Stellar surface structure

A survey by von der Luehe et al. (1995) indicates that about 2000 stars with declinations less than 40 degrees north have apparent diameters of one milli-arcsec and more and therefore are resolved to VLTI baselines in the NIR. Most of these stars are late type giants. Some 50% have apparent diameters of 2.5 mas or more and will permit detailed studies of their surfaces. Figure 2 shows the histogram for the distribution of apparent diameters and spectral class.

The superior imaging capability of VLTI will make possible the study of physical characteristics of surface phenomena and their variation with time. Important surface phenomena are due to hydrodynamic and magneto-hydrodynamic effects and result in large scale convection cells in the outer convection zones and concentrations of magnetic fields. The study of convection through surface temperature and line-of-sight velocity variations provides clues to the fundamental properties of the convection zone. The temporal variation of active regions provides insight in the underlying dynamo processes which generate magnetic fields in stars.

Figure 2: Histogram of CADARS (Catalogue of Apparent Diameters and Absolute Radii of Stars;Fracassini et al.,1988) stars with apparent diameters above 1 mas and declinations south of +40 degrees. The solid line represents the cumulative frequency.

Figure 3 shows the interferometric signal (visibility magnitude) which can be expected for a relatively quiet (solar-type) and an active star. It is important to notice that the signatures in the visibility functions occur at high angular frequencies and have small magnitudes. VLTI will probably be able to resolve well structures only on active giants. Although the target sources are bright (most stars have visual magnitudes between 2 and 8), the low visibility signal and the high spectral resolution required to perform measurements of velocities and Zeeman profiles will make necessary the use of the UTs.

Figure 3: Distribution of visibility magnitude for a quiet (solar-type) star (left) and for an active giant (right). The phenomena shown are convection, active regions (starspots and plages) and small-scale magnetic fields. The large contribution at low frequencies is due to the sharp edge of the stellar disk. Angular frequencies are given in units of ``inverse stellar radii.'' The corresponding scale in arcsec is shown on the top for the Sun at 10pc (left) and for a K giant with 24 solar radius (right).

The detection of surface features on cool giants and supergiants using large single telescopes has been one of the most important successes of interferometric imaging (Busher et al., 1990; Wilson et al., 1992). The best studied example is Betelgeuse. The image shown in Figure 4 is typical of those now being regularly obtained, which show a small number of bright unresolved features containing typically 5-15 % of the stellar flux superimposed on an otherwise uniform disk. The relationship, if any, between these features and the well documented mass loss and variability of Betelgeuse is at present unclear.

Figure 4: Image of the M2 Iab supergiant Betelgeuse at 710nm obtained using optical interferometry with the 4.2-m WHT (adapted from BHB90). This maximum-entropy reconstruction shows a single bright feature that is offset from the centre of an otherwise uniform disk. It represents the first resolved image of a star apart from the suns, and shows a convective hotspot offset from an otherwise uniform disk. This type of feature has recently been rediscovered by HST at ultraviolet wavelengths. The large apparent size of Betelgeuse means that it is one of the few stars than can be resolved with HST. VLTI, with its 100m baseline, will suffer no such limitations and offers the prospect both of investigating stars such as Bbetelguese at much higher spatial resolution, and of extending surface studies to more distant populations and less extended stellar types. Contours are plotted at 5, 10, 20, 30,...90, 95 % of the peak intensity. Note the scale of the axes - this is the one of the largest stars in the sky (in terms of apparent size)

These surface features, which appear as bright ``hotspots'' of emission, are probably the result of large-scale convective upwellings of material from hotter regions of the stellar interior (Schwarzschild, 1975). Their number, evolution timescale and brightness are certainly all consistent with such an hypothesis, but their detection has raised a number of further questions that will likely be amenable to large interferometers like the VLTI. For this reason, cool evolved stars are among the most promising targets for pilot interferometric observations. A brief summary of the possible science goals for such observations are listed below:

Frequency of occurrence

Although now imaged on a handful of massive M supergiants, there is growing evidence that surface inhomogeneities are also present on Mira-type long period variables, i.e., stars of much lower masses (Tuthill et al., 1994; Haniff et al., 1995). Limitations on the resolving power currently attainable from the ground mean that only the nearest and most luminous sources have been observed. The primary goal of an interferometric survey of the local neighbourhood will be to determine the frequency of occurrence of these hotspots as a function of type and luminosity class.

Evolutionary timescale

One of the most useful diagnostics in the study of surface inhomogeneities will be the precise determination of their evolutionary timescales. Predictions exist for convective models, (Schwarzschild, 1975) but there has been little effort to monitor these stellar surfaces using high-resolution imaging methods. A dedicated interferometer offers the possibility of such a program.

Multiplicity

Current ground-based studies of stellar hotspots have been constrained by the limited resolutions of monolithic telescopes. In this sense, observations have only been able to place limits on the sizes and multiplicities of the hotspots seen on these targets. Once again, predictions for these properties exist for a number of models, implying that significant progress could be made if observations at much higher spatial resolution were available.

Location

Another useful diagnostic for elucidating the physical mechanism responsible for surface features will be identification of the precise radial depth at which they occur. Because of the abundance of molecular and atomic species in cool stellar atmospheres, spectrally resolved measurements provide useful information as to the radial stratification of the stellar atmosphere, and so it should be possible, in principle, to map out the vertical locations of the surface inhomogeneities.

Mass loss

Perhaps the most exciting prospect lies in tying together the observed surface features with the prodigious mass loss and variability of cool giants and supergiants. Mass loss from cool stars remains a very poorly understood area, and interferometric observations offer the prospect of imaging circumstellar dust very close to the stellar surface, of monitoring the photospheric radius directly, and of directly relating spatially resolved images with photometric and polarimetric variability.

As well as the main areas listed above, one should not forget more mundane, but equally important, problems that could be addressed by the VLTI. These include precise angular diameter measurements, which lead to effective temperature, and studies of the atmospheric structure. All the questions raised here can be addressed by a combination of programs: (i) detailed studies of selected sources, (ii) monitoring of selected sources every month, and (iii) a survey of the local neighbourhood. In many instances, interferometric data will provide the first direct measurements with which to confront -- and perhaps overturn -- existing theories.