Eric T. --
There is no evidence for an order-of-magnitude overestimation of
absorption in standard XSPEC fits using ACIS data. I infer this
from the attached figure which plots logNH vs visual absorption AV for
829 sources in the Orion Nebula field. The dashed curve shows the
relation NH = 2x10^21 AV for a standard interstellar dust-to-gas
ratio. The result is that NH and AV values are in broad agreement
within expected statistical uncertainties, except for a possible
systematic offset in the direction opposite from the one you report:
NH values appear a factor of ~2 below that expected from AV.
Some details. These are X-rays from magnetically active flaring stars
which undoubtedly have a complicated distribution of plasma
temperatures and possibly weird elemental abundances. We obtain NH
from a 1-temperature wabs+meka model with 0.3 solar abundances, which
is clearly simplistic. The sources with violently discrepant NH-AV
values probably have a wrong spectral model (or incorrect AV). The
smaller circles are sources with 30-500 counts for which the
statistical uncertainty of NH is +/- 0.4-0.2 (1sigma). The larger
circles have 500-10,000 counts with uncertainties around +/- 0.1. The
dataset is a bit weird, half taken at -110C and half at -120C with a
somewhat obsolete treatment of CTI effects (i.e. not the latest Penn
State arf/rmf's). But we suspect calibration errors are not very
important here given the huge range in absorptions among the stars.
The factor of 2 NH offset (or equivalently, 1 mag AV offset) could
either be due to a Chandra problem or to a non-standard gas-to-dust
ratio in the Orion molecular cloud material.
Hope this is useful, and it would be interesting to hear from others
on this topic.
Eric Feigelson
This archive was generated by hypermail 2b29 : Wed May 15 2013 - 01:00:11 EDT