ROSAT PSPC analysis

• Return to the Basic search of the W3Browse page in Netscape.

• Enter CEN x-3 in the Object Name or Coordinates box while selecting ROSAT from the Active Mission Archives section. Click on Submit Query.

• From the three results, select the PSPCB observation with the sequence ID RP900165N00.

• Submit the selection and click on the Basic Science Events file named rp900165n00_bas.fits.Z.

• Retrieve Data Products and Download TAR file into your directory.

• Un-tar the w3browse-$\char93$.tar file and then remove it.

  You will end up with a rp900165n00_bas.fits.Z file which you should uncompress.

• Bring up Xselect and enter a session name (the default xsel would be fine).

• read events < Enter >

  Accept the current directory ./

  Enter the filename rp900165n00_bas.fits. (Xselect allows file completion so typing rp < Tab > should bring up the file, unless you have other files that begin with rp in that dierectory in which case you should make the input unambiguous, then press < Tab >).

• Reset the mission from ASCA to ROSAT.

• The first thing you must do is extract image < Enter > and then saoimage.

  Xselect also allows uniquely abbreviated commands to be used so that you may type ext image < Enter > and sao instead.

• Panning in on the source at the center, select a circular cursor around the source. Save it by pressing the s key on the keyboard. Click on region then write and give it a unique name such as cenx3.reg.

  You may close SAOimage by clicking on etc then on QUIT.

• Using this region, filter region cenx3.reg < Enter >.

  extract events < Enter >.

  clear reg < Enter >.

  ext image < Enter >.

• Now ext curve < Enter > for the timing analysis and plot curve < Enter > to see the light curve.

  Figure 5: This is the lightcurve generated from CEN x-3 as detected by the ROSAT satellite.

• Now filter time cursor to select the part of the lightcurve that is due to the source.

  Type Q at the PLT$>$ prompt to go into filter mode. Make one click to the left of the first high countrate peak, around 2.5$\times$10$^4$, and another click to lthe left of the last dataset, around 5.9$\times$10$^4$. There should be a horizontal cut across most of the high-count data.

  Type x in the plot window to exit the filter mode.

• ext curve < Enter > again now that it has been time filtered and plot the curve to see the dataset that we have selected.

  Figure 6: CEN x-3 time filtered lightcurve

• save curve naming it a unique filename.

• Now ext spectrum < Enter > and plot spec < Enter >.

• Since we can eliminate datapoints in Xspec and there are no glaringly bad points here, there is no need to filter any channels. We can save spectrum < Enter > and allow rebinning of the data.

  Figure 7: Energy spectrum of CEN x-3 in Counts vs. Channel, before rebinning.
  Figure 8: Energy spectrum of CEN x-3 in Counts vs. Channel, after rebinning.

• You are now done with Xselect. Type quit and there is no need to save your session.

• To find the dominant frequencies in CEN x-3's lightcurve, run powspec on your lightcurve generated in Xselect.

• Xselect has a default bintime of 16 seconds. The lightcurve was generated with this bintime and it is the Newbin Time you should use in the power spectrum.

• Allow the Maximum Newbin No. to be the Number of Newbins/Interval.

• Select one interval per frame and 0 for the rebinning. Allow the program to plot your results.

  Figure 9: This is the power spectrum from CEN x-3 using the data collected from the ROSAT satellite

• You should go back and compare the frequencies determined from the EXOSAT data to the frequencies extracted here. The purpose of selecting a 16 second newbin time and a rescaling of the y axis in the power spectrum of the me/rates/c directory data was so you could easily make a comparison between the two observations of the same star.

  Keep in mind that the ROSAT data was taken around 1995 while the EXOSAT data was taken ten years earlier in 1985. Since CEN x-3 is in a binary system, a change in the dominant frequencies of the star's lightcurve could imply changes in its motion over time.

• Now we can compare the energy spectra from ROSAT and its ability to fit our models as compared to the EXOSAT data. Recall that with EXOSAT our fits were not very good. Also, recall that the energy ranges of the two detectors are very different and that ROSAT can only detect X-rays up to 1keV.

  Do you expect the fits to be better or worse?

• First you must run pcarf as you did for GK PER to create an *.arf file. Use the pspcb_gain2_256.rmf and pspcb_v2.spec_resp files you downloaded from the

  http://xray.rutgers.edu/$\sim$matilsky/documents/

  web page as input for the RMF and SPECRESP files, respectively. (If you do not have these files return to the GK PER manual and follow the instructions for copying them into your directory).

• Now that you have the *.arf file you are ready to run Xspec. Type xspec < Enter > at the prompt and set the plotting device cpd /xw < Enter >.

• Enter the data data your_spectrum.pha < Enter >

• Enter the RESP file resp pspcb_gain2_256.rmf < Enter > (if this file is not located in this directory you must give the full pathname of its location).

• Enter the ARF file arf your_arf_file.arf < Enter >

• plot data < Enter >.

• ignore bad < Enter > and plot data < Enter >. There do not seem to be any other bad points to ignore.

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

  Use the flux command to find the flux.

  How do these results compare with the EXOSAT powerlaw fit?

• Repeat the process for a blackbody and a powerlaw spectrum.

  For the fitting of the blackbody spectrum you may want to enter the command fit 100 < Enter >. This will take the fitting process through 100 iterations before asking you if to continue.

  • Which model gave the best fit?

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

  • how did these results compare with the EXOSAT results.