Another required input for SPECTRUM is an atomic and molecular data file. One such file, stdatom.dat, which includes solar atomic abundances from Grevesse & Sauval (1998), is included in the distribution. The format of that file can be seen from the following sample lines from the beginning
code Abund Amass I1/D0 I2/rdmass I3 I4 maxcharge 1 -0.0360 1.008 13.5984 0.000 0.000 0.000 0 2 -1.1060 4.003 24.5874 54.4178 0.000 0.000 1 3 -10.88 6.941 5.3917 75.6400 122.4543 0.000 1 4 -10.89 9.012 9.3227 18.2111 153.8966 217.7187 1 5 -9.25 10.811 8.2980 25.1548 37.9306 259.3752 1 6 -3.49 12.011 11.2603 24.3833 47.8878 64.4939 3
and from the end
606 -7.50 24.02 6.15 6.0055 1.00 00.00 0 607 -7.50 26.01 7.66 6.4627 1.00 00.00 0 608 -7.50 28.01 11.108 6.8604 1.00 00.00 0 813 -7.50 42.981 5.27 10.0436 1.00 00.00 0 814 -7.50 44.06 8.26 10.1897 1.00 00.00 0 822 -7.50 63.879 6.87 11.9921 1.00 00.00 0 840 -7.50 107.223 7.85 13.6118 1.00 00.00 0The first column is the atomic or molecular code, described in the previous section. For instance, ``6'' refers to carbon, and ``606'' to the molecule C
If you want to compute a spectrum with non-standard abundances, this can be accomplished by modifying the relevant abundances in stdatom.dat. However, a number of points must be kept in mind.
First, the overall metallicity in the header for the stellar atmosphere model (see section above) is used to scale the abundances for the metals in stdatom.dat. Thus, if [M/H] = -1.0, SPECTRUM subtracts 1.0 from the logarithmic abundances in stdatom.dat for lithium through uranium. Since the opacity due to the iron peak elements is so important in establishing the structure of the stellar atmosphere, it is advisable to use a stellar atmosphere model with [M/H] as close as possible to the [Fe/H] of the star you are analyzing.3.1
Let us suppose that you want to compute a stellar atmosphere model
with all but a handful of metals having an abundance one tenth that
of the sun. Assuming that iron is not included in that handful,
select the appropriate stellar atmosphere model with [M/H] = -1.0 and
then individually adjust the abundances in stdatom.dat for
the handful. For instance, in the sun (and in stdatom.dat),
the logarithmic abundance of calcium (code = 20)
is -5.68. Let us suppose that you actually want to use an abundance
of -6.38. If the [M/H] of the model is , then the [M/H] scaling
changes the abundance from stdatom.dat to -6.68, so this means that to get an abundance of
-6.38 you will need to adjust the abundance in stdatom.dat
to -5.38. Note that this is equivalent to [Ca/Fe] = +0.30 if [M/H] = [Fe/H].
Another example: Let us suppose that you want to compute a spectrum with iron and all other metals (except for, say, calcium) with abundances 0.15 dex less than found in the sun. Let us say that [Ca/Fe] = +0.30. Select a Kurucz atmosphere model with [M/H] = -0.20 (ATLAS9 models with [M/H] = -0.15 are not available, so -0.20 is the closest). Adjust the abundances of all the metals (except for calcium) in stdatom.dat upwards by 0.05 dex, and calcium upwards by 0.35 dex.
If this all seems to be complicated enough to drive one to tears, v2.76 now has a facility to input a ``fixed abundance'' file that can be used to specify individual fixed elemental abundances which will not be adjusted by the [M/H] scaling factor. Suppose that we want to compute a spectrum with a logarithmic calcium abundance = -6.38, regardless of the value of [M/H] of the model. Create a file (name it anything you like) with the contents:
TOTAL 20 -6.38and run SPECTRUM with the ``x'' switch (see §