The atomic structure and properties of mirror coatings for use in gravitational wave detectors


Gravitational waves are a prediction of Einstein's General Theory of Relativity. They can be regarded as perturbations, or ripples, in the curvature of space-time that travel at the speed of light. Detectable gravitational waves are the result of the asymmetric acceleration of mass that occurs during massive astronomical events, such as coalescing compact binary systems and supernovae. The nature and detection of gravitational waves is the focus of Chapter 1. A direct detection of gravitational waves is still to be made, however, there is strong indirect evidence of their existence through the work Hulse and Taylor. They observed a binary pulsar system over a number of years and found it to have a decaying orbit that followed a decay rate consistent with a model in which energy is lost due to the production of gravitational waves. The most promising method for gravitational wave detection is through the use of long-baseline interferometric gravitational wave detectors, such as LIGO located in the US, GEO600 in Germany and Virgo in Italy. There are planned upgrades to current long-baseline interferometric gravitational wave detectors. These second generation of detectors will aim to improve sensitivity by a factor of around ten, allowing a much greater chance of detecting gravitational waves, particularly from sources such as coalescing compact binary systems. However, the sensitivity of these detectors will still be limited by noise sources, such as photon-shot, seismic and thermal noise, which could be further reduced by the development of new technologies. Chapter 2 discusses the current understanding of thermal noise arising from the mirror coatings in the detector test-masses. This will identify thermal noise as a particularly important noise source, limiting the sensitivity of detectors between the frequency range from a few tens Hertz to several hundred Hertz. There is an international network of scientists working on developing new technologies for future generations of interferometric gravitational wave detectors, which have the aim of increasing detector sensitivity and further reducing the effect of detector noise sources. The research presented in this thesis focuses on investigating the mechanical loss, which is directly related to the thermal noise, of the mirror coatings. In particular the first attempts at correlating changes in atomic structure of the coatings to the mechanical loss where various properties, such as heat-treatment and doping, of the coatings have been systematically changed will be presented. Chapter 3 will focus on the effect of heat-treatment of pure Ta2O5 coatings. The process of heat treating Ta2O5 coatings has observable effects on mechanical loss measured at low temperature, where loss peaks arise in the region of 10s of K and develop as the heat-treatment temperature rises. Heat-treatment also produces subtle changes to the averaged local atomic structure of the coatings where it can be seen that as the heat-treatment temperature is increased, the coatings became more ordered, moving towards crystallisation between heat-treatment at 300-600C coatings before fully crystallising at 800C. Atomic models show Ta2O2 ring fragments which are present in the crystalline phases of similar materials. In general it is observed that as heat-treatment temperature is increased there is an increase in the presence of the Ta2O2 ring fragments and a decrease in the presence of Ta-Ta bonds in the atomic structures. Changing the manufacturing deposition process for the Ta2O5 coatings also creates significant changes in the mechanical loss at low temperatures, where a `low water content' manufacturing processes gives rise to changes in the positions and shapes of the low temperature loss peaks. Preliminary investigations into the local atomic structure at different areas of a heat-treated coating shows that increasing heat-treatment temperature causes more ordered coating material nearer the substrate, compared with areas nearer the surface of the coating. Chapter 4 presents studies on the effect of doping Ta2O5 coatings with TiO2 with doping concentrations of 0, 8.3, 20.4, 25.7, 28.3, 53.8% (cation) TiO2. Mechanical loss measurements of multi-layer SiO2 and Ta2O5 doped with TiO2 coatings show that changing the TiO2 doping concentration reduces the mechanical loss of the coating by up to 40%. It is also shown that changing the TiO2 doping concentration can significantly change the local atomic structure of these coatings. Atomic models created for 20.4% and 53.8% Ti coatings indicate similar inter-atomic bond distances between the 20.4% and 53.8% Ti coatings. The models show that the distributions of Ta-Ti and Ti-O bonds in the atomic structure of the coatings as TiO2 doping is increased. There are also considerable contributions from Ta2O2 ring fragments that are seen in the pure Ta2O5 coatings, with the addition of TaTiO2 ring fragments. Further analysis of the atomic structures of these coatings revealed some preliminary correlations between the atomic structure and mechanical loss, were it is observed that 28.3% Ti coating is the most ordered atomic state out of all the Ti doped coatings and had the lowest measured mechanical loss. This suggests that there may be a link between slightly increased ordering in the atomic structures and a lower measurable mechanical loss. The amount of oxygen in a coating may play a key role important in the level of mechanical loss, as it is observed that the coating with the least oxygen deficiency coating is the coating with the lowest measured mechanical loss. Finally, Chapter 5 explores the material properties and atomic structures of HfO2 coatings, SiO2 coatings and substrates and HfO2 doped with SiO2 coatings. Pure HfO2 are studied as possible alternatives to Ta2O5 coatings. It appears that coatings subject to heat during the manufacturing process of just 100C or above appear part crystallised. Preliminary studies of a HfO2 coating doped with 30% (cation) SiO2 and heat-treated to 600C show that it is a promising coating for future study as it remains amorphous, with a room temperature mechanical loss value comparable to pure HfO2 coatings and therefore Ta2O5 coatings. SiO2 coatings deposited on SiO2 substrates are also studied and they show only subtle changes between them, which appear to lessen as the sample are heat-treated. Changes in the atomic structure of these coatings indicate an increase in order of the structure as heat-treatment temperature is increased, similar to the observed changes in the heat-treated Ta2O5 coatings.


June, 2011