Up to top level
AO15   AO16   AO17   AO18   AO19   Backgrounds   Calibration   Conference   Data   Docs   EPICMOS   EPICpn   Feedback   Gallery   Misc   OM   Pending   PhD_Theses   Publications   RGS   RadMonitor   SAS_Hardware   SAS_WS   SASv16.0   SASv16.0_Installation   SASv16.1   SASv16.1_Installation   SASv17.0   SASv17.0_Installation   SASv18.0   SASv18.0_Installation   SciSim   Simulators_other   Suggestions   Trash   Visibility   XMM-bouncing   XMM-news   XRPS   XSA   esas   incoming  

Logged in as guest

Viewing EPICMOS/8891
Full headers

From: wayne@astro.umd.edu
Subject: EPIC PN/MOS differences
Compose reply
Download message
Move To:
2 replies: 1 2
1 followups: 1

Private message: yes  no

Notes:

Notification:


Date: Tue, 5 Aug 2003 19:39:38 GMT
From: wayne@astro.umd.edu
To: xmmhelp@xmm.vilspa.esa.es
CC: wayne@astro.umd.edu
Subject: EPIC PN/MOS differences
Full_Name: Wayne Baumgartner
Submission from: (NULL) (128.183.19.215)


Question:  Is there a problem with the EPIC calibration at the Oxygen edge?

Our group at Goddard is interested in measuring the abosrption due to
oxygen in the ISM by looking toward clusters with high galactic n_H
columns.  In the course of our investigations, we noticed that fits to
the data led to a very different result for oxygen absorption in the
PN and in the MOSes.

Procedure:

Initially, we fit the cluster data from .3 - 7.2 keV to a
tbvarabs*vapec model in XSPEC.  The data had been rerun through SAS
5.4.1 starting with the chain routines on the ODF data.  Backgrounds
were taken locally from an annulus around the source region.  RMFs and
ARFs were generated using rmfgen and arfgen (with the extended=yes
parameter set for arfgen).

Problem:

The clusters were generally well fit, except for significant residuals
seen from around .5 - .9 keV.  The residuals were similar for the MOS1
and MOS2, but different for the PN.  When the oxygen absorption was
allowed to find separate values for the MOS and PN, significant
discrepancies were found for most of the ~6 clusters we observed.  The
ratio of oxygen absorption in the PN to that in the MOS ranged from
0.1 to 0.7, with small errors.

Diagnosing the problem:

This problem was not observed when using the PPS event files and
spectra, but only with the re-reduced ODF data run through SAS 5.4.1.
We checked to see if the problem was dependent on observation date,
suspecting a time-variable change in the EPIC calibration.  No
correlation was found between O_pn/O_mos and the observation date.  No
correlation was found with cluster temperature, suggesting that
properties of the sources were not responsible.  We also found that
the problem is present in data taken both with the thin filter and the
medium filter.  However, there was a strong correlation with the
galactic column, with lower columns having a larger discrepancy
between the MOS and the PN.

This correlation with n_H column led us to suspect that there was a
problem with the filter calibration, since a wrong value of filter
absorption would manifest itself more at low columns where the
fractional error would be greater.
A solution:

We checked this by looking at observations of the Coma cluster
(seq#0153750101), which has a very low galactic column of 9.2e19.  As
expected, the PN/MOS problem was very visible.  We added an extra edge
at oxygen to the spectral model in order to address the problem and
found that an extra edge in the MOS data moved the measured MOS and PN
absorptions into agreement.  We then used better signal-to-noise data
from 3c273 (which also has a relatively low column of 1.8e20) to
quantify the edge.  With the better 3c273 data (rev 277, thin filter;
using the correct canned RMF and arfgen generated ARF), we found that
an edge with optical depth of 0.19 in MOS1 and 0.12 in MOS2 at an
energy of .556 keV best fit the data.  Putting this extra edge into
our cluster spectral fits solved the PN/MOS discrepancy.

We also found that an extra carbon edge at .288 keV with an optical
depth of .9 in the MOS and .2 in the PN also improved the fits for
3c273.  However, this is at the lower end of the good calibration
range for EPIC, and other things could be causing this problem.

Medium filter 3c273:

The same values for the extra oxygen edge were found. The extra carbon
edge is a little deeper, at 1.2 in the MOS and .4 in the PN.




Reply 1

Resend
From: Matteo Guainazzi <xmmhelp@xmm.vilspa.esa.es>
To: wayne@astro.umd.edu
Subject: Re: EPIC PN/MOS differences (PR#8891)
Date: Thu Aug  7 08:44:01 2003
Dear Dr. Baumgartner,

 I have conveyed your detailed report to the MOS Instrument Team. They basically
confirm your findings. Here you are a summary of the preliminary status of their
study of this issue:

1. a feature at the Oxygen edge is being seen in the MOS exposures, with the
exception of very early observations. The trend is for this feature to deepen up
to around revolution #250, and to stabilize afterwards

2. to our current understanding, most of the discrepancy the you observe between
MOS and pn in this energy range is to ascribe to the MOS cameras deviating from
the "nominal" response in this region

3. in the softest MOS energy range counts are as well missing. This is, however,
**not** due to a residual neutral Carbon edge (as by your analysis), but to the
redistribution function. Anyhow, MOS spectral fitting should be restricted to
the energy band above 0.5 keV

4. study of these feautures are ongoing. A more quantitative assessment, with
new calibrations, will be probably an outcome of the next EPIC calibration
meeting, to be held in September this year

 Please, do not hesitate to ask fur further information and clarifications if
needed.

 Regards,

 Matteo Guainazzi
 XMM-Newton Science Operation Center
 User Support Group




Followup 1

Compose reply
Download message
Date: Thu, 7 Aug 2003 16:29:30 -0400 (EDT)
From: Wayne Baumgartner <wayne@astro.umd.edu>
To: xmmhelp@xmm.vilspa.esa.es
Subject: Re: EPIC PN/MOS differences (PR#8891)
Matteo,

  Thanks for your prompt reply.  I do have a couple of follow-up
  questions.

  1) Does the instrument team have any speculations about what could
      be causing this feature?  An extra oxide layer on the detector?
      Somewhere/something else?  Could the cause of this feature be
      manifesting itself at other places in the spectrum?

  2) You mention in your point #1 the observed trend in this feature.
      This suggests that someone has done some analysis of different
      data sets for many different dates.  Is a simple plot of this
      data available?  Our science observations occur throughout XMM's
      lifetime, and any semi-qualitative information about how this
      feature changes in time would help us model the feature and
      greatly aid our science analysis.

   Many thanks,

     -Wayne Baumgartner


>  Date: Thu, 7 Aug 2003 08:43:56 GMT
>  From: Matteo Guainazzi <xmmhelp@xmm.vilspa.esa.es>
>  X-Loop: xmmhelp@xmm.vilspa.esa.es
>  Content-Length: 1264
>  
>  Dear Dr. Baumgartner,
>  
>   I have conveyed your detailed report to the MOS Instrument Team. They
basically
>  confirm your findings. Here you are a summary of the preliminary status of
their
>  study of this issue:
>  
>  1. a feature at the Oxygen edge is being seen in the MOS exposures, with
the
>  exception of very early observations. The trend is for this feature to
deepen up
>  to around revolution #250, and to stabilize afterwards
>  
>  2. to our current understanding, most of the discrepancy the you observe
between
>  MOS and pn in this energy range is to ascribe to the MOS cameras deviating
from
>  the "nominal" response in this region
>  
>  3. in the softest MOS energy range counts are as well missing. This is,
however,
>  **not** due to a residual neutral Carbon edge (as by your analysis), but
to the
>  redistribution function. Anyhow, MOS spectral fitting should be restricted
to
>  the energy band above 0.5 keV
>  
>  4. study of these feautures are ongoing. A more quantitative assessment,
with
>  new calibrations, will be probably an outcome of the next EPIC calibration
>  meeting, to be held in September this year
>  
>   Please, do not hesitate to ask fur further information and clarifications
if
>  needed.
>  
>   Regards,
>  
>   Matteo Guainazzi
>   XMM-Newton Science Operation Center
>   User Support Group
>  
>  
>  
>  



Reply 2

Resend
From: Matteo Guainazzi <xmmhelp@xmm.vilspa.esa.es>
To: wayne@astro.umd.edu
Subject: Re: EPIC PN/MOS differences (PR#8891)
Date: Mon Aug 11 08:54:08 2003
Dear Wayne,

1. currently, the most likely hypothesis is an Oxygen layer due to a not
complete outgassing in the earliest phase of the mission. As such, the edge at
~0.54 keV should be its most evident feature in a MOS spectrum

2. unfortunately, the study on the MOS Oxygen feature is still ongoing, and no
quantitative report is available yet. A first quantitative assessment is
expected to be available by September this year.

 Regards, Matteo

Up to top level
AO15   AO16   AO17   AO18   AO19   Backgrounds   Calibration   Conference   Data   Docs   EPICMOS   EPICpn   Feedback   Gallery   Misc   OM   Pending   PhD_Theses   Publications   RGS   RadMonitor   SAS_Hardware   SAS_WS   SASv16.0   SASv16.0_Installation   SASv16.1   SASv16.1_Installation   SASv17.0   SASv17.0_Installation   SASv18.0   SASv18.0_Installation   SciSim   Simulators_other   Suggestions   Trash   Visibility   XMM-bouncing   XMM-news   XRPS   XSA   esas   incoming  

Logged in as guest


Please make your (short) question the subject of your request!


Web interface using JitterBug ... back to the XMM home page