Dual Torsion Pendulum Sensor for Measurement of Terrestrial Gravitational Forces

Abstract

The emerging field of gravitational-wave astronomy has provided us with a wealth of new information about our universe. Gravitational-wave detectors have allowed us to observe new astrophysical events which were previously undetectable, with further discoveries expected as the sensitivity of these detectors is improved. In the near future, the sensitivity of terrestrial gravitational-wave detectors will become limited by the effects of gravitational noise from changing mass distributions here on Earth. This influence, known as Newtonian noise, is predicted to become the dominant source of noise for terrestrial detectors at low frequencies.

This thesis presents the development and testing of a novel sensor for measuring fluctuations in local gravity. This sensor is designed to measure the influence of Newtonian noise, and could be used to help develop future noise cancellation schemes for gravitational-wave detectors. These cancellation schemes require an accurate characterisation of Newtonian noise, which is yet to be measured.

The first prototype of this sensor has been constructed at ANU, and has been used to develop a sensing and control scheme for the system. The angular sensitivity of this detector has been measured as 3x10^-11 rad/sqrt{Hz} at 1 Hz and roughly 10^-8 rad/sqrt{Hz} at 0.1 Hz. It was found that the sensitivity of the experiment is currently limited by seismic noise.

This instrument is versatile and has other applications outside of the world of gravitational waves. This thesis presents a model which illustrates the feasibility of using this sensor to detect gravitational forces produced by earthquakes. Since gravitational information travels at the speed of light, these signals will arrive at a remote location significantly faster than any seismic waves. This means that in the future this sensor could be used as part of an early warning system for earthquakes. The response of an upgraded ANU sensor to 54 different earthquakes between magnitude 7 and 7.3 at a distance of 250 km was calculated. The median detection time was found to be 31.3 s for a signal-to-noise ratio of 5, which would provide 46.9 s of warning before the arrival of S waves. The best individual detection time was found to be 16.8 seconds, which would provide 61.4 s of warning time.