Characterisation of the mechanical properties of thin-film mirror coating materials for use in future interferometric gravitational wave detectors

Predicted by Einstein’s General Theory of Relativity, gravitational waves are periodic fluctuations in the curvature of space-time that propagate at the speed of light and are caused by acceleration of asymmetric mass distributions. The first ever direct detection of gravitational waves was a signal originating from the final moments of a binary black hole inspiral and the subsequent merger. Four more black hole inspiral and mergers have been observed since the first detection, as well as the inspiral and merger of a binary neutron star system. These detections have provided a wealth of information about the black holes/binary stars from which the signals originate, and further detections will continue to both test General Relativity and provide ground- breaking insights into previously poorly characterised astrophysical systems. The signals were detected by ground-based interferometric gravitational wave detectors. The sensitivity of interferometric detectors is dependent on reducing numerous sources of noise. A primary limiting noise source is the motion of the front face of test mass mirrors (which reflect the laser light within the detector arms) due to vibration from latent thermal energy in the highly reflective, multilayer coating materials. The power spectral density of this noise source is proportional to operation temperature and coating mechanical loss, which is a property describing energy dissipation within a material. Upgrades to current detectors, as well as proposed next-generation detectors, include operation at cryogenic temperatures to lower this thermally-induced noise. It is therefore crucial to know the low temperature mechanical loss of coating materials of interest for improving thermal noise within gravitational wave detectors. The research presented focuses on mechanical loss and structural characterisation of coating materials with the goal of increasing the observational range of future gravitational wave detectors through reduction of coating thermal noise. An introduction to gravitational waves, overview of astronomical sources, and discussion of the current, worldwide network of detectors and the sources of noise limiting their sensitivities is given in Chapter 1, which also contains a summary of all the detected signals up to this point. A more detailed discussion of coating thermal noise is given in Chapter 2, with a breakdown of the multiple forms of thermally- induced noise within test mass mirror coatings given, along with an explanation of the direct relation between mechanical loss and thermal noise. Accurately characterising coating mechanical loss over a wide range of temperatures is valuable for both estimating thermal noise in next generation detectors and better understanding the links between coating structure and loss. The methods used for characterising coatings are discussed in Chapter 3 in addition to an overview of coating technologies. This chapter provides an explanation for the procedures behind the measurements presented throughout this thesis. The mirror coatings used in current detectors are deposited via ion-beam sputtering and are comprised of alternating layers of high refractive index titania-doped tantalum pentoxide (Ti:Ta2O5) and low refractive index amorphous SiO2. Chapter 4 contains a comprehensive study of the changes in mechanical loss of ion-beam sputtered silica with respect to post-deposition heat treatment, providing updated loss values for thermal noise estimation of current coatings and potential coating designs for future detectors. A peak in dissipation is observed at low temperatures and found to change location and shape with heat treatment, but the activation energy of this thermally activated peak remained constant. The minimum loss at low temperatures is found to occur after heat treatment at 600 C, which is in contrast to the minimum loss at room temperature after 950 C. The loss of a silica deposited by a new coating technique, reactive low voltage ion-plating, was found to be lower as-deposited than the 600 C ion-beam sputtered silica. This technique involves deposition at over twice the temperature, so this loss result could be further evidence of high energy deposition techniques producing coatings with minimal distributions of two-level systems. Chapter 5 contains measurements of Al2O3, a potential replacement for SiO2 as the low refractive index coating material in multilayer coatings. The loss is found to be lower than 600 C heat-treated SiO2 by almost 50% at the low temperature peak but about four times higher at temperatures above 100 K. The loss of Al2O3 deposited at two different thicknesses (505 nm and 2.02 μm) exhibits minimal change in magnitude at temperatures below 40 K with heat treatment up to 800 C. Heat treatment at 300 C reduces the loss above 40 K, but further heat treatment fails to produce significant reduction. In contrast, the coating stress for both thickness went from ∼475 GPa compressive stress to over 140 GPa tensile stress. This could suggest that there is no strong connection between low temperature mechanical loss and coating stress within Al2O3. The high refractive index coating layers (Ti:Ta2O5) used in the advanced detectors are the dominant source of thermal noise within the highly reflective multilayer mirror coating at room temperature and has been shown to exhibit a peak in mechanical loss at cryogenic temperatures. Chapter 6 contains the mechanical loss characterisation of a number of alternate high refractive index coating materials. Increasing the titania doping percentage to 68% results in over an order of magnitude decrease in mechanical loss at temperatures below ∼100 C after heat treatment, but the optical and structural properties require further study. Doping Ta2O5 with zirconium instead (34.5%) produces room temperature loss similar to Ti:Ta2O5 with heat treatments up to 600 C with the added bene t of increased resistance to crystallisation. A decrease in loss with further heat treatment is expected with continued measurement. Measurements of pure TiO2 show loss that decreases with heat treatment up to 300 C, which is unexpected given evidence of crystallisation in titania coatings after annealing at 200 C. The loss of reactive low voltage ion-plated silicon was measured and found to be lower than any previously measured ion-beam sputtered amorphous silicon (aSi) coating. Ta2O5 deposited by this same coating method did not display the same reduction compared to ion-beam sputtering seen with SiO2 or aSi, with loss about 25% greater than ion-beam sputtered. Ta2O5 deposited through direct current magnetron sputtered did exhibit lower loss across all measured temperatures, but, with a measured absorption of ∼85 ppm, it is not competitive from an optical standpoint. All measured Ta2O5 coatings show a low temperature loss peak, but the peaks are not aligned with each other, ranging about 25 to 55 K. Calculations of the activation energies associated with each coating’s peak are valuable for possible correlations between structural properties, mechanical loss, and deposition technique. The mechanical loss characterisation from Chapters 3 through 5 are then summarised in Chapter 7, in which the coating Brownian noise is calculated for dual and multi-material coating stacks. All proposed coating combinations are compared to the estimated thermal noise of the mirror coatings in aLIGO and Advanced Virgo. Dual-material multilayer coating thermal noise estimates confirmed the value of Al2O3 as a low index material for detectors operating at temperatures within the silica loss peak, with a 20% and 25% reduction in Brownian noise at 10 and 20 K, respectively. The greatest reductions were seen with coatings where the high index material is amorphous silicon, which currently has absorption too high to use in a dual-material coating stack. Multi-material coatings have been proposed as a way to take advantage of these mechanical loss gains in light of the high absorption of aSi. The thermal noise estimate for the initial proposed coating design is updated based on new measurements of SiO2 after heat treatment at 450 C (the ideal annealing temperature for aSi absorption) and found to have significantly increased thermal noise. Replacing the SiO2 with Al2O3 reduces the Brownian coating noise by 35% and 39% at 10 and 20 K, respectively, and is an immediately viable option as a low noise coating for future detectors. Continued research of reactive low voltage ion-plated coating materials is important; a speculative multi-material coating deposited by this method and assumed to have loss reductions through doping and heat treatment shows great promise, with estimated Brownian noise reductions of 31%, 33%, and 21% at 10, 20, and 123 K, respectively, compared to the current advanced detector multilayer coating