Probing Ion Exchange in the Triflic Acid-Guanidinium Triflate System: a Solid- State Nuclear Magnetic Resonance Study

Probing ion exchange in the triflic acid-guanidinium triflate system: a solid- state nuclear magnetic resonance study

Citation: Zhu, Haijin, MacFarlane, Douglas and Forsyth, Maria 2014, Probing ion exchange in the triflic acid-guanidinium triflate system: a solid-state nuclear magnetic resonance study, Journal of Physical Chemistry C, vol. 118, no. 49, pp. 28520-28526.

This is the accepted manuscript.

This document is the Accepted Manuscript version of a Published Work that appeared in final form in Journal of Physical Chemistry C, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see http://pubs.acs.org/doi/abs/10.1021/jp5101472

Downloaded from DRO: http://hdl.handle.net/10536/DRO/DU:30070195

DRO Deakin Research Online, Deakin University’s Research Repository Deakin University CRICOS Provider Code: 00113B Page 1 of 19 The Journal of Physical Chemistry

1 2 3 4 Probing Ion Exchange in the Triflic Acid - Guanidinium Triflate System: a 5 6 Solid-State NMR Study 7 8 9 Haijin Zhu,*,a Douglas MacFarlaneb and Maria Forsytha 10 11 a Institute for Frontier Materials and the ARC Centre of Excellence for Electromaterials 12 Science, Deakin University, Geelong, VIC 3216, Australia. E-mail: [email protected]; 13 Fax: 61 (03) 92446868; Tel: +61 (03) 52273696 14 15 b Department of Chemistry and the ARC Centre of Excellence for Electromaterials 16 Science, Monash University, Clayton, VIC 3800, Australia 17 18 19 * Corresponding author, Tel: 61 (03) 52273696; Fax: 61 (03) 92446868; Email: 20 [email protected] (H. Zhu) 21 22 ABSTRACT: Knowledge of ion exchange and transport behaviour in electrolyte materials is 23 24 crucial for designing and developing novel electrolytes for electrochemical device applications 25 26 such as fuel cells or batteries. In the present study, we show that, upon the addition of triflic 27 28 acid (HTf) to the guanidinium triflate (GTf) solid state matrix, several orders of magnitude 29 1 19 30 enhancement in the proton conductivity can be achieved. The static H and F solid-state NMR 31 results show that the addition of HTf has no apparent effect on the local molecular mobility of 32 33 the GTf matrix at room temperature. At higher temperatures, however, the HTf exhibits fast 34 35 ion exchange with the GTf matrix. The exchange rate, as quantified by our continuum T2 fitting 36 37 analysis, increases with increasing temperature. The activation energy for the chemical 38 exchange process was estimated to be 58.4 kJ/mol. It is anticipated that the solid-state NMR 39 40 techniques used in this study may be also applied to other organic solid state electrolyte systems 41 42 to investigate their ion exchange processes. 43 44 Keywords: organic ionic plastic crystal, T2 continuum fitting, proton conducting, 45 46 transport, activation energy. 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 2 of 19

1 2 3 4 1. Introduction 5 6 The rapidly growing demands for clean and sustainable energy sources to replace fossil energy 7 8 has necessitated the research and development on new powerful battery and fuel cell 9 technologies.1 As a vital part of fuel cells, proton-conducting electrolytes are currently 10 11 attracting significant interest from both industry and academic circles. The electrical 12 13 performance of a proton-conducting electrolyte is generally limited by both the transference 14 2 15 number and transport efficiency of the protons in the material. The transference number is 16 influenced not only by the doping concentration of the Brønsted acid, but also by the proton 17 18 dissociation and solvation. Yu et al. have recently studied the proton conductivity of 19 20 bis(trifluoromethanesulfonyl) imide (HTFSI) solutions in 1-butyl-3-methylimidazolium 21 22 bis(trifluoromethanesulfonyl) imide (BMITFSI) ionic liquid at various concentration from ~ 23 3 24 0.1 to as high as ~ 1.0 mol/L. At low HTFSI concentrations, the proton conductivity of the 25 solution was found to initially increase with increasing HTFSI, with a maximum appearing at 26 27 a concentration of approximately 0.1 mol/L at room temperature. Further increase of the acid 28 29 concentration leads to a fast drop of the conductivity. This was attributed to the higher proton 30 - + 31 association with TFSI and therefore lower solvated H concentration at high HTFSI 32 concentrations. Similar trends for the HTFSI/1-(1-Butyl-3-imidazolio)propane-3-sulfonate 33 34 (BIm3S) system and HTFSI/1-ethyl-3methylimidazolium bis-(triflluoromethanesulfonyl) 35 36 (EMImTFSI) have been observed by Yoshizawa et al. in 2004 and Jia et al. in 2009, 37 4-5 38 respectively. Besides the transference number, the proton transport behaviour is another 39 40 important and controlling factor for the conductivity of the electrolyte material. There have 41 been numerous efforts dedicated to elucidating the specific proton conduction mechanisms at 42 43 a molecular level.1-3, 5-8 So far, there are two plausible models for the proton transport 44 45 mechanism: the vehicle-type model where the protons migrate through the medium along with 46 47 a ‘vehicle’ or proton solvent such as H3O+ etc., and the Grotthuss-type model where protons 48 move via local hopping from a proton donating site to a proton accepting site through the 49 50 breaking and formation of hydrogen bonds.1, 9-11 In many cases, both mechanisms contribute 51 52 cooperatively to the overall proton conductivities. For example, in the sulfonic acid containing 53 54 polymer electrolyte membranes (PEMs), the vehicle-type mechanism is dominant in the 55 presence of water, whereas the Grotthus-type mechanism also contributes to the proton 56 57 conductivity through the forming and breaking of the hydrogen bonds among the water 58 59 molecules.9 60

ACS Paragon Plus Environment Page 3 of 19 The Journal of Physical Chemistry

1 2 3 4 Room-temperature ionic liquids (ILs) have been shown to be promising candidates as the 5 proton-conducting electrolyte materials due to their remarkable properties, such as high ionic 6 7 conductivity, high thermal stability, low vapour pressures, non-flammability, good solvation 8 12 9 of many organic and inorganic chemicals, and ability to be used under anhydrous conditions. 10 11 Many families of ionic liquids display rotatory phases and/or plastic crystalline phases in their 12 solid state.13-14 Rotator phases are those where one or more of the ions can rotate on its 13 14 crystallographic site, but without any significant translational motion.15 As well as resulting in 15 16 the plasticity of the materials, this local mobility is highly desirable because it is believed to 17 16 18 result in the creation of vacancies that facilitate fast proton transport in the materials. 19 20 Protic ionic liquids (PILs) represent a proton-conducting sub-class of the ionic liquids family 21 22 and have attracted much attention as next-generation proton conductors for fuel cell 23 + 24 applications. The guanidinium cation (C(NH2)3 ), for example, is formed by protonation of the 25 imine group on the basic guanidine molecule, with subsequent redistribution of the electron 26 27 density of the double bond to yield three equivalent C–N bonds and a resonance stabilization 28 29 of the whole entity as an ion with the three-fold rotational symmetry. It is relatively stable and 30 - 31 can readily form salts, and some ILs, with stable anions such as triflate (Tf, CF3SO3 ), 32 - - 6 dicyanamide (DCA, N(CN)2 ), and thiocyanate (SCN ), etc. The six dissociable protons per 33 34 cation make it an ideal candidate for a proton-conducting electrolyte in both the solid and liquid 35 36 states. 37 38 In our previous work, we have investigated the proton conductivity in the aprotic versions of 39 40 organic ionic plastic crystals (OIPCs) by adding acids of various strengths and compositions.17- 41 18 42 High proton conductivity was achieved in the plastic crystal phase of various acid-containing 43 44 OIPCs. In line with the emergence of these new proton-conducting materials, there is an 45 increasing need to fully understand the nature and mechanisms of the ionic transport and related 46 47 molecular dynamics in these materials. In our most recent work related to this study,19 we have 48 49 investigated the solid-state dynamics of a protic organic solid which we hypothesised may 50 51 display OIPC behaviour, guanidinium triflate (Figure 1), and its mixtures with triflic acid. It 52 was initially hypothesised that diffusion of the additional proton from the doped acid would 53 54 benefit from the six dissociable protons on the guanidinium cations, although we have shown 55 56 that, at room temperature, this behaviour is not evident in this material despite achieving a very 57 58 high conductivity with relatively low acid concentrations. In this study, we will show the 59 - 60 evidence of Tf anions from the doped triflic acid exchanging with the guanidinium triflate

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 4 of 19

1 2 3 4 matrix at higher temperatures. The rate of exchange increases with increasing temperature as 5 suggested by NMR T2 continuum fitting analysis. 6 7 8 9 10 2. Experimental 11 12 2.1. Sample Preparation. The guanidinium triflate (GTf) was synthesized by the reaction of 13 14 triflic acid (HTf) with guanidinium carbonate. The detailed sample preparation procedures 15 19 16 have been described elsewhere. The HTf acid doped samples were prepared by dissolving 17 the doped amount of triflic acid in the GTf, homogenized by adding water and then using a 18 19 rotary evaporator to remove the solvent, followed by further vacuum drying to remove any 20 21 residual water. These samples are labelled as ‘Doped’ GTf samples, although we accept that 22 23 this level of acid is significantly higher than usual doping levels. The neat material is a white 24 25 powder, which easily forms into a rigid solid pellet upon pressing; whereas the doped samples 26 were soft waxy samples for which the pellets were readily deformed under pressure. 27 28 29 30 31 32 33 34 GTf HTf 35 36 37 Figure 1. Structural formulae of guanidinium trifluoromethane sulfonate (GTf) and triflic acid 38 39 (HTf) 40 41 42 43 2.2. Differential Scanning Calorimetry (DSC). DSC was performed on a TA-Q100 instrument. 44 45 The samples were first heated up to 220 oC to completely eliminate any thermal history effects. 46 o o 47 Then the temperature was decreased to –100 C at 10 C/min and subsequently increased to 48 o o 49 220 C at 10 C/min, and the phase transitions were recorded during the second heating scans. 50 51 2.3. AC Impedance Spectroscopy. The ionic conductivity was measured by the method of ac 52 ® 53 impedance spectroscopy using a Solartron SI1260 impedance/gain phase analyser, which was 54 ® 55 connected to a Solartron 1296 dielectric interface with frequency range from 1Hz to 1MHz. 56 A pair of gold coated stainless steel blocking electrodes was used to avoid the etching and 57 58 oxidation from these acid containing materials. The whitish powder samples were pressed into 59 60 pellets and then sandwiched between the electrodes for the conductivity measurement. All samples were packed and air-tightly sealed into a testing cell in a nitrogen atmosphere. Sample

ACS Paragon Plus Environment Page 5 of 19 The Journal of Physical Chemistry

1 2 3 4 temperatures were controlled using a Eurotherm controller (model 2204), and measured with a 5 type T thermal couple attached to one of the electrodes. The temperature was increased from 6 7 25 oC to 30 oC and then up to 140 oC with an interval of 10 oC. At each temperature, samples 8 9 were equilibrated for 15 min before the impedance measurement was taken. 10 11 2.4. Solid-Sate NMR: 1H and 19F Static NMR Experiments. All the 1H and 19F experiments 12 1 13 were performed on a Bruker Avance III 300 MHz wide bore NMR spectrometer ( H Larmor 14 15 frequency of 300.13 MHz). A 4mm double resonance Magic Angle Spinning (MAS) probe 16 head was used to record the spectra from stationary powder samples. For both 1H and 19F 17 18 experiments, the 90o pulse lengths were 2.5 µs, and the recycle delays were 30 s to allow the 19 20 system time to recover to equilibrium. The sample temperatures for the variable temperature 21 20-21 22 experiments were calibrated with lead nitrate, using the method described in literature. 23 24 25 26 3. Results and Discussion 27 28 29 3.1. Thermal Properties and Proton Conductivity. Thermal properties and phase transition 30 31 behaviour of a solid-state electrolyte material are important because they are closely related to 32 the conductivity of the materials.22 Figure 2a shows the DSC melting endotherms of the pure 33 34 GTf and 4 mol% HTf doped GTf samples. The heat flow was normalized by the sample weight 35 36 of the GTf component for comparison. Both samples show similar phase transition 37 38 temperatures and enthalpy. These results imply that the thermal property of the GTf is 39 essentially unaffected by the addition of the acid, which is most probably because they are 40 41 located in different phases. At higher concentrations it is likely that liquidus/eutectic behaviour 42 43 would be more apparent and a eutectic transition would appear at lower temperatures, however 44 45 at these low concentrations these traditional two-component effects do not appear to be 46 - 47 significant. Nonetheless, we will show in the following discussion that the Tf anions from the 48 HTf exhibit a fast exchange process with the GTf matrix at elevated temperatures. Two 49 50 endothermic peaks can be observed from the thermograms of both samples. The main peak at 51 o 52 higher temperatures, with peak temperature of about 160 C is attributed to the melting of the 53 54 GTf. An additional small but distinct solid-solid phase transition peak can be identified at a 55 temperature of about 115 oC. The solid phases of the GTf were labelled as Phase I and Phase 56 57 II from higher to lower temperatures, using the phase nomenclature suggested in the 58 59 literature.11 60

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 6 of 19

1 2 3 4 5 (a) (b) 6 GTf + 4 mol% HTf 7 8 9 10 11 12 13 14 Pure GTf 15 16 17 18 19 20 21 22 23 24 25 Figure 2. (a) DSC thermograms of the neat GTf and 4 mol% HTf doped GTf samples. (b) 26 27 Conductivity of the neat GTf and the HTf doped samples with various concentrations measured 28 o o 29 at temperature range from 25 C to 140 C. 30 31 The solid-solid phase transitions in plastic crystals are generally explained by the onset of 32 33 rotational motions of particular groups of molecules within a slightly expanded crystal 34 lattice.22-24 It is believed that activated rotational motions in the plastic crystal phase are also 35 36 responsible for the increased ionic conductivity through the creation of vacancies and extended 37 38 defects, and therefore make the plastic crystal materials promising candidates for all-solid 39 13, 25 40 electrolyte applications. Figure 2b shows the conductivity of the HTf doped GTf samples 41 measured in the temperature of solid Phase II and Phase I as identified from the DSC 42 43 thermograms in Figure 2a. In phase II where the temperature is relatively low, it is found that 44 45 at low doping concentration of 1 mol%, the composite shows only a slight increase in 46 47 conductivity compared to the pure GTf sample, whereas at doping ratio of 2 mol% or higher, 48 49 the doped samples show a drastic increase (by several orders of magnitude) in conductivity. 50 This behaviour has been discussed in our previous work and was attributed to the percolation- 51 52 dominated conduction mechanism with a percolation threshold of 1.8 mol%, 1.8 mol%, 1.7 53 o o o 19 54 mol% for the temperature of 25 C, 80 C and 120 C, respectively. In this study, we show 55 56 that the conductivity in Phase I exhibits different concentration dependence than in Phase II. 57 At temperatures above 120 oC, the conductivity increases with HTf concentration, but the step 58 59 increase in the conductivity with concentration which is observed at lower temperatures is not 60 obvious, as shown in Figure 2b, suggesting that percolation mechanism is weakened at higher

ACS Paragon Plus Environment Page 7 of 19 The Journal of Physical Chemistry

1 2 3 4 temperatures, and the GTf matrix gradually contributes to the overall conductivity. Moreover, 5 it is also very interesting to notice that for all the doped samples, the conductivity even shows 6 7 a slight decrease when the temperature is increased from 100 oC to 110 oC, and this decrease 8 9 becomes more prominent with increasing HTf concentration. The reason for this behaviour is 10 11 not yet understood. However, it is very likely that this slight drop in conductivity is a result of 12 the HTf doping, and may also be related to the solid-solid phase transition from phase II to 13 14 phase I at 115 oC, as measured from DSC data. The 10 oC temperature difference between the 15 16 DSC and conductivity measurements might be explained by the different dynamic processes 17 18 probed by the different methods. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Figure 3. Selected static 1H (a) and 19F (b) NMR spectra of the neat GTf and 4 mol% HTf 36 37 doped GTf samples measured at different temperatures. 38 39 3.2. Cation and Anion Mobility. The 1H and 19F static NMR spectra of organic solids usually 40 41 consist of a broad and featureless peak due to the strong homonuclear dipole-dipole interactions. 42 43 Since these interactions can be partially or completely averaged out by dynamics, the line width 44 45 of these spectra retains useful information about molecular motions, and is therefore well suited 46 26-27 1 47 for studies of local molecular dynamics. In the GTf sample, H is only present in the cation, 48 and 19F is only present in the anion, which allows us to analyse the motions of cations and 49 50 anions separately. Figure 3a shows the stack plot of the static 1H spectra of the GTf and 4 mol% 51 1 52 HTf doped samples measured at different temperatures. At room temperature (290 K), H 53 54 spectrum of the pure GTf shows a characteristic line shape, known as a Pake doublet, which is 55 generally observed in a rigid system with strong 1H dipolar couplings. This 1H line shape 56 57 suggests that NH2 protons in the pure GTf sample are strongly coupled with each other and that 58 59 these groups are therefore not undergoing any form of fast dynamics that would disrupt this 60 coupling. This result is consistent with a previous crystallographic study on the same system

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 8 of 19

1 2 3 4 which has shown extensive hydrogen bonding between the six protons of each guanidinium 5 cation and the six oxygen lone electron pairs from the triflate anions at room temperature.28 6 7 8 The spectrum of the GTf + 4 mol% sample is composed of a broad component superimposed 9 with a narrow component. This narrow component is attributed to the fast 1H from the doped 10 11 HTf. The integration of the narrow component accounts for approximately 0.7% of the total 12 13 area from the curve deconvolution analysis, which is equivalent to about 4 mol% of HTf 14 15 assuming that the narrow component is purely attributed to the HTf (one mole GTf contains 16 six mole protons); this ratio fits very well with the doping ratio of the HTf. Moreover, the line 17 18 shape of the broad component of the HTf doped sample is essentially identical to that of the 19 20 pure GTf sample. A similar result is observed for the 19F NMR spectra at room temperature 21 22 shown in Figure 3b. The incoming triflate anion from the doped HTf exhibits much faster 23 24 dynamics than the triflate anion in the GTf matrix. These results suggest that the addition of 25 HTf has no apparent effect on the 1H local mobility of the GTf at room temperature. 26 27 28 At higher temperatures, however, the doped system shows apparently different behaviour 29 compared to room temperature. At 360 K, for example, both the 1H and 19F spectra show that 30 31 the broad component of the doped sample exhibits a noticeable difference compare to the pure 32 33 GTf counterpart, suggesting that dynamics of the cations and anions in the GTf matrix are 34 35 modified by the doped HTf. Moreover, the relative ratio between the narrow component and 36 broad component is significantly increased compare to the room temperature for both the 1H 37 38 and 19F spectra. This increase is attributed to the mobilized Tf- in the matrix, as can be identified 39 40 from the narrow components in the 1H and 19F spectra of the pure GTf sample. 41 42 3.3. Anion exchange between the doped HTf and GTf matrix. 43 44 45 In section 3.2, we have shown that the dynamics of the GTf molecular ions is modified by the 46 doped HTf at higher temperatures, which might indicate interaction between the two 47 48 components. The purpose of designing this GTf/HTf composite was based on the hypothesis 49 50 that diffusion of the additional proton from the doped acid would benefit from the six 51 52 dissociable protons on the guanidinium cations. Therefore, proton exchange between the HTf 53 and GTf matrix is highly desirable. Although we have shown that at room temperature, no 54 55 evidence of exchange can be observed in this material, despite achieving a very high 56 57 conductivity, at higher temperatures the two components in this composite do show very 58 59 interesting interactions (as shown in Figure 3). This section aims to probe the exchange process 60 between the two components and to quantify the rate of exchange at different temperatures.

ACS Paragon Plus Environment Page 9 of 19 The Journal of Physical Chemistry

1 2 3 4 There has been many examples in literature using 2D Magic Angle Spinning (MAS) methods 5 to probe ion (molecular) exchange in the solid state.29-30 These methods generally required 6 7 spectra resolution between the two sites, or in other words, the chemical shift for the two spices 8 9 has to be different in order to probe their exchange behaviour. These 2D MAS methods 10 1 19 11 therefore cannot be readily used in the present system because the both the H and F resonance 12 lines are of identical chemical shifts for the GTf matrix and HTf dopant (single 1H and 19F 13 14 peaks was observed for the 4 mol% doped sample). In the present study, we proposed, for the 15 16 first time to the best of our knowledge, an alternative method of using the T2 continuum fitting 17 18 to probe the exchange of the species with identical chemical shifts. In addition to the line shape 19 analysis that has been discussed in section 3.2, the T2 distribution spectra obtained from the 20 21 continuum fitting of the static NMR Free Induction Decay (FID) data can provide us 22 23 quantitative information on the composition and dynamics of the materials. Most of the 24 25 previous T2 fitting analysis of the FID has been performed in terms of discrete two- or three- 26 27 component models. However, this requires a priori assumption of the number of species before 28 analysis. We have recently developed a continuum fitting method based on a mixed Gaussian 29 30 and exponential kernel function which will better reflect the true property of materials 31 32 compared to the traditional methods.26 33 34 Figure 4 shows comparison of the T2 spectra of the pure GTf and the GTf + 4 mol% HTf 35 36 samples obtained from the continuum fitting of the static NMR FIDs. The fast decay 37 38 components were fitted using the Gaussian kernel function, which is based on the assumption 39 40 that the T2 relaxation of the rigid component is dominated by the Gaussian distribution of multi- 41 nuclear couplings with cluster size N>6, and therefore the FID behaviour can be well described 42 43 by a Gaussian function.26 However, in this particular GTf system, 19F chemical shift anisotropy 44 45 (CSA) contributes significantly to the T2 relaxation in addition to dipolar couplings. This 46 47 causes a significant error in the fitting of the rigid components (with short T2s). However, the 48 mobile components, with longer T2s, were fitted using a separate exponential kernel function, 49 26 50 which is a good representation of the T2 relaxation behaviour of the mobile component. The 51 -4 52 fitted T2 spectra in the region of 10 s or longer time scale is therefore reliable. 53 19 54 The area of the peaks in the T2 spectra represents the population of the F in the corresponding 55 56 components. Figure 5 shows the percentage of the integration of the mobile component relative 57 58 to the integration of the overall spectra for both the pure GTf and GTf + 4 mol% HTf samples. 59 -4 -3 60 It can be seen that the percentage of the mobile component (in the region of 10 s to 10 s) in both samples is rather constant at temperatures below 320K, and begins to increase with

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 10 of 19

1 2 3 4 temperature at 330K and above. The difference between the percentages of the mobile 5 components of the two samples allows us to estimate the number of Tf- species which are 6 7 mobilized by the addition of the HTf acid. As indicated by the blue open circles in Figure 5, 8 9 the difference between the mobile fractions for the two samples increases moderately at 10 11 temperature below 350 K. At 360 K, however, it shows an abrupt increase, indicating that a 12 significant amount of Tf- in the matrix was mobilized by the presence of HTf at higher 13 14 temperatures. The total percentage of the mobile component is about 5 mol%, which means 15 16 that 1 mol% of the Tf- was mobilized at this temperature. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 19 Figure 4. F NMR T2 distribution spectra of the pure GTf and GTf + 4 mol% HTf samples at 37 38 different temperatures obtained from continuum T2 fitting of the static NMR FIDs. The red 39 40 arrow indicates the change of T2 of the mobile component (longer T2) with increasing time. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment Page 11 of 19 The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Figure 5. The mobile anion mole concentration in the pure GTf and GTf + 4 mol% HTf samples 29 30 at different temperatures. The dashed line in the figure indicates the doping ratio of HTf which 31 32 is 4 mol%. The black squares and the red triangles are the mobile anion percentage in the GTf 33 + 4 mol% HTf sample and the pure GTf sample respectively; the blue open circles represent 34 35 the difference between the two samples. 36 37 38 39 40 At a fundamental level, T2 relaxation is sensitive to the fluctuation of the local magnetic field 41 42 (molecular motions) at both the low frequencies (around 0 Hz) and/or Larmor frequency, 43 31 * according to the BPP theory. If the T2 , which is defined as the apparent T2 relaxation time, 44 * 45 is dominated by the molecular motions, it is expected that the T2 will increase monotonically 46 47 with increasing temperature (decreasing correlation time). However, it is very interesting to 48 49 observe that, in Figure 4, the T2 of the mobile component in GTf + 4mol% HTf sample 50 decreases with increasing temperature. The population weighted T2 values of the mobile 51 52 component is plotted against temperature and shown in Figure 6. A monotonic decrease with 53 54 increasing temperature is observed. This result suggests that there must be some other 55 56 mechanism that dominates the T2 relaxation instead of molecular dynamics. The only possible 57 58 mechanism in this case would be the molecular chemical exchange between the GTf matrix 59 and the HTf mobile phase. This is very interesting because our previous NMR and DSC data 60

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 12 of 19

1 2 3 4 clearly show that the majority of the guanidinium sublattice remains rigid up to 360 K, which 5 is the highest temperature achievable for our NMR measurements at present. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Figure 6. Plot of the population weighted T2 as a function of temperature. 39 40 In the fast exchange limit k >> ∆ω , where k is the exchange rate, and ∆ω is the chemical 41 ex ex 42 * 32 43 shift difference between the two exchange sites, the apparent relaxation time T2 is given by: 44 45 T * =1/(1/T + k ), [eq. 1] 46 2 2 −1 47 - 48 where k−1 is the first order transition rate constant from mobile to rigid Tf anions, 49 50 k1 − → − 51 Tf Tf , and the exchange rate constant k = k + k , T is the spin-spin rigid ← mobile ex 1 −1 2 52 k−1 53 54 * relaxation time of the mobile component when the exchange rate is zero, and T2 is the time 55 56 constant for both the spin-spin relaxation and chemical exchange. Consider a simple first order 57 58 59 60

ACS Paragon Plus Environment Page 13 of 19 The Journal of Physical Chemistry

1 2 3 k1 4 − → − chemical exchange process Tf Tf , the rate constants k and k are related to the 5 rigid ← mobile −1 1 6 k−1 7 − − 8 concentration of Tfrigid and Tfmobile by the following equation: 9 10 − [Tfrigid ] k 11 = −1 . [eq. 2] 12 − [Tfmobile ] k1 13 14 * 15 Assuming that the exchange is negligible at room temperature (T2 = T2 ), k−1 , k1 , and kex for 16 17 different temperatures was calculated and listed in Table 1. It is seen that the exchange rate is 18 19 rather low at room temperature, and increases abruptly to 223 Hz at 360 K, while the GTf 20 21 matrix is still in the solid phase II. 22 23 24 25 Table 1. The apparent T2 relaxation time and the calculated rate constant for the chemical 26 27 exchange process in the GTf + 4 mol% HTf system at different temperatures. 28 29 30 T/K 290 300 310 320 330 340 350 360 31 32 * 0.54 0.54 0.53 0.52 0.51 0.49 0.45 0.40 33 T2 /mS 34 35 a b 36 k−1 /HZ 0 0 10 22 42 58 126 209 37 38

39 k1 /HZ 0 0 0.4 0.9 1.8 2.7 6.8 14 40 41 42 kex /HZ 0 0 10.4 22.9 43.8 60.7 132.8 223 43 44 45 a k = 0 46 The value of −1 Hz is given based on the assumption that the exchange is negligible at 47 room temperature; 48 49 b 50 The detection limit of the exchange rate k−1 is roughly estimated to be 2 Hz. The value of 0 51 52 here means the exchange k−1 ≤ 2 Hz at 300 K. 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 14 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Figure 7. Plot of the chemical exchange rate as a function of temperature. The black squares 30 31 are the experimental data points, and the red line is a fitted line to the Arrhenius equation. 32 The activation energy as extracted from the fitting is 58.4 kJ/mol. 33 34 35 36

37 The dependence of exchange rate kex on temperature T was empirically described by the 38 39 Arrhenius equation:33 40 41 = −∆ 42 kex Aexp( Ea / RT) , [Eq. 3] 43 44 where A is a pre-exponential factor, R is the gas constant, ∆Ea is the activation energy. The 45 46 temperature dependence of the exchange rate was fitted by Eq. 3 and shown in Figure 7. The 47 48 activation energy, as estimated from the fitting of the experimental data, is 58.4 kJ/mol. This 49 50 value is apparently greater than the ion-exchange activation energy in ordinary aqueous 51 34-35 52 solutions which is about 5 ~ 20 kJ/mol. On the other hand it is comparable to the typical 53 activation energy values for chemical exchange between solids and liquids, which are usually 54 55 larger than 40 kJ/mol.36 Due to this high activation energy, the exchange rate is therefore 56 57 strongly dependent on temperature. 58 59 60

ACS Paragon Plus Environment Page 15 of 19 The Journal of Physical Chemistry

1 2 3 4 From the above observations, we can now suggest the following detailed picture of both the 5 structure and dynamics of the HTf doped GTf system: At room temperature, the H+ and Tf- 6 7 ions from the added HTf most likely exist along the grain boundaries of the bulk GTf matrix, 8 9 and dominate the conductivity. However, this added HTf phase has a negligible effect on the 10 11 ion dynamics in the GTf matrix itself. With increasing temperature the ions in the bulk matrix 12 become more mobile, particularly the Tf- anion, and the presence of the added HTf increases 13 14 the number of matrix ions that are mobilised, as indicated by the difference between the molar 15 16 ratios of the mobile component in the doped and pure GTf samples (see Figure 5). According 17 - 18 to the T2 continuum fitting analysis, approximately 1 mol% of the Tf in the GTf matrix was 19 mobilized by the doped Tf- at 360K. Besides this enhancement in dynamics, it is also observed 20 21 that the mobile Tf- anion exhibits ion exchange (on a time scale of 10-200Hz) with the Tf- 22 23 anions in the GTf matrix. The exchange rate increases with temperature and follows the classic 24 25 Arrhenius behaviour. It is worth mentioning that, although the T2 continuum fitting analysis is 26 19 1 27 very promising for probing the F exchange, unfortunately we cannot do the same for H NMR, 28 1 because the H T2 relaxation is dominated by dipolar couplings, as can be seen from the typical 29 30 Pake doublet spectra. The 1H FID therefore cannot be fitted well with either a Gaussian or 31 32 Exponential kernel function. 33 34 35 36 37 38 39 4. Conclusions 40 41 This study aimed to understand the ionic exchange and transport behaviour in the HTf doped 42 43 GTf system. The main findings of this study can be summarized as follows: 44 45 1. Several orders of magnitude enhancement in the proton conductivity can be achieved upon 46 47 the addition of HTf to the GTf matrix, especially at lower temperatures. 48 49 2. Interaction between the GTf matrix and the doped HTf phase is negligible at room 50 51 temperature, as suggested by the essentially identical broad static 1H and 19F NMR spectra 52 53 components between the pure and the acid doped samples, as well as the exchange rate of 0 Hz 54 55 determind here. 56 - 57 3. An increasing number of Tf anions from the matrix become mobilized, both rotational and 58 - 59 diffusional, by the presence of HTf at higher temperatures. For example, 1 mol% of the Tf was 60 mobilized by HTf at 360 K.

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 16 of 19

1 2 3 4 4. Although the NMR and DSC data clearly show that the majority of the guanidinium 5 sublattice remains rigid up to 360 K, fast Tf- anion exchange was observed at 360 K between 6 7 the GTf matrix and the mobile Tf- anions which are presumably more associated with the HTf. 8 9 The activation energy for the chemical exchange process was estimated to be 58.4 kJ/mol. 10 11 12 13 Acknowledgement: Prof. Maria Forsyth and Prof. Doug MacFarlane wish to acknowledge the 14 15 financial support from the Australian Research Council (ARC) through the Australian Laureate 16 17 program funding FL110100013. ARC is also acknowledged for funding Deakin University’s 18 19 Magnetic Resonance Facility through LIEF grant LE110100141. 20 21 22 23 Reference 24 25 26 Primary Sources 27 Secondary Sources 28 29 Uncategorized References 30 31 1. Rikukawa, M.; Sanui, K., Proton-Conducting Polymer Electrolyte Membranes Based on 32 Hydrocarbon Polymers. Progress in Polymer Science 2000, 25, 1463-1502. 33 2. Graf, R., New Proton Conducting Materials for Technical Applications: What Can We Learn 34 Solid State Nuclear Magnetic Resonance 40 35 from Solid State Nmr Studies? 2011, , 127-133. 36 3. Yu, L.; Pizio, B. S.; Vaden, T. D., Conductivity and Spectroscopic Investigation of 37 Bis(Trifluoromethanesulfonyl)Imide Solution in Ionic Liquid 1-Butyl-3-Methylimidazolium 38 Bis(Trifluoromethanesulfonyl)Imide. The Journal of Physical Chemistry B 2012, 116, 6553-6560. 39 4. Yoshizawa, M.; Ohno, H., Anhydrous Proton Transport System Based on Zwitterionic Liquid 40 and Htfsi. Chemical Communications 2004, 1828-1829. 41 5. Jia, L.; Nguyen, D.; Halleý, J. W.; Pham, P.; Lamanna, W.; Hamrock, S., Proton Transport in 42 43 Htfsi–Tfsi–Emi Mixtures: Experiment and Theory. Journal of The Electrochemical Society 2009, 156, 44 B136-B151. 45 6. Zhao, Z.; Ueno, K.; Angell, C. A., High Conductivity, and “Dry” Proton Motion, in Guanidinium 46 Salt Melts and Binary Solutions. The Journal of Physical Chemistry B 2011, 115, 13467-13472. 47 7. Kreuer, K.-D.; Paddison, S. J.; Spohr, E.; Schuster, M., Transport in Proton Conductors for Fuel- 48 Cell Applications: Simulations, Elementary Reactions, and Phenomenology. Chemical Reviews 2004, 49 104, 4637-4678. 50 51 8. Alberti, G.; Casciola, M.; Massinelli, L.; Bauer, B., Polymeric Proton Conducting Membranes for 52 Medium Temperature Fuel Cells (110–160°C). Journal of Membrane Science 2001, 185, 73-81. 53 9. Zuo, Z.; Fu, Y.; Manthiram, A., Novel Blend Membranes Based on Acid-Base Interactions for 54 Fuel Cells. Polymers 2012, 4, 1627-1644. 55 10. Murch, G. E., The Haven Ratio in Fast Ionic Conductors. Solid State Ionics 1982, 7, 177-198. 56 11. Chezeau, J. M.; Strange, J. H., Diffusion in Molecular Crystals. Physics Reports 1979, 53, 1-92. 57 58 12. Susan, M. A. B. H.; Noda, A.; Mitsushima, S.; Watanabe, M., Bronsted Acid-Base Ionic Liquids 59 and Their Use as New Materials for Anhydrous Proton Conductors. Chemical Communications 2003, 60 0, 938-939.

ACS Paragon Plus Environment Page 17 of 19 The Journal of Physical Chemistry

1 2 3 13. Jin, L.; Nairn, K. M.; Forsyth, C. M.; Seeber, A. J.; MacFarlane, D. R.; Howlett, P. C.; Forsyth, M.; 4 5 Pringle, J. M., Structure and Transport Properties of a Plastic Crystal Ion Conductor: 6 Diethyl(Methyl)(Isobutyl)Phosphonium Hexafluorophosphate. Journal of the American Chemical 7 Society 2012, 134, 9688-9697. 8 14. Pringle, J. M.; Golding, J.; Forsyth, C. M.; Deacon, G. B.; Forsyth, M.; MacFarlane, D. R., Physical 9 Trends and Structural Features in Organic Salts of the Thiocyanate Anion. Journal of Materials 10 Chemistry 2002, 12, 3475-3480. 11 Electrochemical Aspects of Ionic Liquids 12 15. . Second ed.; John Wiley & Sons: New Jersey, 2011. 13 16. Sherwood, J. N., The Plastically Crystalline State: Orientationally Disordered Crystals. Wiley: 14 New York, 1979. 15 17. Rana, U. A.; Vijayaraghavan, R.; MacFarlane, D. R.; Forsyth, M., Plastic Crystal Phases with High 16 Proton Conductivity. Journal of Materials Chemistry 2012, 22, 2965-2974. 17 18. Rana, U. A.; Vijayaraghavan, R.; MacFarlane, D. R.; Forsyth, M., An Organic Ionic Plastic Crystal 18 Electrolyte Based on the Triflate Anion Exhibiting High Proton Transport. Chemical Communications 19 20 2011, 47, 6401-6403. 21 19. Zhu, H.; Rana, U. a.; Ranganathan, V.; Jin, L.; O'Dell, L. A.; MacFarlane, D. R.; Forsyth, M., 22 Proton Transport Behaviour and Molecular Dynamics in the Guanidinium Triflate Solid and Its Mixtures 23 with Triflic Acid. Journal of Materials Chemistry A 2014, 2, 681-691. 24 20. Guan, X.; Stark, R. E., A General Protocol for Temperature Calibration of Mas Nmr Probes at 25 Arbitrary Spinning Speeds. Solid State Nuclear Magnetic Resonance 2010, 38, 74-76. 26 21. Bielecki, A.; Burum, D. P., Temperature Dependence Of207pb Mas Spectra of Solid Lead 27 28 Nitrate. An Accurate, Sensitive Thermometer for Variable-Temperature Mas. Journal of Magnetic 29 Resonance, Series A 1995, 116, 215-220. 30 22. MacFarlane, D. R.; Forsyth, M.; Izgorodina, E. I.; Abbott, A. P.; Annat, G.; Fraser, K., On the 31 Concept of Ionicity in Ionic Liquids. Physical Chemistry Chemical Physics 2009, 11, 4962-4967. 32 23. Fraser, K. J.; Izgorodina, E. I.; Forsyth, M.; Scott, J. L.; MacFarlane, D. R., Liquids Intermediate 33 between "Molecular" and "Ionic" Liquids: Liquid Ion Pairs? Chemical Communications 2007, 0, 3817- 34 35 3819. 36 24. Golding, J.; Hamid, N.; MacFarlane, D. R.; Forsyth, M.; Forsyth, C.; Collins, C.; Huang, J., N- 37 Methyl-N-Alkylpyrrolidinium Hexafluorophosphate Salts: Novel Molten Salts and Plastic Crystal 38 Phases. Chemistry of Materials 2001, 13, 558-564. 39 25. Pringle, J. M.; Howlett, P. C.; MacFarlane, D. R.; Forsyth, M., Organic Ionic Plastic Crystals: 40 Recent Advances. Journal of Materials Chemistry 2010, 20, 2056-2062. 41 26. Zhu, H.; Huinink, H. P.; Magusin, P. C. M. M.; Adan, O. C. G.; Kopinga, K., T2 Distribution Spectra 42 43 Obtained by Continuum Fitting Method Using a Mixed Gaussian and Exponential Kernel Function. 44 Journal of Magnetic Resonance 2013, 235, 109-114. 45 27. Zhu, H.; Graf, R.; Hou, G.; Zhao, Y.; Wang, D.; Spiess, H. W., Solid-State Nmr Characterization 46 of the Multiphase Structure of Polypropylene in-Reactor Alloy. Macromolecular Chemistry and Physics 47 2010, 211, 1157-1166. 48 28. Russell, V. A.; Etter, M. C.; Ward, M. D., Layered Materials by Molecular Design: Structural 49 Enforcement by Hydrogen Bonding in Guanidinium Alkane- and Arenesulfonates. Journal of the 50 51 American Chemical Society 1994, 116, 1941-1952. 52 29. Liu, Q.; Peng, B.; Shen, M.; Hu, B.; Chen, Q., Polymer Chain Diffusion and Li+ Hopping of 53 Poly(Ethylene Oxide)/Liasf6 Crystalline Polymer Electrolytes as Studied by Solid State Nmr and Ac 54 Impedance. Solid State Ionics 2014, 255, 74-79. 55 30. Fischer, G.; Kleinpeter, E., Application of 2d Exsy Nmr Spectroscopy to the Study of the 56 Dynamic Behaviour of Aroylcyanoketene-S,S-Dimethylacetals. Magnetic Resonance in Chemistry 1991, 57 58 29, 204-206. 59 31. Bloembergen, N.; Purcell, E. M.; Pound, R. V., Relaxation Effects in Nuclear Magnetic 60 Resonance Absorption. Physical Review 1948, 73, 679-712.

ACS Paragon Plus Environment The Journal of Physical Chemistry Page 18 of 19

1 2 3 32. Bain, A. D., Chemical Exchange in Nmr. Progress in Nuclear Magnetic Resonance Spectroscopy 4 5 2003, 43, 63-103. 6 33. Laidler, K. J., Chemical Kinetics. Harper & Row: 1987. 7 34. Lee, I. H.; Kuan, Y.-C.; Chern, J.-M., Equilibrium and Kinetics of Heavy Metal Ion Exchange. 8 Journal of the Chinese Institute of Chemical Engineers 2007, 38, 71-84. 9 35. Cabicar, J.; Gosman, A.; Plicka, J.; Štamberg, K., Study of Kinetics, Equilibrium and Isotope 10 Exchange in Ion Exchange Systems. Ii. Journal of Radioanalytical Chemistry 1983, 80, 71-80. 11 12 36. Inglezakis, V. J.; Zorpas, A. A., Heat of Adsorption, Adsorption Energy and Activation Energy in 13 Adsorption and Ion Exchange Systems. Desalination and Water Treatment 2012, 39, 149-157. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment Page 19 of 19 The Journal of Physical Chemistry

1 2 3 4 Table of Content 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Paragon Plus Environment