The use of millimeter-wave carrier frequencies has the potential to revolutionize wireless telecommunications by providing a massive increase in available bandwidth. However, millimeter-wave communications are hindered by poor atmospheric and building penetration, and require complicated RF front-end architectures. Tunable dielectric thin films offer a fast, compact, and cost-effective way to overcome many of the challenges facing the use of millimeter-wave spectrum. Few materials have been characterized in the millimeter-wave regime where measurements become increasingly challenging as test signal wavelengths approach the physical size of devices. The few tunable dielectric materials that have been studied at these frequencies suffered from high dielectric loss or other limitations. In this dissertation, we address both the measurement and materials challenges that have limited the commercial implementation of tunable millimeter-wave devices.

In this work, we describe our implementation of a unified on-wafer approach to measure the relative permittivity of thin films and substrates across a continuous frequency band from 100 Hz to 110 GHz. We achieve this ultra-wide bandwidth by combining electrical measurements of on-wafer planar capacitors and transmission lines, and use finite-element simulations to connect our electrical measurements to material properties.

Motivated by the need for better tunable dielectrics, we also developed a high throughput technique to accelerate the discovery of tunable dielectric thin films. We discuss this technique, which is inspired by the principles of combinatorial materials science and the “Materials Genome Initiative”. Our technique enables the characterization of many unique material compositions using a single 10 mm composition-spread thin film chip. In addition to speeding up the synthesis, fabrication, and measurement steps, the single-sample nature of this approach provides extreme consistency in the processing variables that impact dielectric properties.

Finally, we present another approach to tunable dielectric materials discovery with our development of (SrTiO3)n−1(BaTiO3)1SrO thin films incorporating “targeted chemical pressure”. These atomically-precise, strain-engineered superlattices achieve unparalleled performance, with measured relative tunability of almost 50 % and low dielectric loss even beyond 100 GHz. We discuss our use of the materials-by-design approach, which incorporates collaboration between theory, synthesis, and characterization, to overcome barriers to commercial integration without sacrificing advantageous material properties.