9.4. Stellar Spectra

We examine a high resolution spectrum of a nearby star at a distance of 10 pc. The files corona.spo and corona.res contain the spectrum and instrumental response of this source. This spectrum has been simulated using a RGS-like response. We do not provide real data files, because of their size and because of the time constraints of the course. You can use plot_rgs.com for plotting, but it is more instructive and easier to plot manually!

  1. Read the data into SPEX and plot it in unit Å. The data ranges roughly from 8 to 38 Å. Hint: you may issue successively the commands plot set all and plot line disp t to get a connecting line through the data for better visibility.

  2. Determine the wavelength of the two strongest lines in the spectrum and use this table to identify them: http://www.sron.nl/~kaastra/leiden2018/line_new.pdf. Can you give a rough estimate of the temperature (in keV) of the corona from the fact that you see these lines? Note: The temperatures listed in the line list are in log Kelvin (k_b = 8.617 \times 10^{-8} \mathrm{keV} \ \mathrm{K}^{-1}).

  3. Load a model with just one component called CIE (com cie) which means Collisional Ionization Equilibrium. Fit the spectrum and calculate the error. Was your temperature estimate right?

  4. There is one line that is not fitted correctly. Use the line table from exercise 2 to see which element is emitting this line. Set this element to thawn and fit again. Is the fit better?

  5. Now you can also free the iron and oxygen abundance and fit the spectrum again. What are the values for the abundance? The values in the fit are relative to solar abundances (more details can be obtained with the command asc ter 1 1 abun, see Ascdump: ascii output of plasma and spectral properties). Calculate the errors. Are the fitted abundances consistent with solar abundances?

    These kind of spectra (with a low continuum) often produce wrong abundances because of the low S/N of the continuum and the bias produced by \chi^2 statistics (Humphrey, Liu & Buote, 2008). In SPEX we can switch from C-statistics to \chi^2 statistics to test this. This is accomplished by issuing the command fit stat chi before doing another fit.

  6. Fit the spectrum using fit stat chi. Are the abundances the same as before? Calculate the errors. Are the new values consistent with solar abundances? Switch back to C-statistics after this step.

  7. Fix all abundances and free the ion temperature. Fit the spectrum and calculate the error on the ion temperature. Is it well constrained? Find out what the ion temperature does with the lines for values 1, 10, 100, and 1000 (change the value in the model and calculate the model by typing calc. Do not fit!). What happens?

  8. Plot the spectrum from 21 Å to 23 Å. These lines are called the oxygen triplet. The lines depend strongly on electron density. Play around with the electron density in the same way as you did with the broadening (ion temperature). What happens with these lines? Calculate the error. Can you give an upper limit for the density of the plasma?

  9. There are a lot of lines in this spectrum. You can learn also which lines belong to a certain element by putting its abundance to 0 and calculate. Do this for O, Ne and Fe. If you like, you can also try it out for other elements. Finally, SPEX can also provide a list with all the present ions: asc ter 1 1 icon.

Note: SPEX can provide a lot of information about the plasma or an absorber through the command asc ter. See the SPEX manual for more output options.

Learning goals:

After having done this spectrum, you should know:

  • How to plot a spectrum in wavelength units.

  • How to identify spectral lines.

  • How to get physical parameters from a thermal X-ray spectrum.

  • How to measure abundances from an X-ray spectrum.

  • How to measure electron density from an X-ray spectrum.

  • How to use the “asc” command in SPEX (Ascdump: ascii output of plasma and spectral properties).