Two major anomalies affect the collisional data computed by Tayal et al. (1987) for the forbidden transitions within the
ground configuration of FeXII
values at 4Ry and at 6.6Ry, as
clear from their Table 2 (click here). This feature was pointed out by Mason
(1994).
as a function of E above
6.6Ry for the transitions from the 
up to the
levels.
and 
correlation orbitals in the
radial waves (Fig. 2 (click here)) is the closest to their model.
Regarding the first point, we note that 4Ry is above the highest
excitation threshold in both calculations, so the eventuality of
hitting a true resonance at that energy must be ruled out. This raises
the question whether their unusually high values at 4Ry might be due
to the effect of unphysical resonances in the open channel
energy region, despite their use of a T-matrix smoothing procedure in
that energy range. Our 7 term computation with
correlation orbitals revealed a bunch of resonances, likely to be due to
the correlation orbital, between 4Ry and 7Ry, and
between 15Ry and 20Ry. Our approach does not include T-smoothing
but we still find the same steep drop in the
values between 4Ry and 6.6Ry. This fact seems to suggest that
open channel resonance effects are still affecting their non-resonant
background, causing this behaviour.
As far as the second point is concerned, they ascribe the increase in
to larger contributions from higher partial waves due to the presence
of stronger long-range quadrupole interactions. However we performed a
similar top-up procedure without obtaining such a pronounced effect in
the higher partial waves contributions. Furthermore we point out that,
according to the classification proposed by Burgess & Tully
(1992), the high energy behaviour of 
for forbidden
transitions should follow a constant or E-2 trend,
depending on the role played by electron exchange. An increase in
with E is, on the contrary, typical of optically allowed
transitions.
The use of correlation orbitals in the target description is
problematic, as stated in
Saraph & Storey (1996), because of the introduction of
unwanted spurious resonances in the open channel region and because of the
inaccurate position on the energy scale of the additional resonances brought
in by these non-physical orbitals below the highest excitation
threshold. It is for this reason that we used all spectroscopic
orbitals in our best R-matrix calculation including 19 target terms. A
comparison of our Table 7 (click here) with the Tayal et al. (1987)
collision strengths reveals a situation where their values are, for most
transitions, larger than ours. The use of a different scaling
factor in the geometric series top-up procedure and possible residual
open channel resonance effects in their data might account for this
discrepancy. However, no clear pattern is observed in comparing the
two sets of data for the effective collision strengths (our Table 8 (click here)
and their Table 3 (click here)). Their 
data for the 
transitions show again the same anomalous
behaviour as a function of 
as we found in their
values. Here, probably, in integrating the collision strengths over a
Maxwellian distribution, the presence of their broad resonance
features due to correlation orbitals is balanced by our inclusion of
additional series of physical resonances converging to the twelve
extra thresholds of the 
configuration, which is
lacking in
their target representation. The inclusion of the second
excited configuration in the target expansion is an important feature
of our calculation because it enables us to provide collisional
data for the important transitions up to the
levels. The only set of data previously available for these
transitions was by Flower (1977), who used a very crude target
model and included resonance effects by the approximate method of Petrini
(1970). The radiative and collisional
atomic parameters for FeXII discussed in this paper should
therefore prove a powerful diagnostic tool for future spectroscopic
applications.
Acknowledgements
This research was supported by PPARC grants GR/H94979 and GR/K98506 for the IRON Project meetings and CRAY computing time.