The University of Western Australia
This paper presents an overview of Australian experimental research in gravitational radiation detection and describes the research program of the Australian Consortium for Interferometric Gravitational Astronomy which is constructing the first stage of the Australian International Gravitational Observatory.
AIGO, the Australian International Gravitational Observatory, was first proposed in 1989. It was then recognised that a large scale laser interferometer was required in the southern hemisphere to compliment planned detectors in the USA, Europe and Japan. During the following years four groups in Australia, in collaboration with many overseas groups, have developed advanced techniques for laser interferometry. In the same period the UWA cryogenic resonant mass gravitational wave detector, Niobe, was brought into long term operation and has operated at a burst sensitivity h~ 7 x 10-19 from 1993 to early 1998 when it was warmed up to install improvements.
Figure 1: Map showing location of AIGO in relation to Perth
In 1997 funding was received for the first stage of AIGO, in the form of an advanced research interferometer. This was to be an interferometer located in the AIGO cornerstation on site at Wallingup Plain near Gingin (see Figures 1 and 2), using full scale isolation and suspension systems and advanced interferometric techniques, but with reduced arm length. The interferometer will be upgraded to an observational arm length in the future, but in the interim will be used for extensive evaluation and development of advanced techniques on an 80 metre baseline.
Figure 2: Site Plan for AIGO Stage I and the Gravity Discovery Centre
This paper summarises the work of ACIGA and gives details of the laser system, interferometer system, and the suspension and isolation systems for the new interferometer.
The Australian Consortium for Interferometric Gravitational Astronomy was formed in 1993 to coordinate interferometric gravitational wave research in Australia. The Consortium consisted of ARC funded research groups at the Australian National University (ANU), the University of Adelaide and the University of Western Australia (UWA), along with the CSIRO Optical Technology Centre at Lindfield and members of the Monash University Applied Maths Department.
The Consortium worked on a broad front addressing the breadth of gravitational wave research from the theory of gravity wave sources to high power laser technology, quantum noise and vibration isolation. The research has been highly productive, leading to important new insights into noise mechanisms, prediction and analysis of new sources of gravity waves, a design for a novel high power cw Nd:YAG laser and innovative new vibration isolators and suspension systems. Perhaps the greatest prize of all was that the CSIRO Optical Technology Centre won the contract (under world wide competition) to supply most of the optical substrates for the giant US LIGO project, which is building two 4 km arm length laser interferometer gravitational wave detectors in the USA, one at Hanford, Washington, the other at Livingstone, Louisiana. In addition ANU has formed a data analysis sub-group which, in collaboration with existing data analysis efforts at UWA (focussed on the niobium bar gravitational wave detector), completes Australian all-round expertise in this exciting area of physics.
In 1997 the Consortium was funded to begin construction of the first stage of a laser interferometer gravitational wave detector. Defined as an Advanced Research Interferometer (ARI), the first stage project consists of construction of the cornerstation of a large scale instrument on a site about 80km north of Perth, on Wallingup Plain near the town of Gingin. The site, which has room for a 4km by 4km interferometer, is being provided by the WA Government. The Gingin Shire is providing road access and ARC and University funds are providing the equipment and supporting research costs. Since the project began, several industry groups have offered modest sponsorship funding which has allowed the scope of the project to be expanded somewhat. Gravitational WavesÑA Thumbnail Sketch for Novices
Just as MaxwellÕs Equations predict Electromagnetic Waves, so EinsteinÕs Field Equations of General Relativity predict Gravitational Waves. Gravitational waves are waves in the Riemann curvature tensor. They can be thought of as ripples in the curvature of space-time. They travel at the speed of light. Due to the fact that gravity has only a single positive charge, dipole waves are not allowed. Gravitational waves are quadrupole waves: they arise from time varying mass quadrupole moments. This means that gravitational waves do not produce an acceleration measurable at a single point, but a time varying gravity gradient, that can be sensed like a tidal distortion, as a relative motion between two points. Figure 3 is a schematic diagram of a gravitational wave. From the quadrupole nature of gravity waves, it follows that gravitons are bosons with a spin of two and there exist two independent polarisations which are displaced from each other by 45 degrees, rather than the 90 degree displacement for photons which have a spin of 1. In General Relativity the coupling between space-time curvature and momentum-stress-energy (the sources of mass which create the curvature) is described by the tiny coupling constant 8pG/c4 which means that the curvature created by conventional mass sources, such as planets or stars or lead balls, is tiny. For example, for the Earth it is about 10-9, while at the surface of the sun it is nearer to 10-6. However the fact that gravity has only one charge means that gravitational charge can accumulate indefinitely, and hence it is possible to have arbitrarily large sources (of mass and hence spacetime curvature). Such sources of course are black holes. When large space-time curvature is possible in a single object it is not surprising that ripples with correspondingly large amplitude can be created by the interaction of pairs of such objects.
Figure 3: Schematic diagram of a gravitational wave as it acts on a ring of test particles
A binary pair of orbiting masses such as stars or black holes generates gravitational waves. The energy radiated away makes the system appear viscous. The objects spiral together as energy is lost. However the rate of energy loss goes as the sixth power of the orbital frequency, so the stars coalesce at a steadily increasing rate. If a pair of neutron stars is close enough that their orbit is less than half a day, they will coalesce in less than the age of the universe. However the time for the earth to coalesce with the sun due to gravitational radiation is effectively infinite! When a pair of black holes coalesce a large fraction of the gravitational potential energy can be radiated in gravitational waves. This is a large fraction of the rest mass energy of the system. Similarly, if neutron stars coalesce at least a few percent of the system rest mass energy may be converted to gravitational waves. The majority of the energy is radiated in the last few orbital cycles: for neutron stars or stellar mass black holes these take only milliseconds, so at this moment the gravitational luminosity of such a system approaches the total electromagnetic luminosity of all the stars in the universe.
In spite of the large luminosity of such systems, their detection is not easy due to the small coupling factor we discussed above. Even very large fluxes of gravitational waves equate to tiny changes in curvature, and correspondingly tiny motions. For example, at 1kHz a flux of 30Wm-2 (prodigious by astronomical standards) corresponds to a strain amplitude (fractional change in spacing between test masses due to the spacetime curvature) of 10-20.
In gravity wave detectors we aim to detect induced vibrations with strain amplitude of say 10-22. This may be done acoustically (picking up the vibration induced in large metal bars, such as the niobium bar detector at UWA, (see figure below) or interferometrically by measuring the motions of nearly free test masses with a laser interferometer (the principal of the project discussed here). The laser interferometer may be made almost as long as money can buy, thus providing a relatively straightforward means of increasing sensitivity. Gravity wave technology is about creating measurement systems capable of measuring the small motions required. Some groups are working on designs for space-based instrumentation (5million km space interferometers designed to detect GW in the milliHz range), but the majority of research focuses on creation of detectors capable of detecting audiofrequency gravitational waves on Earth. Acoustic detectors are the best detectors created so far, but they only detect in relatively narrow frequency ranges ? say 700-800 Hz. Laser interferometers in principal are broadband, say 10Hz to 1kHz, and can also be made narrow band and tunable.
Figure 4 : Schematic of the niobium bar gravity wave detector at UWA
Because gravity waves cause small motions, it is interesting to make an analogy between sound waves and gravitational waves. Electromagnetic waves provide us with an extended sense of vision with which to see the universe, and gravity waves provide a sense of hearing. We are at the threshold of a time when humanity will, for the first time, be able to listen to the universe. Gravity wave detectors are the bionic ears with which a new sense - a new spectrum - will become available. Our hopes are that we will be able to hear the roar of the big bang, chirrups of coalescing stars, clicks and ringing from black hole births and continuous whistles from pulsars. But most of us expect that the universe will have surprises in store. Whether or not our predictions are correct, we can certainly anticipate the delight that a deaf child experiences when the bionic ear is first switched on.
Figure 5: Schematic diagram of a laser interferometer gravitational wave detector
Figure 5 shows the basic concept for a laser interferometer gravitational wave detector. It consists of a high power laser injecting light into a Michelson interferometer. All of the components are vibration isolated and placed in a high vacuum to eliminate the effects of air density fluctuations. High optical power is required to minimise photon shot noise (associated with the discrete nature of photons). Because the interferometer must be very long and because optical losses must be minimised, the optical components must be of extremely high quality. Typical precision for polished substrates is ~1nm (that is l/1000, about 100 times better than required in most conventional high quality optics) and coating losses (scattering and absorption) must be correspondingly low. (typically 10-5 ? 10-6).
The suspended interferometer must be servo controlled to maintain correct alignment and spacings without introducing extraneous vibrational noise. This is achieved partly by using local motion sensors and magnetic actuation, and partly by modulating the laser beams and deriving error signals that allow forces to be applied to create a global locking of the interferometer in its desired dark fringe state. However all forces must be applied at frequencies below the low frequency cut off bandwidth of the detector, say 20Hz.
The cw laser must have extremely narrow linewidth, extremely low phase and intensity noise, and be able to operate stably for long periods of time. This requires very careful design of the laser itself (constructed as a cascade of low medium and high power injection locked diode pumped Nd:YAG lasers), as well as the locking of the laser to a reference optical cavity followed by filtering of the laser in a long suspended (vibration isolated) optical cavity called a mode cleaner. Because all these parameters alone are insufficient the interferometer is equipped with additional mirrors to allow resonant buildup of the laser light. This can be achieved in several different ways including the one shown in Figure 5 in which a power recycling mirror causes resonant buildup of the input light and a signal recycling mirror that create a resonant cavity for the signal sidebands.
During 1995-97 ACIGA concentrated on developing the design and the technology for an advanced interferometric gravitational wave detector that would use superior technology to those already under construction in the USA (the LIGO project) and Italy (the Italian-French VIRGO project). The goal was to design a system that had lower mechanical noise, (intrinsic thermal noise and external vibrations), lower optical noise (photon shot noise and laser frequency and intensity noise) and higher intrinsic sensitivity (using newly proposed optical configurations such as the dual recycling Michelson interferometer). At UWA particular efforts went into the design of vibration pre-isolators. These are designed to cut out very low frequency vibrations by mimicking the dynamics of enormous pendulums. An isolator equivalent in frequency and performance to a 6km long pendulum was demonstrated and a design was created that is expected to vastly simplify the operation of an interferometer. In collaboration with LIGO. VIRGO and CSIRO, the UWA group also pursued the development of artificial sapphire optical components to be used as test masses because of the low thermal noise of sapphire. The latter arises because of sapphireÕs very high YoungÕs modulus and very low acoustic losses. Optical polishing was successfully developed by CSIRO which enabled mirrors with losses below 10-5 to be developed. Test mass suspensions with acoustic quality factor above 50 million were also demonstrated.
Adelaide developed a 5W cw diode pumped single frequency Nd:YAG laser and a design for a laser that could produce in excess of 100W. The low noise, reliability and diffraction limited beam quality of the 5W laser makes it suitable for the first stage laser interferometers. Advanced long baseline interferometers, however, will require lasers that can produce much higher powers. The Adelaide design for a high power laser results from applying high power laser principles to the Nd:YAG gain medium and using the latest optical techniques, rather than trying to scale up the low power laser design.
The ANU group has concentrated on analysing advanced optical recycling techniques to determine the most appropriate optical layout for the next generation of suspended mass instruments. First generation detectors , will use a Michelson interferometer modified by the incorporation of Fabry-Perot cavities in the arms in addition to a power recycling mirror. As these devices- are operated on a dark fringe for the carrier frequency, i.e. destructive interference at the output port (see Figure 5), the incoming light which has not been scattered or absorbed is returned from the beam splitter back toward the laser. Power recycling re-uses this light by reflecting it back into the interferometer. In this way the sensitivity of the instrument is improved without changing its bandwidth. So-called advanced recycling techniques involve trading off bandwidth for enhanced sensitivity around a particular gravity wave frequency. This is achieved by placing a mirror, the signal-recycling mirror, at the detector output. In this way signal is resonantly enhanced in an optical cavity formed by this mirror and the interferometer optics. Depending on the actual resonance condition, and on whether the arms contain optical cavities, such recycling is referred to as either dual recycling or resonant sideband extraction.
The ANU group have modelled both techniques and built bench top prototypes to confirm predicted response curves. An important aspect of this work has been the development of RF modulation methods to extract mirror control signals and gravitational wave signals. Optical recycling is only of use in frequency regions where the interferometer is limited by photon noise. The extent of such regions depends on the ability of the isolation and suspension system to minimise thermal noise and seismic noise and on having ultra stable lasers. A critical part of the optical system is a succession of laser frequency and intensity stabilisation stages. The UWA group has developed passive stabilisation reference cavities and precision frequency locking techniques. These can be used to create a low power ultrastable reference laser. The Adelaide high power lasers can then be injection locked to this laser. ANU has developed a detailed theory on noise transfer through injection locked lasers and experimentally verified predictions. This work has shown that a combination of injection locking and electro-optic feedback can deliver a laser of exceptional amplitude stability.
Once photon noise limits the interferometer performance, quantum optical techniques such as the use of squeezed vacuum states can be used to reduce this noise level. The ANU group has recently set a world record in the generation of squeezed vacuum states and has plans to experimentally verify sensitivity improvement in interferometers injected with squeezed light.
Several years ago the WA Government supported a feasibility study into an Australian Gravitational Wave Observatory. The study lead to a range of positive conclusions. First the great significance of an Australian detector to a worldwide array was recognised. An Australian detector has maximal baselines to all the proposed Northern Hemisphere detectors, which means that the directional resolution achievable through signal phase delay measurements is greatly improved. In addition the noise reduction achievable through cross correlation is improved by the number of independent baselines. Finally the probability of coincidence detection between two widely spaced detectors is greatly improved because the Australian detector is nearly coplanar with the northern detectors. (Observation of coincident events depends on source direction and polarisation relative to the orientation of the detector. The detector orientations are largely dictated by geography and are non-optimally located. An adverse direction or polarisation could mean that one detector sees a source while the other is insensitive).
For this project the WA Government has provided the site and funds to
construct the Observatory building. This will be in the form of a 25m by
25 m by 10m high building, mostly consisting of a large insulated and electrically
clean room in the middle of a large expanse of flat, pristine banksia forest
on the sand plain one hourÕs drive from UWA and 20km from the coast.
An artistÕs conception is shown in Figure 6. The cornerstation has
workshop, office and accommodation facilities. Inside, large stainless
steel tanks will house a suspended interferometer. The tanks contain large
vibration isolation structures that will reduce the maximum amplitude of
vibration above 1Hz to about 1nm. This motion is so small compared with
an optical wavelength that alignment and control will be achievable by
minimal forces, derived from a simple PC computer based digital servo system.
The very steep role-off of multistage isolators means that seismic vibration
in the signal band is negligible. Figure 7 shows the target vibration performance.
Figure 6 : Aerial phot of Observatory Cornerstation building and Endstations
The test masses will consist of sapphire masses suspended by delicate niobium suspension foils, which give a very high possible pendulum Q-factor. The predicted thermal noise performance of such a system is also shown in Figure 7.
The isolators and vacuum system for the ARI are currently under construction, while simultaneously Adelaide is completing the 5W laser and developing the 100W laser. A company has agreed to donate two end stations for the interferometer which means that an arm length up to 100m may be possible in the initial instrument.
ANU is developing the detection optics for the ARI. This includes low noise high power optical detectors and modulation and demodulation systems. Installation, testing and integration of all systems will continue during 1999 and into 2000. There will undoubtedly be many problems to solve and difficulties to overcome before the ARI is a fully operational and a reliable instrument. It is intended to use the interferometer to verify the performance of all of the individual systems, and to test various modulation techniques and optical configurations. As soon as feasible it will be worth extending the arm length of the device to at least 1km. While this is expensive it is straightforward, especially now that LIGO has successfully completed 16km of UHV pipe without a single leak! As well as pipe, it requires the use of new test masses and optical surfaces, and modified injection and detection optics. When this is completed early in the next century Australia will have an Observatory which should ensure that it continues to play a significant role in the frontiers of astronomy.
Figure 7 : Vibration isolation performance and thermal noise. (mass = 20kg, Qp=6¥107, Qint = 108)
Figure 8: Conceptual diagram of vibration isolator
It is planned to build a public education facility - the Gravity Discovery Centre - near to the Observatory (but not too near!) It is proposed that this centre would focus on the big questions of science and the universe, but including technology, scientific spin offs, science, art and sculpture. A particular feature that is planned is a series of displays and murals on comparative cosmology: creation myths of various cultures compared with modern cosmology. This concept has the support of the traditional owners of the Observatory site.
The Australian Consortium for Interferometric Gravitational Astronomy wish to thank the bodies who, through grant funding, have made this project possible: The Australian Research Council, The Government of Western Australia Dept of Commerce and Trade, Gingin Shire Council, The University of Western Australia, Australian National University, The University of Adelaide and Monash University. We particularly wish to thank Dr John Barker, Professor Michael Barber, Mr Morrie Moller, Em. Professor John de Laeter, The Hon. Judi Moylan and all members of the GDC Steering Committee for their advice and encouragement.
The University of Western Australia: D.G.Blair, Ju L., M. Notcutt, J. Winterflood, Yang Y., Zhao C.
University of Adelaide: J. Munch, P. Veitch, M. Hamilton, D. Ottaway, D. Mudge, C. Hollitt, P. Kloevekorn
CSIRO Lindfield: C. Walsh (Group Leader), A. Leistner (Head of Fabrication), B Oreb (Head of Metrology), J. Seckold, R. Bulla, E. Pavlovic, G. Davis, W. Stuart, D. Farrant, F. Lesha, C. Sona, R Yin, R. Netterfield, D. Drage and C. Freund.
Monash University: A. Lun, J. Monaghan, , L. Brewin
P. J. Veitch, J. Munch, M. W. Hamilton, D. Ottaway, A. Greentree and A. Tikhomirov, Aust. J. Phys. 48, 999 (1995).
A. Abramovici, W. Althouse, R. Drever, Y. Gursel, S. Kawamura, F. Raab, D. Shoemaker, L. Seivers, R. Spero, K. Thorne, R. Vogt, R. Weiss, S. Whitcomb, M. Zuker, Science, 256, 325 (1992).
B. Caron et al, Nucl.Phys.B (Proc.Sup.) 54B, 167 (1997).
R.W.P Drever, in Gravitational Radiation, N. Deruelle and T. Piran (Eds), North Holland (1983).
B.J. Meers, Phys.Rev.D38, 2317 (1988).
J. Mizuno, K.A., Strain, P.G. Nelson, J.M. Chen, R. Schilling, A. Rüdiger, W. Winkler and K. Danzmann, Phys.Lett.A175, 273 (1993).
M.B. Gray, A.J. Stevenson, H.-A. Bachor and McClelland D.E, Appl.Optics (in press).
D.A. Shaddock, M.B Gray and D.E. McClelland, Appl.Opt. (in press).
C.C. Harb, T.C. Ralph, E.H. Huntington, I. Freitag, D.E. McClelland and H.- A. Bachor, Phys.Rev.A54, 4370-4382 (1996).
E.H. Huntington, C.C. Harb, T.C. Ralph, D.E. McClelland and H.-A. Bachor, Opt.Comm.145, 359-366 (1998).
V. Chickarmane, S. Dhurandhar, T.C. Ralph, M.B. Gray, H.-A. Bachor and D.E. McClelland, Phys.Rev.A57, 3898-3912 (1998).
P.K. Lam, J. Gao, T.C. Ralph, D.E. McClelland and H.-A. Bachor, Internal Report (1998).