Effects of Thermal Annealing on Amorphous Oxides for Gravitational-Wave Detector Mirror Coatings: A Real-Time Spectroscopic Ellipsometry Approach

The direct detection of gravitational waves in 2015 marked a major milestone in physics and opened the era of gravitational-wave astronomy. These spacetime ripples, predicted by Einstein in 1916, are generated by extreme astrophysical events such as black hole mergers and neutron star collisions. Their observation has provided access to phenomena that cannot be studied with electromagnetic telescopes. Current interferometric detectors, such as the European Virgo and the American LIGO observatories, can resolve measurable strain amplitudes of 10⁻²¹, corresponding to length variations of 10⁻¹⁸ meters over kilometer-scale baselines. A major limitation to the sensitivity of current detectors is coating thermal noise, which originates from thermally driven mechanical dissipation within the multilayer dielectric coatings that form the mirror surfaces. Such structures, composed of alternating high- and low-index amorphous oxides, must achieve extremely high reflectivity while minimizing optical absorption and mechanical loss. In current detectors, titania-doped tantala is the standard high-index material, but it remains the dominant source of this noise, motivating the search for alternative high-index oxides that exhibit lower mechanical dissipation while strictly preserving comparable or superior optical properties. Thermal annealing is essential for reducing mechanical losses in amorphous films, yet it induces metastable structural and optical transformations that must be understood to design optimized coatings for next-generation detectors. This thesis introduces a comprehensive method for real-time, in-situ optical characterization of amorphous oxide coatings during thermal annealing. The approach combines spectroscopic ellipsometry with precise thermal control and automated data acquisition. A thorough evaluation of the experimental setup was conducted, including a systematic identification and mitigation of all setup-related and analysis-related sources of error. To establish accurate baseline conditions, temperature-dependent optical models for the substrate and its native silica were developed over the 30–600 °C range, providing a reliable description of the layers underlying each coating. In addition, a suite of automated scripts was implemented to standardize data acquisition and analysis, ensuring reproducibility and allowing the method to be applied across multiple samples. This approach provides direct access to key material parameters — such as refractive index, layer thickness, and optical gap — and to their evolution under heat treatment. It thus offers a rigorous and sensitive route for characterizing the optical and structural evolution of the material during the annealing process. The method was applied to the most relevant high-index oxides used in gravitational-wave detector coatings, namely the current standard (titania-doped tantala) and two promising alternatives, hafnia–tantala and titania–germania. Titania–tantala and hafnia–tantala exhibit similar overall trends, with structural evolution initiating around 200 °C and the largest variations occurring during the heating ramp, followed by comparatively minor changes in the subsequent high-temperature plateau. A more extensive investigation was carried out on titania–germania, with multiple annealing cycles designed to isolate the effects of heating rate, annealing temperature, and plateau duration. The material showed a stronger response to temperature than to the other parameters. While at 500 °C the coating still behaves in close analogy with titania–tantala and hafnia–tantala, at 600 °C it undergoes a thermal response not previously observed in any oxide used for gravitational-wave detector coatings. This regime is characterized by a rapid inversion of both thickness and refractive index at the onset of the high-temperature plateau, accompanied by a pronounced reduction in the optical gap and a substantial release of argon. We propose that this inversion reflects a structural reorganization driven by the enhanced atomic mobility that develops as the material approaches the glass-transition region of germania. At 550 °C, the system enters an intermediate regime whose response is neither fully conventional nor as extreme as the behavior observed at 600 °C. Taken together, these results demonstrate that the thermal response of high-index oxides is strongly material dependent and that effective optimization of coating performance requires material-specific annealing strategies. They also highlight the capability of real-time optical monitoring to reveal relaxation mechanisms that remain inaccessible to ex-situ analyses. The insights obtained are directly relevant to improving the optical and mechanical performance of future interferometric mirrors.