RXTE spectral observations of the 1996-1997 outburst of the microquasar GRO J1655-40

Excellent coverage of the entire 16 month 1996-1997 outburst cycle of GRO J1655-40 was provided by the Rossi X-Ray Timing Explorer (RXTE). We present a full spectral analysis of these data, which includes 52 Proportional Counter Array spectra from 2.5 to 20 keV and High Energy X-Ray Timing Experiment spectra above 20 keV. We also include a nearly continuous All-Sky Monitor light curve with several intensity measurements per day. The data are interpreted in the context of the multicolor blackbody disk/power-law model. The source exhibits two principal outburst states which we associate with the very high and the high/soft states. During the very high state, the spectrum is often dominated by a power-law component with photon index (Gamma) similar to 2.3-2.7. The source exhibits intense hard hares on timescales of hours to days which are correlated with changes in both the fitted temperature and radius of the inner accretion disk. During the high/soft state, the spectrum is dominated by the soft thermal emission from the accretion disk with spectral parameters that suggest approximately constant inner disk radius and temperature. The power-law component is relatively weak with Gamma similar to 2-3. During the last few observations, the source undergoes a transition to the low/hard state. We find that a tight relationship exists between the observed inner radius of the disk and the flux in the power-law component. During intense hard flares, the inner disk radius is observed to decrease by as much as a factor of 3 on a timescale of days. The apparent decrease of the inner disk radius observed during the hares may be due to the failure of the multicolor disk model caused by a steepening of the radial temperature profile in the disk coupled with increased spectral hardening and not physical changes of the inner disk radius. The distortion of the inner disk spectrum by the power-law hares indicates that the physical mechanism responsible for producing the power-law emission is linked to the inner disk region. Assuming that our spectral model is valid during periods of weak power-law emission, our most likely value for the inner disk radius implies a* < 0.7 Such a low value for the black hole angular momentum is inconsistent with the relativistic frame dragging and the "diskoseismic" models as interpretations for the 300 Hz X-ray QPO seen during some of these RXTE observations.