Safety problems associated with rechargeable lithium batteries are now well recognized. Recent spectacular fires involving cell phones, laptops, and (here at LBNL) AA cells have made the news. These events are generally caused by overcharging and subsequent development of internal shorts. Before these batteries can be used in vehicle applications, improvement in cell safety is a must. We have been active in the area of lithium battery safety for many years. For example, a versatile, inexpensive overcharge protection approach developed in our laboratory, uses an electroactive polymer to act as a reversible, self-actuating, low resistance internal shunt that carries overcharge current after the battery is fully charged. Safety can also be improved by using electrode materials that are inherently stable. Olivine-type LiFePO4 is one of the safest cathode materials for Li batteries, owing to the strong covalent bonds between phosphorus and oxygen. The main obstacles to commercial implementation are low rate capability, a direct result of the material's low electronic and ionic conductivities during Li extraction (charging) and insertion (discharging), and low energy density. This two-phase LiFePO4/FePO4 electrode has been the subject of several scientific controversies, including the popular yet unrealistic core-shell model, particle size effects on the behavior of the material, the existence of room-temperature solid solution phases, and their role as the "reaction promoter." Our approach has been to prepare high quality single crystals of LiFePO4 of various sizes and shapes. The solid-state behavior of the crystals during chemical oxidation and reduction, thermal treatment, and electrochemical charge and discharge, were studied in detail using electron microscopy and spectroscopic techniques. The relationship of crystal chemistry and material behavior such as safety and performance, as well as material engineering guidelines to achieve optimum performance of lithium battery electrode will be discussed.