Simulate and Analyze SciSim RGS data

This thread contains a step-by-step recipe to simulate RGS data with SciSim and to proceed analyzing the data using SAS.

In this example we simulate a Raymond-Smith model with a temperature of 0.86 kT(keV), and an abundance of 0.27 in solar units. The column density and the redshift are set to 0. With these values an XSPEC spectrum can be created (following the steps that can be read in Cosmic Simulator guide, chapter 7) and afterwards imported into SciSim.

• Set up Scisim:

  Set the SCISIM_DIR variable to the path where SciSim is located:

  For bash shell execute:

  
  export SCISIM_DIR=/path/to/SciSim 
  
  

  For (t)csh shell execute:

  
  setenv SCISIM_DIR /path/to/SciSim
  
  

  Run the set up script.

  For bash shell execute:

  
  . $SCISIM_DIR/setup.sh 
  
  

  For tcsh shell execute:

  
  source $SCISIM_DIR/setup
  
  

• Load the default configuration for the telescope you are going to use.

  SciSim has three different telescopes configurations:

  • Telescope 1 loads the default configuration for MOS1, RGS1 and OM.
  • Telescope 2 loads the default configuration for MOS2, RGS2 and OM.
  • Telescope 3 loads the default configuration for PN and OM.

    In this example we are running an RGS1 simulation, so we have to select the Telescope-1 default configuration:

    File -> Load Telescope 1

    

  • Introduce the Attitude coordinates:
    • RA=70.453
    • Dec=46.038

      Note that these are J2000 coordinates

      The astronomical position angle (pos) can be left at zero.

• Create a source

  Select the Source menu and the Edit... option. A blank dialogue should appear:

  To create a source, hit the New button at the bottom dialogue. The display now shows default information about the newly created source, positioned at RA,Dec=(0,0).

  The window consists of a General pane, and a pane that contains the tabs denoting the properties of the current source: Optical,Spectrum,Brightness and Shape:

  The Incident Angle and Azimuthal angle in the pane called General are with respect to the pointing of the Spacecraft. As we want the source to be in the Center of View of the Spacecraft, we have to introduce the same coordinates that in the Attitude GUI.

  We now need to address the X-ray properties for our example.

  Select the Brightness tab and enter the following values:

  • Flux=1.7935e-2
  • Energy band=250 12000

  Now select the Spectrum tab. By default a monochrome source is selected. However, we want to specify (as we said before) an XSPEC spectrum that we have created previously, so we pull down the pop-up menu, and select xspec.

  Now enter the following values:

  • FileFile=scisim_threads.qpd
  • Column=0.0

  Your Source Editor window should look like:

  

  Now that we have specified a source, we just simply click on Close button in the Source Editor window.

• Select the background radiation.

  Select the Source menu and the Background... option. A blank dialogue should appear. Introduce the following background value.

  • Background=0.04

    Notice that the background simulation will increase considerably the time needed to perform the simulation.

  • Select the simulator.

    As we said above, in this example our goal is to simulate an exposure of 1000 seconds of the RGS1 instrument. The first step in this thread was to load the default configuration for RGS1. Nevertheless, we will check that all the parameters are properly defined.

    Select the Configure menu and the Configure... option. Another dialogue appears, showing icons for each simulator sub-system.

    By default, the simulators are connected to each other. This means that (from left to right) the output of one simulator is redirected to another. The end of the so-called pipe-lines show a file icon. Such a file icon denotes the name of the file, in which the results will be stored.

    

    In this dialogue, clicking on the left button while pointing the cursor at an icon will select the instrument simulator in question; clicking using the right mouse button will cause a new dialogue to appear which allows the user to configure the specific simulator.

    In this thread we are only interested in a RGS1 simulation, so we can remove the connections to the OM and EPIC simulators. We can do that simply left-clicking the mouse button, while hitting the icons labelled OM and EPIC.

    Using the right mouse button, while pointing at one of the icons, will show the next window:

    

    This dialog is specific to the icon selected, and allows us to modify the simulator or the file, whichever is pointed at. We could change the parameters to meet special needs.

    Hit OK to close the dialogue.

    Since we have loaded the Telescope 1 configuration, we have set up SciSim for a combined MOS1 / RGS1 simulation. This contains by default the right RGA grating, so we can check the RGA dialogue to check that the RGA grating parameters. To do that we just click with the right mouse button in the RGA icon. The next window should appear:

    

    Hit OK to close the dialogue.

    Now we can change the exposure duration. Click on the S/C icon with the right button and change the value in the Duration window:

    • Duration=1000
    

    Hit OK to close the dialogue.

    Finally, hit the Close button to close the Simulator configuration dialogue.

• Run the simulation

  Press the Start button, which is situated near the bottom and should now be highlighted. The Start button should now be greyed-out and the Stop button highlighted.

• Post-Analysis

  Once the simulation has finished, we are left with a file called rfc.out. This binary file has a SciSim specific format and contains the results of the exposure as seen by the RGS camera.

  The file can be converted to an XMM-Newton Observation Data File, with following command:

  
  rodf < rfc.out
  
  

  This command creates a number of FITS files, that can be viewed by the FTOOLS package. Analysis of these files are intended to be done with the XMM-Newton Science Analysis Subsystem package SAS. To do so we need to create a number of files related to attitude and time correlation. This is done with the SAS task odffix, as one of the first commands to be given when running SAS.

• Analysing SciSim data with SAS
  • Set up SAS environment following the steps that you can read in the Start up thread
  • Set up the SAS_CCF environmental variable. In the ccf directory under SCISIM directory you can find a ccf_scisim.cif file that has been produced with the CCF files needed by SciSim. You can use it to analyze SciSim data with SAS.
  • Set up the SAS_CCFPATH environmental variable to your directory where the SCISIM CCF files are located.
  • Set up the SAS_ODF environmental variable to your SciSim FITS directory.
  • Run odffix including the parameter odfdir if you are not directly on the ODF directory: odffix odfdir=$SAS_ODF
  • Run rgsproc with default parameters
  This will take a while... Time for a coffee, perhaps.

  The results of rgsproc will include the spectra and the response files necessary for fitting the simulated first and second order spectra obtained. However, for you to get a feeling about the quality of the simulation, we recomend you to get first a look into the spatial and energy images and the corresponding selection regions. For this you should first generate the spatial and energy images and the "rgsimplot" output as explained in the RGS data analysis thread. The output of our simulation is this:

  

  If everything looks ok, like in this case, you are ready to analyze the spectra with XSPEC

• Analysing the spectrum wiht XSPEC

  Now that we have the spectra and response files, we can run xspec and load the files corresponding to first and second dispersion orders.

  
  xspec
  data 1:1  PxxxxxxyyyyR1eeeeSRSPEC1001.FIT 
  resp 1  PxxxxxxyyyyR1eeeeRSPMAT1001.FIT
  data 2:2 PxxxxxxyyyyR1eeeeSRSPEC2001.FIT
  resp 2  PxxxxxxyyyyR1eeeeRSPMAT2001.FIT
  
  

  We have performed the SciSim Simulation using a Raymond-Smith model with following parameters:

  • nH=0
  • kT=0.86
  • Abundance=0.27
  • redshift=0

  Now we can try to fit our data with an absorbed Raymond-Smith model. After introducing the values and fit the spectrum, we obtain the following plot. The fit is of general good quality, except for the high energy region in second order. The effective area corrections applied in the second order to the RGS data through a fudge function (introduced for correcting effects seen in RGS real data) are obviously wrong for simulated data. Fitting solely the first order simulated data leaving free the Raymond-Smith fundamental parameters yields:

  kT = (0.863 +/- 0.002) keV and Abundance = (0.212 +/- 0.005)

  which are qualitatively acceptable results.

  

  ---

  [SciSimRGSThread.html, v 1.0]

  [Last update: 5-Jan-2005 Carlos Gabriel]