The goal of the CHIANTI project was to develop, to the greatest extent possible, the most complete and accurate database for reproducing astrophysical spectra. In order to determine how well these goals have been met, we have collected lists of spectral lines and their wavelengths observed in the solar spectrum between 50 and 1100 Å. At this stage, we know that the CHIANTI database is incomplete below 50 Å, largely because of the omission of the hydrogen and helium isoelectronic sequences. Beyond 1100 Å, lines of neutral species, which are not included in CHIANTI at this point, begin to dominate the spectrum. Observed spectra of the non-flaring Sun include Behring et al. (1972, 1976), Burton & Ridgely (1970), Feldman & Doschek (1991), Freeman & Jones (1970), Thomas & Neupert (1994), and Widing & Sandlin (1968). The list of Widing and Sandlin has been expanded by Widing (1996) to include unidentified lines not reported in the original paper. Spectra of the flaring Sun include Acton et al. (1985), Dere (1978) and Kastner et al. (1974). The line list of Dere (1978) has been updated to include the additions of Dufton et al. (1983b), Feldman et al. (1985), Jupen (1984), Keenan et al. (1992), Widing (1982) and Widing et al. (1986).
Table 3 lists observed wavelengths, CHIANTI
wavelengths, transitions,
predicted intensity information and the
references for observations of
each line. The first column provides an average
value of the
wavelengths observed in the various reported
spectra. Reported
wavelengths that differed by less than a chosen
wavelength difference
were generally simply averaged to
arrive at the
observed wavelength quoted in Table 3. For
wavelengths below 400 Å,
was set to 0.03 Å, for
wavelengths between 400 and
600 Å, was set to 0.04 Å and
for wavelengths above
600 Å, was set to 0.05 Å.
The second column
contains the wavelength in the CHIANTI data base.
In those cases were
the transition is not in the CHIANTI data base, a
wavelength is
reported and the denoted by a intensity value
of... in the
intensity column. Most of these wavelengths are
from Kelly (1987).
The third column shows the ion giving rise to the
spectral line and the
fourth column provides the transition. The
fourth column gives the
intensity expected from a solar flare using a
differential emission
measure distribution derived by Dere &
Cook (1979). The intensities
are calculated from the differential emission
measure, which spans the
temperature range 3 104 to 2.5 107 K,
with the elemental
abundances of Allen (1973) and the
ionization equilibria of
Arnaud &
Rothenflug (1985). The intensity values
should only be used as a guide to
the identification and the problem of blending
and should be used in a
relative sense. Lines with predicted intensities
less than 10 were not
considered likely to be observed. Some
identifications should be
considered simply coincidental, for example
high-temperature lines seen
in quiet Sun spectra of Behring et al. Predicted
spectral lines with
intensities above 200 were included even if there
was no reported line
at that wavelength. These cases are noted by a
in Col. 1.
Only lines for which there are accurate values of
the energy levels
available based on observed spectra are included.
The last column
provides the reference where the observation was
reported in the
literature.
Table 3 indicates that the CHIANTI data base provides a very nearly complete reproduction of astrophysical emission line spectra between 50 and 1100 Å, at least in terms of observed spectral lines. Many new identifications for previously unknown lines have been found but many lines remain unidentified. The Table also provides information on the likelihood of problems with blending. Based on our experience in putting the observational data together for this Table, the observations between 170 and 600 Å appear to be the most consistent. At shorter wavelengths, lines are often reported by one observer but not another. At longer wavelengths, the observations do not appear to have been made with sufficient sensitivity. Observations with the SUMER instrument on SOHO should help correct the deficiencies in the long wavelength portion of Table 3.
Future papers will describe the comparison between observed spectral line intensities and prediction by CHIANTI. Preliminary work has already been performed on this comparison in order to assure the basic correctness of the CHIANTI data base. For example, EUVE spectra of Procyon have been compared with the spectra computed with the CHIANTI database for a quiet sun differential emission measure and EUVE spectra of Au Mic in a flaring state have been compared with the CHIANTI database for a solar flare differential emission measure. In addition, the SERTS spectra have also been compared with predictions from the CHIANTI database. Detailed comparison with these spectra are in progress but our work to date assures us of the reliability of the CHIANTI database.
Figure 3: Temperature-sensitive line ratio of Si
IV calculated from the CHIANTI database
Figure 4: Density-sensitive line ratio of Si IX
calculated from the CHIANTI database
Figure 5: Synthetic spectrum of the quiet Sun
between 160 and 200 Å