The earliest attempts to detect gravitational waves were with resonant bar detectors. The idea behind these devices is that as a gravitational wave passes through an object, it exerts a tidal force on that object, causing the object to deform. If the object is vibrating at a characteristic resonance, then the deformation will appear as a deviation from its resonant ringing. The first resonant bar detector was created by Joseph Weber in the 1960s, and work has continued with these devices ever since (Details about resonant bar detectors are given in Ju et al. [2000]1).
As the name implies, a resonant bar detector is a massive cylinders of several tons. By measuring the acoustic signal at the end of the bar, researchers hope to detect the bar's interaction with a gravitational wave. As with the Michelson interferometer gravitational wave detectors, the game in developing a resonant bar detector is to suppress the sources of noise. The two fundamental sources of noise for a resonant bar detector are from thermal displacement, because the thermal motion of the atoms in the bar cause the bar to vibrate, and the ability to convert the acoustic signal into electrical signals, which is an issue of electronics and the coupling of the electronics to the bar.
To suppress the thermal noise, the bar is cooled to very low temperature; low in this case means from several degrees Kelvin to 0.1° Kelvin. For such temperatures, one can see the quantum energy states for the bar. This sets a threshold for detecting a gravitational wave, because the energy transferred from the gravitational wave to the bar must be sufficiently large to cause the bar to transition to a new quantum state.
The thermal motion of the bar, even at low temperatures, is normally much greater than the motion induced by a gravitational wave. To see the signal over this motion, the detector is designed so that bar vibrates at a resonance frequency; the signal is then the deviation of vibration from this resonance. To achieve this, the bar is designed to have low acoustic losses, meaning that the bar is designed to ring for a long time. If the timescale for a bar to lose its acoustic energy is τa, and the integration time of the experiment is $tau;i, then the noise in the detector can be supressed by the factor τi/τa. As τa goes to infinity, the detector becomes a perfect harmonic oscillator, which is highly predictable in behavior over time, and therefore sensitive to being forced by a gravitational wave. Typically, the energy loss timescales is about 1ms.
The current resonant bars are sensitive at the frequencies of about 700 and 900 Hz, with band widths of about 50 Hz. Their strain sensitivity is at the 10-18Hz-1/2 level, which is sufficient to detect gravitational wave bursts that may occur within the Galaxy. No resonant bar detector have yet detected a source, and the upper limit they provide on the neutron star binary merger rate is about two orders of magnitude above the theoretical value.
To improve the sensitivity of the detectors, which is limited by quantum mechanical states of the bar, the next generation is moving toward greater masses; future detectors will have bars of several hundred tons.
Currently there are 5 such instruments in operation around the world. The are working in cooperation, with least two detectors in continuous coincident operation since 1993. The experiments are ALLEGRO, AURIGA, EXPLORER, NAUTILUS, and NIOBI. Some of the groups associated with these instruments are planning new experiments to make detectors of much greater sensitivity.
ALLEGRO is an experiment operated by the Louisiana State University in Baton Rouge, Louisiana, USA. This experiment employs an aluminum bar cooled to 6°K. It operates at a frequency of 900 Hz, and has a strain sensitivity of 7×10-19 Hz-1/2. ALLEGRO has had two data runs, with the first covering June 1991 through January 1995, and the second covering 1996 to the present time. The experiment was operating when supernova 1993J occurred; no gravitational waves were detected.
The AURIGA detector is run by the Instituto Nazionale di Fisica Nucleare (INFN) in Legnaro, Italy. It is an aluminum bar cooled to 0.1°K. It operates at a frequency of 900 Hz, with a sensitivity of 3×10-19 over its first data run. AURIGA has been upgraded and between January 13th and May of 2004 it has been undergoing calibration and diagnostics. The sensitivity of the instrument has been improved.
The EXPLORER detector is based at CERN and operated by the University of Rome. It is a 2270kg aluminum bar cooled to 2° K. It operates at the resonance frequencies of 906 and 923 Hz, and has a strain sensitivity of 7×10-19.
The NAUTILUS detector is run by the Instituto Nazionale di Fisica Nucleare (INFN) at the Laboratori Nazionali di Frascati Italy. It is a 2300kg aluminum bar cooled to 0.1°K. It operates at the resonance frequencies of 908 and 924 Hz. NAUTILUS has a sensitivity of 6×10-19.
The university of Western Australia operates the NIOBE detector at Perth, Australia. It is a niobium bar operating at a temperature of 5°K. It is sensitive at the frequency of 700 Hz and has a sensitivity of 5×10-19.
1 Ju, L., Blair, D.G., and Zhau, C. “Detection of Gravitational Waves.” Reports on Progress in Physics 63 (2000): 1317–1427.