On perusing the various threads, documents, and discussions on
analyzing extended sources on the CIAO web pages, I've noticed
a potentially important (and common) oversight. Namely, that the
background observed by the ACIS contains a significant
contribution from (and is dominated by at high energies)
particles. What is important to keep in mind, is that this component
of the background is independent of the telescope optics, and therefore
its spatial (and spectral) distribution is not corrected by exposure
maps and arf calculations. In fact, such things often amplify the
importance of the particle background. In particular, the assumption
that the data D is (crudely speaking) the effective area A times
the incident source spectrum S (plus photon background B_ph) is
wrong. The cosmic ray background is in effect an additive component so
D = A(S+B_ph) + B_cr
So dividing by an exposure map or using separate arfs on your
background and source spectra effectively gives you
EXP_CORRECTED = S + B_ph + B_cr/A
Which, for analysis of hard, faint sources with Chandra, could be
a very misleading amplification of the particle events.
To correct for this, one needs information on both the spectral AND
spatial dependence of the particle background, so the contribution from
it can be subtracted BEFORE corrections for the instrument response are
made. For ASCA, night earth observations could be used, but for
Chandra, the only data set I know of is the closed door observations, and
the memo I've seen on this data set contained no information on the
spatial distribution of events.
Just something to remember during discussions of extended source
analysis.
Mallory
This archive was generated by hypermail 2b29 : Tue Feb 12 2013 - 01:00:09 EST