Development and Characterization of High Temperature, High Energy Density Dielectric Materials to Establish Routes towards Power Electronics Capacitive Devices

Open Access
Shay, Dennis Patrick
Graduate Program:
Materials Science and Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
February 03, 2014
Committee Members:
  • Clive A Randall, Dissertation Advisor/Co-Advisor
  • Clive A Randall, Committee Chair/Co-Chair
  • Gary Lynn Messing, Committee Member
  • Nikolas J Podraza, Special Member
  • Michael T Lanagan, Committee Member
  • Susan E Trolier Mckinstry, Committee Member
  • Dielectric
  • Energy Density
  • Ceramic
  • Impedance Spectroscopy
  • Breakdown Strength
  • Ellipsometry
  • Power Electronics
The maximum electrostatic energy density of a capacitor is a function of the relative permittivity (εr) and the square of the dielectric breakdown strength (Eb). Currently, state–of–the art high temperature (>200 °C), SiC–based power electronics utilize CaZrO3–rich NP0/C0G–type capacitors, which have low relative permittivities of εr ~ 30–40, high breakdown strengths (> 1.0 MV/cm), and are chosen for their minimal change in energy storage with temperature. However, with operating temperatures exceeding the rated temperatures for such capacitors, there is an opportunity to develop new dielectric ceramics having higher energy densities and volumetric efficiencies at high temperatures (>200 °C) by utilizing higher permittivity dielectrics while maintaining high breakdown strengths via doping. The development strategy involves identifying binary linear dielectric compositions that combine medium permittivities with large band gaps, and either substituting end members or identifying doping strategies to minimize high temperature ionic conductivity, achieving large energy densities at elevated temperatures. The solid solution behavior of Ca(TixZr1–x)O3 was characterized in order to determine the optimal composition for balancing permittivity and dielectric breakdown strength to obtain high energy densities at elevated temperatures. Characterization by X–ray diffraction (XRD) showed Vegard’s law behavior across the solid solution with minimal 2nd phases. Microstructural analysis showed little variation in grain size across the solid solution, with grain sizes on the CaTiO3 – rich end of the composition range being ~2.0 – 3.0 μm. Room temperature relative permittivities also showed a linear trend across the solid solution, with permittivities increasing with increasing CaTiO3 content. In order to mitigate ionic conduction, an investigation into various Mn doping concentrations on the Ti4+ site using Ca(Ti0.8Zr0.2)O3 (CTZ) was conducted to determine the optimum doping concentration for minimizing dielectric loss at elevated temperatures. It was determined that doping with 0.5 mol% Mn shows the earliest onset of densification during sintering in air and the lowest dielectric losses with increasing temperature. To determine a Ca(TixZr1–x)O3 composition that will also minimize electronic or band conduction, the optical properties of the Ca(TixZr1–x)O3 solid solution were investigated to identify a composition on the CaTiO3 – rich end of the solid solution with a large band gap. Both ultraviolet–visible diffuse reflectance spectroscopy (UV–Vis) and spectroscopic ellipsometry were utilized to determine the Ca(TixZr1–x)O3 band gaps and optical properties. Both techniques yield an asymmetric Vegard’s law trend in the band gaps across the solid solution, with larger band gaps weighted towards the CaZrO3 – rich end of the solid solution. Band gaps on the CaTiO3 – rich end of the solid solution ranged from 3.7 to 3.8 eV. Divided special range analysis of spectroscopic ellipsometry data was utilized to less ambiguously determine band gaps compared to UV–Vis, for which the analysis contains several assumptions. Additionally, the refractive index and optical dielectric constants were determined via ellipsometry and showed good agreement with calculated values. The resistivity at 250 °C scaled with the band gap energy across the solid solution. Comparing the current–voltage (I–V) behavior at 250 °C for Ca(Tix–yMnyZr0.2)O3 (CTZ + Mn) where x = 0.7, 0.8, 0.9, and y = 0.005, it was found that the Ca(Ti0.795Mn0.005Zr0.2)O3 composition showed the lowest current density and a decrease in current density of 5 orders of magnitude compared to the un–doped composition. The Ca(Ti0.795Mn0.005Zr0.2)O3 composition was selected for single layer, Pt buried electrode capacitor prototyping to evaluate high temperature electrical characteristics. Capacitors of un–doped CTZ were also prototyped to compare the high temperature electrical properties and to determine the effects of Mn doping on high temperature electrical performance. Polarization–field (P–E) hysteresis measurements of CTZ showed a large increase in dielectric loss with increasing temperature, limiting the dielectric breakdown strength and recoverable energy density. When doped with Mn, CTZ + Mn showed a minimization of the temperature dependence of the breakdown strength, and maximum energy densities of 7.00 J/cm3 at a Eb of 1.1 MV/cm at room temperature and 5.36 J/cm3 at Eb = 1.0 MV/cm at 300 °C were observed. Impedance spectroscopy of the CTZ and CTZ + Mn dielectrics showed that doping with Mn resulted in a decrease in ionic conductivity and a subsequent decrease in electronic conductivity. To further increase the high temperature resistivity of the CTZ + Mn system, Zr was replaced with Hf to determine the effects of a heavier ion and higher band gap end member on the high temperature electrical properties. Basic characterization of Ca(Ti0.8Hf0.2)O3 (CTH) and Ca(Ti0.795Mn0.005Hf0.2)O3 (CTH + Mn) showed similar characteristics compared to the CTZ system. High temperature impedance spectroscopy of CTH and CTH + Mn showed similar behavior to the CTZ and CTZ + Mn systems, but with overall decreases in ionic and electronic conductivity. Coupled with thermally stimulated depolarization current measurements (TSDC), oxygen vacancy migration and space charge conduction are dominant and could be minimized with Mn doping. Compared to CTZ + Mn, CTH + Mn showed an additional decrease in current density at 250 °C by an order of magnitude. Systematic energy density characterization showed minimization of the temperature dependence of the dielectric breakdown strength in CTH + Mn, with a maximum energy density of 6.5 J/cm3 and 1.1 MV/cm observed at 300 °C. To gain further insight into how aliovalent Mn controls high temperature conduction in the CTH + Mn system, capacitors were quenched from the sintering temperature and an impedance study was performed. It was observed that ionic conductivity was quenched in due to oxygen vacancies compensating Mn3+, and interfacial features were observed in impedance spectra due to double back–to–back Schottky barriers (depletion layers). As capacitors were re–oxidized, bulk resistivity increased while interfacial resistivity decreased. The hypothesis was supported by the application of dc bias during impedance measurements, which showed similar impedance behavior to the re–oxidation study with a redistribution of oxygen vacancies to the interfaces after the bias was removed. In a final attempt to optimize the CTH + Mn system, a rare earth co–doping study with Mn and Mg was investigated to further minimize ionic conductivity and maximize high temperature energy densities. Co–doping CTH with Mg and Dy yielded the lowest conductivities observed in this work, and the least temperature dependence of conductivity up to 250 °C. Room temperature energy densities of ~10 J/cm3 were observed. However, due to grain sizes approaching 9.0 μm, prototyped single layer capacitors showed a large increase in dielectric loss with increasing temperature due to the minimal number of insulating grain boundaries per dielectric layer.