Nano-network electronic conduction in iron and nickel olivine phosphates

The demonstration of non-carbonaceous-network grain-boundary conduction to be the first in these materials, and that it holds promise for other insulating phosphates.

Abstract

The provision of efficient electron and ion transport is a critical issue in an exciting new group of materials based on lithium metal phosphates that are important as cathodes for lithium-ion batteries. Much interest centres on olivine-type LiFePO(4), the most prominent member of this family. Whereas the one-dimensional lithium-ion mobility in this framework is high, the electronically insulating phosphate groups that benefit the voltage also isolate the redox centres within the lattice. The pristine compound is a very poor conductor (sigma approximately 10(-9) S cm(-1)), thus limiting its electrochemical response. One approach to overcome this is to include conductive phases, increasing its capacity to near-theoretical values. There have also been attempts to alter the inherent conductivity of the lattice by doping it with a supervalent ion. Compositions were reported to be black p-type semiconductors with conductivities of approximately 10(-2) S cm(-1) arising from minority Fe(3+) hole carriers. Our results for doped (and undoped) LiMPO(4) (M = Fe, Ni) show that a percolating nano-network of metal-rich phosphides are responsible for the enhanced conductivity. We believe our demonstration of non-carbonaceous-network grain-boundary conduction to be the first in these materials, and that it holds promise for other insulating phosphates.