The enhancement of lithium ion dissociation in polyelectrolyte gels on the addition of ceramic nano-fillers

Abstract

Nano-particle oxide fillers including TiO₂, SiO₂ and Al₂O₃ have previously been shown to have a significant affect on the properties of both polymer and polymer gel electrolytes. In some cases, conductivity increases of one order of magnitude have been reported in crystalline PEO–base complexes. In this work, we report the effects of TiO₂ and SiO₂ on a poly(Li-AMPS)-based gel polyelectrolyte. Impedance spectroscopy and pfg-NMR spectroscopy indicates an increase in the number of available charge carriers with the addition of filler. An ideal amount of ceramic filler has been identified, with additional filler only saturating the system and reducing the conductivity below that of the pristine polyelectrolyte system. SEM micrographs suggest a model whereby the filler interacts readily with the sulfonate group; the surface area of the filler being an important factor.

---

1 Introduction

Gel electrolytes continue to attract research interest due to the properties which make them suitable for applications including solid state rechargeable lithium ion batteries, supercapacitors, fuel cells and sensors.^1–3 The gel electrolyte has some important advantages over solid and liquid electrolytes in that the risk of leakage is reduced, while electrode interfacial contact can be maintained during volumetric changes associated with charging and discharge cycles of the battery. The ability for the gel electrolyte to act as both the separator and the electrolyte leads to easy fabrication and allows for the possibility of miniaturised structures.^4,5

Before widespread application of gel electrolytes can occur a number of issues still need to be resolved. These materials have a tendency to flow with time,^5,6 which means that leakage is still possible. The lithium ion transfer number in gel and solid polymer electrolytes is usually low, with a large portion of the current being transported by the anions as opposed to the lithium cations.^7–9 This leads to anion concentration gradients within the electrolyte, with lithium ion flow causing polarisation of the battery.

In an effort to eliminate concentration gradients, single ion conducting polyelectrolytes as shown in Fig. 1, with negative charges covalently bound to the polymer chain have been developed. These polyelectrolyte gels have cation transference numbers close to unity, but unfortunately the ambient ionic conductivity of the unfilled polyelectrolyte gel was still quite low, in the order of 10⁻⁵ S cm⁻¹.^4,8

Fig. 1 Typical schematic drawing of a polyelectrolyte.

The addition of ceramic particles to polymer electrolytes has been an area of study for some time. The use of TiO₂ in PEO-based polymer electrolytes has been shown to increase the ionic conductivity by one order of magnitude at the melting point.^10,11 The mechanism proposed for the increase in ionic conductivity in PEO systems involves increased disorder due to interactions between the polymer and the ceramic.^12,13 The use of ceramic fillers in gel electrolytes has the potential to increase the mechanical properties whilst reducing the risk of leakage. Adebahr et al. have recently reported increases in ionic conductivity of nearly a factor of two upon the addition of TiO₂ to poly(methyl methacrylate) (PMMA)-based gels.¹⁴ It has also been proposed for certain systems such as the PMMA-based gel electrolytes that as filler is added an ideal threshold is reached referred to as percolation; this occurs when regions containing filler “connect” leading to an alternative conduction path. This has been suggested to account for a further increase in ionic conductivity. The exact percolation threshold is difficult to determine due to the heterogeneous nature of these gels and the nature of the percolation mechanism being dependent on changes in interfacial tension, which will be affected by gel components.¹⁵

The low ionic conductivity in lithium-based polyelectrolytes is primarily a result of strong ionic association between the lithium and the tethered anion. The use of additives such as boroxine compounds and ionic liquids^16,17 to improve dissociation has shown improvements in ionic conductivity. In this investigation the effect of SiO₂ and TiO₂ addition to a well-known polyelectrolyte system based on poly(Li-AMPS) has been studied by means of pulse field gradient (pfg)-NMR, impedance spectroscopy, thermal analysis and scanning electron microscopy (SEM) to explore the possibility of ionic conductivity enhancement and to better understand the mechanisms of this.

2 Experimental

2.1 Materials

The LiAMPS monomer was obtained by neutralization of 2-acrylamido-2-methyl-1-propanesulfonic acid, AMPS (Sigma), with lithium carbonate, Li₂CO₃ (Pacific Lithium (NZ) Ltd). Polyelectrolyte gels were prepared using free radical polymerisation of the LiAMPS and N,N′-dimethylacrylamide, DMAA (Aldrich), in a molar ratio of 1 ∶ 9. Tetraethyleneglycol diacrylate, TEGDA, was used as the cross-linking agent. Potassium persulfate 5 mol% was used to initiate the reaction. The solvent used was a molar 2 ∶ 1 ratio of ethylene carbonate, EC, and N,N′-dimethylacetamide, DMA. These samples were all prepared under an argon atmosphere. Samples were heated to 50 °C. Samples containing 2, 4, 6, 8, 10, 12 and 15 wt% TiO₂ (Degussa 21 nm) and samples containing 2, 4, 6, 8 and 10 wt% SiO₂ (Degussa 30 nm) were investigated. The filler was added prior to the polymerisation.

2.2 Impedance spectroscopy

The sample was sandwiched between two stainless steel blocking electrolytes in a spring loaded cell. Impedance spectra were measured using a model 1296 Solartron frequency response analyzer (FRA) driven by Solartron impedance measurement software version 3.2.0 over a frequency range of 1 MHz to 1 Hz with a signal voltage 0.1 V. The cell temperature was controlled using a Eurotherm control model 2204. The temperature was measured using a thermocouple Type T with an accuracy of 1 °C.

2.3 Thermal analysis

Differential scanning calorimetry (DSC) measurements were carried out using a Perkin-Elmer DSC-7 analyser. A temperature range of −125 to 70 °C was used. The heating rate used was 20 °C min⁻¹. A two point calibration with cyclohexane (crystal–crystal = −87.54 °C, melting point = 6.54 °C) was used.

2.4 SEM

Low temperature measurements. Gel samples 12 wt% TiO₂, 8 wt% SiO₂, and 4 wt% TiO₂ were carefully loaded by gently clamping into specific SEM mountings. Mounting adhesive was not required. Coating was not necessary for low temperature observation. Samples and holder were immersed into liquid nitrogen and transferred with an air-tight o-ring sealed vacuum attachment through a FISONS Instruments Polaron LT7400 cryo-preparation chamber where the sample was manually cleaved. Fracture was performed under high vacuum (∼10⁻⁶ mbar) and subsequently all samples were freeze-dried at −80 °C again under high vacuum for 1 hour (the effects of sublimation are discussed later). Observation of as-fractured surfaces was performed using a Philips XL30FEG SEM at −192 °C, secondary detector and accelerating voltage of 2 kV at variable magnification.

Room temperature carbon coated measurements and elemental analysis (EDXS). Samples 12 wt% TiO₂, 2 wt% TiO_2, 8 wt% SiO₂ and 2 wt% SiO₂ were immersed into liquid nitrogen and fractured, loaded onto double-sided tape, carbon coated under high vacuum and the fracture surface observed using a backscattered detector on the as-above described instrument. Elemental analysis (EDXS) was performed using accelerating voltage 15 keV and Link Isis software.

2.5 NMR

⁷Li and ¹H pfg-NMR diffusion measurements were performed on a Bruker Avance 300 Spectrometer, operating at proton frequency 300 MHz, using the stimulated spin-echo (STE) pulse sequence.¹⁸ 90 degree pulse lengths were 13 µs (¹H) and 9 µs (⁷Li), respectively with the waiting time between experiments being >5T₁. 12 scans and 16 experiments were performed for each diffusion measurement, in which the gradient strength was varied up to 1800 G cm⁻¹ while keeping all other parameters constant. The diffusion time was 20 ms with a gradient ramp time of 0.30 ms. Measurements were performed over the temperature range of 20 °C to 80 °C at ten-degree intervals with the sample being held at each temperature for 30 minutes to achieve thermal equilibrium. Single lines were observed for both ⁷Li and ¹H. The diffusion coefficient was then determined from the peak amplitude attenuation according to the Stejskal–Tanner diffusion equation, where

A = A₀exp [(γδG)²D(Δ − δ/3)]

¹⁹. Where A is the signal at the applied gradient (G), A₀ is the signal at zero gradient δ is the width of the applied pulse and Δ is time between pulses D is diffusion coefficient γ is the gyro-magnetic spin of nuclei where D is calculated as the slope of ln(A/A₀) vs. [(γδG)² (Δ − δ/3)].

3 Results

3.1 Thermal analysis

Fig. 2 shows the DSC traces for samples containing TiO₂. The T_gs of the samples do not appear affected by filler addition, indicating that the dynamics of the polymer have not undergone any significant change with the addition of TiO₂.²⁰ The peak occurring at −42 °C is an isothermal transition near the eutectic temperature for the binary EC/DMA system (binary eutectic for EC/DMA −52 °C⁷). The onset temperature remains relatively unchanged, however, the size of the eutectic (on a per unit mass of EC/DMA) has been depressed with filler addition. A possible explanation for this behaviour is that the filler suppresses solvent crystallisation.

Fig. 2 DSC traces as a function of TiO₂ filler loading in the poly(Li-AMPS) based system. * denotes T_g, ∧ denotes first crystallisation peak.

Peak broadening appears to be occurring in the final transition and it appears that more than one peak is present. This thermal behaviour is consistent with a phase-separated sample, giving two melting transitions. This peak is becoming more pronounced with TiO₂ additions and given that EC melts at 35 °C (as measured), it can be suggested that the higher melting shoulder represents an EC rich phase. It appears that as filler is added the EC becomes purer in the final melt. The second crystallisation peak at ≈0 °C is also exaggerated with filler addition.

Fig. 3 shows the DSC traces of the samples containing SiO₂ and these appear significantly different to the TiO₂ filled system, once again no change in T_g is observed; however, the first crystallisation peak is smaller in the 1.2 vol% sample and disappears altogether in the higher vol% samples. The onset of the isothermal transition, representative of the EC/DMA eutectic does not change with SiO₂ filler and the decrease in peak area is less than that measured in the TiO₂ filled samples. This indicates that the SiO₂ filled sample crystallises fully on cooling from room temperature.

Fig. 3 DSC traces as a function of SiO₂ filler loading in the poly(Li-AMPS) based system. * denotes T_g, ∧ denotes first crystallisation peak.

One notable difference between the DSC traces of the samples containing the two fillers is the final liquidus. There is no crystallisation before the melt in the SiO₂ filled samples and the shape of the liquidus peak shows typical behaviour of a single phase melt. The SiO₂ acts as a very strong nucleating agent within this system; this is consistent with the absence of the first crystallisation peak. In the SiO₂ filled samples the eutectic has crystallised on freezing prior to the thermal measurements being taken, as the samples are cooled to −130 °C and held there for 20 minutes before measurements are started. The absence of the second crystallisation and the observation of a single melting peak clearly shows no evidence of an EC rich phase in the SiO₂ filled systems; again supporting that SiO₂ may compatibilise the polymer and solvents to a better extent than the TiO₂.

The differences in the thermal behaviour of these filled samples is interesting, however, these differences do not seem to impact on the more crucial transport properties of the system as discussed below.

3.2 Ionic conductivity

Fig. 4 presents an impedance plot, which is representative of the system both with and without filler for sub ambient and ambient temperatures. It can be seen that at sub ambient temperatures an almost complete semi-circle is measured whereas at ambient temperatures only an electrode spike is seen, the latter being typical of a very conductive sample. In contrast to filled and cross-linked polyetherurethane-based polymer electrolytes where multiple arcs could be observed in the Nyquist plots,²¹ the data here can be fitted by a simple equivalent circuit of a resistor in parallel with a constant phase element in series with a double layer capacitance, suggesting a more homogeneous material.

Fig. 4 Impedance plot for sub ambient (semi-circle) and ambient temperatures (spike) for filled and TiO₂ filled poly(Li-AMPS) based system.

The ionic conductivity is presented as a function of temperature in Fig. 5a and b. At all temperatures the ionic conductivity is enhanced by filler addition. This behaviour is more pronounced at elevated temperatures, with conductivity being enhanced by approximately a factor of 10 at 80 °C.

Fig. 5 a: Conductivity as a function of temperature for the poly(Li-AMPS) based system with addition of TiO₂ filler. b: Conductivity as a function of temperature for poly(Li-AMPS) based system with addition of SiO₂ filler.

The dependence of ionic conductivity on filler addition is better shown at a single temperature (T = 80 °C, Fig. 6). Fig. 6 shows how the ionic conductivity reaches a maximum with filler loading, after which, it reduces dramatically. The maximum in ionic conductivity seen in Fig. 6 occurs around 4.8 vol%; this is thought to be the ideal filler loading for this system. The decrease in ionic conductivity seen in the 6 vol% sample may be due to agglomeration of primary particles leading to a lower total surface area interacting with the polyelectrolyte. Thus it appears that the surface area of the filler dictates the conductivity enhancement effect. Fig. 6 also highlights the differences between the two fillers: with the ionic conductivity being lower in the SiO₂ system.

Fig. 6 Conductivity at T =80 °C showing the effect of the two fillers on the poly(Li-AMPS) based system.

3.4 ⁷Li and ¹H NMR characterisation

The temperature dependence of the ⁷Li diffusion coefficient for the samples containing TiO₂ is shown in Fig. 7a. TiO₂ addition leads to an increase in the average diffusivity of the lithium ions over the entire temperature range. Fig. 7b shows the temperature dependence and the affect of SiO₂ on the lithium ion mobility; again with the trend being similar to TiO₂ filled samples.

Fig. 7 a: ⁷Li diffusion coefficient as a function of temperature for the poly(Li-AMPS) based system with addition of TiO₂ filler. b: ⁷Li diffusion coefficient as a function of temperature for the poly(Li-AMPS) based system with addition of SiO₂ filler.

Fig. 8 illustrates the effect of filler as a function of concentration at room temperature. The influence of filler content on lithium ion diffusion can clearly be seen here. For the polyelectrolyte samples with SiO₂ and TiO₂ filler additions, the average mobility of the lithium ions is increased up to filler additions of 4.8 vol%; beyond this threshold a decrease is observed, consistent with the ionic conductivity data. Again the addition of TiO₂ leads to greater enhancement of lithium diffusion when compared to SiO₂.

Fig. 8 Room temperature diffusion coefficients for EC, DMA and Li for samples containing TiO₂.

¹H pfg-NMR was used to probe the transport of the EC and DMA solvent molecules. This data is also included as a function of filler at room temperature in Fig. 8, where it can be seen that filler additions have little to no effect on the diffusion coefficients of either solvent. The DMA has a diffusion coefficient about 10% lower than that of EC. This can be attributed to the greater hydrodynamic radius of the DMA molecules, which leads to a lower diffusion coefficient as indicated by the Stokes–Einstein equation.²²

3.5 SEM Micrographs

Fig. 9a is the low temperature fractured surface image of the 1.6 vol% TiO₂. The image shows how the filler has agglomerated into separate regions, a common occurrence for nano-size particles with large surface area as the high surface energy of these particles promotes agglomeration. Fig. 9b presents the back-scattered electron image of the 0.8 vol% TiO₂ sample. The EDXS data confirms that titanium, sulfur, carbon, oxygen and potassium are all present however; the lighter region (e.g. area 1) appears relatively richer in TiO₂, sulfur and potassium. It should be noted that polymerisation was initiated using potassium persulfate (K₂SO₈), which could therefore account for the presence of the latter element. The darker regions (e.g. area 2) are predominantly carbon and oxygen; this area most likely represents the solvent rich region. This micrograph gives a strong indication that the TiO₂ filler readily interacts with the anion (identified as the sulfonate group via EDXS) on the polymer backbone accounting for the increase in sulfur next to the TiO₂ regions in these images.

Fig. 9 a: Low temperature image of poly(Li-AMPS) based system with the addition of 1.6 vol% TiO₂. b: Back scattered image of poly(Li-AMPS) based system with the addition 0.8 vol% TiO₂.

Fig. 10a presents the low temperature fracture surface of 4.8 vol% TiO₂ sample and it suggests that the filler is dispersed homogeneously throughout the sample in the form of interconnecting fringes. Furthermore the EDXS and micrograph shown in Fig. 10b shows the TiO₂ as being relatively uniformly dispersed throughout the sample. Darker regions are observed in this micrograph, which are due to valleys and peaks on the sample’s surface.

Fig. 10 a: Low temperature image of poly(Li-AMPS) based system with the addition 4.8 vol% TiO₂. b: Back scattered image of poly(Li-AMPS) based system with the addition 4.8 vol% TiO₂.

Fig. 11a is the back-scattered micrograph of 1.2 vol% SiO₂ filled sample at high magnification. In contrast to the TiO₂ images, which showed predominately two regions (dark and light), these samples display three distinct regions (grey, light and dark). The EDXS data shows that silica, sulfur, carbon, oxygen, and potassium element are all present, however, the grey region (e.g. area 1) is predominantly SiO₂, the lighter region (e.g. area 3) is richer in SiO₂, sulfur and potassium while the darker region (e.g. area 2) contains mainly carbon and oxygen. From this micrograph it can be suggested that the SiO₂ does not interact as readily with the anion when compared to the TiO₂ filled sample. This point is more clearly seen in Fig. 11b, which is a low magnification image of the 1.2 vol% SiO₂ sample. Fig. 12 shows the back-scattered image of the ideal filler loading (=4.8 vol%) for SiO₂ and as with the TiO₂ filled sample, the filler seems relatively uniformly dispersed throughout the entire sample. The darker regions in this image correspond to valleys on the samples surface. This was confirmed by EDXS, which showed the presence of TiO₂ or SiO₂ uniformly dispersed throughout the “ideal” composition sample.

Fig. 11 a: Back scattered image of poly(Li-AMPS) based system with the addition 1.2 vol% SiO₂. b: Back scattered image of poly(Li-AMPS) based system with the addition 1.2 vol% SiO₂ low magnification.

Fig. 12 Back scattered image of poly(Li-AMPS) based system with the addition 4.8 vol% SiO₂.

4 Discussion

The conductivity and pfg-NMR diffusion data presented for this polyelectrolyte gel system clearly indicated the enhanced dissociation of the lithium ion from the backbone-tethered anion occurred upon the addition of TiO₂ and SiO₂ nano-filler.

This dissociation leads to a higher number of charge carriers and as per the Nernst–Einstein equation, σ = F∑z_in_iμ_i, higher conductivity is observed. The magnitude of the increase on filler addition is consistent with this hypothesis; that it is conceivable that increased dissociation produces as much as a 3-fold increase in mobile Li ion concentration.

The enhanced diffusivity (which would lead to a higher μ_i) is not likely to be due to an increase in the mobility of the overall system; such an enhancement would be manifested in high solvent diffusion coefficients and likely lowering of T_g. In fact neither of these were affected by the addition of filler. The question then arises as to why the lithium diffusion coefficient increases, thereby leading to an increase in conductivity. If we consider the state of the polyelectrolyte gel there are at least two possible sites for Li⁺—associated to the sulfonate anion on the polymer backbone, or solvated by the EC/DMA solvent. The likelihood is that both sites are present and that exchange occurs between these sites, but in the absence of the filler, the associated species dominates and hence the average diffusion coefficient (which is all that the pfg NMR can measure) is lower. The SEM micrographs, in particular for TiO₂, give direct evidence for SO₃⁻–filler interactions. These interactions are likely to reduce the strength of the lithium ion –sulfonate interaction, consequently enhancing the dissociation of the Li ions from the polymer. This leads to a higher average lithium diffusion coefficient since a higher proportion of lithium ions exists in the solvated state. Furthermore, this model is strengthened by the fact that SiO₂, according to the SEM experiments, does not appear to interact as strongly or as readily with the sulfonate group within the polyelectrolyte thus the effect of this filler on lithium ion diffusivity is lessened. This is then consistent with the lower ionic conductivity measured in the SiO₂ filled system.

Such a model would also require a large surface area of filler, since this would maximize polymer filler interactions. Therefore, as more filler is incorporated and agglomeration of particles occurs, the effective surface area may decrease leading to a diminishment of the diffusion and conductivity enhancement. (The crystallization behaviour of the EC/DMA solvent of the polyelectrolyte gel upon addition of filler also confirms that the filler cannot merely be an inert additive but rather a key component, i.e., interactions between the filler and the other system components must occur which leads to changes in the phase behaviour as discussed previously.)

It is interesting to note previous work on gel electrolytes based on PMMA and lithium salts reported quite different behaviour.¹⁴ In those systems increases of cation–anion dissociation were clearly observed with NMR measurements, however, conductivity enhancements were less obvious, and even showed some decrease at the lower filler contents. In fact, only the percolation concentration of filler (8 wt%) led to high conductivities in contrast to the present polyelectrolyte systems. Whereas the diffusion coefficients of the lithium ion measured by pfg NMR showed similar behaviour as for the present system. In that case the interaction of the filler with the anion lead to a decrease in anion diffusivity as shown Fig. 13. This then explains the contrast in the conductivity behaviour of the two different systems; in the present case the anion is already tethered and so does not contribute to conductivity, whereas in the PMMA based gel case the anion transport number is greater than half and a decrease in anion diffusivity must lead to a decrease in ionic conductivity. In this polyelectrolyte system the same interactions merely free up more lithium ions.

Fig. 13 PMMA-based gel electrolyte highlighting the decrease in anion diffusivity with filler addition.

The results presented in this paper give further insight to the affect filler has in the general field of polymer electrolytes. Even in the PEO and other amorphous polyether based systems,^12,23 anion–filler interactions are likely and may at least in part explain observations of changes in ion transport in those cases. Of course those solvent free systems have additional polymer–filler interactions. In optimising nano-filler polyelectrolyte based systems, conductivities tend to be governed by the nature of the surface groups that might lead to maximum anion–filler interactions.

5 Conclusion

A combination of thermal analysis, impedance spectroscopy, nuclear magnetic spectroscopy and scanning electron microscopy techniques has clearly shown the interaction between two nano-sized fillers and a gel polyelectrolyte. A specific interaction has been identified which explains the enhancement in ionic conductivity by approximately a factor of 10 over the pristine polyelectrolyte. The data suggests that both fillers preferentially interact with the sulfonate group; thus increasing the number of available charge carriers. Consequently, the ⁷Li diffusion coefficient was enhanced with the addition of both TiO₂ and SiO₂ nano-particles. The SEM micrographs support this model and also show how the TiO₂ more readily combines with sulfonate groups when compared to the SiO₂; explaining the higher ⁷Li diffusion coefficients and ionic conductivity values measured in the TiO₂ filled samples. The SEM micrographs show that at the ideal filler concentration both the TiO₂ and SiO₂ are relatively homogeneously dispersed throughout the sample.

Acknowledgements

The authors would like to thank Adam Best for helpful discussions and assistance in this area. This work was sponsored by an Australia Research Council (ARC) grant.

References

1. S. Rajendran, R. Kannan and O. Mahendran, Ionic conductivity studies in poly(Me methacrylate)-PEO hybrid polymer electrolytes with lithium salts, J. Power Sources., 2001, 406–410.
2. B. Scrosati, Application of electroactive polymers, 1993, Chapman and Hall, London.
3. J. Y. Song, Y. Y. Wang and C. C. Wan, Review of gel-type polymer electrolytes for lithium-ion batteries, J. Power Sources., 1999, 183–197.
4. F. M. Gray, Solid polymer electrolytes: fundamental and technological applications, 1991, VCH, New York.
5. G. B. Appetecchi, P. Romagnoli and B. Scrosati, Composite gel membranes: a new class of improved polymer electrolytes for lithium batteries, Electrochem. Commun., 2001, 281–284.
6. Y. K. Yarovoy, H. P. Wang and S. L. Wunder, Dynamic mechanical spectroscopy and conductivity studies of gel electrolytes based on stereocomplexed poly(methyl methacrylate), Solid State Ionics, 1999, 301–310.
7. M. Forsyth, H. A. Every, D. R. MacFarlane, J. Travas-Sejdic and R. Steiner, Compositional effects on structure and transport properties in polyelectrolyte gel, Macromol. Mater. Eng., 2002, 532–531.
8. J. Travas-Sejdic, R. Steiner, J. Desilvestro and P. Pickering, Ion conductivity of novel polyelectrolyte gels for secondary lithium-ion polymer batteries, Electrochim. Acta, 2001, 1461–1466.
9. P. G. Bruce, J. Evan and C. A. Vincent, Conductivity and transference number measurements on polyelectrolytes, Solid State Ionics, 1998, 918–922.
10. F. Croce, G. B. Appetecchi, L. Persi and B. Scrosati, Nanocomposite polymer electrolytes for lithium batteries, Nature (London), 1998, 394(6692), 456–458.
11. A. S. Best, Lithium-ion conducting composites, Ph.D. Thesis. 2001, Monash University, Clayton.
12. A. S. Best, K. M. Nairn, P.J. Newman, D. R. MacFarlane, M. Forsyth and M. J. G. Jak, The effect of additives on ceramic materials for lithium solid electrolytes, J. Aust. Ceram. Soc, 1998, 236–241.
13. F. Croce, L. Persi, B. Scrosati, F. Serraino-Fiory, E. Plichta and M. A. Hendrickson, Role of the ceramic fillers in enhancing the transport properties of composite polymer electrolytes, Electrochimica Acta, 2001, 2457–2461.
14. J. Adebahr, N. Byrne, M. Forsyth, D. R. MacFarlane and P. Jacobsson, Enhancement of ion dynamics in PMMA-based gels with addition of TiO₂, Electrochim. Acta, 2003, in the press.
15. A. S. Best, J. Adebahr, P. Jacobsson, D. R.. MacFarlane and M. Forsyth, Microscopic Interactions in Nanocomposite Electrolytes, Macromolecules, 2001, 4549–4555.
16. F. Zhou, D. R. MacFarlane and M. Forsyth, Boroxine ring compounds as dissociation enhancers in gel electrolytes, Electrochim. Acta, 2003, 48(12), 1749–1758.
17. M. Forsyth, J. Sun, F. Zhou and D. R. MacFarlane, Enhancement of ion dissociation in polyelectrolyte gels, Electrochim. Acta, 2003, 48(14–16), 2129–2136.
18. J. E. Tanner, Use of the stimulated echo in NMR diffusion studies, J. Chem. Phys., 1970, 52(5), 2523–6.
19. E. O. Stejskal and J. E. Tanner, Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient, J. Chem. Phys., 1965, 42(1), 288–92.
20. A. S. Best, A. Ferry, D. R. MacFarlane and M. Forsyth, Conductivity in amorphous polyether nanocomposite materials, Solid State Ionics, 1999, 269–276.
21. M. Forsyth, D. R. MacFarlane, A. S. Best, J. Adebahr, P. Jacobsson and A. J. Hill, The effect of nano-particle TiO₂ fillers on structure and transport in polymer electrolytes, Solid State Ionics, 2002, 203–211.
22. K. Hayamizu, Y. Aihara, S. Arai and W. Price, Diffusion conductivity and DSC studies of a polymer gel composed of cross-linked PEO, γ-butyrolactone and LiBF₄, Solid State Ionics, 1998, 107, 1–12.
23. C. Capiglia, P. Mustarelli, E. Quartarone, C. Tomasi and A. Magistris, Effects of nanoscale SiO₂ on the thermal and transport properties of solvent-free, poly(ethylene oxide) (PEO)-based polymer electrolytes, Solid State Ionics, 1999, 73–79.

---