decreases with alkyl chain length for the imidazolium bistriflamide salts—to 0.53 for [C₈mim]NTf₂, which still has a significantly smaller MW than 3d (475.5 vs. 700.9 g mol⁻¹).^22a They attributed this decrease to increasing dispersive forces.

Despite the high ionicity of 3d, its relatively high MW means that the effective ionic concentration, C_eff (defined as molar concentration times Λ_M/Λ_NE), is still quite low at 1.28 × 10⁻³ mol cm⁻³, compared to 2.9 × 10⁻³ mol cm⁻³ for [C₂mim]NTf₂ and 1.5 × 10⁻³ mol cm⁻³ for [C₈mim]NTf₂.^38a Salt 10b, the smallest RTIL reported here (MW = 476.5 g mol⁻¹), is likely to have an effective ionic concentration at 20 °C of 2.3–2.8 × 10⁻³ mol cm⁻³. Watanabe and co-workers have described a number of correlations between C_eff and various physical properties, however the large difference in MW between 3d and the salts discussed by them makes meaningful comparisons difficult at this stage.^38a,39

The ionicity studies provide further evidence of the weak Coulombic interactions between TDAC cations and their anions. The NMR study has been limited at this stage due to limited NMR access time for these lengthy experiments, but will be pursued further in future work.

2.8 Miscibility and solubility

The miscibility and solubility properties of the salts were investigated at 25 °C. The results are given in Table 3 in order of increasing MW. All of the salts were found to be insoluble or immiscible in water but soluble or miscible in MeOH, EtOH, CH₂Cl₂ and EtOAc. The notable exceptions are the insolubility of 3a in MeOH and EtOH (due to the small and highly-symmetric cation), as well as the insolubility of 4c in MeOH, EtOH and EtOAc. Interestingly, 4c was observed to swell in water but not dissolve. Clearly, the very long C₁₈ chains have interesting effects on the properties of this salt.

Table 3 Miscibility^a and solubility properties of TDAC bistriflamide salts at 25 °C

Compound	MW	Water	MeOH/EtOH	CH2Cl2/EtOAc	Toluene	Et2O	Hexane
a I = insoluble; N = immiscible liquid; Y = soluble or miscible.b insoluble in EtOAc.
[C3(NMe2)3]NTf2 (3a)	448	I	I	Y	N	I	I
[C3(NMe2)2(NEtMe)]NTf2 (10a)	462	I	Y	Y	I	I	I
[C3(NMe2)2(NEt2)]NTf2 (9a)	476	I	Y	Y	I	N	I
[C3(NMe2)2(NPrMe)]NTf2 (10b)	476	N	Y	Y	≥50% IL	≥50% IL	N
[C3(NEtMe)3]NTf2 (4a)	490	N	Y	Y	≥40% IL	≥50% IL	N
[C3(NMe2)2(NBuMe)]NTf2 (10c)	490	N	Y	Y	≥50% IL	≥40% IL	N
[C3(NMe2)2(NPr2)] NTf2 (9b)	505	I	Y	Y	N	≥71% IL	I
[C3(NEt2)2(NMe2)]NTf2 (11a)	505	N	Y	Y	≥50% IL	≥50% IL	N
[C3(NMe2)2(NHexMe)]NTf2 (10d)	519	N	Y	Y	≥40% IL	≥25% IL	N
[C3(NMe2)2(NBu2)]NTf2 (9c)	533	N	Y	Y	≥50% IL	≥33% IL	N
[C3(NEt2)3]NTf2 (3b)	533	N	Y	Y	≥50% IL	N	N
[C3(NEt2)2(NBuMe)]NTf2 (12a)	547	N	Y	Y	≥50% IL	≥50% IL	N
[C3(NBuMe)3]NTf2 (4b)	575	N	Y	Y	≥33% IL	Y	N
[C3(NEt2)2(NHexMe)]NTf2 (12b)	575	N	Y	Y	≥50% IL	≥33% IL	≥77% IL
[C3(NMe2)2(NHex2)]NTf2 (9d)	589	N	Y	Y	≥40% IL	Y	N
[C3(NEt2)2(NBu2)]NTf2 (11b)	589	N	Y	Y	≥50% IL	≥33% IL	≥77% IL
[C3(NPr2)3]NTf2 (3c)	617	N	Y	Y	≥50% IL	≥33% IL	N
[C3(NEt2)2(NHex2)]NTf2 (11c)	645	N	Y	Y	≥33% IL	Y	≥67% IL
[C3(NBu2)3]NTf2 (3d)	701	N	Y	Y	Y	Y	≥67% IL
[C3(NPe2)3]NTf2 (3e)	785	N	Y	Y	Y	Y	≥56% IL
[C3(NHex2)3]NTf2 (3f)	869	N	Y	Y	Y	Y	≥33% IL
[C3(NStMe)3]NTf2 (4c)	1164	I	I	Y^b	Y	N	I
[C3(NDec2)3]NTf2 (3g)	1206	N	Y	Y	Y	Y	Y

---

In toluene, diethylether and hexane, miscibility/solubility was found to increase with MW. Much higher MWs are required for miscibility in hexane such that only 3g is completely miscible in all proportions. For MWs of 575 to 869 g mol⁻¹, the ILs are partially miscible in hexane with the range of miscibility increasing with increasing MW. With partially-miscible samples, miscibility is always observed above a certain proportion of IL (say, above 50% IL), never within a broad intermediate range of proportions (such as 30–70%). This is because it is possible for organic solute molecules to find space in the IL without significantly disrupting the electrostatic attractions between ions whereas it is not possible for low-polarity solvents to separate the ions and create a solution of charged solute molecules. Again, 4c was found to be very unlike 3f and 3g in that it is insoluble in hexane.

Toluene and diethylether exhibit quite similar solubility properties with partial miscibility being exhibited between approximately 476 and 645 g mol⁻¹. Symmetry effects and conformational flexibility appear to be very important in the exceptions to these trends. Greater symmetry decreases solubility/miscibility whereas increased flexibility increases solubility/miscibility. Thus, (i) C_2v-symmetric 9a is insoluble in toluene/diethylether whereas C_s-symmetric 10b, of the same MW, is partially miscible, (ii) C_2v-symmetric 9b is similarly insoluble in toluene and has reduced miscibility in diethylether compared to those of similar MW which have either reduced symmetry (10c and 10d) or three or more flexible ethyl groups (4a and 11a) compared to the two propyl groups of 9b (notably, 9b is also a solid at ambient temperature), (iii) D_3h-symmetric 3b is insoluble in diethylether whereas all those with similar MW have lower symmetry and are at least partially miscible, (iv) 4b (C_3h), 9d (C_2v and four Me groups) and 3c (D_3h) are insoluble in hexane whereas 12b (C_s), 11b and 11c (both C_2v but with no Me groups) are partially miscible in hexane due to lower symmetry or increased flexibility. Salt 4c is again an interesting exception: two immiscible liquid phases are formed in diethylether whereas it is, perhaps surprisingly, soluble in toluene.

Some solid samples were observed to form two immiscible liquid layers (3a with toluene, 9a with Et₂O, 9b with toluene and Et₂O, and 4c with Et₂O), suggesting that at least a small amount of organic solvent must have dissolved into these ILs to break up the crystalline lattice. That 3a should form an immiscible liquid under these conditions is particularly interesting given its high mp (105 °C). This suggests that the mp of this compound can be sharply decreased by the action of a second component.

Water is known to have a significant effect on the properties of ILs,⁴⁰ so we also measured water contents in some water-saturated systems: 3b (5140 ppm), 12b (3270 ppm) 3d (3500 ppm) and 3e (2340 ppm). A lot of data is available for other water-saturated bistriflamide salts.^23,40,41 Although there is significant variation in the reported values with some salts, there is a strong trend of decreasing water content with increasing MW as hydrophobicity increases, as shown in Fig. 11 (data provided in ESI†). Interestingly, water contents appear to have almost reached a lower limit of ca. 2000 ppm with the TDAC bistriflamide salts due to their high MWs.

Fig. 11 Water contents of water-saturated alkyl-substituted bistriflamide ILs at 25 °C (“ammonium” includes pyrrolidinium and piperidinium).

2.9 Chemical stability

For the use of ILs in chemical processes, their stability under a variety of conditions needs to be examined. Tests based on those of Afonso et al.⁴² were carried out on 3b with one (or two) equivalents of reagent (acid (HCl), weak base and nucleophile (NH₃), strong base (NaOH), reducing agent (NaBH₄), oxidising agent (NaIO₄) and Grignard (EtMgI)) at room temperature and at 60 °C for 24 hours. Samples were then analysed by NMR. Salt 3b was found to be stable under all of these conditions except for NaOH at 60 °C, with which it reacts to form diethylamine and the cyclopropenone. In the case of the Grignard reagent, although 3b is stable, the Grignard is not active and presumably the lack of a suitable donor atom in the solvent causes the Grignard to become unstable with respect to MgEt₂ and MgI₂. Afonso and co-workers reported that [BMIM]Cl and [NMe(n-C₈H₁₇)₃]Cl are similarly unstable with NaOH at 60 °C, whereas the guanidinium salt [C(NMe₂)(NHex₂)₂]Cl is stable in NaOH but unstable with 3 equivalents of NaBH₄ at 60 °C.⁴² Noting the decreased thermal stability of the methylated TDAC salts 3a, 9 and 10, the chemical stability of these classes will be reported in due course. Nonetheless, these results suggest that TAC salts will be suitable solvents for a variety of reaction conditions.

2.10 Electrochemical stability

The electrochemical stability of an IL is an important property for any electrochemical application. It is the range of voltage over which the IL is electrochemically inert, i.e. neither oxidized nor reduced. The upper limit is the resistance of the IL to oxidation, and is generally determined by the more electron rich species: the anion. The lower limit is the resistance of the IL to reduction, and is generally, though not always, determined by the more electron poor species: the cation. The difference between these limits is the electrochemical window.

Cyclic voltammetry was carried out on a recrystallized sample of 3b as a representative example of the structures synthesised in this work. This was dried under vacuum at 50 °C for 72 hours beforehand, due to water also having a negative effect on the size of the electrochemical window and the introduction of reduction peaks associated with the water. The final water content was 71 ppm.

The cyclic voltammogram is shown in Fig. 12, measured with a platinum working electrode and referenced to the Fc/Fc⁺ redox couple. If a cut-off limit for current density of 1 mA cm⁻² is used, then the oxidation limit of 3b is 1.2 V. The negative peak on the return sweep around 1.2 V indicates that the oxidation process is to some extent reversible. Oxidation is usually ascribed to the anion, the more electron rich species, however, TFSA anions often resist oxidation up to 2.5 V, so in this case we are seeing oxidation of the cation. This agrees well with results reported by Johnson, who observed reversible oxidation of [C₃(NMe₂)₃]⁺ in acetonitrile at 1.3 V.⁴³ This low oxidation limit further reinforces the electron rich nature of the TAC cation due to the extensive π donation from the amino substituents. The oxidation limit is lower than all the other common cation classes, Table 4, however, the potential reversibility of the process suggests possible applications of these materials in energy storage applications. We will explore this possibility in our future work.

Fig. 12 Cyclic voltammogram of 3b.

Table 4 Electrochemical potentials of 3b and selected ILs

IL	E(red) V vs. Fc/Fc^+	E(ox) V vs. Fc/Fc^+	EW V	Ref.
3b	−2.4	1.2	3.6	—
[P14,6,6,6]^+	−2.7	2.5	5.2	44
[C4mim]^+	−2.5	1.8	4.3	44
[N6,2,2,2]^+	−2.6	2.1	4.7	44
[C4py]^+	−1.4	2.4	3.8	45

---

The reduction limit of 3b was found to be −2.4 V. Reduction is usually ascribed to the cation, and this is the case here as well, as some bistriflamide ILs have been observed to have a reduction limit of less than −3.0 V. It is noted that Johnson⁴³ observed reduction of [C₃(NMe₂)₃]⁺ at less than −3.0 V, however, that was in solution. When comparisons are made to other classes of cations with bistriflamide anions, the reduction limit of 3b is higher than [C₄py]NTf₂ by 1.0 V, but similar to [P_14,6,6,6]NTf₂, [C₄mim]NTf₂ and [N_6,2,2,2]NTf₂.^44,45

The oxidation and reduction limits mean that 3b has a moderate electrochemical window of 3.6 V, largely due to the low oxidation potential of TAC cations to the dication radical. The electrochemical window is smaller than imidazolium, phosphonium and ammonium ILs, but comparable to pyridinium ILs, which have a more positive reduction limit for the pyridinium cations.

2.11 X-ray crystallography

Crystals of 3a and 3c were investigated by X-ray diffraction studies. Salt 3a packs in the P2₁/c space group with three independent cations and anions in each unit cell. Each cation is sandwiched between two bistriflamide anions which are oriented in one of two ways: (mode A) with the N atom, two O and two F atoms near the cation, or (mode B) by two O atoms and four F atoms near the cation. One independent cation interacts via two mode A, one via two mode B and the other by one of each mode. The latter combination is shown in Fig. 13. The C–C distances range 1.371(3)–1.382(4) Å with an average of 1.377 Å. The exocyclic C–N distances range 1.320(3)–1.328(3) Å with an average of 1.325 Å while the N–Me distances range 1.446(4)–1.458(3) Å with an average of 1.452 Å. At least 16 salts containing [C₃(NMe₂)₃]⁺ have been crystallographically characterised and the distances found here for the bistriflamide salt are typical.^16,46

Fig. 13 ORTEP of 3a illustrating one cation and the anions above (mode A) and below (mode B) the cation plane.

Salt 3c packs in the I₂ space group with two independent cations and anions in the unit cell. Fig. 14 shows the two cations along with the bistriflamide anions above and below the ring. In each case, fewer anion atoms are in the vicinity of the cation coplanar atoms than is found in 3a. This appears to be a combination of steric effects from protruding Pr groups as well as the presence of more anion–alkyl chain interactions. Both cations have adopted a conformation with four alkyl groups on one side and the other two on the other side. A mode A interaction occurs on the side with two alkyl groups, but with the anion displaced by the alkyl groups such that the amido atom lies over a N(CH₂)₂ group, rather than the C₃ ring as seen in 3a. The four protruding alkyl groups on the other side of the cation force a mode B type of interaction in which the anion is displaced such that only one O and two F atoms are near the coplanar C₃N₃ atoms. The ring C–C bond distances are slightly longer and range 1.381(4)–1.389(4) Å with an average of 1.385 Å. The N–Pr distances are also slightly longer (average 1.329 Å) than the corresponding N–Me distances, whereas the exocyclic C–N bond distances are quite similar (an average of 1.466 Å). The CH₂–CH₂ distances in 3c average 1.509 Å whereas the CH₂–CH₃ distances average 1.520 Å. The bistriflamide anions in both structures are found in the more stable trans configuration (C–S⋯S–C ranges 164.4(2)°–174.5(2)°).

Fig. 14 ORTEP of 3c illustrating the two cations and their anions (mode A above and mode B below the cation plane).

3. Conclusions

We have prepared a number of TDAC bistriflamide salts with a range of MWs and a variety of symmetry classes. The melting points can be nicely rationalised by consideration of MWs, symmetry and conformational flexibility. TGA studies found that dimethylamino (and ethylmethylamino) groups significantly decrease thermal stability while an increase in stability is found with increasing alkyl chain length when there are Me groups present. This indicates that steric factors, rather than electronic factors, control the rate of decomposition, and that the mechanism is S_N2. Density was found to be strongly dependent on MW and temperature, but not symmetry. Density fitting parameters a and b are strongly correlated with MW whereas the thermal expansion coefficient α_P was found to be relatively poorly correlated, although similar in magnitude to other ILs. Viscosity similarly is highly MW dependent, but subtle differences occurs whereby higher symmetry increases viscosities and increased alkyl flexibility lowers viscosities. Somewhat surprisingly, conductivity does not appear to be obviously affected by these viscosity differences and is largely MW dependent. A Walden plot revealed that most of the ILs can be classified as “good ILs”. Similarly, diffusion coefficient measurements confirmed that ion correlations are low, at least for short alkyl chains. Significant ion-pairing occurs when there are at least two alkyl chains of C₆ or longer. With only one dihexylamino group (9d), the ionicity was found to increase with increasing temperature. As is common with most bistriflamide salts, the TDAC salts are insoluble/immiscible in water but soluble/miscible in small polar organic solvents. Solubility/miscibility in non-polar solvents was found to be largely dependent on MW, but also subtly dependent on symmetry and conformational flexibility. Stability to common reagents (acid, base, redox) is very good with the exception of NaOH at elevated temperatures. Grignard reagents, however, appear to be unstable in these ILs. The addition of ether functional groups may be beneficial, as donor solvents are known to stabilise Grignard reagents. The electrochemical window was found to be quite small and similar to pyridinium salts, however, this is due to a low oxidation potential rather than a high reduction potential, so their use in reducing environments is still feasible while the reversibility of the oxidation process may be beneficial. The solid state structures of 3a and 3c further illustrate the steric protection afforded to the coplanar TAC atoms by addition of longer alkyl chains.

4. Experimental

All operations were performed using standard Schlenk techniques with a dinitrogen atmosphere in order to reduce exposure to water. ¹H-, ¹³C{¹H}-NMR spectra were collected on a Varian Unity-300 operating at 300 and 75 MHz, respectively, an Agilent DD2-400MR operating at 400 and 100 MHz, respectively, or on a Varian INOVA-500 operating at 500 and 126 MHz, respectively, in CDCl₃, referenced to residual solvent peaks. Electrospray mass spectrometry was carried out on a Micromass LCT, with samples dissolved in acetonitrile. Water contents were determined by Karl Fischer titration using a Metrohm 831 KF coulometer. Chloride contents were determined using a AutolabEco Chemie, with associated GPES software, under a dinitrogen atmosphere. The electrodes were either a glassy carbon (3 mm diameter) or platinum (1 mm diameter) working electrode, a platinum wire counter electrode and a silver reference electrode. Mircoanalysis was performed by Campbell Microanalytical Laboratory, Dunedin. Pentachlorocyclopropane (Acros), dimethylsulfate, diethylamine (Merck), dipropylamine (Acros), dibutylamine (Koch-Light), dipentylamine, didecylamine, butylmethylamine, hexylmethylamine, stearylmethylamine, ethylamine, propylamine, and triethylamine (Sigma-Aldrich) were used as obtained commercially. LiNTf₂ (3 M) was kindly provided by Prof. Ken Marsh. Ethylmethylamine was prepared by a modification of methods described by Lucier and Wawzonek.⁴⁷ The following salts were prepared by previously published methods: [C₃(NMe₂)₃]NTf₂ (3a),¹⁷ [C₃(NEt₂)₃]NTf₂ (3b),¹⁷ [C₃(NPr₂)₃]NTf₂ (3c),¹⁷ [C₃(NBu₂)₃]NTf₂ (3d),¹⁷ [C₃(NBuMe)₃]NTf₂ (4b),¹⁷ [C₃(NEt₂)₂NMe₂]NTf₂ (11a),¹⁸ [C₃(NEt₂)₂NBu₂]NTf₂ (11b),¹⁸ [C₃(NEt₂)₂NHex₂]NTf₂ (11c),¹⁸ [C₃(NEt₂)₂NBuMe]NTf₂ (12a)¹⁸ and [C₃(NEt₂)₂NHexMe]NTf₂ (12b).¹⁸

Syntheses of the starting materials [C₃(NMe₂)₃]Cl (1a), C₃(NMe₂)₂O and [C₃(NMe₂)₂(OMe)]MeSO₄ (7) are provided in the ESI† along with the syntheses and characterisation details of the TDAC salts: 1e, 3e, 1f, 3f, 1g, 3g, 4a, 2c, 4c, 9a, 9b, 9c, 9d, 13a, 10a, 13b, 10b, 10c and 10d.

DSC was performed on a Perkin Elmer Q100: samples of mass 5–20 mg were sealed in a vented aluminium pan and placed in the furnace with a 50 mL min⁻¹ nitrogen stream; the temperature was raised at 10 °C min⁻¹. TGA data were collected on dried samples using a TA Instruments SDT Q600 at 10 °C min⁻¹ after further drying at 100 °C for one hour in the instrument. Density measurements were carried out on an Anton Parr DMA 5000 instrument, an oscillating U-tube density meter, from 20 to 90 °C in 10 °C steps. Viscosities were measured on an Anton Parr AMVn falling-ball viscometer or a Brookfield-Wells cone-and-plate viscometer operating at 0.005–0.2 s⁻¹ rotation speed range. Conductivity of 3e, 3g and 12b were measured using a Schott LF4100 + probe and an impedance bridge conductivity meter. The instrument was calibrated with 0.1 mol L⁻¹ KCl solution. All other conductivities were measured by AC impedance spectroscopy on a Solatron SI 1296 frequency response analyser, at ranges up to 0.01 Hz to 10 MHz. Measurements were carried out with a dip cell probe containing two platinum wires covered in glass. The resistance was identified using a Nyquist Plot, and conductivity then calculated using κ = l/AR, where l/A is the cell constant, which was determined using 0.01 mol L⁻¹ KCl solution at 25 °C. All samples were measured from 20 °C (or above the melting point if solid) to 80 or 90 °C, and performed sealed or under a dinitrogen gas flow.

Solubility and miscibility studies were carried out by taking 0.5 mL of sample and adding step-wise 10 × 0.05 mL of solvent followed by 9 × 0.5 mL of solvent. After each addition of solvent the sample was mixed and allowed to equilibrate at 25 °C to determine whether the sample was miscible or immiscible. In the case of solid samples, a 0.1 g sample was taken and 2.5 mL of solvent was added and the sample was equilibrated at 25 °C. In some cases, the solid sample was observed to form two immiscible liquid layers. Water contents of water-saturated ILs were measured by Karl-Fischer titration after equilibration at 25 °C for 24 hours followed by centrifugation.

Chemical stability tests were conducted by placing a 250 mg sample with an equimolar amount of HCl (38%, aq), NH₃ (40%, aq, 2 equivalents), KOH, NaBH₄, NaIO₄ or EtMgI, and stirring at ambient temperature or at 60 °C. After 24 hours, ¹H-NMR was used to assess whether any degradation of the cation had occurred.

Cyclic voltammetry was carried out using an Eco Chemie Autolab PGSTAT 302N potentiostat running GPES 4.9 software. A platinum working electrode (1.2 mm diameter), platinum wire secondary electrode and silver wire reference electrode were used. Recrystallized [C₃(NEt₂)₃]TFSA was dried under vacuum at 50 °C for 72 hours, which reduced water content to 71 ppm. Sample was degassed by bubbling with argon for 60 min before the experiment, and kept under an argon atmosphere for the experiments. Once the electrochemical window was measured, ferrocene was added to the IL as an internal reference.

X-ray crystallography: single crystals of 3a and 3c formed in the neat liquid. A suitable crystal of each was selected and mounted on a SuperNova, Dual, Cu at zero, Atlas diffractometer. Using Olex2,⁴⁸ the structures were solved with the XS structure solution program⁴⁹ using Direct Methods and refined with the XL refinement package⁴⁹ using Least Squares minimisation. Crystal data and structure refinement and structural details are given in the ESI† along with the atom numbering schemes.

Acknowledgements

Dr Matthew Polson and Dr Chris Fitchett for X-ray data collection and refinement.

Notes and references

1. N. V. Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008, 37, 123.
2. For example: (a) Ionic Liquids in Synthesis, ed. P. Wasserscheid and T. Welton, Wiley-VCH, 2nd edn, 2008; (b) Ionic Liquids as Green Solvents: Progress and Prospects, ed. R. D. Rogers and K. R. Seddon, ACS Symp. Ser., American Chemical Society, Washington, D.C., 2003, vol. 856; (c) Green Solvents II: Properties and Applications of Ionic Liquids, ed. A. Mohammad and Dr Inamuddin, Springer, Netherlands, 2012.
3. For example: Ionic Liquids IIIB: Fundamentals, Progress, Challenges, and Opportunities – Transformations and Processes, ed. R. D. Rogers and K. R. Seddon, ACS Symp. Ser., American Chemical Society, Washington, D.C., 2005, vol. 902; Electrochemical Aspects of Ionic Liquids, ed. H. Ohno, Wiley-Interscience, Hoboken, 2005.
4. See: (a) H. Kunkel and G. Maas, Eur. J. Org. Chem., 2007, 3746 and references therein; (b) P. S. Kulkarni, L. C. Branco, J. G. Crespo, M. C. Nunes, A. Raymundo and C. A. M. Afonso, Chem.–Eur. J., 2007, 13, 8478.
5. (a) W. Kantlehner, E. Haug, W. W. Mergen, P. Speh, T. Maier, J. J. Kapassakalidis and H.-J. Bräuner, Synthesis, 1983, 904; (b) W. Kantlehner, E. Haug, W. W. Mergen, P. Speh, J. J. Kapassakalidis, H.-J. Bräuner and H. Hagen, Liebigs Ann. Chem., 1984, 108.
6. R. Breslow, J. Am. Chem. Soc., 1957, 79, 5318.
7. (a) Z. Yoshida and Y. Tawara, J. Am. Chem. Soc., 1971, 93, 2573; (b) Z. Yoshida, Top. Curr. Chem., 1973, 40, 47.
8. K. Komatsu and T. Kitagawa, Chem. Rev., 2003, 103, 1371.
9. J. S. Bandar and T. H. Lambert, Synthesis, 2013, 45, 2485.
10. R. Gompper and K. Schönafinger, Chem. Ber., 1979, 112, 1514.
11. J. R. Butchard, O. J. Curnow, R. J. Pipal, W. T. Robinson and R. Shang, J. Phys. Org. Chem., 2008, 21, 127.
12. (a) J. R. Butchard, O. J. Curnow, D. J. Garrett and R. G. A. R. Maclagan, Angew. Chem., Int. Ed., 2006, 45, 7550; (b) J. R. Butchard, O. J. Curnow, D. J. Garrett, R. G. A. R. Maclagan, E. Libowitzky, P. M. B. Piccoli and A. J. Schultz, Dalton Trans., 2012, 41, 11765.
13. C. Wilcox and R. Breslow, Tetrahedron Lett., 1980, 21, 3241.
14. (a) R. Weiss and C. Priesner, Angew. Chem., Int. Ed. Engl., 1978, 17, 445; (b) Y. Yagyu, N. Matsumura, H. Tanaka, Z. Maeda and H. Inoue, J. Chem. Res., Synop., 1995, 420.
15. R. Weiss and K. Schloter, Tetrahedron Lett., 1975, 3491.
16. (a) R. Weiss, T. Brenner, F. Hampel and A. Wolski, Angew. Chem., Int. Ed. Engl., 1995, 34, 439; (b) R. Weiss, M. Rechinger, F. Hampel and A. Wolski, Angew. Chem., Int. Ed. Engl., 1995, 34, 441; (c) R. Weiss, O. Schwab and F. Hampel, Chem.–Eur. J., 1999, 5, 968.
17. O. J. Curnow, D. R. MacFarlane and K. J. Walst, Chem. Commun., 2011, 47, 10248.
18. O. J. Curnow, M. T. Holmes, L. C. Ratten, K. J. Walst and R. Yunis, RSC Adv., 2012, 2, 10794.
19. (a) M. J. Taylor, P. W. J. Surman and G. R. Clark, J. Chem. Soc., Chem. Commun., 1994, 21, 2517; (b) G. R. Clark, P. W. J. Surman and M. J. Taylor, J. Chem. Soc., Faraday Trans., 1995, 91, 1523.
20. Z. Yoshida, H. Konishi, Y. Tawara, K. Nishikawa and H. Ogoshi, Tetrahedron Lett., 1973, 2619.
21. (a) J. S. Bandar and T. H. Lambert, J. Am. Chem. Soc., 2012, 134, 5552; (b) H. Heydt, P. Eisenbarth, K. Feith, K. Urgast, G. Maas and M. Regitz, Isr. J. Chem., 1986, 27, 96; (c) H. Bruns, M. Patil, J. Carreras, A. Vazquez, W. Thiel, R. Goddard and M. Alcarazo, Angew. Chem., Int. Ed., 2010, 49, 3680; (d) R. Weiss, K. G. Wagner, C. Priesner and J. Macheleid, J. Am. Chem. Soc., 1985, 107, 4491; (e) R. Weiss and R. H. Lowack, Bull. Soc. Chim. Belg., 1991, 100, 483; (f) R. Gompper and K. Schönafinger, Chem. Ber., 1979, 112, 1535.
22. (a) H. Tokuda, K. Hayamizu, K. Ishii, M. A. B. H. Susan and M. Watanabe, J. Phys. Chem. B, 2005, 109, 6103; (b) J. G. Huddleston, A. E. Visser, W. M. Reichert, H. D. Willauer, G. A. Broker and R. D. Rogers, Green Chem., 2001, 3, 156.
23. J. G. Huddleston, A. E. Visser, W. M. Reichert, H. D. Willauer, G. A. Broker and R. D. Rogers, Green Chem., 2001, 3, 156.
24. M. Montanino, M. Carewska, F. Alessandrini, S. Passerini and G. B. Appetecchi, Electrochim. Acta, 2011, 57, 153.
25. C. Maton, N. De Vos and C. V. Stevens, Chem. Soc. Rev., 2013, 42, 5963.
26. T. J. Wooster, K. M. Johanson, K. J. Fraser, D. R. MacFarlane and J. L. Scott, Green Chem., 2006, 8, 691.
27. M. Tariq, A. P. Serro, J. L. Mata, B. Saramago, J. M. S. S. Esperança, J. N. C. Lopes and L. P. N. Rebelo, Fluid Phase Equilib., 2010, 294, 131.
28. C. Ye and J. M. Shreeve, J. Phys. Chem. A, 2007, 111, 1456.
29. Confusingly, a is frequently used in this equation for the slope, with b for the intercept, which is similar but contrary to the use of a and b in a plot of ρ versus T. Thus, we have used c and d here to eliminate this confusion.
30. G. Adamová, R. L. Gardas, L. P. N. Rebelo, A. J. Robertson and K. R. Seddon, Dalton Trans., 2011, 40, 12750.
31. G. Adamová, R. L. Gardas, M. Nieuwenhuyzen, A. V. Puga, L. P. N. Rebelo, A. J. Robertson and K. R. Seddon, Dalton Trans., 2012, 41, 8316.
32. M. Tariq, P. A. S. Forte, M. F. C. Gomes, J. N. C. Lopes and L. P. N. Rebelo, J. Chem. Thermodyn., 2009, 41, 790–798.
33. J. M. S. S. Esperança, H. J. R. Guedes, M. Blesic and L. P. N. Rebelo, J. Chem. Eng. Data, 2006, 51, 237.
34. Z. Y. Gu and J. F. Brennecke, J. Chem. Eng. Data, 2002, 47, 339.
35. Z. J. Chen and J.-M. Lee, J. Phys. Chem. B, 2014, 118, 2712.
36. (a) C. A. Angell, J. Non-Cryst. Solids, 1991, 131–133, 13; (b) C. A. Angell, in Molten Salts and Ionic Liquids: Never the Twain?, ed. M. Gaune-Escard and K. R. Seddon, John Wiley & Sons, Hoboken, 2010, pp. 1–24; (c) J.-P. Belieres and C. A. Angell, J. Phys. Chem. B, 2007, 111, 4926.
37. D. R. MacFarlane, M. Forsyth, E. I. Izgorodina, A. P. Abbott, G. Annat and K. Fraser, Phys. Chem. Chem. Phys., 2009, 11, 4962.
38. (a) H. Tokuda, S. Tsuzuki, M. A. B. H. Susan, K. Hayamizu and M. Watanabe, J. Phys. Chem. B, 2006, 110, 19593; (b) H. Tokuda, K. Hayamizu, K. Ishii, M. A. B. H. Susan and M. Watanabe, J. Phys. Chem. B, 2004, 108, 16593; (c) H. Tokuda, K. Ishii, M. A. B. H. Susan, S. Tsuzuki, K. Hayamizu and M. Watanabe, J. Phys. Chem. B, 2006, 110, 2833; (d) H. Tokuda, K. Hayamizu, K. Ishii, M. A. B. H. Susan and M. Watanabe, J. Phys. Chem. B, 2005, 109, 6103; (e) A. Noda, K. Hayamizu and M. Watanabe, J. Phys. Chem. B, 2001, 105, 4603.
39. K. Ueno, H. Tokuda and M. Watanabe, Phys. Chem. Chem. Phys., 2010, 12, 1649.
40. K. R. Seddon, A. Stark and M.-J. Torres, Pure Appl. Chem., 2000, 72, 2275.
41. (a) K. Machanová, A. Boisset, Z. Sedláková, M. Anouti, M. Bendová and J. Jacquemin, J. Chem. Eng. Data, 2012, 57, 2227; (b) M. G. Freire, C. M. S. S. Neves, K. Shimizu, C. E. S. Bernardes, I. M. Marrucho, J. A. P. Coutinho, J. N. C. Lopes and L. P. N. Rebelo, J. Phys. Chem. B, 2010, 114, 15925; (c) K. Řehák, P. Morávek and M. Strejc, Fluid Phase Equilib., 2012, 316, 17; (d) J. Jacquemin, P. Husson, A. A. H. Padua and V. Majer, Green Chem., 2006, 8, 172; (e) C. M. S. S. Neves, P. J. Carvalho, M. G. Freire and J. A. P. Coutinho, J. Chem. Thermodyn., 2011, 43, 948; (f) Y. Cao, Y. Chen, X. Wang and T. Mu, RSC Adv., 2014, 4, 5169; (g) E. J. Gonzalez and E. A. Macedo, Fluid Phase Equilib., 2014, 383, 72; (h) M. G. Freire, P. J. Carvalho, R. L. Gardas, I. M. Marrucho, L. M. N. B. F. Santos and J. A. P. Coutinho, J. Phys. Chem. B, 2008, 112, 1604; (i) M. G. Freire, C. M. S. S. Neves, P. J. Carvalho, R. L. Gardas, A. M. Fernandes, I. M. Marrucho, L. M. N. B. F. Santos and J. A. P. Coutinho, J. Phys. Chem. B, 2007, 111, 13082; (j) J. Salminen, N. Papaiconomou, R. A. Kumara, J.-M. Lee, J. Kerr, J. Newmana and J. M. Prausnitz, Fluid Phase Equilib., 2007, 261, 421; (k) N. Papaiconomou, J. Salminen, J.-M. Lee and J. M. Prausnitz, J. Chem. Eng. Data, 2007, 52, 833.
42. (a) P. S. Kulkarni, L. C. Branco, J. G. Crespo, M. C. Nunes, A. Raymundo and C. A. M. Afonso, Chem.–Eur. J., 2007, 13, 8478; (b) N. M. M. Mateaus, L. C. Branco, N. M. T. Lopurenço and C. A. M. Afonso, Green Chem., 2003, 5, 347.
43. R. W. Johnson, Tetrahedron Lett., 1976, 8, 589.
44. A. M. O'Mahony, D. S. Silvester, L. Aldous, C. Hardacre and R. G. Compton, J. Chem. Eng. Data, 2008, 53, 2884.
45. H. Matsumoto, M. Yanagida, K. Tanimoto, T. Kojima, Y. Tamiya and Y. Miyazaki, Proc. Electrochem. Soc.: Molten Salts XII, 2000, p. 186.
46. (a) C. D. Cowman, J. C. Thibeault, R. F. Ziolo and H. B. Gray, J. Am. Chem. Soc., 1976, 98, 3209; (b) R. Weiss, M. Rechinger and F. Hampel, Z. Kristallogr., 1995, 210, 71; (c) H. N. Schaefer, H. Burzlaff, A. M. H. Grimmeiss and R. Weiss, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1992, 48, 795; (d) J. R. Butchard, O. J. Curnow, R. J. Pipal, W. T. Robinson and R. Shang, J. Phys. Org. Chem., 2008, 21, 127; (e) A. T. Ku and M. Sundaralingam, J. Am. Chem. Soc., 1972, 94, 1688; (f) H. N. Schaefer, H. Burzlaff, A. M. H. Grimmeiss and R. Weiss, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1992, 48, 912; (g) H. N. Schaefer, H. Burzlaff, A. M. H. Grimmeiss and R. Weiss, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1991, 47, 1808.
47. (a) S. Wawzonek, W. McKillip and C. J. Peterson, Org. Synth., 1964, 44, 75; (b) J. J. Lucier, A. D. Harris and P. S. Korosec, Org. Synth., 1964, 44, 72.
48. O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339.
49. G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008, 64, 112.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis and characterisation details of starting materials and TDAC salts; tables of typical ¹H- and ¹³C-NMR chemical shift ranges, density data (with fit parameters), molar volumes, viscosity, viscosity fit parameters, conductivity, conductivity fit parameters, water contents of water-saturated bistriflamide ILs, crystallographic data for 3a and 3c, bond lengths and angles for 3a and 3c; figures of density parameters versus MWs and atomic numbering schemes for 3a and 3c; a fragility plot; and an adjusted Walden plot with associated discussion. Crystallographic data in CIF format is available: CDCC 1055725 and 1055726. See DOI: 10.1039/c5ra05254h

---