Gravitational-wave bursts with memory and experimental prospects

Experimenters usually divide the gravitational waves which they hope to detect into three classes: 'bursts' in which the wave field h^TT _ij rises from zero, oscillates for only a few cycles and then returns to zero; 'periodic waves', and 'stochastic waves'¹. There is, however, a fourth class, 'bursts with memory'^2-6 (BWM), in which h^TT _ij rises from zero, oscillates for a few cycles, and then after a burst of duration Δt settles down into a non-zero final value δh^TT _ij. Here we show that for any kind of detector the best way to search for a BWM is to integrate up the signal for an integration time t̂ 1/f_opt, where f_opt is the frequency at which the detector has optimal amplitude sensitivity to ordinary bursts (bursts without memory). In such a search the sensitivity to BWM with duration Δ <= 1/f_opt is independent of the burst duration Δt and is approximately equal to the sensitivity to ordinary bursts one cycle long with frequency f_opt (see Fig. 1). It is possible, though not highly probable, that BWM will be among the earliest kinds of gravitational waves detected; therefore experimenters should take them into account when planning their search strategies and data analyses.