Normal Lines

SPECTRUM includes code for all the normal spectral line-broadening mechanisms, including natural, Doppler, Stark, resonance and van der Waals. For most metallic lines, van der Waals broadening dominates. SPECTRUM includes code to implement the Anstee & O'Mara van der Waals broadening theory (see Anstee & O'Mara, 1991; Anstee & O'Mara, 1995). This broadening theory is automatically implemented for neutral atomic species if the transition type is indicated in the linelist file (see §  for how this is done) for s-p, p-s, p-d, d-p, d-f and f-d transitions. For transitions involving higher angular momentum quantum numbers, or for ionized species, the classical van der Waals theory is implemented, with the mean square radii of the states given by the hydrogenic approximation:

$$\langle r^2 \rangle = \frac{n^{*2}}{2}[5n^{*2}+1-3l(l+1)]$$

where $n^*$ is the effective principal quantum number. If the transition type is not indicated, then the classical van der Waals broadening theory is used, but with the term involving the angular momentum quantum number neglected in the expression for the mean square radius.

However, for ionized species, the Anstee-O'Mara theory can be invoked if the ``AO'' transition type is specified in the linelist. This requires the user to include the $\sigma$ and $\alpha$ parameters of that theory in a packed format immediately after the ``AO'' in the linelist file. For instance, Barklem & O'Mara (1998) have calculated these parameters for certain lines of ionized species; for the Ca II $\lambda$8498 line, $\sigma = 291$, $\alpha = 0.275$. Below is an example of how those data can be included in the linelist file:

8498.023  20.1   13650   25414   -1.416  1.000 AO  291.275 T89

Of course, it is also possible to use the ``AO'' transition type with neutral species if, for some reason, you are not satisfied with the tabulated values of $\sigma$ and $\alpha$ built into SPECTRUM.

Finally, if the user wants complete control over natural, quadratic Stark and van der Waals broadening, it is possible to input the relevant half-widths using the ``GA'' transition type. These half-widths may either be calculated by the user with her favorite broadening theory, or extracted from the original Kurucz linelist files (see http://kurucz.harvard.edu) or from, for instance, the VALD database (http://ams.astro.univie.ac.at/~vald/). As an example, in the Kurucz hyperfine linelist file, for the line Fe I $\lambda$4620.128 it is given that $\log(\Gamma_{\rm rad}) = 8.32$, $\log(\Gamma_{\rm Stark}) = -6.20$ and $\log(\Gamma_{\rm vdw}) = -7.78$. These values may be entered into the linelist file in the following way:

4620.140   26.0  24772   46410   -3.739  1.000 GA 8.32 -6.20 -7.78  NIST

The Stark half width is per electron number (i.e. SPECTRUM multiplies the input half width by the electron number density) and the van der Waals half width is per neutral hydrogen number and is for $T = 10000$K. The van der Waals half width is further modified by SPECTRUM to take into account van der Waals broadening due to neutral helium and H$_2$ collisions. Bear in mind that these half-widths are logarithmic quantities. If you truly want to enter a zero Stark or van der Waals half width, don't enter 0.00! Enter a large negative quantity such as -15.0. SPECTRUM interprets a 0.00 entered for either the Stark or the van der Waals half-width as an indication that the user would like the program to make its own estimate for those half-widths, and that is exactly what it does.