Thermoelectric thin films are a promising power source for wearable devices, an alternative to batteries offering a continuous, long-lasting flow of electricity with improved mechanical flexibility compared to its bulk counterparts. They are quiet and reliable off-grid solution for generating electricity when exposed to a temperature difference. The exposure to numerous kinds of stress is expected for many applications. As a flexible device is deformed in various ways, its transport properties are required to remain relatively unaffected. Furthermore, depending on the synthesis method, intrinsic stress can be significant. Magnetron sputtering allows to control the magnitude of the intrinsic stress to a certain extent. Stress can be further influenced through adjusting variables such as the substrate, synthesis temperature and deposition rate. It can be also externally induced by bending or stretching the sample.

Despite numerous studies showing the potential benefits of controlling stress, it is often overlooked when reporting thermoelectric properties of a material. Depending on which aspect of stress is of interest, a suitable substrate can be chosen; one with minimal lattice mismatch for epitaxial growth, a selection based on the thermal expansion coefficients to vary the thermal stress, or a flexible substrate for researching the effects of tensile stress. Because thin films have small cross-sectional area relative to bulk materials, studying large stresses is possible applying relatively small forces. This makes them not only convenient to work with when it comes to creating microdevices and sensors for Internet of Things, but also the perfect objects to study effects of stress.

This work explores the effects of stress on the thermoelectric properties of Mg3Bi2-based thin films. To create a benchmark, a single crystal Mg3Bi2 thin film was synthesized, with its composition, structure and transport properties characterized. Later, small defects were introduced into the lattice by substituting some Bi atoms with Sb, creating polycrystalline films. The films had highest power factor around the room temperature, which became the focus of this work. The thermal stress was adjusted by selecting various substrates with wide range of thermal expansion coefficients, and synthesizing the Mg3Bi2 thin films at different temperatures - the estimated compressive stress reached up to 250 MPa on thermoplastic substrates, while the tensile stress went up to over 150 MPa on crystalline substrates. High compressive stress was observed to change the nature of major charge carriers from positive to negative. The measured power factor varied by four orders of magnitude due to stress, which could explain the wide range of reported values for same material systems in the literature.

This work also focuses on the thermomechanical stability of the amorphous TiNiSn thin films, which is flexible when amorphous, but brittle when crystalline. Unlike Mg3Bi2, this material is resistant to oxidation, and therefore suitable for studies even at high temperatures. The tensile force acting on the sample was altered at various temperatures, observing the changes in the structure of the material using synchrotron radiation. Up to 250◦C, the sample remained stable and virtually unchanged under tensile stress of approximately 3.2 GPa and below. The crystallization occurred at 300◦C when stress of 1.0 GPa was applied- which is over 200◦C below the reported crystallization temperature for TiNiSn without stress. The degree of the crystallization appears to be proportional to the applied stress and quite stable under constant forces. This is promising for creating crystalline-amorphous composites at lower temperatures, using stress as an additional control parameter. Altering the Sn content changes the stress required for inducing the crystallization.

The effects of stress were shown to be quite significant when it comes to thermoelectricity, as shown on Mg3Bi2 thin films. Furthermore, stress has allowed to observe crystallization in amorphous TiNiSn thin films at significantly lower temperature than otherwise required. It is often feasible to manipulate the stress, at least to a certain extent, by small alterations to the experiment, such as choosing a substrate and growth temperature. While at times it can be challenging or time consuming to measure the stress in the thin films, there are ways to estimate it. This work highlights various ways to induce, estimate and measure stress, along with the benefits of doing so.