Properties and nature of Be stars

J. Nemravová; P. Harmanec; J. Kubát; P. Koubský; L. Iliev; S. Yang; J. Ribeiro; M. Šlechta; L. Kotková; M. Wolf; P. Škoda

doi:10.1051/0004-6361/200913885

Free Access

Issue		A&A Volume 516, June-July 2010


Article Number		A80
Number of page(s)		10
Section		Stellar structure and evolution
DOI		https://doi.org/10.1051/0004-6361/200913885
Published online		19 July 2010

A&A 516, A80 (2010)

27. Orbital and recent long-term variations of the Pleiades Be star Pleione = BU Tauri

J. Nemravová¹ - P. Harmanec¹ - J. Kubát² - P. Koubský² - L. Iliev³ - S. Yang⁴ - J. Ribeiro⁵ - M. Slechta² - L. Kotková² - M. Wolf¹ - P. Skoda²

1 - Astronomical Institute of the Charles University, Faculty of Mathematics and Physics, V Holesovickách 2, 180 00 Praha 8, Czech Republic
2 - Astronomical Institute of the Academy of Sciences, 251 65 Ondrejov, Czech Republic
3 - Institute of Astronomy, Bulgarian Academy of Sciences, 1784, 72 Tsarigradsko Chaussee Blvd., Sofia, Bulgaria
4 - Physics & Astronomy Department, University of Victoria, PO Box 3055 STN CSC, Victoria, BC, V8W 3P6, Canada
5 - Observatório do Instituto Geográfico do Exército, R. Venezuela 29, 3 Esq. 1500-618, Lisboa, Portugal

Received 16 December 2009 / Accepted 19 February 2010

Abstract
Radial-velocity variations of the H $\alpha$ emission measured on the steep wings of the H $\alpha$ line, prewhitened for the long-time changes, vary periodically with a period of 218 $.\!\!^{\rm d}$ 025 $\pm$ 0 $.\!\!^{\rm d}$ 022, confirming the suspected binary nature of the bright Be star BU Tau, a member of the Pleiades cluster. The orbit seems to have a high eccentricity over 0.7, but we also briefly discuss the possibility that the true orbit is circular and that the eccentricity is spurious owing to the phase-dependent effects of the circumstellar matter. The projected angular separation of the spectroscopic orbit is large enough to allow the detection of the binary with large optical interferometers, provided the magnitude difference primary - secondary is not too large. Since our data cover the onset of a new shell phase up to development of a metallic shell spectrum, we also briefly discuss the recent long-term changes. We confirm the formation of a new envelope, coexisting with the previous one, at the onset of the new shell phase. We find that the full width at half maximum of the H $\alpha$ profile has been decreasing with time for both envelopes. In this connection, we briefly discuss Hirata's hypothesis of precessing gaseous disk and possible alternative scenarios of the observed long-term changes.

Key words: stars: early-type - binaries: close - stars: emission-line, Be - stars: individual: BU Tau

1 Introduction

Pleione (BU Tau, 28 Tau, HD 23862) is a well-known Be star and a member of the Pleiades cluster. It underwent several phase transitions between B, Be, and Be shell phases, accompanied by pronounced light variations; see, e.g. Gulliver (1977), Sharov & Lyuty (1976), Iliev et al. (1988), Sharov & Lyutyj (1992), Hirata & Kogure (1976), Hirata & Kogure (1977), Hirata (1995), Doazan et al. (1988), Iliev et al. (2007), and Tanaka et al. (2007).

There is a rather complicated history of attempts to study the radial-velocity (RV hereafter) variations of this star. Struve & Swings (1943) measured RVs on the photographic spectra taken in the years 1938-1943 and tentatively concluded that the RV of BU Tau varies with a possible period of 142 days or - less likely - 106 days. Merrill (1952) studied RVs from 1941 to 1951 and found clear long-term variations with some overlapping changes on a shorter time scale. Gulliver (1977) analyzed a large collection of digitized photographic spectra from 1938-1954 and from 1969-1975 and concluded that there are no significant RV changes. Ballereau et al. (1988) carried out an analysis of a homogeneous series of Haute Provence high-dispersion photographic spectra from 1978-1987 and once more concluded that the shell RVs vary with periods of 136.0 and 106.7 days. Katahira et al. (1996b,a) analyzed shell RVs from the two consecutive shell phases separated some 34 years, using published as well as new RVs and concluded that BU Tau is a spectroscopic binary with an orbital period of 218 $.\!\!^{\rm d}$ 0, semi-amplitude of 5.9 km s^-1, and a large orbital eccentricity of 0.60. However, Rivinius et al. (2006) - analyzing a series of electronic spectra - were unable to confirm the 218-d period and concluded that BU Tau is not a spectroscopic binary. Hirata (2007) analyzed a long series of polarimetric observations and presented a model of a slowly precessing disk to explain the long-term B - Be - Be shell phase transition. He argued that the disk precession is caused by the attractive force of the secondary in the 218-d binary. Harmanec (1982) compiled the majority of at that time available RVs of BU Tau and averaged them over about 100 days. This resulted in a smooth RV curve with a period of about 13 000 days (35.6 years), in phase with the recorded shell episodes. Harmanec (1982) speculated that BU Tau could be a long-periodic binary with shell phases occurring always at the same orbital phases. A more distant companion with an angular distance of 0 $.\!\!^{\prime\prime}$ 22 was indeed discovered from speckle interferometry by McAlister et al. (1989). Gies et al. (1990) studied a sequence of low-dispersion H $\alpha$ spectra of BU Tau taken with a sampling rate of 7 ms during a lunar occultation on 1987 March 6. They detected an asymmetry of the envelope in agreement with the observed long-term V/R changes. They speculated that the speckle-interferometric component could have an eccentric orbit and that the recurrent shell phases could be caused by its periastron passages. Luthardt & Menchenkova (1994) compiled RVs from the years 1938-1990 and confirmed a period of 12 450-12 860 days. They advocated an eccentric orbit and mass transfer resulting in a release of a new shell during periastron passages, but the gaps in their RV curve do not allow one to conclude that the orbit has a high eccentricity. Finally, using the technique of adaptive optics photometry and astrometry, Roberts et al. (2007) report discovery of a new companion to BU Tau at a separation of 4 $.\!\!^{\prime\prime}$ 66 with a spectral type M5. They also confirm a companion at 0 $.\!\!^{\prime\prime}$ 24 and discuss other suggested companions.

Table 1: Journal of new spectroscopic observations for BU Tau.

We succeeded in collecting a rich series of electronic spectra at several observatories, covering many cycles of the suspected 218-d period. The main goal of this study is, therefore, to resolve the issue of whether BU Tau is a spectroscopic binary. Katahira et al. (1996b,a) based their orbit on the RV measurements of shell lines that may be affected by possible asymmetries in the circumstellar matter. Moreover, their RV curve has a rather small amplitude and is based on a collection of heterogeneous data. It naturally shows a rather large scatter around the mean curve. The spectra at our disposal all cover the red spectral region near H $\alpha$ . They were taken over the time interval when the star had fairly strong H $\alpha$ emission. Therefore, our study is based on the RV measurements of the steep wings of the emission, which is a procedure that turned out to be successful for detecting the duplicity of several other Be stars (Harmanec et al. 2000; Koubský et al. 2000; Bozic et al. 1995; Miroshnichenko et al. 2002,2001).

Since very pronounced long-term spectral variations occurred over the time interval covered by our spectra, we also briefly describe these changes and discuss them, especially in relation to the model put forward by Hirata (2007).

2 Spectroscopic observations and their reductions

The red spectra at our disposal were obtained at five observatories and their overview is in Table 1. Details about the instruments and data reduction can be found in Appendix A where also Table A.1 with our RV measurements of the steep wings of the H $\alpha$ emission and of the H $\alpha$ absorption core is provided. The latter was measured for comparison with the RVs collected and analyzed by Katahira et al. (1996b), but only for those spectra where the absorption was clearly visible.

Over the interval of the more than 5000 days covered by our observations, the strength of the H $\alpha$ emission gradually declined and the shape of the H $\alpha$ profile underwent notable changes. Typical examples for several distinct stages are shown in Fig. 1, and the whole development of a new shell and metallic-shell phase is shown as a gray-scale representation of all usable H $\alpha$ profiles in Fig. 2.

$\begin{figure} \par\includegraphics[width=8cm,clip]{13885fg1.EPS} \end{figure}$	Figure 1: Comparison of H $\alpha$ profiles from different stages of the long-term changes.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=9cm]{13885fg2.EPS} \vspace*{-2mm} \end{figure}$

Figure 2:

A complete series of our H $\alpha$ profiles in a gray representation (only a few saturated or underexposed spectra were omitted). Abscissa shows the wavelength scale in Å, while the time on ordinate is shown in JD-2400000. Each horizontal strip represents an average of spectra secured within 200 days, and dark horizontal belts correspond to time intervals from which no spectra are available. At the bottom, there is a scale showing the correspondence between the flux level in the units of continuum and the gray scale.

Open with DEXTER

The fading of the H $\alpha$ emission was accompanied by a light decrease in the J, H, K, and L IR photometric bands that started around JD 2451500 (Taranova et al. 2008). This clearly corresponds to the gradual development of the hydrogen shell spectrum according to our spectra - see also Tanaka et al. (2007). Emission has been slowly fading from JD 2453000 until now, when its peak intensity represents only about 30% of the intensity seen in our earliest spectra. During the transition from a single-peaked to double-peaked emission, there is some time interval when the H $\alpha$ profile has a characteristic wine-bottle shape. The occasional presence of additional absorption components has been already noted by Iliev et al. (2007) or Tanaka et al. (2007) and is typical of all recorded shell phases of BU Tau. Besides the occasional presence of one or more additional absorptions, extended red emission wings are seen on some H $\alpha$ profiles. This makes the emission wings asymmetric and hard to measure for RV. We also note that all double-peaked profiles recorded prior to about JD 24540000 always have a red peak stronger than the violet one. Figure 2 shows that the width of the H $\alpha$ emission has remained more or less constant over the whole time interval covered by our observations. The same figure also shows that the metallic shell phase appeared rather abruptly.

Figure 3 shows the gradual development of the He I 6678 Å line profile. It illustrates well how shallow the line is at the beginning of a new shell phase. A very interesting finding is that, even for the B8 star, a presumably photospheric He I line can develop a shell component. The profile clearly gets stronger and narrower as the hydrogen shell line gets deeper. The additional absorption at the blue wing of the line seen on more recent spectra is the Fe II 6677.305 Å shell line.

3 Radial-velocity changes

Figures 4 and 5 are the time plots of the measured RVs vs. time for the H $\alpha$ emission wings and the absorption core. In the later, we also included all shell RVs used and published by Katahira et al. (1996b) and Rivinius et al. (2006). One can see systematic RV changes on at least two distinct time scales: a smooth change on a longer time scale and overlapping more rapid changes, especially the occasional steep decreases in RV.

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg3.EPS}} \vspace*{-2mm}\end{figure}$	Figure 3: Selected He I 6678 Å line profiles, ordered in time, with corresponding HJDs.
Open with DEXTER

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg4.EPS}} \vspace*{-2mm}\end{figure}$	Figure 4: Measured RVs of the H $\alpha$ emission wings plotted vs. time. Prewhitening for the long-term changes, carried out with the help of the program HEC13, is shown by a line. Empty squares show the alternate way to remove long-term RV changes via individual $\gamma$ velocities for subsets spanning no more than a year. See the text for details.
Open with DEXTER

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg5.EPS}} \vspace*{-2mm}\end{figure}$	Figure 5: Measured H $\alpha$ absorption core plotted vs. time. We also included the RV measurements of shell lines by Katahira et al. (1996b) and Rivinius et al. (2006) to this plot. Prewhitening for the long-term changes, carried out with the help of the program HEC13, is shown by a line. See the text for details.
Open with DEXTER

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg6.EPS}}\\ [5mm] \resizebox{9cm}{!}{\includegraphics{13885fg7.EPS}} \vspace*{-2mm}\end{figure}$	Figure 6: A comparison of two pairs of the H $\alpha$ line profiles from the locally recorded velocity extrema (HJDs of the profiles are indicated).
Open with DEXTER

Considering the uncertainties in accurate RV measurements combined with the fact that the full amplitude of the changes is low, it was deemed useful to convince readers that the RV changes are not only a result of changing asymmetry of the profiles, but they also represent a real shift of the whole line. To this end, we compare in Fig. 6 two pairs of the H $\alpha$ line profiles obtained near the local RV extrema. The upper pair comes from the beginning of a new shell phase and the bottom one from a more recent time when a weaker emission and deeper shell cores are present in the profiles (note a large difference in the flux scale of the two plots). The RV shift of the whole emission and absorption core is seen beyond any doubt. We, therefore, conclude that our RV measurements reflect real RV variations of BU Tau.

In accordance with Katahira et al. (1996b), we find that the evolution of the emission episode is accompanied by long-term RV changes that need to be removed prior to a search for possible periodic RV changes. To also make this step as objective as possible, we used two different procedures.

One is that we smoothed the long-term changes using the program HEC13, written by PH and based on a smoothing technique developed by Vondrák (1977,1969). For both emission and absorption RVs, optimal smoothings were obtained for the smoothing parameter $\varepsilon=10^{-16}$ fitted through 200-d normals. (Inspecting the time plots of RVs, we identified $\sim$ 200 days as a time scale on which more rapid changes were observed, and this was the reason for the choice of 200-d normals. We have verified, however, that the result of smoothing is not sensitive to the particular choice of the averaging interval for the smoothing within reasonable limits). The RV residuals from the smoothing were subjected to a period search based on the Stellingwerf (1978) PDM technique over a period range from 5000 down to 0.05 d. The dominant frequency found in both searches was 0.004587 c d^-1 and its integer submultiples. The one-day aliases were largely supressed thanks to having data from observatories, that have a large difference in their local time, producing much shallower minima in the $\theta$ statistics ( $\sim$ 0.75-0.82) and scattered phase diagrams. To make the diagrams readable, we show the corresponding $\theta$ statistics in Fig. 7 for the emission (top) and absorption (bottom) RVs only for a limited frequency interval down to 0.1 c d^-1. The result seems to confirm the 218-d periodicity discovered by Katahira et al. (1996b).

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg8.EPS}}\par\resizebox{9cm}{!}{\includegraphics{13885fg9.EPS}} \vspace*{-2mm}\end{figure}$	Figure 7: Stellingwerf (1978) PDM $\theta$ statistics for all emission-wing RVs ( top) and shell absorption-core RVs including Katahira et al. (1996b). The dominant frequency of 0.004587 c d^-1 corresponds to the 218-d period.
Open with DEXTER

$\begin{figure} \par\rotatebox{-180}{\resizebox{9cm}{!}{\includegraphics{13885fg10.EPS}}} \vspace*{-2mm}\end{figure}$	Figure 8: Orbital RV curves of the H $\alpha$ emission shown for subsets of data spanning less than a year. For all plots, period 218 $.\!\!^{\rm d}$ 053 was used, with phase zero at HJD 2452041.11, which corresponds to the RV minimum (see Table 2).
Open with DEXTER

As another demonstration that the 218-d period is real, we show phase plots in Fig. 8 for the original RVs (without prewhitening for the long-term changes) for several subsets of data covering time intervals no longer than one year. Clearly similar RV curves, with sharp minima, rather flat maxima, and a mutual phase coherence, are seen in all cases. The first subset is based solely on the RVs from the Ondrejov spectra secured with Reticon detector, which were already investigated by Rivinius et al. (2006).

4 BU Tau as a spectroscopic binary

Our findings, and especially the fact that the H $\alpha$ Balmer emission line moves in RV as a whole in spite of very large secular changes of its strength, indicate that BU Tau is indeed a single-line spectroscopic binary that moves in a highly eccentric orbit. We therefore used the program SPEL (written by the late Dr. Jirí Horn and never published) to derive the orbital elements. For comparison with Katahira et al. (1996b), we first derived orbital elements for Balmer absorption RVs, using the data from their study, RVs published by Rivinius et al. (2006), and our own H $\alpha$ absorption RVs, prewhitened with HEC13 as shown in Fig. 5. The resulting orbital elements are given as solution 1 in Table 2 and the corresponding phase plots are shown in Fig. 9. For more clarity, we plot there the photographic RVs, Heros RVs from Rivinius et al. (2006), and our H $\alpha$ absorption RVs in three separate panels. Although Rivinius et al. (2006) write that the suspected binary nature of BU Tau could not be confirmed on the basis of their data, their RVs also nicely follow the 218-d period. This constitutes yet another support for the reality of this period. Our solution 1 agrees well with the result of Katahira et al. (1996b).

Next, we analyzed the emission RVs that we consider as most realistically describing the true orbital motion. To see how sensitive the result is to the manner of prewhitening the data we derived the elements not only for the RVs prewhitened with the help of HEC13 (see above) but also from the original data. To this end, we divided the data into subsets spanning no more than one year and allowed SPEL to derive separate $\gamma$ velocities for individual data subsets. The results are summarized in Table 2, and the corresponding RV curves compared in Fig. 10.

$\begin{figure} \par\rotatebox{-180}{\includegraphics[width=9cm]{13885fg11.EPS}} \vspace*{-2mm}\end{figure}$

Figure 9:

Top: the phase plots of all available Balmer absorption RVs, prewhitened for the long-term RV variations with HEC13 (as shown in Fig. 5). Elements from solution 1 of Table 2 were used, with phase zero at minimum RV. For clarity, we show three different data subsets separately: Top panel: photographic RVs from Katahira et al. (1996b); Central panel: RVs from electronic Heros spectra published by Rivinius et al. (2006); Bottom panel: RVs from electronic spectra used in this paper.

Open with DEXTER

Table 2: Several sets of orbital elements.

$\begin{figure} \par\includegraphics[width=8cm,clip]{13885fg12.eps}\\ [5mm] \includegraphics[width=8cm,clip]{13885fg13.eps} \vspace*{-2mm} \end{figure}$

Figure 10:

The orbital RV curves of BU Tau based on the H $\alpha$ emission RVs plotted for the solutions 2 and 3 of Table 2. Phase zero corresponds to the respective epoch of minimum RV and the O-C deviations from the solutions are shown by small circles in separate panels. Top two panels: RVs prewhitened via HEC13 (solution 2). Two bottom panels: original RVs minus locally derived systemic $\gamma$ RVs (solution 3). See the text for details.

Open with DEXTER

The inspection of Fig. 10 shows that even the H $\alpha$ emission-wing RVs are indeed indicative of an orbit with high eccentricity but that there is also an alternative possibility that the observed deep RV minimum could be a consequence of some unspecified effect of circumstellar matter, reminiscent of ``an inverse rotational or Rossiter effect''. In this case, the true orbit could essentially be circular. To this end, we derived yet another, a circular-orbit solution for the H $\alpha$ emission RVs prewhitened for long-term changes via HEC13, omitting all RVs from the phase interval around phase zero with the most negative RVs. This resulted in the following elements: $P=218\hbox{$.\!\!^{\rm d}$ }34\pm0.62$ , $T_{\rm super.c.}={\rm HJD}~2452009.9\pm4.8$ , $K_1=1.72\pm0.21$ km s^-1.

Using the eliptical-orbit elements for the H $\alpha$ emission RVs from Table 2, we estimated the basic properties of the binary from the mass function f(m)=0.00165 $M_{\odot}$ for several plausible orbital inclinations, assuming a normal mass of the primary corresponding to its spectral type after Harmanec (1988) to be M₁=2.9 $M_{\odot}$ .

The results of Table 3 show that the binary properties, especially the low mass ratio, are quite similar to other binaries discovered so far with Be primaries. For the estimates, we only considered higher orbital inclinations since BU Tau is one of the cases of an inverse correlation between the brightness and emission-line strength, which indicates that we see the system roughly equator-on - cf, e.g., Harmanec (1983).

If we adopt the distance to Pleiades d=138 pc after Groenewegen et al. (2007), we estimate that the projected angular distance of the binary components should be $\theta=0\hbox{$.\!\!^{\prime\prime}$ }0075$ , dropping down to $0\hbox{$.\!\!^{\prime\prime}$ }0018$ at periastron. This angular separation is certainly within reach of existing large optical interferometers. The only problem is the luminosity ratio primary/secondary. If the secondary would be a normal late M dwarf corresponding to its mass, it would be fainter in the visual region by more than 10 mag and the only chance to search for it would be in the far IR region, where, however, the IR excess from the Be envelope can complicate the detection. However - if it were a hot subdwarf, similar to the one found for another Be binary $\varphi$ Per by Gies et al. (1998) - it might be observable in the optical region since the absolute visual magnitude of BU Tau is fainter for some 2 mag than for the $\varphi$ Per B0.5e primary. Finally, a cool Roche-lobe filling secondary seems improbable since it would probably produce binary eclipses.

In any case, attempts to resolve the 218-d binary system with some large interferometer are very desirable since a visual orbit would help not only to estimate the true orbital inclination but also to clarify whether the orbit has a high eccentricity or is nearly circular.

5 Comments on Hirata's model

Table 3: Basic physical properties of BU Tau as a single-line binary based on elliptical-orbit solution for the H $\alpha$ emission RVs - (cf. Table 2).

$\begin{figure} \par\includegraphics[width=9cm,clip]{13885fg14.EPS} \vspace*{8mm} \end{figure}$	Figure 11: A time development of the FWHM (in Å) of the H $\alpha$ emission. The rapid increase is caused by the formation and a fast strengthening of another double emission due to a newly formed envelope.
Open with DEXTER

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg15.EPS}} \vspace*{5mm} \end{figure}$

Figure 12:

A series of the H $\alpha$ profiles over the time interval of the formation of a new shell. The HJDs-2400000 of individual spectra are shown and the time runs from the top to the bottom. One can see how the new broad emission gradually rises in intensity and how its blending with the decaying previous double (but narrower) emission creates a profile with four emission peaks for some time. Then the new emission gets so strong that it merges with the original one.

Open with DEXTER

We have postponed a detailed study of the long-term changes for a later work (Iliev et al. in prep.), but we wish to comment briefly on the hypothesis put forward recently by Hirata (2007). He obtained systematic spectroscopy and polarimetry of BU Tau from 1974 to 2003 and finds a change in the polarization angle from about 60 $^\circ$ to 130 $^\circ$ over that time interval. He interprets this change as evidence of the precession of the circumstellar disk that is responsible for the observed H $\alpha$ emission. He further argues that also the change in the H $\alpha$ profiles from a weak double emission with a strong central absorption core to a strong emission with a wine-bottle shape indicates that the disk was first seen more or less edge-on and later more face-on. Tanaka et al. (2007) studied the spectra of BU Tau from Nov. 2005 until April 2007, which cover the period of a formation of the new shell phase. They argue that a new disk was formed in the equatorial plane of the B star while the old disk was decaying but still present. According to their interpretation, the old disk was precessing in space as suggested by Hirata (2007). Our spectra cover a much longer time interval, including the one studied by Tanaka et al. (2007), and as Fig. 2 shows, the change of the H $\alpha$ profile was smooth. We thus measured the full width at half maximum (FWHM) of a representative selection of our H $\alpha$ emission-line profiles and the variation in FWHM with time is shown in Fig. 11. It was already demonstrated by Struve (1931) in his first model of Be stars as rapidly rotating objects that there is a clear correlation between the width of presumably photospheric He I lines and the width of the Balmer emission lines, which is preserved during the long-term changes. This correlation has been confirmed by a number of later studies - see, e.g., Fig. 5 of Slettebak (1979). One would therefore expect that, if the appearance of a new shell phase of BU Tau is primarily a consequence of a geometrical effect, namely a gradual precession of a flat disk that becomes to be seen equator-on, the FWHM should gradually grow as the new shell phase is approaching. In contrast, Fig. 11 shows that the FWHM of H $\alpha$ was slowly decreasing during the last 15 years. Its dramatic increase is related to the formation of a new envelope, which our spectra clearly confirm - see Fig. 12. The apparent discontinous increase in the FWHM occurs at the moment when the strength of the broader emission from the new envelope rises to a half of the peak intensity of the original emission. All this indicates that the observed variations are primarily due to physical changes in the circumstellar matter and cannot be reduced to a simple geometrical cause - a precession of the original gaseous disk. There has been a rather widespread tendency in recent years to intepret the presence of shell absorption lines as evidence of an equator-on view, since many investigators are picturing the Be star disk as a flat structure located at the stellar equator with a (rather small) opening angle (Bjorkman & Cassinelli 1993; Waters 1986; Hanuschik 1996,1995). It is true that this model can lead to theoretical Balmer profiles similar to the observed ones, see, e.g., the 3D radiative line transfer models by Hummel (1994). One should be aware, however, that there is no unique proof of a specific geometry on the level of various simplifications of current models. For instance, Höflich (1987, 1988) succeeded in modeling several Balmer emission-line profiles of particular Be stars with his model consisting of an NLTE atmosphere and a spherical envelope. It is then conceivable that strong shell lines could also develop in the spectrum of a Be star seen more or less pole-on in situations where a very extended spheroidal envelope forms around it. Similarly, it might be worth considering whether the asymmetry detected by the gradual change in the polarimetric angle is indeed caused by the precession of a flat disk or by some other effect, e.g. by a slowly revolving elongated (non-axisymmetric) disk.

Acknowledgements

We profited from the use of the program SPEL, written by our late colleague Dr. Jirí Horn. We acknowledge the use of the publicly available Elodie spectra from the electronic archive of the Haute Provence Observatory. Our thanks go to Drs. M. Ceniga, P. Hadrava, A. Kawka, D. Korcáková, J. Krticka, M. Netolický, S. Stefl, and V. Votruba, who secured some of the Ondrejov spectrograms used in this study. We also thank the referee, Dr. A.F. Gulliver, for his comments on the first version of the paper. The research of the Czech authors was supported by the grant 205/06/0304 and 205/08/H005 of the Czech Science Foundation and also from the Research Programs MSM0021620860 Physical study of objects and processes in the solar system and in astrophysics of the Ministry of Education of the Czech Republic, and AV0Z10030501 of the Academy of Sciences of the Czech Republic. The research of PK was supported by the ESA PECS grant 98058. In its final stages, the research of J.N., P.H., and M.W. was also supported by the grant P209/10/0715 of the Czech Science Foundation. We acknowledge the use of the electronic database from the CDS, Strasbourg and electronic bibliography maintained by the NASA/ADS system.

Appendix A: Overview of available spectroscopic observations

Here, we provide some details on the spectra used in this study and listed in Table 1 and on their reduction:

1.

Ondrejov spectra: All 101 electronic spectrograms were obtained in the coudé focus of the 2.0-m reflector and have a linear dispersion of 17.2 Å mm^-1 and a 2-pixel resolution 12600 (11-12 km s^-1 per pixel). The first 35 spectra were taken with a Reticon 1872RF linear detector and cover a spectral region from 6300 to 6730 Å. Complete reductions of these spectrograms were carried out by JN with the program SPEFO, written by the late Dr. J. Horn and further developed by Dr. P. Skoda and more recently by Mr. J. Krpata - see Horn et al. (1996) and Skoda (1996). The remaining spectra were secured with an SITe-5 $800 \times 2000$ CCD detector and cover a slightly longer wavelength interval 6260-6760 Å. Their initial reductions (bias subtraction, flatfielding, creation of 1-D images, and wavelength calibration) were carried out by MS in IRAF.

2.

DAO spectra: These spectrograms were obtained in the coudé focus of the 1.22-m reflector of the Dominion Astrophysical Observatory by SY, who also carried out their initial reductions (bias subtraction, flatfielding, and creation of 1-D images). Their wavelength calibration was carried out by JN in SPEFO. The spectra were obtained with the 32121H spectrograph with the IS32R image slicer. The detectors were UBC-1 $4096 \times 200$ CCD for data before May 2005 and SITe-4 $4096 \times 2048$ CCD for data after May 2005. They cover a wavelength region from 6150 to 6750 Å, have a linear dispersion of 10 Å mm^-1 and 2-pixel resolution of 21700 ( $\sim$ 7 km s^-1 per pixel).

3.

OHP spectra: The public ELODIE archive of the Haute Provence Observatory (Moultaka et al. 2004) contains 30 spectra listed as BU Tau, but some of them are actually spectra of 27 Tau. We were able to recover 21 usable spectra. For the purpose of this study, we extracted, rectified, and measured only the red parts of these spectrograms.

$\begin{figure} \par\includegraphics[width=7.3cm,clip]{13885fg16.EPS}\\ \includegraphics[width=7.3cm,clip]{13885fg17.EPS} \end{figure}$	Figure A.1: A comparison of independent RV measurements of the steep H $\alpha$ emission wings ( upper panel) and shell core absorption ( bottom panel).
Open with DEXTER

4.

Rozhen spectra: All 23 spectra from Rozhen observatory were obtained in the coudé spectrograph of the 2-m RCC telescope. A CCD camera Photometrics AT200 with SITe SI003AB $1024 \times 1024$ chip was used. The spectrograph was used in a configuration providing high-resolution spectra suitable for revealing fine details and the structure of the spectral lines. A Bausch & Lomb 632/22.3 grating was used in its 2nd order, giving a linear dispersion of 4.2 A/mm with 2-pixel resolution of 33 000 ( $\sim$ 4.5 km s^-1 per pixel). Wavelength coverage is about 100 Å around H $\alpha$ . The initial reduction (bias subtraction, flatfielding, creation of 1-D images and wavelength calibration) was carried out by LI in MIDAS.

Table A.1: Radial velocities of the H $\alpha$ emission wings and shell absorption core obtained via averaging the independent measurements by J. Nemravová and P. Harmanec; DAO = Dominion Astrophysical Observatory, Victoria; ROZ = Rozhen National Observatory; OND = Ondrejov Observatory; LIS = IGeoE-Lisbon; OHP = Haute Provence Observatory.

5.

Lisboa spectra: these 4 CCD spectra were obtained with the IGeoE 0.356-m SC telescope working at F/11. The spectrograph is a Littrow LHIRESIII with a 2400 grooves per mm grating and a spectral resolution of about 14.000. The initial reduction (bias subtraction, flatfielding, creation of 1-D images, and wavelength calibration) of the spectra was made by JR.

The rectification and removal of cosmics and flaws of all spectrograms were carried out in a uniform way by JN in SPEFO. The program SPEFO was also used to RV measurements, based on a comparison of direct and flipped images of the spectral line profiles. Since we were searching for small RV variations and since the setting on the steep wings of the emission-line profiles was not always straightforward (see below), these RV measurements were carried out independently by JN and PH. Besides the settings on the steep wings of the H $\alpha$ emission, we also measured the H $\alpha$ absorption core on all spectra where such absorption was present to have a comparison with the results of Katahira et al. (1996b). We also tried to measure RV of the He I 6678 Å absorption wings but due to weakness of this line and its possible structure, these measurements turned out to be useless so we did not use them. Following Horn et al. (1996), we also measured selected stronger and unblended telluric lines in all spectra and used them to a correction of the RV zero point. Thanks to that, the spectra from all observatories can be treated as coming from one instrument for all practical purposes.

A comparison of the two sets of independent RV measurements is shown in Fig. A.1. In general, the agreement is good. A formal regression between the measurements of PH and JN was derived. Its slope is $0.98 \pm 0.01$ for the emission and $0.94 \pm 0.01$ for the absorption. For the absorption line, it is conceivable that in specific cases one or the other measurer was confused by a telluric line blended with the stellar absorption core. For analysis, we used the mean RVs of the two independent measurements. All our RVs with the corresponding HJDs of their mid-exposures are provided in Table A.1.

References

Ballereau, D., Chauville, J., & Mekkas, A. 1988, A&AS, 75, 139 [NASA ADS] [CrossRef] [EDP Sciences] [MathSciNet] [PubMed] [Google Scholar]
Bjorkman, J. E., & Cassinelli, J. P. 1993, ApJ, 409, 429 [NASA ADS] [CrossRef] [Google Scholar]
Bozic, H., Harmanec, P., Horn, J., et al. 1995, A&A, 304, 235 [NASA ADS] [Google Scholar]
Doazan, V., Bourdonneau, B., & Thomas, R. N. 1988, A&A, 205, L11 [NASA ADS] [Google Scholar]
Gies, D. R., McKibben, W. P., Kelton, P. W., Opal, C. B., & Sawyer, S. 1990, AJ, 100, 1601 [NASA ADS] [CrossRef] [Google Scholar]
Gies, D. R., Bagnuolo, Jr., W. G., Ferrara, E. C., et al. 1998, ApJ, 493, 440 [NASA ADS] [CrossRef] [Google Scholar]
Groenewegen, M. A. T., Decin, L., Salaris, M., et al. 2007, A&A, 463, 579 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
Gulliver, A. F. 1977, ApJS, 35, 441 [NASA ADS] [CrossRef] [Google Scholar]
Hanuschik, R. W. 1995, Be Star Newsletter, 30, 17 [NASA ADS] [Google Scholar]
Hanuschik, R. W. 1996, A&A, 308, 170 [NASA ADS] [Google Scholar]
Harmanec, P. 1982, in Be Stars, IAU Symp., 98, 279 [Google Scholar]
Harmanec, P. 1983, Hvar Observatory Bulletin, 7, 55 [NASA ADS] [Google Scholar]
Harmanec, P. 1988, Bull. astr. Inst. Czechosl, 39, 329 [NASA ADS] [Google Scholar]
Harmanec, P., Habuda, P., Stefl, S., et al. 2000, A&A, 364, L85 [NASA ADS] [Google Scholar]
Hirata, R. 1995, PASJ, 47, 195 [NASA ADS] [Google Scholar]
Hirata, R. 2007, in Active OB-Stars: Laboratories for Stellare and Circumstellar Physics, ed. A. T. Okazaki, S. P. Owocki, & S. Stefl, ASP Conf. Ser., 361, 267 [Google Scholar]
Hirata, R., & Kogure, T. 1976, PASJ, 28, 509 [NASA ADS] [Google Scholar]
Hirata, R., & Kogure, T. 1977, PASJ, 29, 477 [NASA ADS] [Google Scholar]
Horn, J., Kubát, J., Harmanec, P., et al. 1996, A&A, 309, 521 [NASA ADS] [Google Scholar]
Hummel, W. 1994, A&A, 289, 458 [NASA ADS] [Google Scholar]
Iliev, L., Kovachev, B., & Ruusalepp, M. 1988, Information Bulletin on Variable Stars, 3204, 1 [NASA ADS] [Google Scholar]
Iliev, L., Koubský, P., Kubát, J., et al. 2007, in Active OB-Stars: Laboratories for Stellare and Circumstellar Physics, ed. A. T. Okazaki, S. P. Owocki, & S. Stefl, ASP Conf. Ser., 361, 440 [Google Scholar]
Katahira, J.-I., Hirata, R., Ito, M., et al. 1996a, in Rev. Mex. Astron. Astrofis., ed. V. Niemela, N. Morrell, P. Pismis, & S. Torres-Peimbert, Rev. Mex. Astron. Astrofis. Conf. Ser. 27, 5, 114 [Google Scholar]
Katahira, J.-I., Hirata, R., Ito, M., et al. 1996b, PASJ, 48, 317 [NASA ADS] [CrossRef] [Google Scholar]
Koubský, P., Harmanec, P., Hubert, A. M., et al. 2000, A&A, 356, 913 [NASA ADS] [Google Scholar]
Luthardt, R., & Menchenkova, E. V. 1994, A&A, 284, 118 [NASA ADS] [Google Scholar]
McAlister, H. A., Hartkopf, W. I., Sowell, J. R., Dombrowski, E. G., & Franz, O. G. 1989, AJ, 97, 510 [NASA ADS] [CrossRef] [Google Scholar]
Merrill, P. W. 1952, ApJ, 115, 145 [NASA ADS] [CrossRef] [Google Scholar]
Miroshnichenko, A. S., Fabregat, J., Bjorkman, K. S., et al. 2001, A&A, 377, 485 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
Miroshnichenko, A. S., Bjorkman, K. S., & Krugov, V. D. 2002, PASP, 114, 1226 [NASA ADS] [CrossRef] [Google Scholar]
Moultaka, J., Ilovaisky, S. A., Prugniel, P., et al. 2004, PASP, 116, 693 [NASA ADS] [CrossRef] [Google Scholar]
Rivinius, T., Stefl, S., & Baade, D. 2006, A&A, 459, 137 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
Roberts, Jr., L. C., Turner, N. H., & ten Brummelaar, T. A. 2007, AJ, 133, 545 [NASA ADS] [CrossRef] [Google Scholar]
Sharov, A. S., & Lyuty, V. M. 1976, in Be and Shell Stars, ed. A. Slettebak, IAU Symp., 70, 105 [Google Scholar]
Sharov, A. S., & Lyutyj, V. M. 1992, AZh, 69, 544 [NASA ADS] [Google Scholar]
Skoda, P. 1996, in Astronomical Data Analysis Software and Systems V, ASP Conf. Ser. 101, 187 [Google Scholar]
Slettebak, A. 1979, Space Sci. Rev., 23, 541 [NASA ADS] [CrossRef] [Google Scholar]
Stellingwerf, R. F. 1978, ApJ, 224, 953 [NASA ADS] [CrossRef] [Google Scholar]
Struve, O. 1931, ApJ, 73, 94 [NASA ADS] [CrossRef] [Google Scholar]
Struve, O., & Swings, P. 1943, ApJ, 97, 426 [NASA ADS] [CrossRef] [Google Scholar]
Tanaka, K., Sadakane, K., Narusawa, S.-Y., et al. 2007, PASJ, 59, L35 [NASA ADS] [Google Scholar]
Taranova, O., Shenavrin, V., & Nadjip, A. D. 2008, Peremennye Zvezdy Prilozhenie, 8, 6 [NASA ADS] [Google Scholar]
Vondrák, J. 1969, Bull. Astron. Inst. Czechosl., 20, 349 [NASA ADS] [Google Scholar]
Vondrák, J. 1977, Bull. Astron. Inst. Czechosl., 28, 84 [NASA ADS] [Google Scholar]
Waters, L. B. F. M. 1986, A&A, 162, 121 [NASA ADS] [Google Scholar]

Footnotes

... stars: Based on new spectral and photometric observations from the following observatories: Dominion Astrophysical Observatory, Herzberg Institute of Astrophysics, National Research Council of Canada, Haute Provence, IGeoE-Lisbon, Astronomical Institute AS CR Ondrejov, and Rozhen.
...: The program HEC13 with brief instructions how to use it is available to interested users at http://astro.troja.mff.cuni.cz/ftp/hec/HEC13

All Tables

Table 1: Journal of new spectroscopic observations for BU Tau.

Table 2: Several sets of orbital elements.

Table 3: Basic physical properties of BU Tau as a single-line binary based on elliptical-orbit solution for the H $\alpha$ emission RVs - (cf. Table 2).

All Figures

$\begin{figure} \par\includegraphics[width=8cm,clip]{13885fg1.EPS} \end{figure}$	Figure 1: Comparison of H $\alpha$ profiles from different stages of the long-term changes.
Open with DEXTER
In the text

$\begin{figure} \par\includegraphics[width=9cm]{13885fg2.EPS} \vspace*{-2mm} \end{figure}$	Figure 2: A complete series of our H $\alpha$ profiles in a gray representation (only a few saturated or underexposed spectra were omitted). Abscissa shows the wavelength scale in Å, while the time on ordinate is shown in JD-2400000. Each horizontal strip represents an average of spectra secured within 200 days, and dark horizontal belts correspond to time intervals from which no spectra are available. At the bottom, there is a scale showing the correspondence between the flux level in the units of continuum and the gray scale.
Open with DEXTER
In the text

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg3.EPS}} \vspace*{-2mm}\end{figure}$	Figure 3: Selected He I 6678 Å line profiles, ordered in time, with corresponding HJDs.
Open with DEXTER
In the text

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg4.EPS}} \vspace*{-2mm}\end{figure}$	Figure 4: Measured RVs of the H $\alpha$ emission wings plotted vs. time. Prewhitening for the long-term changes, carried out with the help of the program HEC13, is shown by a line. Empty squares show the alternate way to remove long-term RV changes via individual $\gamma$ velocities for subsets spanning no more than a year. See the text for details.
Open with DEXTER
In the text

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg5.EPS}} \vspace*{-2mm}\end{figure}$	Figure 5: Measured H $\alpha$ absorption core plotted vs. time. We also included the RV measurements of shell lines by Katahira et al. (1996b) and Rivinius et al. (2006) to this plot. Prewhitening for the long-term changes, carried out with the help of the program HEC13, is shown by a line. See the text for details.
Open with DEXTER
In the text

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg6.EPS}}\\ [5mm] \resizebox{9cm}{!}{\includegraphics{13885fg7.EPS}} \vspace*{-2mm}\end{figure}$	Figure 6: A comparison of two pairs of the H $\alpha$ line profiles from the locally recorded velocity extrema (HJDs of the profiles are indicated).
Open with DEXTER
In the text

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg8.EPS}}\par\resizebox{9cm}{!}{\includegraphics{13885fg9.EPS}} \vspace*{-2mm}\end{figure}$	Figure 7: Stellingwerf (1978) PDM $\theta$ statistics for all emission-wing RVs ( top) and shell absorption-core RVs including Katahira et al. (1996b). The dominant frequency of 0.004587 c d^-1 corresponds to the 218-d period.
Open with DEXTER
In the text

$\begin{figure} \par\rotatebox{-180}{\resizebox{9cm}{!}{\includegraphics{13885fg10.EPS}}} \vspace*{-2mm}\end{figure}$	Figure 8: Orbital RV curves of the H $\alpha$ emission shown for subsets of data spanning less than a year. For all plots, period 218 $.\!\!^{\rm d}$ 053 was used, with phase zero at HJD 2452041.11, which corresponds to the RV minimum (see Table 2).
Open with DEXTER
In the text

$\begin{figure} \par\rotatebox{-180}{\includegraphics[width=9cm]{13885fg11.EPS}} \vspace*{-2mm}\end{figure}$	Figure 9: Top: the phase plots of all available Balmer absorption RVs, prewhitened for the long-term RV variations with HEC13 (as shown in Fig. 5). Elements from solution 1 of Table 2 were used, with phase zero at minimum RV. For clarity, we show three different data subsets separately: Top panel: photographic RVs from Katahira et al. (1996b); Central panel: RVs from electronic Heros spectra published by Rivinius et al. (2006); Bottom panel: RVs from electronic spectra used in this paper.
Open with DEXTER
In the text

$\begin{figure} \par\includegraphics[width=8cm,clip]{13885fg12.eps}\\ [5mm] \includegraphics[width=8cm,clip]{13885fg13.eps} \vspace*{-2mm} \end{figure}$	Figure 10: The orbital RV curves of BU Tau based on the H $\alpha$ emission RVs plotted for the solutions 2 and 3 of Table 2. Phase zero corresponds to the respective epoch of minimum RV and the O-C deviations from the solutions are shown by small circles in separate panels. Top two panels: RVs prewhitened via HEC13 (solution 2). Two bottom panels: original RVs minus locally derived systemic $\gamma$ RVs (solution 3). See the text for details.
Open with DEXTER
In the text

$\begin{figure} \par\includegraphics[width=9cm,clip]{13885fg14.EPS} \vspace*{8mm} \end{figure}$	Figure 11: A time development of the FWHM (in Å) of the H $\alpha$ emission. The rapid increase is caused by the formation and a fast strengthening of another double emission due to a newly formed envelope.
Open with DEXTER
In the text

$\begin{figure} \par\resizebox{9cm}{!}{\includegraphics{13885fg15.EPS}} \vspace*{5mm} \end{figure}$	Figure 12: A series of the H $\alpha$ profiles over the time interval of the formation of a new shell. The HJDs-2400000 of individual spectra are shown and the time runs from the top to the bottom. One can see how the new broad emission gradually rises in intensity and how its blending with the decaying previous double (but narrower) emission creates a profile with four emission peaks for some time. Then the new emission gets so strong that it merges with the original one.
Open with DEXTER
In the text

$\begin{figure} \par\includegraphics[width=7.3cm,clip]{13885fg16.EPS}\\ \includegraphics[width=7.3cm,clip]{13885fg17.EPS} \end{figure}$	Figure A.1: A comparison of independent RV measurements of the steep H $\alpha$ emission wings ( upper panel) and shell core absorption ( bottom panel).
Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.