Cumming, Alan V.
Aspects of mirrors and suspensions for advanced gravitational wave detectors.
PhD thesis, University of Glasgow.
Full text available as:
Gravitational waves were first predicted by Albert Einstein's Theory of general relativity, published in 1916. These waves are perturbations in the curvature of space-time. Indirect evidence of their existence has been obtained via observations of binary pulsar system inspirals by Hulse and Taylor. Research is now focussed on achieving direct detection of gravitational waves, giving a new way of observing astronoomical events in the universe.
Gravitational waves are quadrupole in nature, causing tidal strains in space. The weak nature of gravity means that the magnitude of these strains is very small. Only astronomical scale sources are likely to produce waves of sufficient amplitude to be detected on Earth. In the frequency band of a few Hz to a few kHz, the expected strain amplitude for violent sources is of the order of 10[superscript -22]. Detection is most likely to be achieved using long baseline interferometer detectors.
Currently several such detectors are in operation worldwide, including the GEO600 detector, built in a collaboration involving the Institute for Gravitational Research at the University of Glasgow, the Albert Einstein Institute (Hannover and Golm), and the University of Cardiff. In America the LIGO detector network has three large interferometric detectors - two of 4 km arm length and one with 2 km arms. In Italy a European collaboration has constructed the 3 km VIRGO detector. Currently GEO600 and LIGO have undertaken 5 data taking science runs with the most recent year long run, also involving VIRGO, concluding in November 2007. No detections have yet been confirmed, but analysis on the results of the most recent GEO600/LIGO/VIRGO run is ongoing.
These detectors are now operating at, or close to their design sensitivities, so research is focussed on reduction of various noise sources by upgrading of the detectors. One important noise source is thermal noise (both Brownian and thermo-elastic) - a limiting factor at midband frequencies. Reduction of mechanical loss in mirrors and their suspensions will help lessen the impact of thermal noise in future detectors.
The research detailed in this thesis was aimed at reducing thermal noise. In particular, it covers work undertaken to investigate the mechanical loss of suspension ribbons and fibres, test mass mirror coatings and also diffractive surfaces on test masses to evaluate their suitability for employment in future advanced gravitational wave detectors.
Upgrade of LIGO to "Advanced LIGO" will aim to reduce thermal noise by implementing mirror suspension techniques pioneered in GEO600. Specifically, it was initially proposed that test masses be suspended from silica ribbon fibres, a key choice that will be re-evaluated in this thesis. Ribbons (or fibres) will be fabricated by a CO[subscript 2] laser pulling machine being developed in Glasgow, with control programming being undertaken by the author.
Characterising the dimensions, strength and vertical bounce frequencies of the ribbons is important to confirm their suitability for use in detector mirror suspensions. A dimensional characterisation machine was constructed to measure the ribbon's cross sectional dimensions, with emphasis being placed on achieving high resolution in the ribbon neck regions, where the most bending occurs. Also, a bounce testing machine was constructed to experimentally measure the ribbon's vertical bounce frequency. Finally a proof load test was constructed to verify that ribbons could support the required weight. Results showed that ribbons could be fabricated successfully with the required strength and bounce frequency, though shaping of the cross section still requires further research to achieve the optimum.
In a pendulum system most of the energy is stored as gravitational potential energy rather than bending energy of the suspension fibres or ribbons. Thus the effective loss of the suspension fibres/ribbons is reduced or "diluted" and thermal noise is lower than may be naively expected. Dilution of the mechanical loss of the pendulum suspensions was investigated using finite element modelling. Methods for importing data from the dimensional characterisation machine were developed, and it was observed that the dilution resulting from ribbon suspensions was not as high as had been initially expected, with bending in the neck region of the ribbon being seen to significantly reduce dilution. It was observed that the rectangular ribbons had inferior dilution to equivalent cross section circular fibres for necks of the length typically being produced. A typical 7.5 mm necked ribbon was seen to have a dilution 1.5 times lower than an equivalent fibre, despite the ribbons having 3.3 times greater dilution with no necks. Ribbons were only seen to have this superior dilution for very short necks. Bending in the necks resulted in an increased amount of bending strain energy occurring which caused the lower dilution factors. Additionally, bending occurring in the ears that join the fibres or ribbons to the masses was seen to further reduce the dilution. In the light of low dilution factors, reduction (ideally nulling) of thermoelastic noise was studied. Reduction in thermal noise in this way is proposed through the use of tapered fibres, which showed that a lower overall noise level than that from the baseline ribbons planned for Advanced LIGO can be achieved, despite lower dilution factors. In the light of this work tapered fibres have now been adopted as the baseling for Advanced LIGO. Measurement of test mass mirror samples showed that the mechanical loss of mirror coatings can be significantly reduced by doping the high refractive index layer, with reduction up to a factor of 2.5 in measured mechanical loss observed, when compared to equivalent undoped coatings. In order to perform these measurements an interferometric read out system was constructed.
Future detectors will use higher laser powers which may cause thermal distortions in transmissive optical components. Use of all reflective components may be required to reduce this problem, possibly via diffractive mirrors. Measurements were undertaken on samples to discover if introducing a diffraction grating to an optic's surface increased the mechanical loss. However, the grating was not seen to do this, and also did not increase the mechanical loss of an optical coating applied on top of its surface, which verified that diffractive optics are viable for use in future detectors.
Actions (login required)