Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.mrao.cam.ac.uk/projects/OAS/publications/fulltext/nam04.pdf Дата изменения: Wed Feb 20 16:41:46 2013 Дата индексирования: Sat Mar 1 03:04:10 2014 Кодировка: Поисковые слова: millennium

Unveiling Orionis
J. S. Young, A. G. Basden, N. A. Bharmal, R. C. Boysen, C. A. Haniff, B. O'Donovan, E. B. Seneta, H. Thorsteinsson, N. D. Thureau
Astrophysics Group, Cavendish Laboratory Cambridge, UK

E. Pedretti, J. D. Monnier
University of Michigan, USA

COAST UV Coverage 800 600 400 v (10^4 lambda) 200 0 750nm 782nm 905nm

Observing Programme
Here we present preliminary results from a campaign to characterise the wavelength- and time-dependent surface structures of late-type supergiant stars. These data are the outcome of a collaboration between the Optical Interferometry Group in Cambridge, UK, and a team based at the University of Michigan. We have used the Cambridge Optical Aperture Synthesis Telescope (COAST) and Infrared Optical Telescope Array (IOTA) to image a Orionis in seven optical and near-infrared colours contemporaneously. Previous measurements have shown both a general decrease in apparent asymmetry going from the optical to the near-infrared (Young et. al. 2000), and correlations between the degree of asymmetry and molecular opacity, when crossing spectral features (Young et. al. 2002). However, the two effects have not been observed simultaneously. Our latest data offer the potential to investigate these related effects, and to test whether a convective origin for the asymmetry is plausible.

-200 -400 -600 -800 -800 -600 -400 -200 0 200 400 600 800 u (10^4 lambda)

Figure 1: uv plane coverage for COAST (left) and IOTA (right) observations. The angular resolution was 25x100 milliarcseconds (mas) at 750nm wit h COAST, and 50mas at 1.5 µm with IOTA. The typical size of the a Orionis stellar disc is 50mas.

Data Analysis
Interferometers measure Fourier components of the sky brightness distribution. Specifically, the basic observables for optical/IR stellar interferometers are the (squared) visibility amplitude and the closure phase. The raw interference fringe data recorded at COAST and IOTA were reduced by standard methods to obtain a set of estimates of these quantities for each observing waveband. Two methods have been employed to infer the appearance of a Orionis from the Fourier data: (a) reconstruction of images with or without using a priori information, using a Maximum-Entropy (MEM) based algorithm; and (b) fitting simple parameterisations of the sky brightness distribution to the visibility and closure phase data. Prior knowledge was incorporated in some of the MEM reconstructions by using a moderately limb-darkened (similar to predictions from stellar atmospheres) symmetric disc as the default image. This has the effect of favouring brightness distributions that have little or no flux beyond the radius of the default image disc. In this initial report we only present results obtained from the measurements with COAST.

Figure 2: Squared visibilities (top) and closure phases (bottom) measured for a Orionis at wavelengths of 782nm (left) and 905nm (right) with the COAST interferometer. The visi bilities on shorter baselines have been omitted ­ these are insensitive to structure on scales finer than the stellar disc. The error bars on the majority of the closure phases are <10 °. The closure phases differ significantly zero or 180 ° in both wavebands, indicating that the source cannot be centro-symmetric. The green points are the predictions of disc plus "hotspot" models for the two wavelengths (similar to those in Table 1). The best -fit hotspot position is similar in the two models, but the size and limb-darkening of the disc component differ.

Preliminary Results


Figure 3: From left to right, image reconstructions from COAST data at 750, 782, and 905nm. North is up and East is to the right. Contours are drawn at 2, 10, 20, 30, ..., 90% of the peak flux. The images were generated using a Maximum Entropy -based algorithm, with the reconstructions constrained to be somewhat disc-like as described above. These reconstructions are not unique. However all reconstructed images do exhibit asymmetries, with similar SE -NW directions.

Waveband Reduced 2 750/13nm 2.8 2.9 782/5nm 905/50nm 2.5 2.4 0.8 0.9

Disc Diam. /mas L-d par. (9.0) 91.0±0.9 (9.0) 90.7±0.7 96.8±0.8 97.1±0.7 59.0±0.4 59.8±0.3 (9.0) (9.0) (2.5) (2.5)

r /mas 4.4±5.6 (6.6) 6.9±3.1 (6.6) 13.1±2.4 (6.6)

Hotspot /deg Frac. flux 340±14 .039±.005 (333) .038±.005 335±8 (333) 344±3 (333) .065±.004 .066±.004 .062±.004 .068±.004

Waveband Reduced 2 750/13nm 2.7 2.8 782/5nm 905/50nm 2.1 2.5 0.8 0.8

Disc Diam. /mas L-d par. (9.0) 95.8±1.6 (9.0) 95.8±1.4 101.9±1.0 101.3±0.9 60.0±0.6 60.9±0.4 (9.0) (9.0) (2.5) (2.5)

r /mas 6.9±4.4 (6.6) 13.1±1.5 (6.6) 11.0±1.6 (6.6)

Hotspot /deg Frac. flux 342±12 .22±.03 (326) .24±.02 326±3 (326) 339±3 (326) .37±.02 .26±.01 .26±.02 .27±.02

Table 1: Fits of models with an unresolved hotspot. The models consisted of a limb -darkened disc with a single unresolved bright feature (modelled as a 5mas FWHM circular Gaussian) superimposed. The limb-darkening parameterisation used for the disc component was that of Hestroffer (1997). To make a comparison of the hotspot fluxes more meaningful, a second set of fits were made with the hotspot position fixed and wavelength-independent. As the table shows, this common location gives only slightly worse fits than the best -fit position obtained for each wavelength separately.

Table 2: Fits of models with a 20mas FWHM Gaussian hotspot. Otherwise as Table 1. These parameterisations are qualitatively similar to the image reconst ructions in Figure 3.

Discussion and Conclusions
Neither the model -fitting results nor MEM images unambiguously characterise the appearance of a Orionis. Nevertheless, we definitely detect an asymmetry in the appearance of the stellar disc in all three optical wavebands. These include both continuum and TiO bands, so the asymmetry is not simply due to spatial variations in TiO opacity. However, TiO is likely to play a role in determining the wavelength-dependence of the strength of the asymmetry. If we model the 2004 asymmetry as a single circular feature superimposed on the stellar disc:
­ The COAST data are consistent with such a model, with a "hotspot" location that is independent of observing wavelength ­ The size of the hotspot is unconstrained, but larger hotspots gi ve slightly better fits to the 750nm and 782nm data ­ Hotspot flux: · If the hotspot is unresolved, it contributes ~5% of the total flux. It contributes more flux at 782nm (TiO) and 905nm (near -continuum) than at 750nm (continuum) · If the hotspot is resolved, the hotspot flux can be as high as 3 7%. The fraction of the flux in a resolved hotspot depends somewhat on its radial positon, but is roughly independent of wavelength ­ Orionis appears much more limb -darkened at 750nm and 782nm ( 6) than at 905nm ( ~ 2.5). However this is unlikely to be representative of the underlying stellar disc in the absence of asymmetries ­ We intend to use the near -infrared data from IOTA to further constrain the wavelength-dependence of the asymmetry

References
· D. Hestroffer (1997). A&A 327, 199. · J. S. Young, J. E. Baldwin, R. C. Boysen, C. A. Haniff, P. R. La wson, C. D. Mackay, D. Pearson, J. Rogers, D. St. -Jacques, P. J. Warner, D. M. A. Wilson, and R. W. Wilson (2000). New views of Betelgeuse: multi-wavelength surface imaging and implications for models of hotspot generation. MNRAS 315, 635. · J. S. Young, J. E. Baldwin, A. G. Basden, N. A. Bharmal, D. F. Buscher, A. V. George, C. A. Haniff, J. W. Keen, B. O'Donovan, D. Pearson, H. Thorsteinsson, N. Thureau, R. N. Tubbs, and P. J. Warner (2002). Astrophysical results from COAST. Proc. SPIE 4838, 369.