Experimental and Modelling Study of Ionic Conductivity in Lithium Phosphate and Silicate Glasses

This study is an attempt to investigate ionic conductivity and structural components of lithium containing glasses for their use as possible electrolytes in solid state Li-ion batteries. 33.3Li 2 O-66.7SiO 2 , Li 2 O-P 2 O 5 (multiple compositions) and Li 3 Fe 2 (PO 4 ) 3 glasses were synthesised using conventional melt quenching and were characterised using various techniques including XRD, DSC, TGA, pycnometry and 31 P MAS-NMR. It was found a small (5mol%) addition of Nb 2 O 5 to the Li 3 Fe 2 (PO 4 ) 3 glass was necessary for the formation of a glass 37.5Li 2 O-20Fe 2 O 3 -5Nb 2 O 5 -37.5P 2 O 5 . The glasses conductivity was then measured using impedance spectroscopy over a frequency range of 200kHz – 100Hz and a temperature range of 300K – 525K. It was observed that an increasing temperature corresponded to an increasing conductivity, as E expected from the Arrhenious equation: ? = ? 0 exp (? k a T ). Conductivity values were B compared to published values and the first reported conductivity values for 37.5Li 2 O- 20Fe 2 O 3 -5Nb 2 O 5 -37.5P 2 O 5 glass were obtained. Activation energies were also calculated and compared to published data. MD models of lithium disilicate, lithium metaphosphate and 37.5Li 2 O-20Fe 2 O 3 -5Nb 2 O 5 - 37.5P 2 O 5 were made. These models were then analysed and compared to experimental diffraction results. It was found that the lithium disilicate and the lithium metaphosphate model structures compare well to experimental data (X-ray and Neutron diffraction). Conductivity was estimated from mean squared displacement. All models gave an overestimation of conductivity compared to experimental data, due to limitations of the simulation method. However, MD did predict that lithium disilicate and lithium metaphosphate glasses have a similar conductivity as observed experimentally. It was found that whilst the addition of Fe (and Nb) to the lithium phosphate glass improved chemical durability, its conductivity was reduced. In addition to this the conductivity for the 37.5Li 2 O-20Fe 2 O 3 -5Nb 2 O 5 -37.5P 2 O 5 glass was found to be significantly lower than its crystalline counterpart which will be a disadvantage for its use as a solid electrolyte.