Dependence of Conductivity on the Interplay of Structure and Polymer Dynamics in a Composite Polymer Electrolyte Robert L. Karlinsey, Lyudmila M. Bronstein, and Josef W. Zwanziger* Department of Chemistry, Indiana UniVersity, Bloomington, Indiana 47405 ReceiVed: September 5, 2003; In Final Form: NoVember 10, 2003 The conductivity and transport properties of a composite polymer electrolyte were studied by comparing the structure and dynamics as a function of both salt and organic-inorganic composite content. The system consisted of poly(ethylene oxide) (PEO), an organic-inorganic composite (OIC) prepared from aluminum tri-sec-butoxide and [(3-glycidyloxy)propyl]trimethoxysilane (GLYMO) and lithium triflate (LiCF 3 SO 3 ). The systems with and without salt yielded strikingly different physical properties when the OIC content exceeded 50%. Through analysis of 29 Si NMR spectra, it was found that the lithium ion of LiCF 3 SO 3 (LiTf) promotes the condensation of GLYMO, which peaks near 50% OIC content. Also, short-range structural evidence for PEO-OIC blending at high OIC content was observed in the salt-free system through comparisons of the line shapes of the 27 Al NMR spectra. This blending is absent in the ternary system due to prominent PEO- LiTf interactions, as confirmed by X-ray, DSC, and impedance spectroscopy experiments. Furthermore, the glass transition temperature exhibits a linear increase as a function of OIC content, whereas the conductivity over this range first shows a sharp increase followed by a mild decrease. The dielectric constant also was found to vary nonlinearly with OIC content, indicating that ionic screening is modulated by OIC. Because in this system the conductivity and the glass transition temperature do not show a significant correlation, although structurally it is clear that PEO and salt are intimately mixed, a model was developed for the transport that focuses principally on the density of mobile lithium ions. The model predicts relatively constant ion mobilities and diffusion constants but a strongly varying mobile ion number density as a function of OIC content, which then explains the dependence of conductivity on OIC content in this electrolyte. 1. Introduction Solid polymer electrolytes are of interest to replace existing liquid and paste electrode separators in secondary lithium batteries, due to advantages realized only in the solid state. For example, thin-film, solvent-free batteries eliminate leakage concerns, reduce the overall battery weight, and permit various thin-film geometries suitable for packaging or technological applications, such as power sources for the microprocessors in “smart cards”. 1,2 In addition, electrolyte-electrode compliance is improved with polymer-based interfaces rather than those formed, for example, with solid inorganic electrolytes. 3 All these features add to the appeal for developing polymer-based electrolytes with improved material properties. The prototypical polymer electrolyte consists of a coordinating polymer (typically poly(ethylene) oxide, PEO) and an alkali salt, typically of the form LiX. 4 Ionic conductivity is active in this system at temperatures above the glass transition, where it typically follows a Vogel-Tamann-Fulcher temperature de- pendence. 5,6 The frequency dependence of the conductivity shows a marked plateau at lower frequencies. 7,8 These observa- tions have been explained with a dynamic bond percolation model. 7-11 Furthermore, in most cases the conductivity is suppressed by crystalline regions. Thus, because the conductivity of the prototypical system is in fact not very high, much research has focused on optimizing it by lowering the glass transition temperature and stabilizing the resulting amorphous material against crystallization. This general approach has yielded some successes but often at the expense of mechanical properties, and in any event has not produced materials with sufficiently high conductivity at ambient temperature. 12 Inorganic additives of various types have been employed in attempts to improve properties of the base polymer electrolyte. In one approach, various inorganic nanoparticles have been used, including silica and alumina; 13,14 this approach appears primarily to stabilize the amorphous phase of the polymer system but does not in general improve the transference number. 15 An improve- ment in this latter respect comes from the use of layered inorganic materials, such as clay particles, which serve as both hosts for the polymer and the source of ions. 16-19 The resulting materials have unity transference numbers, and markedly low activation barriers to conductivity. Finally, the inorganic material can be grafted directly to the polymer, resulting again in single- ion conductors. 20,21 In an attempt to combine many of the best properties of the composite systems described above, we have developed a composite electrolyte composed of PEO, a lithium salt, and an organic-inorganic composite (OIC) prepared from aluminum tri-sec-butoxide and [(3-glycidyloxy)propyl]trimethoxysilane (GLYMO). 22-24 This system shows marked improvement in conductivity and transference numbers over pure PEO-based polymer electrolytes, and some striking differences as well. In particular, although every indication is that in this composite the conductivity is related to polymer motion, and shows the same plateau as a function of frequency as does the prototypical system, the conductivity is not well-correlated with the glass * Corresponding author. Current address: Department of Chemistry, Dalhousie University, Halifax, NS B3H 4J3 Canada. Email: jzwanzig@dal.ca 918 J. Phys. Chem. B 2004, 108, 918-928 10.1021/jp036655e CCC: $27.50 © 2004 American Chemical Society Published on Web 12/17/2003