Astron. Astrophys. 322, 511-522 (1997)

5. The optical spectrum of IN Comae

5.1. Activity indicators

Fig. 4a-h displays eight regions of the optical spectrum of IN Comae relevant for magnetic activity:

Fig. 4a and b. Active-chromosphere indicators of IN Comae at optical wavelengths (thick lines); panel a Ca II H&K; b He I D3 and Na I -D doublet (the right two tick marks indicate the interstellar NaD absorption features); c Ca II 8498; d Ca II 8542; e Ca II 8662; f H ; g the 6430-Å region used for Doppler imaging (the tick marks identify the three mapping lines); and h Li I 6707. The overplotted thin lines are shifted and broadened spectra of the G5III-IV M-K standard  CrB, while the thin lines plotted just below the respective IN-Comae spectra represent the residual spectra of IN Comae after subtraction of the  CrB spectrum.

Fig. 4a shows the broad Ca II H and K emission lines from a spectrum taken in 1992 (Strassmeier 1994a). Their line profile shapes are, at first glance, contrary to the other chromospherically influenced lines in that they have no central (H₃ and K₃) absorption dip but, instead, a central emission peak (compare with, e.g., Ca II 8498 in Fig. 4c).

Fig. 4b is a close up of the He I D₃ and the neutral sodium D₁ and D₂ dublett. Notice the emission "shoulders" of these lines as compared to the broadened reference star. The two upper tick marks indicate the sharp interstellar Na D absorption features in the IN Comae spectrum that, however, are not obvious in the reference star spectrum.

Figs. 4c-d show the emission in the three infrared triplet lines of Ca II. Again, we see similar emission shoulders as for He I D₃ and Na D₁ D₂ but once the photospheric contribution is removed (by subtraction of a broadened and shifted spectrum of  CrB) the residual emission line profiles appear similar to the Ca II H and K profiles. Using the method described in Linsky et al. (1979) and the measured index of 0.955 0.02 mag, we obtain absolute emission-line fluxes in a 1-Å band around the line center of 4.62 10⁶, 3.84 10⁶, and 3.85 10⁶ erg cm^-2 s^-1 for the lines at 8498 Å , 8542 Å , and 8662 Å , respectively.

Fig. 4f shows the broad and double-peaked H emission line profile with a full width at the continuum level of the residual profile of 18 Å ( 410 km s^-1), that likely excludes a chromospheric origin. In fact, the width of the central absorption feature is significantly broader than the rotationally broadened H absorption line of the reference star, resulting in the triple -peaked residual emission. We suspect that part of the H profile of IN Comae is due to a strong stellar wind or is from the planetary nebula itself. If so, part of the emission "shoulders" in the other chromospheric lines should also have a similar cause.

5.2. The distance to IN Comae

The existence of Ca II H and K emission allows us to obtain a new and independent determination of the distance to IN Comae by using the Wilson-Bappu relation in the updated form of Lutz (1970). This calibration relates the FWHM of the K line in km s^-1 to the absolute magnitude of the star,

The measured FWHM must be corrected for instrumental and rotational broadening before being inserted in Eq. (2) and here we use the form

as recommended by Lutz (1970). The emission-line width measured from our K-line spectrum in Fig. 4a is 1.79 0.10 Å , which comes out to 72 km s^-1 for the corrected FWHM and to and, finally, to a distance of 710 parsecs adopting an unspotted magnitude of mag (Strassmeier et al. 1997) and neglectable interstellar reddening (Feltz 1972). This distance is in good agreement with Kaler's (1983) result of 620 200 pc and still consistent with the 400 200 pc value computed by Longmore & Tritton (1980) despite that the absolute magnitude is too bright by a full magnitude for a normal G5III-IV star.

One could argue that the G5III-IV component is a foreground star and not physically related to the subdwarf nor to the planetary nebula. But Feibelman & Kaler (1983) showed that the UV spectrum of IN Comae is a composite of a sdO and a G5 component. Also, we see sharp "interstellar" absorption lines in the NaD doublet (Fig. 4b) that would be hard to explain at the high galactic latitude of if IN Comae were a foreground star. Further evidence comes from the broad H emission line (see previous section). So, it seems that there is good evidence that IN Comae is indeed within the planetary nebula LoTr-5.

5.3. The spectral classification

We used several of our 6430-Å region spectra of IN Comae to compare them with a series of similar spectra of M-K standard stars of well-known spectral classification (Strassmeier & Fekel 1990). The standard star spectra were first rotationally broadened to match the value for IN Comae ( km s^-1) and then shifted by the appropriate radial velocity. The color is used as an additional constraint.

Fig. 4g compares a 6430-Å spectrum of IN Comae with  CrB, the latter listed by Keenan & McNeil (1989) as a M-K G5III-IV standard star. We note, however, that  CrB is a mildly active star with a relatively long rotation period of 59 days seen in chromospheric data (Fernie 1991) and in photospheric data (Choi et al. 1995). Further template stars used in this procedure were HR 4255 (G4III, =0.83 mag), HR 6140 (G4III, =0.82 mag), o  UMa (G5III, =0.84 mag), HR 5161 (G6III, =0.85 mag), HR 7071 (G5III, =0.82 mag),  Aql (G8IV, =0.85 mag), and  Eri (K0.5V, =0.81 mag) but resulted in less satisfactory fits than with  CrB as the template. From the 6430-Å wavelength-region templates we can exclude a plain class-III giant classification for IN Comae. Note that our (spectral) classification is identical to that obtained by Malasan et al. (1989) from broad-band color indices.

5.4. Radial velocities

Our new radial velocities were obtained from cross-correlating the entire spectrum of IN Comae at 6430-Å with the radial-velocity standard stars 16 Vir (K0.5III, km s^-1) and/or  Aur (K0III, km s^-1). The average velocities are listed in Table 1. The standard error of an observation of unit weight is slightly less than 5 km s^-1. This is still anomalously high for red-wavelength spectra at the given resolution ( = 35,000) but is due to the broad and variable line-profile shape of IN Comae.

Our velocities range from -7.2 to -11.3 km s^-1, with an unweighted average of km s^-1 and fit well into the range of the CORAVEL velocities of Jasniewicz et al. (1994) from 1988 to 1991 (-2.2 to -20.4 km s^-1), but do not show the same (much larger) range than the metallic-line velocities from the work of Malasan et al. (1989) from 1984 through 1989 (-8.1 to -60.0 km s^-1) as well as from Acker et al. (1985) from 1983 and 1984 ( to -101 km s^-1). We suspect that most of the discrepancy between the Acker et al. and some of the Malasan et al. data and the current CCD velocities is likely due to the fact that these authors observed mostly at blue wavelengths where immense line blending by the broad and variable line profiles of IN Comae occurs, probably resulting in spurious radial velocities. Although CORAVEL was also used to observe the blue wavelength regions (3600-5200 Å) its velocities do not suffer from blending by the same amount as the direct spectra and are therefore more precise.

The cross-correlation functions of our 14 spectra show a variable and flat-peaked profile with mostly two asymmetric maxima but sometimes even three maxima at changing positions. This "doubled" structure originally led Jasniewicz et al. (1987) to conclude that IN Comae is a double-lined spectrocopic binary with a 1.99-day period. However, if we measure the stronger peaks of the cross-correlation function separately, we can not confirm a periodic behavior. Therefore, we agree with the suggestion of Jasniewicz et al. (1994, 1996) and conclude that these deformations are indeed due to the complex influence of starspots rotating in and out of view.

© European Southern Observatory (ESO) 1997

Online publication: June 5, 1998

helpdesk.link@springer.de