The Variable CaII Absorption in $\beta$
Pictoris during 1998
S.I. Barnes, William Tobin, K.R. Pollard
, PASA, 17 (3), 241.

Next Section: The CaII absorption in
Title/Abstract Page: The Variable CaII Absorption
Previous Section: The Variable CaII Absorption
Contents Page: Volume 17, Number 3

Introduction

Attention was drawn to the bright, southern A5V star $\beta$ Pictoris by its large IRAS infrared excess. Imaging soon revealed that $\beta$ Pictoris is surrounded by an edge-on dust disc extending over 1000 AU from the star. Spectroscopy revealed the presence of circumstellar absorption features, some of which are variable with velocities up to a few x 10² kms^-1 in both low-ionization species such as CaII and high-ionization species like AlIII and CIV. It is believed the system may be a transitional object between a young star surrounded by a proto-planetary disc and more-evolved systems where giant planets (at least) have formed. Recent reviews include those by Artymowicz (1997) and Vidal-Madjar, Lecavelier des Etangs & Ferlet (1998).

The FEB hypothesis

The variable absorption features have been ascribed to material evaporating from planetesimals or comet-like bodies on star-grazing orbits. This `Falling Evaporating Bodies' (FEB) hypothesis has principally been developed by workers in Grenoble (e.g. Beust et al. 1998 and references therein). The FEB model explains several observed phenomena. For example, shocks can produce the AlIII and CIV ions not expected from photoionization by an A5V star, and lower-velocity absorption features are longer-lived than higher-velocity ones because the FEB crosses the line of sight further from the star. However, puzzles remain, such as the extended lifetimes of lower-velocity features (possibly attributable to an infalling stream of FEBs) and the comparative rarity of blue-shifted features, which may be indicative of the presence of planetary resonances perturbing FEBs from circular into star-grazing orbits at selected azimuths. The behaviour of the variable absorption requires better observational characterization.

The CaII H & K absorption

**Figure 1:** The predictions of Beust & Lissauer (1994) for the prograde, equatorial passage of an FEB across the disc of $\beta$ Pictoris (their Figure 7). Bold line: evolution of CaII K. Narrow line: evolution of CaII H. The closed and open circles are our K and H observations of the $\sim$ 120 kms^-1 feature on 1998 November 27, arbitrarily offset in abscissa and arbitrarily scaled in ordinate to emphasize the similarity in form, as explained in Section 2.2.
$\begin{figure} \begin{center} \psfig{file=barnesf1.ps,height=10cm} \end{center} \end{figure}$

The CaII H & K absorption lines have different oscillator strengths (that of K is twice that of H) and this makes it possible to determine the optical depth of the obscuring ionic cloud,

$\tau_{\rm K} (= 2\tau_{\rm H})$ , and the fraction $\alpha$ of the stellar photosphere that it covers. Under the crude model of a uniform cloud and photosphere, these parameters are obtained simply by solving the equations for the line depths,

$p_{\rm K} = \alpha(1 - e^{-\tau_{\rm K}})$ and

$p_{\rm H} = \alpha(1 - e^{-\tau_{\rm K}/2}),$ whose observed values should fall within the limits

$p_{\rm H}= \frac{1}{2}p_{\rm K}$ (optically thin) and

$p_{\rm H}= p_{\rm K}$ (optically thick). Simple analyses of this sort have shown that the covering factors, $\alpha$ , are smaller for higher-velocity features, as expected if radiation pressure close to the star inhibits expansion of the cloud of Ca⁺ ions. In reality, the physical situation is more complicated. The photosphere is limb darkened, so an ionic cloud of a given size blocks proportionately more light when its projected location is closer to the centre of the stellar disc. More importantly, the photosphere is rotating fast (

$V\sin i \simeq 140$ km s^-1). Since the FEB absorption is detected against the spectral background of the photospheric CaII lines, whose profiles are a strong function of wavelength, the intensity of the photospheric light at any wavelength and at any point on the projected stellar disc is altered because of rotational Doppler shifts. A concrete example will make this clearer. If planetary perturbations cause planetesimals to become star-grazing, the FEB orbits are likely to be equatorial. The photospheric K profile increases in intensity away from its central wavelength. An FEB located in front of either the blue- or the red-shifted equatorial limb will therefore experience enhanced flux at the central wavelength of the absorption compared to what is would experience if the star was not rotating. Because of blending, however, the profile of the photospheric H line in $\beta$ Pictoris is not symmetric about the wavelength corresponding to the atomic transition. The intensity initially decreases on the long-wavelength side, where it is blended with Balmer H $\varepsilon$ . An equatorial FEB therefore experiences enhanced intensity if it is located in front of the approaching limb, and reduced intensity if it is located in front of the receding limb. As Beust & Lissauer (1994) and Hubeny & Heap (1996) realised, the joint evolution of the differently-behaving H & K absorption features should permit tracking of the passage of FEBs across the stellar disc and thus provide further evidence for the FEB hypothesis. Figure 1 shows predictions from Beust & Lissauer for the prograde, equatorial passage an FEB of a certain size and orbit; it is even possible for the H line to be stronger than the K line, despite its weaker oscillator strength.

**Figure 2:** Sample reduced and normalized spectra of $\beta$ Pictoris obtained on 1998 November 27. The reference profiles used for the normalization are also shown. The normalized spectra have been fitted by four Gaussians whose central velocities are indicated by vertical bars. One corresponds to the deep `stable' central feature at the stellar radial velocity of $\sim$ 20 km s^-1. One weak blueshifted and two stronger redshifted features have also been fitted. Had the multiple Gaussian fits been plotted, discrepancies between them and the normalized profiles would have been too small to see at the scale of this diagram.
$\begin{figure} \begin{center} \psfig{file=barnesf2.ps,height=10cm} \end{center} \end{figure}$

Previous observations

Observations of the CaII lines in $\beta$ Pictoris are thus of great interest, and intermittent échelle spectroscopy of this star has been underway from the Mt John University Observatory (MJUO) since 1992, sometimes as part of multi-observatory campaigns. Results from 1992 December have been published by Lagrange et al. (1996); data from 1994-96 (intensive during 1995) have been presented by Petterson & Tobin (1999). For these observations, the detector was too small to capture both the H and K lines, which were therefore observed sequentially, but difficulties arise in their joint interpretation because the absorption can evolve over the same timescale as required to obtain a spectrum. For this reason a spectrograph focal reducer was built (Tobin et al. 1998) and a larger CCD detector acquired in order to permit simultaneous observations of the H & K lines. A two-year programme of intensive monitoring began in 1997 (Persson 1998).

Next Section: The CaII absorption in
Title/Abstract Page: The Variable CaII Absorption
Previous Section: The Variable CaII Absorption
Contents Page: Volume 17, Number 3

ASKAP

Public

Australia Telescope National Facility

Introduction

The FEB hypothesis

The CaII H & K absorption

Previous observations