EXOSAT analysis

• Go to the HEASARC website at http://heasarc.gsfc.nasa.gov

• Go to the W3Browse link and select the Basic search.

• Select the EXOSAT satellite from the Past Mission Archives section and enter CEN x-3 in the Object Name or Coordinates box.

• Click on Submit Query.

• Go to the EXOSAT ME Spectra and Lightcurves (HEASARC_ME) section and select the second to last observation (with the longest exposure).

• Submit the entry and select the following files to download:

  Argon 1-15keV Spectrum (FITS, 9 kbytes) s56059.pha.Z

  0.8-8.9keV 1-10s Lightcurve (FITS, 62 kbytes) c56050.lc.Z

  0.8-8.9keV 1-10s Lightcurve (FITS, 43 kbytes) c56071.lc.Z

  0.8-8.9keV 1-10s Lightcurve (FITS, 43 kbytes) c56081.lc.Z

  0.8-8.9keV 1-10s Lightcurve (FITS, 83 kbytes) c56059.lc.Z

  Argon 1-15keV Response Matrix (FITS, 44 kbytes) s56059.rsp.Z

  0.8-8.9keV Lightcurve (FITS, 6 kbytes) d56050.lc.Z

  0.8-8.9keV Lightcurve (FITS, 6 kbytes) d56071.lc.Z

  0.8-8.9keV Lightcurve (FITS, 6 kbytes) d56081.lc.Z

  0.8-8.9keV Lightcurve (FITS, 7 kbytes) d56059.lc.Z

• Click on Retrieve Data Products and click on Download TAR file to place your w3browse-$\char93$.tar file the appropriate directory.

• Un-tar and uncompress your files.

  You should end up with a small directory tree containing files:

  me/rates/d, me/rates/c, and me/spectra/.

• Remove the w3browse-$\char93$.tar file.

• Change into the me/rates/d directory and run the lcurve routine on each of the *.lc files using the default values as input to all the commands.

  For the output file name use the same prefix as the input file with a .flc as a suffix.

• create a file list.con and it listing the *.flc files one on each line (four lines all together) then run lcurve with list.con as your input (remember to enter it as @list.con).

  When asked for the Number of Newbins/Interval, choose the Maximum Newbin No., in this case 1589 so that you will have 1 interval with all your data on it.

  Figure 1: This is the lightcurve generated from CEN x-3 as detected by the EXOSAT satellite.

• Run powspec using the output from your concatenated lcurve output. Again use the Minimum Newbin Time when asked for Newbin Time or negative rebinning (30 sec) and the Maximum Newbin No. for the the Number of Newbins/Interval.

  The plot of your power spectrum appears on the following page. You may wish to rescale and examine the peaks in the plot to determine the dominant frequencies in the curve.

  Figure 2: This is the fourier transform (powerspectrum) of the CEN x-3 lightcurve shown on the previous page.

• Now change into the me/rates/c directory.

• Run lcurve on each of these *.lc files.

  Notice that the Minimum Newbin Time is 0.5 seconds. This allows you to choose smaller time bins and detect higher frequencies. If you choose the 0.5 sec for the Newbin Time or negative rebinning you will get lightcurves with thousands of points and very fine resolution. The routine will also take very long to process. select a timing of 16 second bins and process all four *.lc files using the same binsize. Continue to enter the Maximum Newbin No. for the Number of Newbins/Interval since this will vary from file to file.

• Again create a list.con file, listing all four generated *.flc files in a column. Run lcurve on the list of files again using the same Newbin Time or negative rebinning as the one you chose for the individual *.lc files and follow the same procedure as in the last concatenated lightcurve you produced. The following is a plot of the resultant lightcurve using 5 second Newbin Time for all four *.lc and the final, combined lightcurve. Notice the fine resolution of the points.

  Figure 3: This is the lightcurve generated from CEN x-3 using the minimally binned data at 0.5 seconds, and rebinned into 16 second Newbins.

• Now run powspec on this lightcurve.

• When you have generated a plot of your power spectrum and are in the PLT$>$ mode, type rescale y 0 4000 < Enter >. You will have changed the y axis to a lower maximum value. Notice the peaks in the higher frequencies come up. Note the location and height of the peaks. The resultant plot appears on the following page.

  Figure 4: This is the power spectrum of CEN x-3 displayed using a logarithmic scale in frequency.

• Change into the it me/spectra directoy. We will now try to analyze the spectrum of CEN x-3 and fit it to the three models that we have used.

• Type xspec < Enter > to call upXspec.

• Enter data s56059.pha at the prompt. Notice the message

     using response (RMF) file...       s56059.rsp
  

  Recall that this is not the case with ROSAT data which we will be suing shortly.

• Call up the plotting device cpd /xw and plot the data any time you make a change to it.

• Get rid of bad data points with the ignore bad command.

• If you plot ldata < Enter > you will examine the plot with a logarithmic scale in the y axis. Notice that around channel 55 the curve loses its smoothness and shows a low signal to noise ratio.

  It would be useful to eliminate the points above channel 54. Type ignore 55-** < Enter >. Return to a linear plot.

• Change to units of energy.

• Select a powerlaw spectrum, renormalize and fit the data.

  Use the flux command to find the flux and, if you know the distance, determine the luminosity of the source.

• Repeat the process for a blackbody and a powerlaw spectrum.
  • Which model gave the best fit?

  • What were the temperatures extracted for the blackbody and the thermal bremsstrahlung models?