Detection of gravitational waves from a coalescing binary system - effect of thermal noise on the efficiency of the detector

Photon-counting (shot) noise should be the main source of noise in a long-baseline interferometric gravitational wave detector above a frequency of about 100 Hz. Since seismic isolation prevents reasonable sensitivity below that frequency, it is accurate to consider photon-counting noise as the only noise present. But there are prospects to improve the 100-Hz seismic cut-off and reduce it to a value between 10 and 50 Hz. In such a frequency range, however, thermal noise in the suspension of the test masses and in the masses themselves becomes dominant. We analyse quantitatively its effects and show that by reducing the seismic cut-off frequency by a factor of 2 an improvement of about $$\sqrt{2}$$ in the signal-to-noise ratio can be achieved, under reasonable conditions, for the detection of a signal coming from a coalescing binary. Further improvement in seismic isolation is shown not to yield any significantly better performance.