Understanding and Mitigating Defects in Barium Strontium Titanate Ruddlesden-Popper Superlattices for High Frequency Tunable Dielectrics

Abstract

Materials challenges are a substantial obstacle for gigahertz frequency technologies such as high-bandwidth 5G cellphone networks. These challenges are especially prominent in the development of tunable dielectric materials, essential for tunable filters and antennas. Large loss of the electric field arises in the most common tunable dielectrics at high frequencies from entropically-favored point defects and polar inhomogeneities. Strained (SrTiO3)nSrO Ruddlesden-Popper superlattices are tunable dielectrics with unprecedented low loss up to 120 GHz and the highest reported figure of merit (FOM) of all known tunable dielectrics. This low loss is thought to be due to the defect-mitigating nature of the Ruddlesden-Popper structure. In this thesis, I improve on previous (SrTiO3)nSrO tunable dielectrics with the incorporation of BaTiO3 into the metastable structure, forming (SrTiO3)n 1(BaTiO3)1SrO. A single layer of BaTiO3 in the superlattice provides targeted chemical pressure to polarize the neighboring SrTiO3 for thicker, more tunable films, giving a record-breaking 200% improvement in the FOM for tunable dielectrics. Preliminary work on higher barium-concentration SrTiO3 Ruddlesden-Popper superlattices is shown for out-of-plane ferroelectric/tunable devices. These structures were made using molecular-beam epitaxy and required advancements in the calibration of shutter times to deposit precisely one monolayer. This thesis discusses a refinement of the standard SrTiO3 calibration by reflection high-energy electron diffraction (RHEED) oscillations, with an emphasis on oscillation curvature and codeposition oscillations. This thesis also describes efforts to understand the loss mechanisms in these materials by probing the defect-mitigating nature of (SrTiO3)nSrO via positron annihilation lifetime spectroscopy (PALS) of a compositional series of n = 6 (Sr1+δTiO3)nSrO films. Throughout this series the dominant positron lifetime shows little variance and is associated with TiOx vacancies. These TiOx vacancies are likely charge neutral and may be the (SrO)2 layers themselves, demonstrating the (SrTiO3)nSrO structure’s ability to accommodate off-stoichiometry and point defects. To date, the (SrTiO3)n−m(BaTiO3)mSrO films have only been integrated with non-industry-standard substrates like DyScO3. I integrated (Ba,Sr)TiO3 and (BaSrTiO3)n-1(SrTiO3)1SrO materials with silicon in MIM devices, showing the highest reported film quality and lowest current leakage in the literature. One to three layers of (SrO)2 were inserted into 20 nm thick (Ba,Sr)TiO3 films and found to further lower current leakage by an order of magnitude, demonstrating the Ruddlesden-Popper’s ability to dramatically lower loss. Finally, in a separate study, (SrTiO3)nSrO superlattices were examined for how their atomically precise interfaces suppress thermal transport for applications including thermal barrier coatings and thermoelectrics.