3 Results

The Ritter observation obtained in 1994 revealed a weak emission in the H$\alpha$ line similar to that detected by Coté & van Kerkwijk (1993) in 1990. The line profile is shown in Fig. 1a along with a theoretical one, calculated with the code SYNSPEC (Hubeny et al. 1995) for the atmospheric parameters from Heasley et al. (1982), the rotational velocity from Brown & Verschueren (1997), and the solar chemical composition. The equivalent width (EW) of the emission part (after the theoretical profile subtraction) is 0.24 Å, while the peak separation is 350 kms^-1. The emission component in the SAAO 1998 spectrum is hardly recognizable.

The presence of an emission component in the He I 5876 Å line is dubious in 1994 and is not seen in 1998 (see Fig. 1b). Its absorption component is deeper than the theoretical one for the described atmospheric parameters, as was noticed by Heasley et al. (1982) for other early B-type stars (including $\delta$ Sco). This difference might also be due to the presence of a shell component or to an effect of microturbulence (see Leone & Lanzafame 1998). In fact, the He I line EW in the 1994 and 1998 spectra is almost the same as the value reported by Heasley et al. (1982), $0.73\pm0.01$ Å vs. 0.69 Å. Thus, there is no evidence for a growth in the amount of circumstellar matter around $\delta$ Sco during a significant portion of the last binary cycle.

Figure 1: The H$\alpha$ and He I 5876 Å lines in the 1994 Ritter and the 1998 SAAO spectra. The observational data are shown by solid lines, while the theoretical profiles for $T_{\rm eff}=27\,500$ K, log g = 4.0, and $v\sin i = 150$ kms-1 are shown by dashed lines. The intensity scale is normalized to the level of the underlying continuum, while the heliocentric RV are given in kms-1.

All our spectra of $\delta$ Sco obtained during the 2000/1 campaign contain emission lines. All emission lines detected in our high-resolution spectra have double-peaked profiles (Fig. 2) with a mean separation of the blue and red peak of 250 kms^-1 (He I and Si II lines) and 150 kms^-1(H$\alpha$). Such a profile shape suggests that the lines are formed in a circumstellar disk, typical for classical Be stars. The peak separation difference can be explained by the fact that the He and Si lines are formed closer to the star, where the disk rotates faster.

The normalized H$\alpha$ profiles obtained at CAO are systematically weaker and narrower than those obtained at Ritter. This might be due to the narrower spectral region observed at CAO, which contains a small area useful for continuum determination, and a stronger contamination of the CAO data by the telluric lines. However, this does not affect the RVs and has no impact on our conclusions. Parameters of the H$\alpha$ line profiles are presented in Table 3.

Our ESO spectrum covers the whole optical range and contains information about both photospheric and circumstellar features. The double-peaked emission is present in all hydrogen lines up to H₈ and in most of the He I lines (see Fig. 3). Other species detected (e.g., O II, He II, N  II, N  III, Si  IV) show typical photospheric absorption profiles, which are consistent with the mentioned fundamental parameters. No clear signs of the secondary were found in the spectrum. This might imply that the components' temperature difference is not large enough to produce additional detectable spectral features. However, this suggestion needs to be verified by higher-resolution future observations.

Our results obtained for the H$\alpha$, He I 5876 Å, and Si  II 6347 and 6371 Å lines showed that their intensities were increasing until September 12 and decreasing after that time. The H$\alpha$ line FWHM displays a similar behaviour. The EWs of all 4 mentioned lines showed anti-correlation with the visual brightness of the system. Their RVs, measured by matching the direct and reversed line profiles, were shifting to the blue when the system was fading and vice versa (Figs. 2a,b and 4). At the same time, the He  I 6678 Å line exhibited a small variability of the peak strengths between August 9 and 13, but was not detected at all on August 20 (Fig. 2c). No sign of the Fe II 5317 Å line was detected in the 2000 Ritter spectra. It appeared first in the ESO spectrum and is seen in the 2001 Ritter spectra with a larger strength (Fig. 2d). Other Fe II lines, most of which have weak emission components, are also seen in the ESO spectrum. The Na I D lines are purely interstellar in origin and exhibit no change in strengths or positions.

 

 

Table 3: Characteristics of the H$\alpha$ line in the spectra of $\delta$ Sco.

Date	HJD	Obs.	H$\alpha$
2000	2451000+		EW	$I_{\rm b}$	$I_{\rm r}$	$I_{\rm c}$	$RV_{\rm mean}$	$\Delta$ $V_{\rm peak}$	FWHM
			Å				kms-1	kms-1	kms-1
07/19	745.264	Ski	2.92	1.41	1.40
07/20	746.302	Ski	3.15	1.41	1.41
07/25	751.313	BOL	3.00	1.35
07/28	754.286	CAO	2.59	1.46	1.41	1.30	-17.0	150	272
07/29	755.320	BOL	3.40	1.42
08/04	761.575	Rit	3.37	1.55	1.53	1.39	-19.5	152	295
08/07	764.246	CAO	2.64	1.46	1.45	1.33	-18.5	151	283
08/07	764.568	Rit	3.28	1.55	1.52	1.37	-21.0	155	293
08/09	766.258	CAO	2.76	1.51	1.47	1.33	-19.5	154	277
08/10	767.592	Rit	3.59	1.57	1.55	1.39	-19.5	154	298
08/12	769.269	CAO	2.99	1.52	1.49	1.37	-25.0	152	278
08/20	777.225	CAO	3.03	1.50	1.53	1.37	-28.0	154	278
08/20	777.546	Rit	4.18	1.61	1.64	1.46	-28.5	151	305
08/24	781.577	Rit	4.35	1.64	1.62	1.47	-30.0	153	309
08/30	787.566	Rit:	4.62	1.65	1.60	1.50	-35.0
08/31	788.548	Rit	4.68	1.64	1.63	1.54	-38.5	165	333
09/04	792.538	Rit	5.16	1.73	1.67	1.57	-48.5	160	321
09/06	794.541	Rit	5.01	1.72	1.71	1.57	-48.0	160	321
09/12	800.527	Rit	5.26	1.77	1.69	1.57	-54.0	160	320
09/13	801.525	Rit	4.98	1.71	1.68	1.56	-49.0	155	320
09/16	804.532	Rit	4.99	1.69	1.68	1.57	-46.0	155	323
09/17	805.524	Rit	4.82	1.67	1.68	1.56	-40.5	155	321
09/18	806.508	Rit	4.90	1.68	1.68	1.55	-38.5	150	324
09/19	807.506	Rit	4.60	1.71	1.61	1.55	-41.5	152	316
09/21	809.506	Rit	4.82	1.73	1.66	1.60	-33.5	150	308
09/26	814.494	Rit	4.06	1.65	1.59	1.49	-30.5	145	292
09/27	815.502	Rit	3.89	1.63	1.54	1.49	-31.0	150	292
09/28	816.503	Rit	3.90	1.64	1.59	1.48	-26.0	145	283
10/03	821.497	Rit	3.32	1.62	1.50	1.44	-24.5	145	271
10/23	841.493	ESO	2.75	1.48	1.49	1.37	-18.0	150	275
02/10$^{\rm a}$	951.948	Rit	3.80	1.59	1.59	1.47	-8.5	135	282
02/20$^{\rm a}$	961.904	Rit	4.00	1.60	1.61	1.49	-8.0	135	282
02/27$^{\rm a}$	968.978	Rit	4.00	1.61	1.62	1.49	-7.5	135	282
03/10$^{\rm a}$	978.944	Rit	4.34	1.60	1.65	1.51	-7.5	130	300

The intensities of the blue and red emission peaks normalized to the underlying continuum are listed in Cols. 5 and 6, respectively; the intensity of the central depression is given in Col. 7; the RV of the overall profile is given in Col. 8; the peak separation and full width at half-maximum are given in Cols. 9 and 10, respectively.
The low-resolution Skinakas data are not suitable for the RVmeasurements. The BOL data do not resolve the double-peaked H$\alpha$structure.
The Ritter data marked with a colon have the lowest S/N ratio.
$^{\rm a}$ The spectrum was obtained in 2001.

Figure 2: The emission line profiles of $\delta$ Sco. a) The H$\alpha$ profiles with extreme RVs obtained at Ritter (solid lines). The CAO H$\alpha$ profile obtained on 2000 July 28 is shown by a dash-dotted line. b) The He I 5876 Å line profiles obtained at Ritter. c) The He I 6678 Å line profiles obtained at CAO. d) Time evolution of the Fe II 5317 Å line. The intensity and heliocentric RV are given in the same units as in Fig. 1.

Figure 3: Portions of the ESO spectrum of $\delta$ Sco obtained on 2000 October 23. The observational data are shown by a solid line, while the theoretical profiles for $T_{\rm eff}=27\,000$ K, log g = 4.0, and $v\sin i = 150$ kms-1 are shown a dashed line. The intensity is normalized to the underlying continuum level, while the heliocentric wavelengths are given in Å.