next previous
Up: The visual line

3. Description and analysis of the emission line spectrum

3.1. The Balmer lines

The lines of Balmer series are visible as far as H31. The profiles of the first of them consist of two components: a central narrow component with FWHM equal to about 95 kms-1 and a broad component, indicating stellar wind and having FWZI (full width of zero intensity) of about 2000 kms-1. Only the broad component of Htex2html_wrap_inline1418 is confidently seen. The profiles of the rest of the Balmer members consist of one component, i.e. they are ordinary nebular lines. Some of the Balmer profiles have asimmetry.

The line H8 is badly blending with the line HeI 3889 and is inappropriate for investigation.

We obtained radial velocities, measuring only the narrow components (Table 1 (click here)). The radial velocities of all lines were the same within the range of the error, i.e. we didn't ascertain the existence of Balmer progression. The velocities of both spectra differ appreciably from the values of the radial velocity curve of the hydrogen lines at our phases in Fig. 3 (click here). These differences exceed the level of our error and that is why we consider they are due to a change of dynamics of the nebula.

The line fluxes (Table 2 (click here)) obtained at the present time are compared with the fluxes obtained in 1986 (Tomov 1993a). Each of the latter of them is an arithmetical mean of three spectra taken at the moments JD 2446660.53, 2446752.18 and 2446755.27, which are at phases 0.838, 0.950 and 0.954. We compare spectra taken approximately at the same phases since the intensity of most of the emission lines vary during the orbital cycle. The line fluxes in the work of Tomov (1993a) have been systematically reduced by a factor of about 1.5. The data used in the present work have been improved and then deredened in the way described in the last section. The data from the table show that the fluxes of the Balmer lines have decreased on average by a factor of 3.2.

3.2. The helium lines

The FWHM of the singlets is 45-50 kms-1 and those of the triplets - about 60 kms-1. Some of the lines of both groups have asymmetry in their profiles. The two groups have the same velocity (Table 1 (click here)) practically equal to the value at our phases of the radial velocity curve of the helium lines in Fig. 3 (click here). The line fluxes of the singlets have decreased on average by a factor of 2.1 and those ones of the triplets - by a factor of 1.9 (Table 2 (click here)).

3.3. The lines of elements of high degree of ionization

Lines of different highly ionized elements namely HeII, NIII, OIII and CIII are present in the blue region of the spectrum of AG Peg.

Like the Balmer lines the line of HeII 4686 consist of two components: a central nebular component with FWHM equal to 80 kms-1 and a broad component, indicating stellar wind. The profile of the broad component, observed at the moment JD 2446755.27 when the line was more intensive is compared with the one, observed at JD 2449969.38 in Fig. 4 (click here). The error of the local continuum in the first case is equal to tex2html_wrap_inline18445% as Kodak IIaO emulsion having a lower noise was used. This error in the second case is about tex2html_wrap_inline184410%. The observed spectrum in the region of the HeII 4686 broad component was corrected through removing the strongest absorption lines of the red giant spectrum. Then the individual profiles were analysed by fitting with a sum of two gaussian components. The FWHM of the broad component in the first case is equal to 1210 tex2html_wrap_inline1844 50 kms-1 and in the second one tex2html_wrap_inline1852 kms-1. Taking into account the observational errors we have no reason to suppose that the line width has changed. This procedure allows us to obtain the equivalent width and the line flux with errors equal to about 30% in the first case, and about 50% in the second one. These errors are due first of all to the errors of the local continuum. If we take a velocity of the wind at a distance 2tex2html_wrap_inline1856 from the center of the line this velocity is obtained to be 1030 kms-1 and 1110 kms-1 in the two cases. We are inclined to adopt the first one of these values made up to 1000 kms-1, as it is based on emulsion with better quality.

Figure 4: The broad component of the HeII 4686 line, observed at the moments JD 2446755.27 (more intensive one) and JD 2449969.38. The level of the local continuum is marked with a dashed line. The radial velocities are in units kms-1

The velocity of the narrow component of HeII 4686 (Table 1 (click here)) is in agreement with the radial velocity curve of this line in Fig. 3 (click here). In a period of nine years its flux has decreased by a factor of 3.0 (Table 2 (click here)).

We considered only the lines HeII 4200 and HeII 4542 of the Pickering series as the rest of them are badly blended with Balmer lines. The profiles of the lines HeII 4200 and HeII 4542 are asymmetric and their FWHM are equal to 55 kms-1 and 48 kms-1 respectively. The FWHM of the NIII lines are about 45-50 kms-1.

Since there was a great difference between the velocities of the OIII lines of the two spectrograms, their values were excluded from further consideration. The lines of the rest of the elements have practically the same velocities whose arithmetical mean is shown in Table 1 (click here). However, they differ from the values of the radial velocity curve of the ionized elements at the some phases in Fig. 3 (click here).

The fluxes of the Pickering lines of HeII have decreased on average by a factor of 1.1 and those of NIII - by a factor of 1.7.

3.4. The forbidden lines

Only the lines of the elements [OIII] and [NeIII] are included in this group. In contrast to the middle of eighties (Tomov & Tomova 1992), now the lines of [OIII] tex2html_wrap_inline1876 4959 and 5007 have disappeared entirely and [OIII] 4363 has considerably weakened - its flux has been reduced by a factor of 4.9 (Table 2 (click here)). Its FWHM is equal to 73 kms-1.

There was a great difference between the [NeIII] 3868 line radial velocities in the two spectrograms and that is why they were excluded from further consideration. Like HI, HeII, NIII and CIV lines, the velocity of the [OIII] 4363 line (Table 1 (click here)) differs considerably from the one obtained previously at the same phases (Fig. 3 (click here)).

3.5. The lines of ionized metals

Besides all groups of lines examined up to now, emission lines of singly ionized and neutral metals as FeII, TiII, SiII, CaII, CrII, MgII, VII, NiII, AlII, FeI, SiI, CaI and [FeII] present in the visual spectrum of AG Peg. Only the FeII and TiII lines are intensive enough to make investigation possible. Their FWHM is equal to 45 kms-1. The K and H lines of CaII are inappropriate for investigation because of blending with the lines NIII 3934 and HeII 3968, [NeIII] 3968 respectively.

The radial velocity of the FeII and TiII lines (Table 1 (click here)) was obtained to be equal to the velocity of the lines of the singly ionized metals (Tomov & Tomova 1992) which doesn't vary with the orbital phase and is equal to -18.1 kms-1. These facts are probably due to the following reason: in contrast to the rest of nebular lines the lines of the singly ionized metals are excited in a part of the giant chromosphere, which faces the hot companion (Boyarchuk 1966; Kenyon et al. 1993) and is located very close to the mass center of the system, whose velocity is derived in the interval 16-18 kms-1 (Hutchings et al. 1975; Tomov & Tomova 1992; Kenyon et al. 1993).

The line fluxes of FeII have decreased on average by a factor of 1.7 (Table 2 (click here)). We assume the part of the giant atmosphere which is ionized by the hot companion has decreased approximately in the same ratio.

next previous
Up: The visual line

Copyright by the European Southern Observatory (ESO)