The basis of the RVS concept is being able to detect signals averaging less than a single electron from faint stars in each transit. Careful calibration of each transit spectrum and averaging of multiple transit spectra allows the noise to be reduced and the sub-electron signal to be revealed.
Astrium Radiation Campaigns have provided RVS test spectra that have propagated through the RVS CCDs, both in unirradiated regions and regions irradiated to the end of mission (EoM) damage levels. Radiation Campaign #4 (RC4) was the first time RVS faint spectra were acquired in Low Resolution (LR) mode with EoM radiation damage (GAIA.ASF.TCN.PLM.00670). Astrium's analysis (GAIA.ASF.TCN.PLM.00665) demonstrated for the first time that 1.5 electrons per pixel survives, even in EoM radiation-damaged CCDs. The faintest spectra consist of only 0.5 electrons per pixel, which corresponds to GRVS = 15.75 (V~17).
Ben Dryer and David Hall at The Open University (OU), supervised by George Seabroke and Mark Cropper at MSSL-UCL, have independently analysed the RC4 RVS tests on behalf of DPAC (GAIA-C6-TN-OU-BJD-001 and GAIA-C6-TN-OU-BJD-002). This work has verified the Astrium result that spectra consisting of 1.5 electrons per pixel survive and extended the analysis to the sub-electron regime. The rightmost plots in the figure above shows the sub-electron signals also survive in EoM radiation-damaged CCDs.
The RC4 GRVS = 15.75 spectrum (rightmost plots in the figure above) was obtained by running the test set-up 6000 times and averaging the resulting 6000 individual spectra. Accounting for typical RVS dead time at the faint end, Astrium estimate that RVS will see each source 32 times during the mission. There are three strips of RVS CCDs so on average there will be 96 useable RVS spectra per source at the end of the mission. Using this EoM spectrum, Astrium cross-correlated it with a synthetic template to obtain the EoM radial velocity (RV) accuracy of 15 km/s. |