Mesoscopic aggregation in aprotic ionic liquids due to the microphase separation of polar and non-polar components is expected to correlate strongly with the physicochemical properties of ionic liquids and therefore their potential applications. The most commonly cited experimental evidence of such aggregation is the observation of a low-q pre-peak in the x-ray and neutron scattering profiles, attributed to the polarity alternation of polar and apolar phases. In this work, a homologous series of phosphonium ionic liquids with the bis(trifluoromethylsulfonyl)imide anion and systematically varying alkyl chain lengths on the phosphonium cation are investigated by small and wide-angle x-ray scattering, dynamic-mechanical spectroscopy, and broadband dielectric spectroscopy. A comparison of the real space correlation distance corresponding to the pre-peak and the presence or absence of the slow sub-α dielectric relaxation previously associated with the motion of mesoscale aggregates reveals a disruption of mesoscale aggregates with increasing symmetry of the quaternary phosphonium cation. These findings contribute to the broader understanding of the interplay of molecular structures, mesoscale aggregation, and physicochemical properties in aprotic ionic liquids.

Ionic liquids (ILs) are valued for their unique characteristics such as low vapor pressure, low flammability, wide liquidus ranges, and electrochemical stability. The large number of potential IL molecular structures with different functional groups makes them promising designer solvents with applications in energy storage, nanoparticle growth, biomass processing, and organic synthesis.1–3 Rational design of ILs capable of use in these applications requires the development of detailed structure-property relationships. Considerable progress has been made along these lines through detailed experimental and computational work on a wide variety of cations such as imidazolium, pyrrolidinium, piperidinium, phosphonium, and ammonium combined with a range of anions.4–8 An emerging obstacle in this endeavor is the finding that certain ILs aggregate and form long-lived mesostructures that extend over a few nanometers, while others do not show clear evidence of aggregation.2,9 This mesoscale organization presumably arises due to the solvophobic separation of polar and non-polar moieties on the cation charge center. Non-polar alkyl groups are excluded from the regions occupied by cations and anions and form extended aggregates surrounded by ionic shells.2,11 The existence of these distinct regions provides ILs with the ability to solvate both polar and non-polar molecules, an advantage critical for their applications as solvents in synthesis and material processing.2,12

The primary experimental evidence of the formation of mesoscale structures is the emergence of a low momentum transfer, q, pre-peak in the x-ray and neutron scattering profiles. Computer simulations which reproduce the experimental scattering profiles have also provided snapshots showing the existence of three-dimensional mesoscale organization present in the liquid phase of ILs.2,13–19 Despite the initial uncertainty that the pre-peak may indicate only a local ordering due to cation anisotropy, it is now almost universally attributed to a long-range order induced by hydrophobic aggregation.1,2,20–22 This assignment is strengthened by recent results from neutron spin echo (NSE), broadband dielectric spectroscopy (BDS), and dynamic-mechanical spectroscopy (DMS) which reveal aggregate dynamics at timescales considerably longer than the primary structural relaxation.9,24–27 In addition, these slow sub-α relaxations contribute to the increases in the zero-shear viscosity and static dielectric permittivity, highlighting the influence of aggregation on physicochemical properties.27 However, experimental data on the dynamics of mesoscale aggregates are currently limited to the well-studied imidazolium-based ionic liquids. There are numerous open scientific questions regarding the nature and lifetimes of mesoscale aggregates in other classes of aprotic ILs.

In this work, x-ray scattering, dynamic-mechanical spectroscopy, and broadband dielectric spectroscopy are utilized to investigate the influence of chemical structures on the formation and dynamics of mesoscale aggregates in a series of tetraalkylphosphonium bis(trifluoromethylsulfonyl)imide ILs. Unexpectedly, it is observed that increasing the volume fraction of non-polar functional groups in the phosphonium ILs does not necessarily promote mesoscale aggregation. A detailed analysis of the results reveals a disruption of the aggregates with increasing lengths of the shorter alkyl chains as evident from the absence of the sub-α relaxation as well as a substantial reduction in the real-space distance corresponding to the low-q peak. The combination of insights from the experimental techniques capable of probing both the structure and the dynamics of mesoscale aggregates enables us to make this distinction.

Four phosphonium-based ionic liquids, triethyl-alkyl-phosphonium and tributyl-alkyl-phosphonium with alkyl chain lengths of octyl and dodecyl, with a common bis(trifluoromethylsulfonyl)imide anion, are the focus of the work reported here. Molecular structures and acronyms of the ionic liquids are provided in Fig. 1. The phosphonium ionic liquids were obtained from the Nippon Chemical Industrial Co. The ILs were dried under vacuum (10−6 bar) at 50 °C for 24 h prior to experiments. Small-angle and wide-angle x-ray scattering measurements were conducted at room temperature using a SAXSLab Ganesha x-ray scattering system. The samples were encased in a button cell with Kapton windows. An empty cell was also measured to enable subtraction of the Kapton background. Broadband dielectric spectroscopy (BDS) measurements were conducted in the frequency range of 10−1–107 Hz and temperature range 180-400 K using a Novocontrol High Resolution Dielectric Alpha Analyzer with a QUATRO liquid nitrogen temperature control system with temperature stability better than ±0.1 K. The samples were measured in a parallel plate capacitor geometry with 20 mm diameter gold-plated brass electrodes. A sample thickness of 0.5 mm was maintained using three Teflon spacers. The dynamic-mechanical spectra were obtained via oscillatory shear measurements over the frequency range 0.1–100 Hz with 0.05-2 strain % on Hybrid Rheometer 2 (TA Instruments) using parallel plate geometry with diameters of 8 mm and 3 mm. The temperature was controlled by an Environmental Test Chamber with nitrogen as the gas source with temperature stability ±0.1 K. Differential scanning calorimetry (DSC) measurements were performed on a TA Instruments Q2000 calorimeter at a cooling rate of 10 K/min. The calorimetric glass transition temperature, Tg,DSC, was determined at the midpoint of the step in heat flow corresponding to the maximum in the temperature derivative of the heat flow.

FIG. 1.

Structures of the studied phosphonium ionic liquids.

FIG. 1.

Structures of the studied phosphonium ionic liquids.

Close modal

The x-ray scattering profiles are presented in Fig. 2 as intensity (I) of scattered x-rays versus momentum transfer (q). Three distinct peaks are observed for each IL. Based on numerous experimental and computational studies of a wide variety of ionic liquids, the highest and middle-q peaks are assigned to adjacency and charge-alternation correlations, respectively.11,29 The origin of the adjacency peak is the inter- and intramolecular correlations of neighboring atoms.11 The charge-alternation peak arises, as the name implies, from the ordering (alternation) of cations and anions mediated by Coulombic interactions and is typical of molten salts. It corresponds to the distance separating two like-charge ions, that is the anion-anion or cation-cation separation distance.11 In the triethyl-alkyl-phosphoniums, like ionic groups are separated by an average of d = 7.3 Å, regardless of the length of the longer alkyl chain, where d = 2π/qpeak. Increasing the short-chain length to four carbons in the tributyl-alkyl-phosphoniums increases the spacing between the like ionic groups to 8.3 Å.

FIG. 2.

X-ray diffraction spectra of the phosphonium-based ionic liquids: (a) triethyloctylphosphonium (TEOP NTf2) and triethyldodecylphosphonium (TEDP NTf2) bis(trifluoromethylsulfonyl)imide; (b) tributyloctylphosphonium (TBOP NTf2) and tributyldodecylphosphonium (TBDP NTf2) bis(trifluoromethylsulfonyl)imide. Approximate distances in real-space, d = 2π/qpeak, for each major peak are indicated at the arrows. Lines correspond to fits with two Gaussian functions.

FIG. 2.

X-ray diffraction spectra of the phosphonium-based ionic liquids: (a) triethyloctylphosphonium (TEOP NTf2) and triethyldodecylphosphonium (TEDP NTf2) bis(trifluoromethylsulfonyl)imide; (b) tributyloctylphosphonium (TBOP NTf2) and tributyldodecylphosphonium (TBDP NTf2) bis(trifluoromethylsulfonyl)imide. Approximate distances in real-space, d = 2π/qpeak, for each major peak are indicated at the arrows. Lines correspond to fits with two Gaussian functions.

Close modal

The lowest q peak, known as the pre-peak or first sharp diffraction peak, is found in aprotic ionic liquids which have sufficiently long non-polar alkyl chains substituted on the cation charge center.1,2,18–21,30,31 The origin of this peak is universally assigned to the existence of alternating polar and non-polar regions resulting in mesoscale heterogeneity; however, there is some debate as to whether it is indicative of a pseudomicellar three-dimensional nanostructure or only local ordering due to cation anisotropy.20,21 The formation of hydrophobic aggregates relies on the segregation of neighboring alkyl chains into a non-polar region. In this view, the real-space correlation distance given by the polarity-alternation peak corresponds to the average distance separating ionic regions on the opposite sides of the non-polar inclusions. This distance will therefore depend on the alkyl chain length and the degree of interdigitation of opposing alkyl chains. A significant shift in the length scale corresponding to the pre-peak occurs when the shorter alkyl chains are lengthened from ethyl to butyl. Despite having the same length of the longest alkyl chain, the distance is decreased from 14.1 Å to 10.5 Å for triethyloctylphosphonium (TEOP) and tributyloctylphosphonium (TBOP) and from 22.1 Å to 14.6 Å for triethyldodecylphosphonium (TEDP) and tributyldodecylphosphonium (TBDP), as indicated in Fig. 2. The dependence of aggregate size, d, on the length of the longer alkyl chain, nc, for the studied phosphonium ionic liquids is compared with other aprotic ionic liquids found in the literature in Fig. 7 and discussed later. In accordance with common practice, the existence of the x-ray pre-peak in each of the studied phosphonium ILs might be taken as an indication that they each contain similar mesoscale aggregates. If this is the case, a slow relaxation associated with aggregate dynamics should be present in their dielectric spectra.27 

Broadband dielectric spectroscopy (BDS) of ionic liquids has previously been used to probe ion dynamics and charge transport over broad temperature and frequency ranges.32 By comparison with other experimental techniques, the ion dynamics of these purely ionic materials are found to occur at the same timescale as structural α-relaxations. In a recent article, we demonstrated that an additional slow sub-α relaxation emerges with the onset of mesoscale aggregation in two series of 1-alkyl-3-methylimidazolium ionic liquids with bis(trifluoromethylsulfonyl)imide and tetrafluoroborate anions.27 The timescales of this relaxation were found to correspond with the decay time of the pre-peak as obtained by neutron spin echo (NSE) spectroscopy.24–27 The additional dielectric relaxation was therefore attributed to fluctuations of the mesoscale aggregates. It is now our intention to apply a similar analysis presented in that paper to the current series of phosphonium ionic liquids and to relate the trends observed for the slow sub-α relaxation to the length scale of the polarity-alternation peak as well as the molecular structure.

The real and imaginary parts of complex permittivity, ε*ω=εωiεω, and conductivity, σ*ω=σω+iσω, are presented in Fig. 3 as functions of the radial frequency, ω = 2πf, and over a range of temperatures for TEOP NTf2. The lines correspond to fits obtained by a combination of the Havriliak-Negami function and Debye equation with a power law to account for low frequency dispersion due to electrode polarization, as given in Eq. (1), where ε is the high frequency limiting permittivity, σ0 is the dc ionic conductivity, Δεe and Δεaggregate are the dielectric strengths, τe and τaggregate are the relaxation times, β and γ are the stretching parameters, A is the pre-exponential factor, and n is the exponent:33 

(1)

The rate of the faster relaxation, ωe = 1/τe, corresponds closely to the rate of the structural relaxation as measured by dynamic-mechanical spectroscopy (see Fig. 6), as well as the frequency of the peak in the electric loss modulus, M″. It is therefore attributed to the ion hopping dynamics previously shown to correspond to the structural relaxation in aprotic ionic liquids.32 The slow sub-α relaxation is most readily observed as a peak in εder at frequencies below that of the ion dynamics, where εder=π/2ε/lnω. The derivative representation is utilized to suppress the dominant contribution of dc ionic conductivity to the dielectric spectra and to enable the observation of dynamics slower than the conductivity relaxation. The existence of the slow relaxation in TEOP NTf2 and its ammonium homologue was previously reported by Griffin et al.34 A clearer representation showing the existence of the two relaxations is made by plotting εder against the frequency normalized by the frequency of the ion dynamics, ωe, as presented in Fig. 4, for all the studied phosphonium ionic liquids. Here, the dashed lines represent the Havriliak-Negami function used to describe the α-relaxation and the dotted-dashed line is the Debye equation used to describe the slow sub-α relaxation. The most obvious difference in the dielectric spectra is the absence of the slower dielectric process in the tributyl-alkyl-phosphonium ionic liquids. In addition, the width of their α-relaxation is noticeably narrower (see Table I) for shape parameters of the Havriliak-Negami fit function. The disappearance of the sub-α relaxation dynamics indicates that long-lived mesoscale aggregates do not form in the two tributyl-alkyl-phosphoniums, in apparent contradiction to the picture from the measured x-ray scattering profiles.

FIG. 3.

Real, ε and σ, and imaginary, ε and σ, parts of complex permittivity, ε*ω=εωiεω, and conductivity, σ*ω=σω+iσω, for TEOP NTf2. A slow sub-α relaxation emerges with increasing temperature, as observed in εder=π/2ε/lnω. Lines correspond to fits obtained by Eq. (1). The static dielectric permittivity, εs, is obtained as the low frequency limit of the fit without the contribution of the power law. The long-range ionic (dc) conductivity, σ0, corresponds to the value of the frequency independent plateau in σ as indicated.

FIG. 3.

Real, ε and σ, and imaginary, ε and σ, parts of complex permittivity, ε*ω=εωiεω, and conductivity, σ*ω=σω+iσω, for TEOP NTf2. A slow sub-α relaxation emerges with increasing temperature, as observed in εder=π/2ε/lnω. Lines correspond to fits obtained by Eq. (1). The static dielectric permittivity, εs, is obtained as the low frequency limit of the fit without the contribution of the power law. The long-range ionic (dc) conductivity, σ0, corresponds to the value of the frequency independent plateau in σ as indicated.

Close modal
FIG. 4.

Derivative representation of the real part of complex dielectric function, εder=(π/2)[ε/ln(ω)], of the two series of phosphonium bis(trifluoromethylsulfonyl)imide ionic liquids: (a) triethyloctylphosphonium, TEOP NTf2; (b) triethyldodecylphosphonium, TEDP NTf2; (c) tributyloctylphosphonium, TBOP NTf2; and (d) tributyldodecylphosphonium, TBDP NTf2.

FIG. 4.

Derivative representation of the real part of complex dielectric function, εder=(π/2)[ε/ln(ω)], of the two series of phosphonium bis(trifluoromethylsulfonyl)imide ionic liquids: (a) triethyloctylphosphonium, TEOP NTf2; (b) triethyldodecylphosphonium, TEDP NTf2; (c) tributyloctylphosphonium, TBOP NTf2; and (d) tributyldodecylphosphonium, TBDP NTf2.

Close modal
TABLE I.

Shape parameters from fits using the Havriliak-Negami function [Eq. (1)].

CationT (K)β ± 0.1γ
TEOP 200 0.5 0.8 ± 0.2 
TEDP 225 0.4 1.0 ± 0.2 
TBOP 200 0.8 0.4 ± 0.1 
TBDP 200 0.7 0.5 ± 0.2 
CationT (K)β ± 0.1γ
TEOP 200 0.5 0.8 ± 0.2 
TEDP 225 0.4 1.0 ± 0.2 
TBOP 200 0.8 0.4 ± 0.1 
TBDP 200 0.7 0.5 ± 0.2 

The real and imaginary parts of the complex shear modulus, G*=G(ω)+iG(ω), of TEOP NTf2 and TBOP NTf2 are presented in Fig. 5; shift factors are provided in the supplementary material. The lines correspond to a fit with a single Cole-Davidson function as given in Eq. (2), where G is the high frequency limiting shear modulus, τα is the relaxation time of the structural α-relaxation, and γ is a parameter associated with the spectral shape.35 

(2)
FIG. 5.

Real (open symbols) and imaginary (closed symbols) parts of the complex shear modulus of TEOP NTf2 and TBOP NTf2 as measured over the temperature range 190-205 K. Lines correspond to fits by Eq. (2).

FIG. 5.

Real (open symbols) and imaginary (closed symbols) parts of the complex shear modulus of TEOP NTf2 and TBOP NTf2 as measured over the temperature range 190-205 K. Lines correspond to fits by Eq. (2).

Close modal

In our previous article, a slow sub-α relaxation was also observed by dynamic-mechanical spectroscopy (DMS) for 1-octyl-3-methylimidazolium tetrafluoroborate and attributed to the motion of mesoscale aggregates. However, such a relaxation is not found in either of the phosphonium ionic liquids investigated here. DMS measurements of the longer chain TEDP NTf2 and TBDP NTf2 were not successful due to their high propensity to crystallize (see the supplementary material). The broader frequency range of BDS enables the investigation of the dynamics of interest at temperatures above the crystallization point. The relative magnitude and sharpness of the x-ray pre-peak is much lower in the phosphonium ILs compared to the imidazolium systems. This indicates that even in the triethyl-alkyl-phosphoniums the degree of correlation is much lower than the corresponding imidazolium systems.20 The sub-α dynamic-mechanical relaxation may be less sensitive to lower extents of aggregation than the dielectric relaxation, as is suggested by its response to alkyl chain length in the imidazolium ILs. It is also possible that the mesoscopic domains are disrupted at the low temperatures probed by DMS, as indicated by the disappearance of the slow sub-α dielectric relaxation with decreasing temperature. A decrease in organization at low temperatures has previously been observed in molecular dynamics simulations and x-ray measurements of trihexyl-tetradecyl-phosphonium bis(trifluoromethylsulfonyl)imide and attributed to a reduced ordering of the polar phase.13,36,37 The structural α-relaxation rates, ωα, obtained by DMS are presented with the relaxation rates obtained by BDS in Fig. 6. As stated previously, the ion hopping relaxation rate corresponds closely to the structural relaxation rate, while the slow sub-α relaxation is approximately 20 times slower at the higher temperatures.

FIG. 6.

(a) Structural relaxation, ωα, obtained by DMS (crossed symbols), ion hopping rates obtained, ωe, by BDS (closed symbols), and the relaxation rate of the slow sub-α relaxation, ωaggregate, obtained by BDS (open symbols). (b) Relaxation rates versus temperature scaled by the calorimetric glass transition temperature obtained by DSC, Tg,DSC.

FIG. 6.

(a) Structural relaxation, ωα, obtained by DMS (crossed symbols), ion hopping rates obtained, ωe, by BDS (closed symbols), and the relaxation rate of the slow sub-α relaxation, ωaggregate, obtained by BDS (open symbols). (b) Relaxation rates versus temperature scaled by the calorimetric glass transition temperature obtained by DSC, Tg,DSC.

Close modal

The real-space correlation distance corresponding to the low-q, polarity alternation-peak in x-ray scattering of the phosphonium series is presented as a function of the longest alkyl chain length in Fig. 7. These values are compared with literature data for the ILs 1-alkyl-3-methylimidazolium,1 1-alkyl-1-methylpyrrolidinium,10 1-alkyl-1-methylpiperidinium,23 trialkyl-methylammonium,28 and trihexyl-tetradecyl-phosphonium13 bis(trifluoromethylsulfonyl)imide. The mesoscale aggregates are characterized by the formation of a non-polar domain consisting of interdigitated alkyl chain tails surrounded by a polar domain of anions and cations. The pre-peak is taken as the distance, d = 2π/qpeak, separating the polar domains on the opposite sides of the non-polar inclusion. The dependence of d on the number of carbons in the alkyl chain, nc, will depend on the degree of interdigitation, the ratio of trans/gauche isomers, and the location of ions within the polar domain.1,38 A further insight into this correlation distance is obtained by comparing d with the length of an extended, all-trans alkyl chain. This distance is given by the Tanford equation, lmax=1.5Å+1.265Å/CH2nc, and is represented in Fig. 7 by the dotted-dashed line.39 The maximum possible size, dmax, of a mesoscale aggregate will approximately correspond to an aggregate in which opposing alkyl chains are completely extended and non-interdigitated, dmax=2lmax=3.0Å+2.53Å/CH2nc. This case is represented by the dashed line in Fig. 7. The majority of ILs have values of d intermediate of these two extremes, indicating varying degrees of interdigitation as well as possible changes in the organization of the polar phases and trans/gauche ratios. However, the d-spacing of the tributyl-alkyl-phosphonium and trihexyl-tetradecyl-phosphonium is significantly lower. In fact, they fall below the length scale of a single fully extended alkyl chain, revealing that aggregation is no longer necessary to explain the origin of this distance. In addition, the slope is significantly lower for the tri-butyl-alkyl-phosphonium series, see Table II. The d-spacing in these systems is well approximated by 0.87lmax, which is shown as a dotted line in Fig. 7, indicating that the only possible mesoscale aggregate capable of producing this length scale would consist of fully interdigitated alkyl chains with some7t 2b degree of trans/gauche isomerism.39 Recent atomistic simulations on a series of similar quaternary phosphonium chloride-based ILs reveal even shorter length scales associated with the x-ray scattering pre-peak.40 Therefore, the dramatic change in the d-spacing upon lengthening the shorter alkyl chains on quaternary phosphonium ILs from ethyl to butyl is attributed to a disruption of the mesoscale hydrophobic aggregates, in agreement with the atomistic simulations which show a breaking of the polar network in the tributyl-alkyl-phosphonium chlorides.40 

FIG. 7.

Real-space distance of the polarity alternation peak, d = 2π/qpeak, as a function of the number of carbons in the alkyl chain for imidazolium1 (Im), pyrrolidinium10 (Pyr), piperidinium23 (Pip), trialkyl-methyl-ammonium28 (TAlkylMA), triethyl-alkyl-phosphonium (TEAlkylP), tributyl-alkyl-phosphonium (TBAlkylP), and trihexyl-tetradecyl-phosphonium13 (THAlkylP) ionic liquids with the bis(trifluoromethylsulfonyl)-imide anion. The dotted-dashed line is lmax given by the Tanford equation, the dashed line is 2lmax, the dotted line is 0.87 lmax, and the solid lines are linear fits (parameters are listed in Table II).

FIG. 7.

Real-space distance of the polarity alternation peak, d = 2π/qpeak, as a function of the number of carbons in the alkyl chain for imidazolium1 (Im), pyrrolidinium10 (Pyr), piperidinium23 (Pip), trialkyl-methyl-ammonium28 (TAlkylMA), triethyl-alkyl-phosphonium (TEAlkylP), tributyl-alkyl-phosphonium (TBAlkylP), and trihexyl-tetradecyl-phosphonium13 (THAlkylP) ionic liquids with the bis(trifluoromethylsulfonyl)-imide anion. The dotted-dashed line is lmax given by the Tanford equation, the dashed line is 2lmax, the dotted line is 0.87 lmax, and the solid lines are linear fits (parameters are listed in Table II).

Close modal
TABLE II.

Parameters for the linear fits, corresponding to the solid lines in Fig. 7.

IL seriesSlope (Å/CH2)
Im 2.0 
Pyr 2.3 
Pip 1.2 
TAlkylMA 1.3 
TEAlkylP 2.0 
TBAlkylP 1.0 
IL seriesSlope (Å/CH2)
Im 2.0 
Pyr 2.3 
Pip 1.2 
TAlkylMA 1.3 
TEAlkylP 2.0 
TBAlkylP 1.0 

The loss of the mesoscale aggregates is corroborated by the absence of the slow sub-α relaxation, which is present in the more highly aggregating imidazolium and triethyl-alkyl-phosphonium ILs.27 The existence of a pre-peak is therefore insufficient evidence that similar types of long-lived hydrophobic aggregates are present across ranges of ionic liquids. Rather, the actual distances corresponding to the pre-peak must also correspond to anticipated length scales based on the chemical structure of the IL. In addition, techniques capable of probing the dynamics of mesoscale aggregates, such as dielectric spectroscopy, provide valuable insight into the existence of aggregates as well as their influence on physicochemical properties. The sensitivity of the mesoscale organization to the lengths of the alkyl chains on quaternary phosphonium ILs distinguishes them from other classes of aprotic ionic liquids. These results show that a high degree of tunability of the mesoscale structures is achievable with minor changes to the chemical structure. The mechanism of the disruption is attributed to a weakening of Coulombic interactions within the polar phase due to the increased distance between neighboring ionic groups. This larger separation is evidenced by the increase in charge adjacency distances corresponding to the middle-q x-ray peak (see Fig. 2). Studies over a wider range and combination of alkyl chain lengths are currently underway to more fully elucidate the role of chemical structures in the transition from aggregating to non-aggregating phosphonium ILs.

Elucidating the link between the chemical structure and physicochemical properties is a necessary step for the full realization of ionic liquids as truly designer solvents. Accordingly, the influence of mesoscale organization on such properties must be identified. With increasing alkyl chain length and the onset of mesoscale aggregation, several authors have reported a reduction in ion mobility, leading to a decrease in the ionic conductivity and a simultaneous increase in the zero-shear viscosity.41–44 In addition, we have previously reported an increase in the static dielectric permittivity and viscosity due to the additional slow sub-α dielectric and dynamic-mechanical relaxation in aggregating imidazolium ILs. As seen in Figs. 8 and 9, the transition from triethyl-alkyl-phosphonium to tributyl-alkyl-phosphonium results in a decrease of the static dielectric permittivity as well as the ionic conductivity. The reduction in ionic conductivity might be interpreted as an aggregation induced effect. However, considering the x-ray and BDS evidence presented earlier, this interpretation is not favored. Elucidating whether this decrease is due to the reduced ion mobility or effective number density of charge carriers requires a comparison of diffusivities obtained by pulsed field gradient nuclear magnetic resonance spectroscopy and charge diffusivities as previously accomplished for imidazolium and ammonium ILs.32,44 The reduction in static permittivity in the tributyl-alkyl-phosphoniums is consistent with the absence of aggregate induced dynamics. These results highlight the importance of utilizing multiple experimental approaches to investigate the formation of long-lived, mesoscale aggregates, so that changes in physicochemical properties of ionic liquids are not misinterpreted as aggregation induced effects.

FIG. 8.

Static dielectric permittivity, εs, is reduced in the non-aggregating tributyl-alkyl-phosphoniums relative to the aggregating triethyl-alkyl-phosphonium series.

FIG. 8.

Static dielectric permittivity, εs, is reduced in the non-aggregating tributyl-alkyl-phosphoniums relative to the aggregating triethyl-alkyl-phosphonium series.

Close modal
FIG. 9.

Long-range ionic (dc) conductivity as a function of (a) inverse temperature and (b) temperature normalized by the calorimetric glass transition temperature, Tg,DSC. The ionic conductivity is reduced in the tributyl-alkyl-phosphoniums.

FIG. 9.

Long-range ionic (dc) conductivity as a function of (a) inverse temperature and (b) temperature normalized by the calorimetric glass transition temperature, Tg,DSC. The ionic conductivity is reduced in the tributyl-alkyl-phosphoniums.

Close modal

In conclusion, the mesoscopic organization, dynamics, and charge transport properties of a series of tetraalkylphosphonium ionic liquids were investigated by small and wide-angle x-ray scattering, broadband dielectric spectroscopy, dynamic-mechanical spectroscopy, and differential scanning calorimetry. A comparison of estimated aggregate size from the Tanford equation with the aggregate size obtained from the x-ray scattering pre-peak indicates a disruption of mesoscale aggregates in tributyl-alkyl-phosphonium ILs. The absence of aggregation is corroborated by the loss of the slow sub-α dielectric relaxation previously linked to the dynamics of mesoscale aggregates. The combination of techniques capable of probing the mesoscale structure with those capable of measuring the dynamic signature of such structures allows us to distinguish ionic liquids that exhibit an x-ray scattering pre-peak due to the formation of long-lived mesoscale aggregates. This distinction is an important step in elucidating the influence of mesoscale aggregation in the physicochemical properties of ionic liquids.

See supplementary material for differential scanning calorimetry heat flow curves and DMS shift factors.

Z.V. and J.S. acknowledge support by the National Science Foundation, the Division of Materials Research, Polymers Program through No. DMR-1508394. T.C. gratefully acknowledges financial support from the U.S. Army Research Office under Contract No. W911NF-17-1-0052. K.T. gratefully acknowledges financial support from the Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan (No. 26410104). The rheology measurements were conducted at the Oak Ridge National Laboratory’s Center for Nanophase Materials Sciences, which is a DOE Office of Science User Facility. The x-ray scattering measurements were performed at the Duke University Shared Materials Instrumentation Facility (SMIF), a member of the North Carolina Research Triangle Nanotechnology Network (RTNN), which is supported by the National Science Foundation (Grant No. ECCS-1542015) as part of the National Nanotechnology Coordinated Infrastructure (NNCI).

1.
O.
Russina
 et al.,
J. Phys. Chem. Lett.
3
,
27
(
2011
).
2.
R.
Hayes
,
G. G.
Warr
, and
R.
Atkin
,
Chem. Rev.
115
,
6357
(
2015
).
3.
R. D.
Rogers
and
K. R.
Seddon
,
Science
302
,
792
(
2003
).
4.
B.
Kirchner
,
Ionic Liquids
, Volume 290 of Topics in Current Chemistry, 1st ed. (
Springer-Verlag Berlin Heidelberg
,
2009
).
5.
M.
Paluch
,
Dielectric Properties of Ionic Liquids
, Advances in Dielectrics (
Springer International Publishing
,
2016
).
6.
C.
Austen Angell
,
Y.
Ansari
, and
Z.
Zhao
,
Faraday Discuss.
154
,
9
(
2012
).
7.
T.
Cosby
 et al.,
J. Chem. Phys.
147
,
234504
(
2017
).
8.
J. R.
Sangoro
,
Colloid Polym. Sci.
292
,
1933
(
2014
).
9.
O.
Russina
 et al.,
Top. Curr. Chem.
375
,
58
(
2017
).
10.
C. S.
Santos
 et al.,
J. Chem. Phys.
134
,
121101
(
2011
).
11.
J. C.
Araque
,
J. J.
Hettige
, and
C. J.
Margulis
,
J. Phys. Chem. B
119
,
12727
(
2015
).
12.
M. D.
Fayer
,
Chem. Phys. Lett.
616-617
,
259
(
2014
).
13.
J. J.
Hettige
 et al.,
J. Chem. Phys.
144
,
121102
(
2016
).
14.
L.
Gontrani
 et al.,
J. Phys. Chem. B
113
,
9235
(
2009
).
15.
C. S.
Santos
 et al.,
J. Chem. Phys.
134
,
064501
(
2011
).
16.
H. K.
Kashyap
 et al.,
J. Phys. Chem. B
117
,
15328
(
2013
).
17.
J. N.
Canongia Lopes
and
A. A.
Padua
,
J. Phys. Chem. B
110
,
3330
(
2006
).
18.
K.
Shimizu
 et al.,
J. Mol. Struct.: THEOCHEM
946
,
70
(
2010
).
19.
K.
Shimizu
and
J. N.
Canongia Lopes
,
Fluid Phase Equilib.
418
,
181
(
2016
).
20.
C.
Hardacre
 et al.,
J. Chem. Phys.
133
,
074510
(
2010
).
21.
H. V. R.
Annapureddy
 et al.,
J. Phys. Chem. B
114
,
16838
(
2010
).
22.
K.
Fujii
 et al.,
J. Chem. Phys.
135
,
244502
(
2011
).
23.
A.
Triolo
 et al.,
J. Chem. Phys.
130
,
164521
(
2009
).
24.
O.
Russina
 et al.,
J. Phys. Chem. B
113
,
8469
(
2009
).
25.
M.
Kofu
 et al.,
J. Phys. Chem. B
117
,
2773
(
2013
).
26.
O.
Russina
 et al.,
J. Phys. Chem. Lett.
8
,
1197
(
2017
).
27.
T.
Cosby
 et al.,
J. Phys. Chem. Lett.
8
,
3544
(
2017
).
28.
T.
Pott
and
P.
Meleard
,
Phys. Chem. Chem. Phys.
11
,
5469
(
2009
).
29.
T.
Murphy
,
R.
Atkin
, and
G. G.
Warr
,
Curr. Opin. Colloid Interface Sci.
20
,
282
(
2015
).
30.
D. W.
Bruce
 et al.,
J. Phys. Chem. B
121
,
6002
(
2017
).
31.
K.
Shimizu
,
C. E.
Bernardes
, and
J. N.
Canongia Lopes
,
J. Phys. Chem. B
118
,
567
(
2014
).
32.
J. R.
Sangoro
and
F.
Kremer
,
Acc. Chem. Res.
45
,
525
(
2012
).
33.
F.
Kremer
and
A.
Schönhals
,
Broadband Dielectric Spectroscopy
(
Springer-Verlag
,
Berlin
,
2003
).
34.
P. J.
Griffin
 et al.,
J. Chem. Phys.
142
,
084501
(
2015
).
35.
C.
Gainaru
 et al.,
Colloid Polym. Sci.
292
,
1913
(
2014
).
36.
H. K.
Kashyap
 et al.,
Faraday Discuss.
154
,
133
(
2012
).
37.
J. J.
Hettige
,
H. K.
Kashyap
, and
C. J.
Margulis
,
J. Chem. Phys.
140
,
111102
(
2014
).
38.
Faraday Discuss.
154,
189
(
2012
).
39.
C. A.
Tanford
,
The Hydrophobic Effect: Formation of Micelles and Biological Membranes
(
Wiley
,
1980
).
40.
Y. L.
Wang
 et al.,
J. Chem. Phys.
147
,
224502
(
2017
).
41.
P. J.
Griffin
 et al.,
J. Chem. Phys.
144
,
151104
(
2016
).
42.
J.
Sun
,
M.
Forsyth
, and
D. R.
MacFarlane
,
J. Phys. Chem. B
102
,
8858
(
1998
).
43.
D. R.
McFarlane
 et al.,
Electrochim. Acta
45
,
1271
(
2000
).
44.
P. J.
Griffin
 et al.,
J. Phys. Chem. B
118
,
783
(
2014
).

Supplementary Material