Tuning the Ionic and Electronic Properties of Metal Phosphates by Atomic Layer Deposition

Abstract

Atomic layer deposition (ALD) can be used in nano-scaling of solid-state batteries for microelectronics applications due to its compositional tunability and atomic scale control of thin films enabled by the sequential nature of this technique. The Li+ ion conductivity of the electrolyte is a key factor in the performance of Li-ion battery systems. In bulk Li-ion SSBs, solid state electrolytes have been developed with ionic conductivities approaching that of liquid electrolytes (10-2 S/cm) by utilizing fast-ion conducting crystalline phases. However, ALD Li+ conductors have been limited to the range of 10-7-10-6 S/cm with the primary strategy being to develop amorphous Li+ conductors (e.g., amorphous phosphates and silicates doped by glass formers). This work utilizes ALD processing techniques to develop previously unexplored chemistries and crystalline structures in metal phosphate-based Li-ion conductors inspired by the fast-ion conducting NASICON phase Li-Al-Ti-P-O family of materials – a complex system that can contain 5+ elements. This work leans on the layer-by-layer quality of ALD to tune the composition of the system by alternating between layers of Li2O, TiO2, Al2O3 and POx in different sequences to control the composition of the mixed metal phosphates with varying degrees of compositional and structural complexity. The amorphous route to fast-ion conduction is explored by development of ALD processes for the solid-state electrolytes: Al-doped Li3PO4 (LAP), lithium titanium phosphate (LTP), and Al-doped LTP (LATP). In LAP (5.3 x 10-7 S/cm) and LATP (1.8 x 10-6 S/cm) the Al-doping strategy was found to have an effect of increasing ionic conductivity by acting as a glass former in small doping amounts. In the LAP system, higher concentrations of Al % lead to glass-modifying behavior with large Al2O3 domains that interrupt the LAP network and are detrimental to Li+ ion conduction. Additionally, this doping strategy is limited in its ability to improve the conductivity in already well-connected amorphous material like LTP (1.5 x 10-6 S/cm), meaning that it is most effective in systems like Li3PO4 that lack strong cross-linking agents. Thermal annealing was used to study the effects on composition and initial phosphate domain separation on the crystallization route towards the NASICON crystalline structure. In the thermal LTP process, an electron conducting impurity (TiO2) was discovered that provided an external electron conduction path necessary for lithiation of the NASICON structure and enabled the study of the NASICON thin film as an electrode material as well. A semicrystalline nanocomposite of NASICON LTP and TiO2 anatase was found to have an ionic conductivity of 9.3 x 10-7 S/cm and electronic conductivity of 2.3 x 10-7 S/cm. This nanocomposite electrode demonstrated to have a reversible capacity of 328 mAh/g at 1 C and 71 % retention at 20 C owed to a) the fast kinetics of the NASICON phase, and 2) the intimate interfacial contact provided by the amorphous LTP matrix between crystalline grains. The fast kinetics of this system offer routes for further exploration of this material with applications beyond batteries in microelectronics (i.e., supercapacitors). The TiO2 impurity was removed from the LTP process by forming trimethyl phosphate (TMP) radicals with an argon plasma step to allow the TMP to fully saturate the surface and disallow excess TiO2 formation. This led to a phosphate network with higher degree of crosslinking and no electronic conduction – thus making it a suitable amorphous electrolyte material. However, thermal annealing studies showed that the changes to the phosphate network altered the crystallization pathway – forming γ -Li3PO4 and TiP2O7 rather than NASCION LTP. Understanding the nature of these changes to the phosphate network requires additional characterization techniques of these amorphous phosphates that probe the local bonding environment and molecular structure with respect to the changed process conditions. This work explores a new ALD mixed metal phosphate chemistry and how composition-property and structure-property relationships are affected at different scales depending on ALD process conditions and post-processing. A highlight of this work is the demonstration of the control of the TiO2 impurity in the ALD process that allows control over the LATP materials properties as either an electrode or electrolyte – providing an opportunity to study this system as a single material battery or supercapacitor. This dissertation marks the beginning of an exploration of a much larger system of Li1-xMxM’2-x(PO4)3 where introduction of dopants (M = Al+3, Cr+3, Ga+3, Sc+3, Y+3, In+3, etc) or alternate metal phosphates (M’ = Ti+4, Ge+4, Sn+4, Hf+4, or Zr+4, etc) can be modified in the layer-by-layer approach to alter the materials and electrochemical properties of both thin film electrodes and electrolytes within these chemistries.