[Date Prev][Date Next][Thread Prev][Thread Next][Search] [Main Index] [Thread Index] [HEASARC Archives]

Memo on the new version of XRT response

Dear ASCA users,

Following I am sending a memo on the summary of problems in the ASCA XRT 
response tables and a fine tuned new version summarized by M. Ishida.

F. Nagase

- -------------

Response Status and Near Future
by M. Ishida

The XRT response tables (xrt_ea/psf_v1.1) now in use for ASCA data
analysis in general contain several problems.
  Before the launch, reflectivity of gold was not wellknown especially
above 4 keV. The XRT team measured the effective area of all flight
telescopes at 4.51 keV (titanium K-alpha), and 8.04 keV (copper
K-alpha) at ISAS 30m X-ray pencil beam line facility, and obtained the
reflectivity in the entire 0.3-12.0 keV band by interpolation. 

Accuracy of the interpolation is, however, not enough because of
limited number of measurements. This causes discrepancy between
estimated spectra and those observed in flight.
  As a result, in fitting Crab spectrum obtained by the GIS, one finds
(1) a hard tail which starts at around 8 keV, gradually increases with
X-ray energy, and amounts to about 8 % at 10 keV in data-to-model
ratio, (2) a broad enhancement peaked at 5.7 keV with the width of
300-400 eV in Gaussian sigma and the equivalent width of 50 eV. The
excess of the data point relative to the power-law model is some 5 %
at the peak. This feature is due to over-correction in the process of
producing current response functions. Also, (3) the photon index
obtained is 2.15, which is a little larger than the conventional value
2.08-2.12. Note that all these problems are common between the GIS and
the SIS.
  Another problem is in the encircled energy function which is given
only within 6 arcmin in radius. Since energy dependence of PSF is not
substantial at the core, single PSF common for all energies is used in
xrt_psf_v1.1. These, together with insufficient calibration of the
absolute efficiencies of the detectors, have made it difficult to
obtain correct absolute flux from an X-ray source, and the flux from
the SIS is usually greater than that from the GIS by about 15-20 %.
It is known that the Crab flux obtained by the GIS is about 18%
smaller than the conventional value.

  The XRT team has been tuning the effective area (EA) and the point
spread function (PSF) with inflight calibration data of Cyg X-1 and
Crab, and the new EA/PSF tables (v2.0) has just been transfered from
the XRT team to GSFC Guest Observer Facility. These tables will be
used in the ARF generator ASCAARF, and one will be able to create new
energy response functions of the SIS and the GIS just by setting
EA/PSF_v2.0 as XRT calibration files in running ASCAARF.
  In the new version of the energy response with EA/PSF v2.0, all the
artificial spectral structures described above will disappear, because
of fine tuning of EA table above 4 keV. In addtion, owing to the
revision of PSF table which reflects flux leakage outside 6 arcmin as
well as the energy dependence of PSF, the flux obtained by the GIS
increases by about 13%, and the GIS has become to provide correct
absolute flux for Crab after the dead time correction (a slight
modification in the detector efficiency is also made). The SIS flux is
then tuned so that it gives the same flux value as that of the GIS
with 3C273, Mkn421, EXO2030+375 etc (Note that Crab is too strong to
be used as a calibration source for the SIS). This SIS efficiency
correction is reflected also in the ARF generator ASCAARF. As a
result, all the four detectors will give correct absolute fluxes
within +/-2% in the new energy responses.

  Until the release of new ASCAARF with EA/PSF v2.0 from GSFC GOF, the
Guest Observers are recommended to be careful on the following points;
(1) the iron line profile; the 5.7 keV structure may alter the line
    profile from the optically thick accretion disk. The data to model
    ratio at 5.7 keV in Crab spectrum from the GIS is 5 %. The same
    level of discrepancy is expected for the SIS.
(2) The photon index will be estimated to be larger than real by
(3) If the spectrum is thermal, the difference between the result of
    the spectral fitting and the real value depends upon the
    temperature itself. Thermal bremsstrahlung temperature of 8 keV is
    evaluated to be 8.5 keV. The difference is reduced if the
    temperature is lower.
(4) The absolute flux obtained by the SIS is reliable within +/- 5%.
    The flux obtained from the GIS is usually smaller by 15-20% than
    that obtained by the SIS.

------- End of Forwarded Message