Low ionization stages of iron, owing to their rich and complex
structure and their relatively high abundance, dominate certain
wavelength regions in the spectra of a variety of stellar and
non-stellar objects, as well as the interstellar medium.
For example,
Fe I contributes significantly to the UV opacity in the solar atmosphere
(e.g. Bell et al. 1994) and it is possible that the presence of
autoionizing resonances of this ion might be a major contributor to observed
features in the absorption spectra.
The interpretation and modeling of these observations rely almost
entirely on theoretical calculations which, until the recent advances
made under the Opacity Project (OP, Seaton et al. 1994) and the IRON
Project (IP, Hummer et al. 1993), could not be
obtained either with sufficient accuracy or on the large scale needed for
a full determination of the relevant astrophysical parameters.
However,
some of the OP calculations, particularly for the low
ionization stages of IRON, including Fe I, are of insufficient accuracy.
For example, earlier R-matrix calculations for Fe I were carried out by
Sawey & Berrington (1992), but they included only
terms dominated by the ground configuration of Fe II
and therefore do not accurately represent the coupling effects (for
example, their calculations did not obtain the ground state of Fe I).
Other calculations of photoionization cross sections for Fe I have been
carried out in central
field type approximations (Reilman & Manson 1979; Verner et al.\
1993), and using the many-body perturbation method by Kelly & Ron
(1972) and Kelly (1972), but are limited to just the ground
state. Moreover, these central field calculations
ignore the coupling effects and resonances
which led to an underestimation of the
photoionization cross section by more than three orders of magnitude
as shown in a previous paper (Bautista & Pradhan 1995).
The many-body perturbation calculation by Kelly considers a limited
number of coupling effects including those that result from the
configuration
and reproduces some of the overall structure
of the photoionization cross section, but neglects correlations
from the 
configuration that are of considerable importance
(see Sect. 4.3). Kelly's computations also did not allow for
most of the narrow
autoionization resonances that converge onto each of the ionization
thresholds and make an important contribution to the cross section in the low energy region.
In the absence of reliable and extensive radiative data for Fe I, theoretical
calculations and modeling work that involve this ion, like computations of
stellar opacities and modeling of the ionization equilibrium of iron,
have been severely limited and some authors even prefer to
exclude this ion from their models (e.g. LeBlanc & Michaud 1995).
It is one of the aims of the IP to carry out improved calculations for the low ionization stages of iron (Bautista et al. 1995; Pradhan 1995) and results for some of them have already been reported, e.g. Fe II (Nahar & Pradhan 1994; Nahar 1995a), Fe III (Nahar 1995b), and Fe V (Bautista 1996), and are expected to become available soon from TOPbase at CDS (Cunto et al. 1993; Mendoza 1995, private communication).