δ Scuti pulsations in β Pictoris (original) (raw)

Journal Article

South African Astronomical Observatory, PO Box 9, Observatory, 7935 Cape, South Africa

Search for other works by this author on:

Received:

28 January 2003

Accepted:

07 February 2003

• PDF
• Split View
• • Article contents
  • Figures & tables
  • Video
  • Audio
  • Supplementary Data

Navbar Search Filter Mobile Enter search term Search

Abstract

The discovery of very low-amplitude periodic brightness variations of β Pictoris is reported. Three frequencies in the approximate range 47–53 d−1 could be extracted from four observing runs spanning a week. The character of the variations, together with the high incidence of δ Scuti pulsations in mid A type stars, suggests that β Pic is a member of this class of variable stars.

1 Introduction

β Pictoris is famous for its dusty circumstellar disc which might provide the raw material for planet formation; see Artymowicz (1997) and Vidal-Madjar, Lecavelier des Etangs & Ferlet (1998) for reviews. Much has been written about spectroscopic variability of the star, which is ascribed to cometary infalls (e.g. Beust & Morbidelli 2000). Photometric variability has also been observed (Lecavelier des Etangs et al. 1995, 1997). Both long time-scale (∼1000 d) and short time-scale (∼30 d) variations are claimed, with amplitudes ∼0.011 and ∼0.06 mag respectively. Explanations which have been offered include planetary occultations (Lecavelier des Etangs et al. 1997) and scattering and occultation by an orbiting cloud of dust (Lamers, Lecavelier des Etangs & Vidal-Madjar 1997).

In this brief note we report rapid periodic microvariability of β Pic, which we ascribe to δ Scuti pulsations. The next section of the paper describes the data acquisition and analysis. Concluding remarks are made in Section 3.

2 The Data

The observations were made with the South African Astronomical Observatory 0.5- and 0.75-m telescopes, with the Modular Photometer and University of Cape Town Photometer respectively. Both are photoelectric instruments, with GaAs photomultipliers. Neutral density filters were used to bring the photon count rates from the very bright star (_V_= 3.85) down to manageable levels.

Initially differential photometry of the star was attempted. When it became clear that the level of variability, if any, was very low, we resorted to continuous monitoring. This procedure has the obvious drawbacks that nothing is learned about the wavelength dependence of any variability, and that stellar variations are contaminated by atmospheric transparency changes. On the other hand, the amount of data collected is maximized, hence improving the likelihood of finding small systematic brightness variations.

The observing strategy was therefore simply to obtain measurements at 10-s intervals, through a single filter (Johnson U or B). Background sky counts were negligible, and were only obtained two or three times during the night. Observing was interrupted at frequent intervals to check 0.5-m guiding, and to correct the rather poorer tracking of the 0.75-m telescope. An observing log is given in Table 1. Only data obtained under perfectly photometric conditions were retained.

Table 1.

Log of the observations of β Pic; see text for further details.

The reduced _B_-band observations are plotted in Fig. 1. The data from the four nights were individually de-trended by subtracting fitted parabolas. There is some residual trend: it is of course impossible to tell whether its origin is atmospheric, stellar, or both.

Figure 1.

Partially de-trended _B_-band light curves of β Pic. Each panel is labelled with the last four digits of the appropriate HJD. Each plotted point is the average of three consecutive 10-s integrations.

Fig. 2 shows amplitude spectra of the data in Fig. 1, over the frequency interval [16, 90] d−1. The top panel spectrum is that of the raw data; the next three panels are spectra of the residuals after prewhitening by one, two and three frequencies respectively. Prewhitening was performed by (i) finding the frequency at which the amplitude spectrum is a maximum; (ii) fitting (by linear least squares) a sinusoid with that frequency to the data; and (iii) subtracting the sinusoid from the data.

Figure 2.

Amplitude spectra of the raw _B_-band data (top panel), and of the residuals after prewhitening by one, two and three frequencies (next three panels). Note that the vertical scales on the different panels are not the same.

Aliasing is very obvious in the spectra, and this should be borne in mind when studying the results in Table 2. The table gives the results of simultaneously fitting three sinusoids to the data, using an iterated non-linear least squares procedure. Starting guesses for the algorithm were obtained from the prewhitening procedure illustrated in Fig. 2.

Table 2.

The results of fitting three sinusoids simultaneously to the raw data. Formal uncertainties in the last digits of the numbers are shown in brackets.

Although somewhat extreme (see below), the periods and amplitudes are consistent with δ Scuti type pulsations. Furthermore, the reason for testing β Pic for relatively rapid (time-scale of hours or shorter) variations was that its spectral type of A5V (e.g. Kondo & Bruhweiler 1985) places it well inside the δ Scuti instability region (e.g. Rodríguez & Breger 2001). It therefore seems very plausible that, in addition to its other accomplishments, β Pictoris is a δ Scuti star.

The two _U_-band data sets are too short to derive trustworthy results, but the longer run shows an excess of power in the range 42–49 d−1, at somewhat higher amplitude than seen in the B waveband. The frequency range includes the highest amplitude frequency seen in B, and the higher amplitude in U is consistent with δ Scuti pulsation behaviour.

A useful formula for calculating the pulsation constants of the three periodicities is

(Breger & Bregman 1975). Lanz, Heap & Hubeny (1995) found _T_eff= 8200 ± 150 K, log _g_= 4.25 ± 0.1 and L/ L⊙= 11.3 ± 3.5, from which we derive _M_bol= 2.11 ± 0.34 mag. With the benefit of the Hipparcos parallax of 51.87 ± 0.51 mas (ESA 1997), we find _MV_= 2.42 ± 0.02 mag. Applying the bolometric correction of −0.15 mag appropriate for an A5V star (Cox 2000), _M_bol= 2.27 mag, in good agreement with the photometrically deduced value.

The calculated pulsation constants are listed in Table 2. Comparison to the model calculations of Stellingwerf (1979) shows that (in order of increasing frequency) the values correspond to third, fourth and fifth radial overtones (or ℓ > 0 non-radial modes). Assuming the derived stellar parameters to be correct, and taking _Q_= 0.033 for the fundamental radial mode, the latter would have a frequency of 19.7 d−1.

3 Concluding Remarks

The information in the catalogue of Rodríguez, López-González & López de Coca (2000) may be used to place the δ Scuti properties of β Pic in perspective. There are 636 stars in the catalogue, of which only six have pulsation periods shorter than the principal mode of β Pic. Three stars have pulsation amplitudes that are similarly small (∼2 mmag), no doubt pointing to the difficulty of detecting such low-amplitude variability. β Pic is also of course one of the brightest known δ Scuti stars: only six stars in the Rodríguez et al. (2000) catalogue are brighter.

The position of β Pic in the δ Scuti instability strip is shown in Fig. 3, which is based on fig. 8 in Rodríguez & Breger (2001), where details can be found. The de-reddened Strömgren (b_−_y) index is used as a temperature indicator in this diagram. The measured value for β Pic is (b_−_y) = 0.094 (Crawford, Barnes & Golson 1970). International Ultraviolet Explorer observations have been used to show that reddening is probably negligible (Kondo & Bruhweiler 1985), so that (b_−_y)0≈ 0.094.

Figure 3.

The δ Scuti instability strip in the Hertzsprung–Russell diagram. The position of β Pic is shown by the open circle on the zero-age main sequence.

The position of β Pic near the zero-age main sequence is not a surprise, as it is known to be young, and may in fact be a pre-main-sequence (PMS) star (Lanz et al. 1995; Barrado y Navascués et al. 1999). If this is the case, then β Pic is the tenth of the rare PMS δ Scuti stars to be discovered (Kurtz 2002; Rodríguez 2002). Kurtz (2002) discussed the importance of studying these stars: amongst other reasons, the data could be used to constrain stellar structure models of stars in the relevant part of the Hertzsprung–Russell diagram. Marconi & Palla (1998) calculated the position of the instability strip for PMS stars, and found it to lie between ∼6500 and ∼7500 K; clearly, β Pic is well beyond the blue border. More encouraging is that the authors' models showed that overtone pulsations are preferentially excited in hotter stars.

Acknowledgments

The author is grateful to Professor Don Kurtz for discussions of previous attempts to detect photometric variability in β Pic, and to Dr Eloy Rodríguez for very kindly supplying the data used to draw Fig. 3. The author also thanks Erica and Thea Koen for assisting with the observations on the 0.75-m telescope.

References

,

1997

,

Ann. Rev. Earth Planet. Sci.

,

25

,

175

, ed.,

2000

,

Allen's Astrophysical Quantities

, 4th edn.

Springer-Verlag

,

New York

ESA

,

1997

,

The Hipparcos and Tycho Catalogues

, ESA SP-1200.

ESA Publications Division

, Noordwijk

,

2002

, in , eds, ASP Conf. Ser. Vol. 256,

Observational Aspects of Pulsating B and A Stars

.

Astron. Soc. Pac.

San Francisco

, p.

31

,

2002

, in , eds, ASP Conf. Ser. Vol. 256,

Observational Aspects of Pulsating B and A Stars

.

Astron. Soc. Pac.

,

San Francisco

, p.

137

,

1998

,

Planet. Space Sci.

,

46

,

629

© 2003 RAS

Citations

Views

Altmetric

Metrics

Total Views 486

263 Pageviews

223 PDF Downloads

Since 11/1/2016

Citations

21 Web of Science

×

Email alerts

Citing articles via

• Latest

• Most Read

• Most Cited

More from Oxford Academic