3 Results

3.1 Photometric characteristics

The available photometric data show that the optical brightness of AS 314 varied within 0.25 mag during the last 40 years. The HIPPARCOS photometry (67 observations in 17 dates between March 1990 and March 1993) from Table 2 along with the results from Table 1 give the following range of the variations: 9.97 $\le V \le$ 9.73 mag. The V-band light curve is shown in Fig. 1. The amount of the data obtained so far is insufficient for a periodicity analysis. Note that our BV data are close to those published before, while our U-B is nearly 0.1 mag bluer. The latter may be due in part to the U-B difference of the object and its comparison star (HR6989, U-B=-0.19 mag). However, in the sky they are located close enough to each other (0.5 $\hbox{$^\circ$ }$) and were usually observed with an air mass difference not exceeding 0.1. Besides that our instrumental photometric system is very close to the Johnson one (Bergner et al. [1988]). Thus, we do not expect systematic errors in the observed U-B to be larger than 0.02-0.03 mag. Perhaps, the difference is mainly due to the deviation of the U-band transmission curve in our and other authors' filter sets.

   

Table 2: HIPPARCOS photometry of AS314

JD	V	n	JD	V	n
2447963.83	9.85 $\pm$ 0.01	4	2448492.08	9.87 $\pm$ 0.01	3
2447990.58	9.91 $\pm$ 0.02	4	2448504.85	9.82 $\pm$ 0.02	2
2448143.67	9.75 $\pm$ 0.02	2	2448532.82	9.90 $\pm$ 0.02	9
2448166.70	9.82 $\pm$ 0.01	2	2448533.29	9.92 $\pm$ 0.01	2
2448191.02	9.94 $\pm$ 0.01	2	2448537.35	9.78 $\pm$ 0.02	4
2448311.40	9.75 $\pm$ 0.01	7	2448703.71	9.73 $\pm$ 0.02	4
2448329.24	9.86 $\pm$ 0.01	3	2448749.75	9.81 $\pm$ 0.02	4
2448357.58	9.81 $\pm$ 0.02	4	2449060.63	9.78 $\pm$ 0.01	4
2448491.91	9.90 $\pm$ 0.02	4

The data are averaged within the observing date and translated from the HIPPARCOS photometric system to the Johnson V-band system using formulae from Harmanec ([1998]). The mean observing dates are given in Cols. 1 and 4, the number of observations during each date are given in Cols. 3 and 6.

Figure 2: SEDs of AS314 and the LBV HRCar (open triangles) in the range from 0.3 to 100 $\mu$m. Our ground-based data for AS314 are shown by filled circles, those of MSX by open circles, while those of IRAS by filled squares

The spectral energy distribution (SED) of AS314 in the region $0.3-1.2\;\mu$m constructed from our averaged optical and near-IR data is consistent with what is expected for a B9-type supergiant (e.g., Wegner [1994]) affected by the interstellar reddening A_V = 2.8 mag assuming the total-to-selective extinction ratio R=3.1. The assumption that the star is a dwarf would result in a much earlier spectral type (B2-B3) which is not consistent with the earlier spectroscopic determinations described above (see Sect. 1) and our own results (see Sect. 3.2). There is a small excess radiation in the HKL-bands ( E_H-K = 0.1 mag, E_K-L = 0.2 mag) which can be explained by free-free emission from the stellar wind. However, the excess radiation is much larger at longer wavelengths ( $\lambda \ge 10\; \mu$m) indicating the presence of cool circumstellar dust in the system. In addition to the IRAS data quoted above (see Sect. 1), we idenitified AS314 with the source 019.2440-03.6636 from the recently released mid-IR survey carried out by the MSX satellite in 1996/7 (Egan et al. [1999]). The star was detected in two bands centered at 8.28 and 21.34 $\mu$m with the fluxes 0.12 and 4.31 Jy, respectively. The MSX fluxes are in good agreement with those of IRAS indicating no noticeable variations of the dusty envelope radiation on a time scale between the surveys (13 years). Furthermore, the MSX result provides more confidence in identification of the optical and IR objects because of a high accuracy of the MSX positional measurements ( $\sim 2-3\hbox{$^{\prime\prime}$ }$). Despite both the IRAS and MSX positions are very close to the optical one (4.8 $\hbox{$^{\prime\prime}$ }$ and 4.1 $\hbox{$^{\prime\prime}$ }$ offset respectively), the IRAS positional error box is much larger $(35\hbox{$^{\prime\prime}$ }\times 6 \hbox{$^{\prime\prime}$ }$) than the MSX one. Dereddened SEDs of AS314 and of a galactic LBV HRCar (e.g. de Winter et al. [1992]) shown in Fig. 2 are very close to each other. The same kind of far-IR excesses have also planetary nebulae, but below we will show that AS314 is not related to this particular type of objects.

   
3.2 Spectrum

The spectrum of AS314 is not dominated with emission lines. It contains a rather strong and narrow $\rm H\alpha$ (equivalent width, EW, $\sim 14$ Å) line showing a PCyg-type profile. The line appeared slightly variable in our spectra which is in part due to different dispersions. However, a physical reason for the variations, such as variable mass loss rate can not be excluded as well. The H$\beta$ and H$\gamma$ lines also show noticeable emission components, while in H$\delta$ only weak signs of emission are seen.

Identification of the spectral lines was done on the basis of a catalog of Coluzzi (1993) and a compilation of Johansson (1978) for Fe II lines. In total 351 lines were identified in the spectral range between 4002 and 7743 Å. Characteristics of the Balmer lines are presented in Table 3, averaged radial velocities of the lines of different species in Table 4, while the overall information including laboratory wavelengths, identifications, peak intensities, equivalent widths, and radial velocities of all identified lines (except for those of the Balmer lines) are given in Table 5. The measured line characteristics were averaged for those lines detected in different spectra. The most informative parts of the 1998 spectrum are shown in Figs. 3 and 4.

Many Fe II lines of low excitation with PCyg-type profiles are present in the spectrum. At the same time, nearly 30 Fe II lines with excitation energy higher than 10 eV are seen in absorption. The only forbidden emission lines seen in the spectra ([O I] 5577 and 6300 Å) are presumably telluric.

   

Table 3: Characteristics of the Balmer lines. $I_{\rm max}$ and $V_{\rm max}$ are the intensity and heliocentric radial velocity of the emission peak, respectively; $I_{\rm abs}$ and $V_{\rm abs}$are the minimum intensity of the absorption component and its radial velocity; $V_{\rm sec}$ is the radial velocity of the secondary emission peak. All intensities are normalized to the continuum

Line	$I_{\rm max}$	$I_{\rm abs}$	$V_{\rm max}$	$V_{\rm abs}$	$V_{\rm sec}$
H $\alpha^{\rm a}$	11.84	0.18	-21	-105	-150
H $\alpha^{\rm b}$	7.88	0.16	-23	-120	-175
H $\alpha^{\rm c}$	10.48	0.20	-28	-105	-150
H$\beta$	2.66	0.09	-24	-105	-170
H$\gamma$	1.17	0.05	-28	-100
H$\delta$	-	0.01	-	- 96

$^{\rm a}$ - July 1997, $^{\rm b}$ - June 1998, $^{\rm c}$ - July 1999.

   

Table 4: Radial velocities for groups of lines in the spectrum of AS314. The averaged velocity in kms^-1 and the number of lines (N) used to derive these values are listed for each spectrum. In the last column $\chi$is the excitation potentials in eV, while r is the residual intensity with respect to the continuum

Element	1997		1998		1999		Comment
	R.V.	N	R.V.	N	R.V.	N
abs He I, C II	$-41\,\pm\, 3$	4	$-60\,\pm\, 2$	8	$-43\,\pm\, 1$	3
abs S II, Ne I	$-46\,\pm\, 4$	8	$-60\,\pm\, 2$	12	$-44\,\pm\,2$	15
abs Fe II			$-61\,\pm\,2$	29	$-42\,\pm\,3$	16	$\chi \ge$ 10
abs Si II, Mg II	$-54\,\pm\,3$	5	$-57\,\pm\,3$	8	$-58\,\pm\,1$	2
em Fe II	$-57\,\pm\,2$	18	$-57\,\pm\,2$	35	$-48\,\pm\,2$	18	$\chi \le 4$, $r \ge 0.85$
abs Fe II	-97:	8	$-98\,\pm\,3$	25	$-85\,\pm\,3$	16	$\chi \le 4$, $r \ge 0.85$
em Fe II			$-28\,\pm\,2$	4	-28	1	$\chi \le 4$, $r \le 0.55$
abs Fe II			$-106\,\pm\,2$	6	$-91\,\pm\,1$	2	$\chi \le 4$, $r \le 0.55$
abs Na I	$-68\,\pm\,1$	2	-88	2	$-66\,\pm\,1$	2
I.S. Na I	-6	2	-6	2	$-3\,\pm\,1$	2
I.S. DIB	-8:	5	$-8\,\pm\,2$	12	$-4\,\pm\,3$	14

Absorption lines of He I, C II, N I, N II, Ne I, Na I, Mg I, Mg II, Si II, Si III, S II, Ti II, Cr II, and Fe II were found in the spectrum. Comparison of the strengths of some of them with those of other supergiants gives a spectral type of $\rm A0.2 \,\pm\, 0.3$, which is consistent with the above photometric estimate. The classification criteria and the basic system of equivalent widths were taken from Kopylov ([1960a], [1960b]) and Chentsov & Luud ([1989]) for the blue and red spectral region, respectively. The absortion lines, which are usually used for the quantitative classification of supergiants (e.g., He I 4471, Mg II 4481 Å), in the spectrum of AS314 look the same as those in those of normal supergiants. The resulting spectral type is the average of the estimates derived from using different criteria (EW ratios). For example, the EW ratio of the He I 4471 and Mg II 4481 Å lines gives the spectral type B9.8, while that of Si II 4128/30 and He I 4471 Å gives A0.5.

The absolute visual magnitude estimate, $M_v = -7.3\,\pm\,0.5$ mag, is less reliable, since the H$\delta$ and H$\gamma$ profiles are distorted by the stellar wind emission, while the O I 7771/5 Å triplet was beyond our spectral range. A direct comparison of the spectra of AS314 and HD223960 (A0.1, Bartaya et al. [1994]) suggests that the temperature of the former is not higher and the luminosity is not lower than those of the latter.

Since only a few lines of O I and no O II lines have been found in our spectra of AS314, one can suggest a N/O overabundance. This is also found for the LBV's AGCar, HRCar (Hutsémekers & van Drom [1991]), and the high-luminosity supergiant/LBV candidate MWC314 (Miroshnichenko et al. [1998]).

Figure 3: The blue part of the 1998 spectrum of AS314

Figure 4: The yellow-red part of the 1998 (a,c,d) and 1999 (b) spectra of AS314. a) all noticeable unmarked lines seen are those of Fe II; b) the absorption components of the sodium D lines are marked by solid lines, telluric water vapour lines are marked by dots; d) all the lines with PCyg-type profiles are those of Fe II, while the pure absorption lines are telluric

The sodium D_1,2 lines consist of three absorption components. The strength of the interstellar components is consistent with the high reddening evident from the photometry. The other two components have radial velocities of about -90 and -121 kms^-1 (see Fig. 4b) and are most likely of circumstellar origin. A significant number of diffuse interstellar bands (DIB) is found in the spectrum of AS314. The strength of some DIBs in stellar spectra display a good correlation with the star's reddening (e.g. Herbig [1993]). We use this property to obtain an independent estimate of the latter. The equivalent widths of the DIBs at 5780 and 5797 Å, 0.39 and 0.15 Å, respectively, correspond to E_B-V = 0.9 (Herbig [1993]), which is in good agreement with the dereddened optical color-indices. The interstellar features (DIBs and Na I D lines) have radial velocities of about -6 kms^-1, which is most likely due to absorption in a nearby interstellar cloud Lynds 410 (Lynds [1962]).

The mean radial velocities of the He I, C II, S II, and Ne I pure absorption lines in the 1997 and 1999 spectra differ significantly ( $+16\,\pm\,3$ kms^-1) from those in the 1998 spectrum. The difference exceeds the measurement errors and points to the existence of real variations in the radial velocity of the star. The same effect was found for the high-excitation Fe II lines and circumstellar components of the Na I lines (see Table 4).

Another peculiar property detected is that all the radial velocities are negative, being rather unusual for this direction in the Galaxy. In general, stars in the vicinity of AS314 show positive radial velocities as it is expected from the galactic rotation curve (Dubath et al. [1988]). For example, for the two LBV candidates, HD168607 and HD168625, that are both located within 5 $\hbox{$^\circ$ }$ from the object, the radial velocities are about +10 kms^-1 (Chentsov & Luud [1989]). However, there is a discrepancy for AS314 ranging from 70 kms^-1assuming a distance D = 1 kpc to 130 kms^-1 for D= 10 kpc. These estimates are based on the averaged radial velocity of AS314 of $-50\pm10$ kms^-1 (see Table 4) and on calculations using formulae from Dubath et al. ([1988]) for the direction towards the star. Such a difference indicates that the star has a large peculiar velocity.

Some emission line profiles were found slightly different in our spectra. This could be due to different dispersions and/or possible variations of the stellar wind (see Fig. 5). The observed pure absorption line profile shape variations are mainly instrumental, as the line equivalent widths are essentially the same in all three spectra.