eso9706

The Biggest Star in the Sky

11 March 1997

Click to Enlarge

An international team of astronomers has used large telescopes in Chile and Australia to measure the biggest star in the sky. The star, designated R Doradus , is of the so-called red giant type and is located in the southern constellation of Dorado. Its apparent diameter (i.e., the size which the star appears to have when seen from the Earth) is larger than any other so far observed, except for the Sun. In particular, it exceeds by more than 30 % that of Betelgeuse , which for the past 75 years has held the title of star with the largest apparent size.

Measuring sizes of stars

Measuring the sizes of stars is very difficult due to their enormous distances. For example, if our Sun were placed at the distance of the next closest star (four light-years away), it would have about the same apparent size as a DM 1 (or US quarter-dollar) coin placed at a distance of 500 km (about 0.01 arcsec). Even for the most powerful astronomical telescopes, it is a very challenging task to measure such small angles.

Ideally, the angular resolution of a telescope (its capability to resolve fine details in celestial sources) increases with its diameter. In practice, although ground-based optical telescopes now have diameters up to 10 metres, their actual resolution of visual light is that of a telescope of only about 20 centimetres aperture. This is because of the constant turbulence in the Earth's atmosphere. This turbulence causes the stars to twinkle in a way which delights the poets but frustrates the astronomers, since it blurs the fine details of the images.

The first, and largest, star apart from the Sun to have its diameter measured was Betelgeuse, the brightest star in the constellation of Orion. Its angular diameter was found to be 0.044 arcsec by Albert Michelson and his team who used the Hooker telescope on Mt. Wilson in California in the early 1920s, pioneering interferometry techniques. Betelgeuse kept its title as the star with the largest apparent size for the next 75 years. This title has now been taken by R Doradus.

R Doradus is a variable star in the constellation of Dorado (the Swordfish), located in the far southern sky. At a distance of about 200 light years it is relatively nearby. R Doradus is a variable star with a period of about 338 days, changing its magnitude from approximately 4.8 at maximum (when it is visible with the unaided eye) to 6.6 at minimum (when it requires a small telescope).

 Interferometry at the NTT

In August 1993, the team of astronomers [1] pointed the ESO 3.5-metre New Technology Telescope (NTT) towards R Doradus. For these observations, the NTT was covered with an opaque mask with seven holes arranged on a 3.3-metre diameter circle. Each of these holes had a diameter of 25 cm, which was smaller than the cells of turbulence in the atmosphere above. The main motivation for using the mask was to suppress the effects of the turbulence and in this way restore the full resolution capability of the NTT [2].

The seven light beams from a star were brought to interfere with each other at the telescope's focus. Each pair of holes in the mask produced a fringe pattern in the image of the star, so at any moment there were 21 distinct fringe patterns. A camera in the focal plane recorded these fringes, their contrast being determined during subsequent computer analysis.

A star which is very far away will appear too small for its disk to be resolved by the telescope. All of the 21 fringes will then have approximately the same contrast. On the other hand, if the star is closer by and has a perceptible size, the contrast of the fringe patterns will be reduced for widely separate mask holes. By comparing the fringe contrast of the target star with that of a more distant, unresolved star, it is then possible to estimate the size of the target.

The present NTT observations were made at infrared wavelengths (1.25 microns) with the SHARP camera, developed by the Max-Planck Institut for Extraterrestrial Physics (Garching, Germany). Several hundred very short exposures of R Doradus were made, each lasting 0.1 second (this is short enough to freeze the 21 fringe patterns in each exposure). Immediately thereafter, a similar series of observations was made of an unresolved calibrator ' star (Gamma Reticuli). This procedure was repeated several times, producing thousands of images to be analysed.

Additional observations were made in 1995 with the NTT as well as the 3.9-m Anglo-Australian Telescope at Siding Spring (Australia). These observations, and the application of different interferometric data analysis techniques to similar data sets, confirmed the results of the earlier ones.

The results

The results clearly showed that R Doradus is extended, having an angular diameter of 0.057 +- 0.005 arcsec (assuming that the star appears as a uniform disk). This apparent size is 30% larger than Betelgeuse!

The bigger a star's apparent diameter, the more easily it can be resolved. The surprise is therefore not only the large diameter of R Doradus, but also the fact that this was not discovered earlier.

Many of the larger stars were already measured by Albert Michelson and his team. The reason for the late discovery is most likely the southern latitude of R~Doradus, which makes it inaccessible to the stellar interferometers predominantly located on the northern hemisphere. R Doradus is an inconspicuous star at visible wavelength but is one of the brightest in the sky in the infrared. This led Robert Wing (Ohio State University) to predict in 1971 that R Doradus should have a large angular size. Only now has this prediction been confirmed.

The NTT observations were made in the infrared. At first sight it may seem more sensible to observe at shorter, visible wavelengths because this would result in better angular resolution. However, measurements in the infrared - although more difficult to perform - result in a better estimate of the diameter of the underlying atmosphere (photosphere) of a star. The combination of a high-quality telescope and a high-quality infrared camera made this result possible.

R Doradus is approximately 200 light years away. The measured size implies that it has a physical diameter of 370 +- 50 times that of the Sun, or well over 515 million km! If R Doradus would be placed at the centre of the Solar System, its surface would be outside of the orbit of Mars. Although even bigger stars are known - Betelgeuse for one - none appears as large in the sky because they are all at greater distances. The very large apparent size of R Doradus is due to the combination of its relative proximity and large physical size.

R Doradus has about the same mass as the Sun, but it is 6500 times brighter [3].

Interferometry with the VLT

Although much more difficult, interferometry can also be done combining light from different telescopes. This has been successfully demonstrated by teams in France, UK and USA. As the telescopes can be some distance apart, the separation of the collecting apertures can be much increased, simulating a telescope with a diameter of a hundred metres, and the angular resolution can reach the level of the milli-arcsec.

This will be the case when the ESO Very Large Telescope Interferometer (VLTI) becomes operational some years from now. The VLTI is able to combine the light of four telescopes with 8.2-m diameter, and also from several smaller, movable auxiliary 1.8-m telescopes, at separations of up to 200 metres.

The VLTI will be a very powerful tool for studying small details in many astronomical objects. The team has already made observations of R~Doradus with the Anglo-Australian Telescope which show the star to have structure on its surface, analogous to (but many times larger than) Sun spots. The VLTI would provide forty times more resolution, allowing such structures to be studied in incredible detail.

Notes

[1] Tim R. Bedding, J. Gordon Robertson and Ralph G. Marson (School of Physics, University of Sydney, Australia), Albert A. Zijlstra and Oskar von der Lühe (ESO), John R. Barton (Anglo-Australian Observatory, Epping, Australia) and Brian S. Carter (South African Astronomical Observatory, Observatory, South Africa).

[2] For more details, see the article First light from the NTT Interferometer on page 2 of the December 1993 issue of the ESO house journal The Messenger.

[3] This number (6500) was incorrected given (as 180) on the printed version of this Press Release. Sorry for the inconvenience!.

More information

An article describing the results will appear in the April 21, 1997 issue of the British scientific journal Monthly Notices of the Royal Astronomical Society.