UC Berkeley

The simultaneous metal-insulator and structural phase transitions of vanadium dioxide (VO2) makes this material attractive for a variety of applications. Heating VO2 above 67 °C, the electronic conductivity increases by a factor of 105, and the lattice shrinks along the c-axis by 1%, a very large spontaneous strain. As a strongly correlated electron material, VO2 is extremely interesting from a condensed matter physics perspective, with ongoing research into the nature of the metal-insulator transition (MIT). Vanadium dioxide began drawing significant interest as early as the 1950s, but studies were often hindered by the tendency of bulk samples to crack upon passage across the phase transition. Recent advances in sample synthesis allow us to study VO2 nanobeams and thin films, which are extremely high quality and mechanically robust. This dissertation focuses on applications of VO2 in actuation and sensing, tuning the phase transition behavior by doping, and understanding the behavior of point defects in response to the structural phase change.

First, we develop bimorph actuators with VO2 nanobeams utilizing the phase transition as solid engine converting heat into motion. Microscale mechanical motion is typically actuated by mechanisms including electrostatic interaction, thermal expansion, and piezoelectricity, as well as more exotic types like shape memory, electrochemical reactions, and thermal responsivity of polymers. These mechanisms typically offer either large-amplitude or high-speed actuation, but not both. We demonstrate a microscale solid engine (SE) based on the phase transition of VO2 at 67 °C with large transformation strain (up to 2%), analogous to the steam engine imparting large volume change in a liquid-vapor phase transition. Compared to poly-crystal thin films, single-crystal VO2 nanobeam-based bimorphs deliver higher performance of actuation both with high amplitude (> bimorph length) and at high speed (> 4 kHz in air and > 60 Hz in water). The energy efficiency of the devices is calculated to be equivalent to thermoelectrics with figure of merit ZT = 2 at the working temperatures, and much higher than other bimorph actuators. The bimorph SE can be easily scaled down to the nanoscale, and operates with high stability in near-room-temperature, ambient or aqueous conditions. Based on the μSE, we demonstrate a macroscopic smart composite of VO2 bimorphs embedded in a polymer, producing high-amplitude actuation at the millimeter scale.

To implement this powerful phase-transition actuation technology in a more controllable and reliable manner, we next develop patternable bimorph actuators using VO2 thin films and photolithography. Existing actuator materials cannot simultatneously deliver high amplitude, work output, and speed, but we show that our microactuators excel in all of these metrics. Our VO2 bimorph actuators are driven by the structural phase transition structural in response to heat, electric current, or light and bend with exceedingly high displacement-to-length ratios up to 1 in the sub-100 µm length scale, work densities over 0.63 J/cm3, and at frequencies up to 6 kHz. These devices operate in ambient and aqueous environments, and their microfabrication process enables integrated designs of planar as well as three-dimensional geometries using common process steps for Si-based MEMS. Combining the superior performance, high durability, diversity in responsive stimuli, versatile working environments, and microscale manufacturability, these actuators offer potential applications in micro-electromechanical systems, microfluidics, robotics, drug delivery, and artificial muscles.

Adapting VO2 to a wider range of applications and environments requires control of the MIT behavior, which can be controlled by doping the material with transition metals. Often, lower operating temperature is desired for achieving the desired functionality with less thermal or other energy input. We study the effect of tungsten doping in VO2 films, which lowers the MIT temperature by 23 K / % W, but increases the electron concentration and dramatically reduces the carrier mobility in the material. This characterization of the electronic properties will assist future design of electronic devices utilizing doped VO2. Within nanobeams, we incorporate tungsten with a spatially varying profile to broaden the MIT behavior and increase the range of applications. At room temperature, the graded doped nanowires show metal phase on the tips and insulator phase near the center of the nanowire, and the metal phase grows progressively toward the center when temperature rises. As such, each individual nanowire acts as a micro-thermometer that can be simply read out with an optical microscope. The nanowire resistance decreases gradually with the temperature rise eventually reaching two orders of magnitude drop, in stark contrast to abrupt resistance change in un-doped VO2 wires. This novel phase transition yields an extremely high temperature coefficient of resistivity ~ 10 %/K, simultaneously with a very low resistivity down to 0.001 Ωcm, making them a promising infrared sensing material for un-cooled micro-bolometers. Lastly, they form bimorph thermal actuators that bend with an unusually high curvature, ~900 m-1K-1 over a wide temperature range (35 – 80 °C), significantly broadening the responding temperature range of previous VO2 bimorph actuators.

Finally, we study point defects in VO2 and their interaction with the structural phase transition. Several potential mechanisms are discussed for enhanced point defect motion in a material undergoing a significant structural distortion (like the 1% expansion VO2 experiences in the MIT). From this structural viewpoint, strain fields and anelasticity may encourage point defects to migrate near the phase transition boundary, regions of large strain. Additionally, local energy supplied by the domain wall between phases or by the latent heat may also activate point defect motion. We investigate the potential for a MIT-based anneal in VO2 by introducing point defects into thin films by ion irradiation. By monitoring the crystal quality of the film via Hall effect and RBS channeling, we compare two different anneals: a cyclic anneal (thousands of times across the phase transition) and a conventional single-step anneal, holding the VO2 film just above the MIT temperature. The results suggest that there is no enhanced point defect motion with the phase change, as the defective thin films recover similarly whether cycled thousands of times across the MIT or held at a constant T above TC for an equivalent amount of time.