Dear Suzaku users,
As shown in the in-orbit performance paper of HXD (Kokubun et al. 2007), there are about 20% residual in Crab spectra observed by the GSO when fitted with the single power-law model. Through the clarification of the current response (ae_hxd_gso[hx,xi]nom_20080129.rsp), the HXD team realized that the GSO discrepancy is mainly caused by the diagonal elements of the response matrix, which is connected with difficulties in data handling around lower threshold, and so on. Since these problems cannot be resolved easily and it is difficult to construct a new response matrix promptly, we first release an correction "arf" file ae_hxd_gsohxnom_crab_20070502.arf and ae_hxd_gsoxinom_crab_20070502.arf for each aim point, to reduce the discrepancy.
The correction factor C(E) is obtained as to reproduce the 70-300 keV Crab spectra smoothly with the single power-law model of which the parameters were determined by the 12-70 keV PIN spectra. Then, the factor becomes,
C(E) = 1.36 * (E/100)^0.65 * exp(-E/230) @ XIS nominal
1.18 * (E/100)^0.55 * exp(-E/320) @ HXD nominal ,
where E is the energy in the unit of keV.
The correction arf files can be simply used with the current response files in XSPEC as follows (in case of xis nominal dataset):
data 1 gso_cl.pha backgrnd 1 gso_bgd.pha response 1 ae_hxd_gsoxinom_20080129.rsp arf 1 ae_hxd_gsoxinom_crab_20070502.arfPlease be careful that 1) the correction factor was obtained only in the 70-300 keV energy band, and 2) there might still remain calibration uncertainties with the level of about 10%.
An example without/with the arf file is shown below. With utilizing the correction arf, the 12-300 keV Crab spectrum at the XIS nominal position (2005 September 15 observation) is well represented with the single power-law model (photon index = 2.10 +/- 0.01, normalization @ 1 keV = 11.5 +/- 0.2). The best-fit parameters obtained at the HXD nominal position are almost the same (photon index = 2.09 +/- 0.01, normalization @ 1 keV = 10.8 +/- 0.3).
