As has been noted firstly by Liller (2000), CI Aql suffers
from a significant amount of interstellar reddening. Since the most
interesting parameters of a nova system depends criticially on the
inferred luminosity, and consequently, the distance, an
accurate determination of the colour excess and visual extinction
is highly desirable.
Several spectroscopic methods using absorption features
originating from the interstellar matter in the line of sight
exist (see, e.g. Jenniskens & Désert 1994;
Munari & Zwitter 1997; Oudmaijer et al. 1997), though these
methods suffer from significant limitations. For instance, the Na I D
doublet provides reliable excesses only in the moderately
reddened region (up to
E(B-V)=0
4, Munari & Zwitter 1997).
Some of the diffuse interstellar bands (DIBs) provide good reddenings
(e.g. DIB 5849, Oudmaijer et al. 1997), while other give only rough
estimates. The internal consistency is in the order of 0
1-0
2 even
in the best cases, therefore, a certain amount of uncertainty
cannot be exceeded.
We have surveyed all of the medium resolution spectra to identify
possible DIBs taken from the list of Jenniskens & Désert (1994).
We have unambiguously found DIB 5849 and 6613. Their equivalent
widths (W) was measured using the IRAF task splot. The resulting
values are:
Å and
Å.
Unfortunately, the latter is slightly affected by a telluric line at
6612 Å, that is why its width has a larger uncertainty.
Jenniskens & Désert (1994) gave the following ratios for
the
W/E(B-V): 0.048 (err. 0.008) for 5849 and 0.231 (err. 0.037) for
6613. The corresponding reddenings are E(B-V)
and E(B-V)
.
The interstellar line Ca II 3933.66
can be also used through the empirical relationships between the width of
DIB 5780 and interstellar lines (see Table 2 in Jenniskens & Désert
1994). Therefore, although we have not detected DIB 5780, we could
convert the measured Ca II equivalent width to
which
resulted in an E(B-V)=0
30.
In the case of CI Aql the Na I D doublet is of lower significance due to
the saturation effects. Interestingly, as has been noted by the
referee, there is an apparent difference between the strength of Na I D
lines obtained by us and that of Greiner et al. (1996), i.e. the latter
data suggest a weaker, unresolved doublet. A real difference would
query the interstellar origin of this resonance line and the whole reddening
estimation should be reconsidered. We attribute this
phenomenon to the lower resolution of that spectrum by Greiner et al.
(1996), because a close inspection of their Fig. 1 reveals a broad
(15 Å) single Na I D line. We could reproduce this kind of
appearance with a resampling and Gaussian convolution of our spectrum
mimicking the same resolution as quoted by Greiner et al.
(about 1 Å FWHM). However, we cannot solidly exclude the
possibility of other origin, e.g. some kind of circumstellar absorption.
The measured equivalent widths are
Å and
Å.
Their ratio is 1.10, far from the theoretically expected 2.0 at the
lowest optical depths, but exactly what is found for the asymptotic
behaviour at high reddenings (Munari & Zwitter 1997).
If one checks the relation between equivalent
width and reddening presented in Fig. 2 in Munari & Zwitter (1997),
only a weak and approximative conclusion can be drawn as
0
8
<E(B-V)<1
5.
Further constraints on the reddening are provided by the published colour
measurements. Hanzl (2000) gave B-V=0
on May 7.03 UT
(
= +2 d), while Jesacher et al. (in Wilson et al. 2000) presented
B-V=0
82 on May 10.97 UT (
= +6 d). The B-V colour
of novae around maximum tends to be about B-V=0
with a
significant dispersion of
(Warner 1995).
The resulting reddening lies between 0
46-0
59 mag (with 0
06 formal
error). Two magnitudes down from maximum the dispersion decreases, therefore
the relation
can be also
used. There are a few BV CCD photometric measurements in the
VSNET database obtained in early June, resulting in again a reddening
about 0
5. However, three prominent emission complexes are
covered by the standard V passband (N II 5679, He I 5876/Na I D, H
),
therefore we consider these colour measurements to be heavily affected
by the presence of such strong emission.
It is difficult to say which reddening is more reliable. In the
following discussion we adopt the unweighted mean of spectroscopic
values which is 0
3 (formal error). Despite
the limitations of using interstellar lines, their observational
data are much less affected by the nova itself as in the case of
multicolour photometry.
Copyright ESO 2001