5. Structures present in the lines of different elements

Despite the fact that our material was taken at several different dates and is not large enough to permit a follow up of the variations, we provide in what follows a detailed description of what is seen on our spectra. The reasons are twofold: the history of the star is a very complex one and even isolated bits of information are valuable for the general picture and second, because we present for the first time a collection of equivalent widths. Much of the previous material was based upon photographic plates, which are difficult to calibrate accurately.

Hydrogen. Practically all Balmer lines show structures. H 7 (from 04.04.93) appears as a sligthly asymmetri íc line, as if an unresolved emission component were present on the red border. H 6 (from 02.10.91) has an asymmetric profile, with unresolved emission components on both the violet and the red of the absorption line center. The wing is clearly steeper on the red border. In 1996 a sharp line core is visible, with a small red emission component.
H 5 (from 23.10.91) shows a violet and a red emission, separated respectively 166 and 110 km/s from the absorption line center, with V>R. This corresponds well with what is also observed in H 6, taken both in October 1991. In 1996 one sees very well the central sharp absorption superimposed upon the broad absorption line. The central absorption is flanked by two emission borders, with .

H 4 (25.10.91) shows a rather broad absorption line extending over 50 A (3000 km/s) upon which two intense emission components are seen, separated by a sharp central autoreversal. The violet and red peaks are separated by 150 and 100 km/s respectively from the autoabsorption line center. One sees V<R with equivalent widths of both emission components, measured from the continuum level, of 0.96 and 1.59 respectively. The emission line structure seem to be displaced toward the violet of the absorption line center. Since the absorption line is ill defined, the displacement can only be estimated to lie around 100 km/s In 1997 the profile is rather similar, with . Within the uncertainties of the absorption line, the emission structure seems more centered upon the absorption line profile than formerly.

H 3 (from 25.10.91) is seen only as an emission feature, with two peaks; the autoabsorption does not reach the continuum level. and the separation between the two peaks is of 220 km/s. The equivalent width of the e mission is of 61 A. The total width of the line seems to be of the order of 3700 km/s. The total width of the line is even larger in 1996, reaching 6000 km/s.

 

Paschen 15.11. 27.12. 30.12. 21.2. 31.3.

1989 1990 1990 1992 1993

P 21 - s 200 130

P 20 s 100 - 170 130

P 19 70 85 s 143^* 130^*

P 18& V>R V<R s s V<R

P 17 150^* 85^* 60^* 55^* 20^*

V>=R V=R V=R

P 16&& 170 115 s s,nm 95

A A A

P 15&& 115 115 s s 95

A A A A A

P 14 140 110 60 165 95

A V>R,A V=R^*,A V=R^*,A V=R^*,A

P 13&& 120 85 s 85^* 125^*

A A A V=R,A V>R,A

P 12 110^* 85^* 125^* 125^*

V>R,A V=R,A V<R,A V<R,A

The table provides for each Paschen line the separation between the components, in km/s, s = single. The second line provides the relation between the violet(V) and red(R) components. An asterisk means that the two peaks are well visible. A = stands for an underlying absorption line. & in the first column stands for a blend with O I, && for a blend with Ca II.

 

Paschen series.

The phenomena in the Paschen series (see Fig. 3 (click here)) are best summarized in tabular form. Table 4 (click here) groups the principal features. We add some comments which do not fit in the table. In 1990 the profile of the underlying absorption line is much stronger on the red side, on the 27.12. Three days later the violet wing has almost disappeared, so that one has the impression of an inversed P Cyg profile, and this is also true for the spectra of 1992 and 1993.

P 9 in 1990 has a double peak, with V = R, and a separation of 30 km/s. It must be mentioned that the existence of the double peak is somewhat doubtful, in view of the numerous components of atmospheric absorption.

P 7 in 1989 has a double structure, with V < R and peak separation of 95 km/s. In 1992, , with 145 km/s of peak separation.

As can be seen from Table 4 (click here) and Fig. 4 (click here), there exists no clear progression in the peak separations with the Paschen number, even within the same spectrum. We have also listed the equivalent widths (see Table 5 (click here)). It should be stated that these values are somewhat uncertain in the higher numbers, because of the confluence of the Balmer lines. In the lower numbers we see the underlying absorption and we have then the difficulty in defining what to measure. Whenever possible we have made both the measures from the interpolated continuum and from the underlying absorption line. Usually the measures differ considerably, up to almost a factor of two, the measures from the absorption profile being of course the larger ones.

 

Paschen 15.11. 27.12. 30.12. 21.2. 31.3. Notes

1989 1990 1990 1992 1993

P 25 - - 0.10 - -

P 23 - - 0.25 - -

P 22 - - 0.09 - -

P 21 - 0.92 nm 1.14 1.01

P 20 0.89 1.02 0.26 1.15 1.19 1.09

P 19 1.02 0.99 0.36 1.18 1.23 1.17

P 18 nm nm 0.62 nm nm Blend O I

P 17 1.21 1.45 1.20 1.88 1.73 1.24

P 16 1.64 2.23 1.54 3.54 2.15 Blend Ca II

1.98 2.30 4.22 2.43

P 15 2.36 2.82 1.61 3.96 2.61 Blend Ca II

2.76 3.25 4.57 3.22

P 14 1.39 1.60 1.21 2.17 1.61

1.94 2.36 2.46 2.31

P 13 1.97 2.24 1.94 3.74 2.05 Blend Ca II

2.61 3.38 4.22 3.21

P 12 1.16 1.60 - 1.96 1.28

2.21 3.03 3.04 2.74

The table provides in the first line the equivalent width measured from the interpolated continuous spectrum and in the second line the equivalent width measured from the lower points of the underlying absorption line. For the spectrum from 30.12.90, see text. nm = non measured.

 

As already said, the present material is much too limited in time coverage as to permit to draw useful conclusions, except the obvious one that the structures vary strongly over very short time intervals and that the variations can be up to 100% in equivalent width. We stress the fact that these are the first systematic measures of equivalent widths for this object.

Helium. Structures are present in some lines. 4471() observed on October 21, 1991, has the violet wing in emission, at a distance of 320 km/s from the center of the absorption line. In 1996 the broad underlying photospheric line is well visible and the violet emission line is found at about 280 km/s.

In 4921, of the ¹D series, (25.10.1991) the absorption line is flanked by two emission peaks, which are symmetrically placed around the absorption line, at a distance of 100 km/s, with V<R. In 1996 the appearance is about the same, but the violet emission is much broader than the violet one and structure is perceptible in both emission components. Also the forbidden line at 4920 is visible.

The 5015 line of the serie (25.10.1991), is flanked by two emissions separated 190 km/s, V<R. In 1996,we find a broad emission feature divided by an absorption peak, which leaves a narrower violet component and a broader red component, with V>R.

The 5875 line has its violet wing in emission in 1996. We have examined also the other He I lines present at this date and we find about normal absorption profiles, although in a number of lines one suspects a broad violet emission feature.

Oxygen. The variations in the profiles of O I 7774 and 8446 are also very large. The variations are illustrated in Figs. 2 (click here) and 3 (click here).

The 7774 line is an absorption line, with adjacent emission structures. For 7774 we find on 7.1.90 a double emission structure with a central absorption, the emission peaks being about equal and located at -285 and +165 km/s from the absorption line center, WE/WA = 1.65. On the 29.12.90 the red emission has disappeared, whereas the violet is found at -245 km/s, WE/WA = 0.22. On 29.10.91 we have two symmetric emissions, with V = R, displaced -155 and +25 km/s from the average of the two absorption lines, which are interrupted by a central emission peak., with WE/WA = 0.44. On 16.3.92 the structure is similar, except that now , with separations of -210 and +245 km/s, with WE/ WA = 0.19. On 7.4.93, the profile is similar to that of the 29.12.90, with a similar displacement of -245 km/s, with WE/WA = 0.42. In the preceding WE/WA stands for the ratio of the equivalent widths of the emissions over the absorption, counted from the interpolated continuum. As can be seen, in general the absorption predominates, except in the first spectrum. In 1996 (Jan. 9) the line presents a rather complicated structure. We have first the absorption line at 7776, accompanied by two broad emission features, with V>R. The violet one is about 380 km/s broad and the red, 250 km/s. On the violet wing one sees a further absorption, displaced 150 km/s from the main absorption line. The variations of the 8446 line are more difficult to follow, because of the blend with P 18. The line is however always in strong emission. On 15.11.89, we have , peak separation 100 km/s, W = 4.0 A. On 27.12.90, we find , peak separation 115 km/s, W = 4.5 A. On 30.12.90 (three days later) we find no double structure, W = 4.08 A. On 21.2.92 we find , peak separation 170 km/s, W = 5.1 A. On 31.3.93 we find two peaks of similar height, with separation of 95 km/s and W = 5.2 A. In all cases we have discounted the extrapolated equivalent width of P 18. It should be added that the structure of O I does not run parallel to that of the Paschen lines- for instance on 27.12.90 we find for O I and V> R for the Paschen lines. On other dates they are roughly parallel.

The profiles of the other O I lines present on the 1996 spectrum do not show structures.

Calcium. On the spectrum from 07.03.96 one observes 3933 as a sharp absorption lines, accompanied by an emission structure, broader on the V than on the red, and with V>R. The 3973 line shows a sharp absorption core, accompanied by two sharp satellite absorption lines, located respectively at +850 and +1200 km/s. We have attributed these two lines to Ca II, because of the similarity of the line profiles and because we were unable to find a satisfactory identification for the two lines The less displaced line is more intense than the other.

Heavier elements. Some lines of ionized iron (4232 and 4303 from M.27; 4549 and 4583 from M.38 and 4629 from M.37) and of ionized chromium (4558 and 4588 from M.44) do show structures. The structures appear as emission wings accompanying the absorption line, with V>R, the red emission being sometimes very weak or invisible. The emission separation is of the order of 200 km/s. All lines in which this structure is visible are the strongest of their respective multiplets, so that it cannot be discarded that with better resolution, all ionized metal lines could show such a structure.

The phenomenon of the variable emission structures has been described in the past by several observers, first by Doazan (1960) and then, among others, by Hubert-Delplace & Hubert (1979).

We have also looked into the variations of equivalent width of the few lines which happen to be present on two spectra taken on different dates. We find variations of up to a factor two in the equivalent widths, but the material is too scanty to permit to draw more conclusions. We had already found variations of such an amount in the Paschen lines of this star and moreover variations of such an amount are also present in other B[e] objects.