---

XMM-Newton EPIC Response and Background File Page

---

The SOC page

Old version of page

XMM-Newton EPIC Response and Background File Page

Back to Calibration Homepage

Downloads

The most recent Canned Response Matrices can be downloaded here. See notes below.

The EPIC Background Files can be downloaded here. See notes below.

Canned response matrices

General Information

The PN response matrices are available for spectra selected from single-pixel events only, doubles only and single+doubles combined, with event threshold 20 adu for most observing modes. This allows one to make use of the additional signal-to-noise improvement with double event patterns. However the user should be warned to compare spectral fitting results with and without the double events, as the effects of pile-up can make the results a little variable.

These files have filenames of the form;

epn_ff20_sY9_<ver>.rmf	for singles, where ver is the version number, eg; v6.7
epn_ff20_dY9_<ver>.rmf	for doubles, where ver is the version number, eg; v6.7
epn_ff20_sdY9_<ver>.rmf	singles plus doubles, where ver is the version number, eg; v6.7

For the PN the responses are provided at intervals of CCD 20 rows, due to changing energy resolution from charge transfer effects. That is, a nomenclature Y0 can be used from rows 1-20, and Y9 from rows 181-200 (i.e. on-axis).

• Y0 CCD lines 1-20 (i.e farthest from centre)
• Y1 CCD lines 21-40
• ...
• Y9 CCD lines 181-200 (i.e. nearest centre)

The PN response varies with observation mode and different ready-made matrices are available for each mode.

• ff - fullframe mode
• ef - extended full frame mode
• sw - small window mode
• lw - large window mode
• ti - timing mode
• bu - burst mode

The mode is coded into the filename such that epn_sw20_dY9.rmf is a file for small window mode.

MOS response files now include all available observation modes:

The canned MOS RMFs are observing mode and observation date dependent. The files are called;
	<camera>_<rev>_<mode>_<patterns>_<ver>.rmf
	camera m1 for MOS-1 or m2 for MOS-2 rev The revolution number of the observation. mode im for imaging mode or tu for timing mode. patterns p0 for pattern 0 only or pall for patterns 0 - 12 ver version number of the release, eg; v1.2
For example;
	m1_360_im_pall_v1.2.rmf m2_439_tu_p0_v1.2.rmf

Users should, in general, use the rmf with the revolution number closest to the revolution in which their observation was performed. The revolution 534 matrices are, however, an exception to this rule. In this revolution the MOS detectors were cooled, resulting in a much improved spectral response. Therefore, observations with a revolution number greater than 534 should use the revolution 534 response files but earlier revolutions must use the previous response from e.g. revolution 439.

In timing mode (tu) we are currently only supporting pattern 0.

Update 2004-12-03

PN response files for all modes have been updated

This release improves the agreement between the MOS and PN responses below 1 keV. For further details please refer to XMM-SOC-SRN-0183.

Update 2004-10-15

MOS response files for all modes have been updated

The timing mode response has been much improved below 1 keV where large discrepancies with model fits were previously seen.

The tail of the redistribution function has been adjusted for all modes. This particularly affects very hard sources which previously showed an unnatural deficit at low energies wher e there were no source photons.

The naming convention has changed to include the revolution number. The general description above has been modified to reflect these changes.

Update 2004-07-28

Response files for both MOS and pn have been updated.

• pn: further adjustments of the redistribution parameters in the Partial Event Model at low energies (<600 eV) were done.
• MOS: Observations of calibration targets confirm a significant change in the low energy redistribution characteristics of the MOS cameras with time. This change is probably due to an increase in the surface charge loss property of the CCDs which degrades the low energy resolution. Epoch dependent calibration files have been produced which reflect these changes. For detailed information see XMM-SOC-CAL-SRN-169.

Update 2003-01-29

PN response files for all modes have been updated:

• Quantum efficiency QE (concerns all readout modes)
  New QE function is based on measurements of the thickness of the SiO2 layer on top of the CCD. This improves spectral fits around the O-edge.
  
  
• Spectral redistribution (concerns all readout modes)
  Data from the isolated neutron star RXJ1856.5-3754 were used to adjust the redistribution parameters in the Partial Event Model at low energies (<600 eV). The parameters were tuned to make the spectral fit parameters of RXJ1856.5-3754 consistent with those obtained by the  Chandra LETG in high resolution. In particular the absorption column density agrees now with that obtained by the LETG, while before the pn value was too low.
  
  
• Charge Transfer Inefficiency (Small Window mode)
  The CTI correction for SW mode was further adjusted at low energies to further improve the spectral fits around the O-edge. At the O-edge the events were CTI-overcorrected before.
  

For further information please visit the Update Details Page

Update 2002-05-28

PN response files for timing and burst mode have now been added to the repository. Due to the lack of point sources with strong emission lines the energy resolution assumed to produce the RMF files for pn timing and burst mode has not been checked over the whole energy band. Be careful when interpreting derived line widths. Lines in spectra of extended sources are expected to be broadened by CTI correction which assumes a certain position of the source on the detector.

Update 2002-04-29

A set of CCD-only redistribution matrix files (RMF) are available here. These may be used in combination with the SAS task arfgen to provide the full response for the EPIC detectors on and off axis.

They are virtually identical to the files produced by the SAS task rmfgen but can be used to save time.

It is likely that they will be updated from time to time to give the user access to the current best knowledge independently of the SAS release schedule. Changes to the RMF tend to come from improvements to generation algorithms which imply a code change. Conversely, improvements in the knowledge of the instrument effective areas (produced by arfgen) can usually be included in updates to the CCF which avoids the need for a new software release.

The MOS files use the nomenclature ****all**** and *****p0*** which mean patterns 0-12 and pattern 0 only respectively. Pattern 0 is currently the only supported pattern type for timing mode observations. In imaging mode pattern 0 spectra can be selected to improve the spectral resolution for bright sources.

The PN response matrices are available for spectra  ...

The corresponding information has been moved upwards, since it is still valid ...

How to use canned responses and/or arfgen and rmfgen

A cookbook giving advice about the use of arfgen and rmfgen, alone or in conjunction with the standard matrices, is available in a technical note for a variety of source types. This note gives a direct comparison of the responses created by the SAS tasks with the standard canned matrices. Cookbooks are available for the following versions; SAS 5.3, SAS 5.1 & SAS 5.0.

Usage

Briefly to extract spectra using the SAS xmmselect GUI to examine calibrated event lists, the user should use a flag selection #XMMEA_EM&&(PATTERN<=12) (MOS) or (FLAG==0)&&(PATTERN<=4) (PN), and use an extraction radius ~1 arcmin radius.

With the PI column check button selected, extract OGIP spectrum, and then request binning of 15eV (MOS) or 5eV (PN), and invoke "withenergyrange".

The energy range appropriate for these responses is 0-11999 eV (MOS), and 0-20479 eV (PN) Using the grppha tool (for example), one can change/verify that the RESPFILE is set to the appropriate response set, that the EXPOSURE has been correctly filled, BACKFILE and BACKSCAL can be manually adjusted, and appropriate GROUPING applied.

The old matrices (corresponding to the update done on 28/05/2002) can still be downloaded from the SOC ftp server:
MOS, CCD Only & PN, CCD Only.

EPIC Background files

Please refer to the explanatory notes regarding downloading the EPIC background files.

The following sources of information are also available;

1. University of Birmingham
   
   The XMM-Newton EPIC Background Analysis webpage from the University of Birmingham.
   
   The old files and notes are still available for download:
   
   • Old background files.
   • Old explanatory notes. (ps.gz)
   
   
2. Helsinki University Observatory
   
   Supporting data for astro-ph/0504362 "XMM-Newton EPIC background modeling for extended sources" (ApJ accepted), J. Nevalainen, M. Markevitch and D. Lumb available at:
      http://www.astro.helsinki.fi/~jnevalai/XMMbkg/.

Ad hoc EPIC files

1. January 29, 2003: Updated canned response matrices. Responses updated during December 2002 [link] include ALL EPIC public available observation modes.
2. May 28, 2002: Canned response matrices and how to use them with the SAS. Responses updated 2002-05-28 include PN timing and burst mode.
3. January 10, 2002: Background files updated together with the corresponding explanation notes.
4. June 1, 2001: Ready-made response matrices and how to use them with the SAS. Responses updated 2002-04-29.
5. May 7, 2001: Background files