Upper Limit Server


Version 3.0: November 12 2013

Purpose and Scope

The Upper Limit Server (ULS) is designed to return the upper limit to a count rate and flux which could be produced by a source at the input position in a given image. It is not intended to be a source detection system and will return an upper limit rather than a count rate for faint sources. For stronger sources it will return a count rate but the automatic processing means that these will not be as accurate as count rates available in the mission source catalogues. To find the strengths of detected sources it is strongly recommended to use a data archive such as the XMM-Newton Science Archive (XSA), LEDAS or HEASARC.

How It Works

The routine works by extracting images which contain the entered celestial position from an image repository. Each image is then processed by the XMM-Newton SAS task eregionanalyse which finds the background subtracted count rate at the position. The count rate is calculated by:

  • Extract the counts from a circle of radius 30" about the position.
  • Extract the background counts from an annulus of inner radius 45" and outer radius 75" about the position.
  • Perform a background subtraction, normalising for the areas of the source and background regions.
  • Calculate a background subtracted count rate at the source position.
  • Calculate an upper limit at the source position, using the Bayesian statistics of Kraft, Burrows and Nousek 1991 (ApJ 374, 344), with a uniform prior. The significance (1,2 or 3-sigma) is user selectable with the default set at 2-sigma (95.4%).

    NB: The count rates are corrected for instrumental effects such as the point spread function and vignetting.

    If the background subtracted count rate is greater than twice its error then the output shows the measured count rate with the error. Otherwise, the upper limit is displayed. This is conservative and means that low-significance, real sources, will be shown as an upper limit rather than a measured count rate. It should be stressed that the source catalogues should be consulted to find the presence and strength of detected sources.

  • Convert the count rate to a flux. The conversion uses the PIMMS software (v 3.9b) with a fixed spectrum of a power-law of slope 1.7 absorbed by a column of NH=3x1020 cm-2.

How to use it

  • Select missions to use (only XMM-Newton slew and pointed data are currently available)
  • Enter either a target name or the RA, DEC of the required sky position. Alternatively enter a list of coordinates from an ASCII file.

    The coordinate list should be an ASCII file with one position per line and optionally a source name or identifier, e.g.

    54.260694 +47.646995 AGN1
    23.6223 -34.567

    Coordinates may be entered in decimal degree or sexigesimal formats but the declination MUST contain the sign (+/-). Lists of target names without coordinates are not supported. Acceptable formats are:
    • 54.2606 -27.4534
    • 54.2606 -27.4534 AGN 345
    • 12:02:30.3 +37:15:26
    • 12 02 30 +37 16 26.4 New Galaxy
    The current implementation takes about 30-60 seconds execution time per source. It is recommended to limit lists to around 20 sources.


The output is a block of results per position. Each block has, as a header, the coordinates of the position and the identifier or source name.

There is an output line for each observation containing the input position. The columns are:

XMM-Newton slew data

  • OBSDATE : the data of the observation in YYYY-MM-DDTHH:MM:SS format, e.g. 2004-03-05T14:00:00 being 14.00 UTC on 5th March 2004.
  • COUNT RATE (0.2-12.0) : The count rate or upper limit to the count rate in the full (0.2-12 keV) XMM-Newton energy band. Photons with event pattern 0 only are used from 0.2-0.5 keV and event patterns 0-4 are used at energies above 0.5 keV.

  • COUNT RATE (0.2-2.0) : the soft energy band count rate or upper limit.

  • COUNT RATE (2.0-12.0) : the hard energy band count rate or upper limit.

  • EXP TIME (s) : the exposure time in the full (0.2-12 keV) energy band at the given position

  • FLUX (0.2-12.0) : the flux in the full energy band, calculated from the count rate assuming a spectral model of a power-law of slope 1.7 absorbed by a Hydrogen column of 3x1020 cm-2, units of ergs s-1 cm-2

  • FLUX (0.2-2.0) : the flux in the soft energy band

  • FLUX (2.0-12.0) : the flux in the hard energy band

  • QFLAG : a quality flag. For the XMM-Newton slew data this indicates whether the count rate is so high that the source is piled-up. If this is set to 'N' then the given count rate is likely to be an underestimate of the true value.

XMM-Newton pointed data

The same output as XMM-Newton slew data except that only the 0.2-12 keV count rate and flux are shown.


If the mouse is placed over the count rate values, a box is displayed showing the number of counts detected in the source box, the number of background counts calculated from the number of counts in the background box multiplied by the scaling factor (BSF) for the respective box sizes and the encircled energy factor (EEF) which was used to correct for photons scattered outside the finite source box.

count rate = ( source_counts - backgnd_counts ) * EEF / exposure_time


XMM-Newton slew survey

XMM-Newton slew data is taken by the EPIC-pn instrument with the 'Medium' filter. Data are available for all slews which have been processed by the XMM-Newton processing pipeline and ingested into the archive. This usually includes everything up to about two weeks before the current date.

Flux conversion

The XMM-Newton count rate is converted to a flux using the PIMMS v3.9b software. PIMMS v3.9b, assumes that the EPIC pn count rates contain events of pattern 0-4 which is nearly compatible with the pattern selection used in the creation of the slew images (pattern 0 from 200-500 eV and patterns 0-4 at energies from 500-12000 eV). The conversion is performed with a spectral model of a power-law of slope 1.7 absorbed by a Hydrogen column of 3x1020 cm-2. To convert to flux using a different model; take the count rate from the output of the tool as the input count rate for a 'Medium' filter EPIC-pn observation using event patterns 0-4. If using PIMMS or WebPimms please remember that they currently require that the input count rate is from an extraction radius of 15" (roughly 72% of the total counts).

Limitations and approximations

Due to space restrictions in the slew image archive the routine currently uses the broad band (0.2-12 keV) exposure maps to find the exposure time at a point. The exposure time of the soft band and hard band images will be slightly different to this value. An algorithm should be implemented to correct for this in a future version as described in section 6.3 of Saxton et al., 2008, A&A, 480, 611.


XMM-Newton pointed observations

The system retrieves "all-EPIC" images, which consist of a single, 0.2-12 keV image, prepared by combining the EPIC-pn, MOS-1 and MOS-2 individual camera images. This gives a global source measurement for each observation. For a more detailed upper limit, or flux, from each camera, use the FLIX tool developed by Clive Page at the University of Leicester.

Change Log

  • VERSION 3.0
    • Solved a problem where multiple results were returned from a single XMM slew
    • Added a target name resolver
    • Switched to use XMM slew images produced by the standard pipeline. These contain all useful slew data taken up to about two weeks before the current date.
    • Support hms,dms coordinate format for input file. NB: Target lists now MUST include the sign '+/-' in front of the declination.
  • VERSION 2.0
    • Added in support for XMM-Newton pointed data.
    • Allow the user to choose the significance level for the upper limit

Future Plans

  • Add in more missions
  • Allow the spectral model for flux conversion to be user selectable
  • Allow user selection of type of statistics.
  • Support output to a text file
  • Plot results as a time series