Communiqué de presse

VLT Spectra "Resolve" a Stellar Disk at 25,000 Light-Years Distance

Unique Observations of a Microlensing Event

25 avril 2001

Using the FORS1 multi-mode instrument at the 8.2-metre VLT ANTU telescope on Paranal during a microlensing event, the team was able to obtain detailed spectra of the different parts of the remote star. In doing so, they managed to probe its gaseous atmosphere at different depths. This is the first time that it has been possible to obtain detailed, spatially resolved spectra across the full face of a normal star other than the Sun [4].

Like our Sun, stars are large gaseous spheres. However, while we are able to perceive the Sun's disk, all other stars are so far away that they normally appear as points of light. Only specialized observing techniques, like interferometry [1], are able to "resolve" the images of nearby stars and to show them as extended balls of fire. But opportunities may sometimes arise that allow amazing observational feats in this field. Indeed, an international team of astronomers [2] has just "resolved" a single, normal star some 25,000 light years away, or about 1.6 billion times more distant than the Sun [3], by taking advantage of a multiple microlensing event.

During such a rare event, the light from the remote star is amplified by the gravity of a faint object that passes in front of it, as seen from the Earth. In fact, this gravitational lens acts as a magnifying glass that focusses different parts of the star's image at different times.

A many-faceted success story

The following story is about a most unusual astronomical observation and also shows how modern astrophysics works .

It combines the study of stellar atmospheres with the intricate optical effects produced by the gravitational field of a binary star in the Milky Way. The successful outcome was dependent on diligent observers in various regions of the world and ultimately on the critical timing of spectral observations with the ESO Very Large Telescope (VLT) at the Paranal Observatory in Chile.

Thanks to the effective collaboration among the scientists and a certain measure of good luck, unique data were obtained that are now providing fundamental new insights into stellar astrophysics.

The face of a star

Distant stars appear as small points of light, even to the largest telescopes on Earth. They are simply too far away to be "resolved" by normal telescopes, and no information can therefore be obtained about what the stellar surfaces look like. This is a fundamental obstacle to the detailed study of stars other than the Sun.

We know, however, that the disk of a star does not present itself as a uniform surface. As is the case of the Sun that exhibits variable structures like sunspots (in particular at the time of the present solar maximum), other stars may also have "star-spots".

Another general feature of solar and stellar disks is that they appear fainter towards the periphery. This phenomenon is known as "limb darkening" and is actually a matter of the viewing angle. When we look towards the middle of the solar disk, we see into rather deep and hot layers of its atmosphere. Contrarily, when we view the very edge of the solar disk, we only see the upper, cooler and dimmer parts.

Thus, by looking at different areas of its disk, we are able to probe different depths of the solar atmosphere. This in turn permits to determine the structure (temperature, pressure, chemical composition, etc.) of the upper layers of the Sun.

For more distant stars, however, their disks appear much too small for this kind of detailed observation. Despite much instrumental progress, therefore, fundamental observational information about stars is still lacking, especially for stars different from the Sun.

This is one of the main reasons why the astronomers are thrilled by a new series of spectra from the FORS1 multi-mode instrument at the 8.2-m VLT ANTU telescope at Paranal. They "resolve" for the first time the surface of a normal star some 25,000 light-years away.

This amazing observational feat has been possible with some help from a natural "magnifying glass". The road leading to this remarkable result is an instructive and interesting one.

Gravitational microlensing

The light from a distant star is affected by the gravity of the objects it passes on its way to us. This effect was predicted by Albert Einstein early last century and observationally confirmed in 1919 when a solar eclipse allowed the study of stars close to the line of sight of the Sun. Accurate positional measurements showed that the light from those remote stars was bent by the Sun's gravitational field.

However, the light may not only be deflected, it can also be amplified . In that case, the massive object works like a giant "magnifying lens" that concentrates the light from the distant source.

Effects of gravitational optics in space were first observed in 1979. When produced by extended, very heavy clusters of galaxies, they may take the form of large, spectacular arcs and well-separated multiple images. Less massive lenses, however, produce images with extensions that are too small to be distinguished directly.

Such "microlensing" effects occur when a compact body (usually a Milky Way star moving in its galactic orbit) passes almost directly between the observer and a luminous background object (usually also a star). One then sees that the brightness of that object rises and falls as the lens passes across the line-of-sight. The observed light intensity is described by a so-called "light curve". Normally, the lensing object is a faint low-mass star, one of the most common objects in the Milky Way.

Microlensing events

In most cases, these low-mass stars are too faint to be directly observed. This is especially so in crowded sky fields in which there are many much brighter stars - including the luminous giant stars that are monitored for microlensing effects. However, the gravity of a low-mass star is strong enough to produce a lensing effect if the geometrical alignment is sufficiently precise. This happens rarely, but by looking at a large number of background stars, it has been possible to detect a fair number of microlensing events during the past few years.

International collaborations like Experience pour la Recherche d'Objets Sombres (EROS), Optical Gravitational Lensing Experiment (OGLE) and Microlensing Observations in Astrophysics (MOA) scan the skies continuously for such microlensing events which typically last from a few weeks to some months. When a star is found to brighten in a way that looks like what is expected from microlensing, they send electronic alerts to other teams like Probing Lensing Anomalies NETwork (PLANET) and Microlensing Planet Search Project (MPS) who then intensively monitor the possible lensing events.

One of the main goals of these research programmes is to search for "dark matter" . Indeed, microlensing effects are excellent tools for learning more about this mysterious component of the Universe, as they provide information about lensing objects that otherwise are too faint to be observed.

However, microlensing events may also provide very useful information about the background object (the "source"), the light of which is amplified and magnified. When more light is available, more detailed (e.g., spectroscopic) observations can be made. In particular, on rare occasions, it can also help to "resolve" the surface of a distant star.

Using distorted lenses

If the lensing object is multiple, e.g., a binary star or a star with a planet, the gravitational lens will give rise to interesting phenomena. Whenever the gravitational fields from the two (or more) objects "co-operate", the lensing effect may become distorted and/or unusually strong.

Depending on the exact geometry of the lens, i.e. the momentary, relative positions in the sky of the lensing objects and the background object, it is possible that the background source may at some moment be very sharply magnified. In fact, this effect may be so "sharp", that the light from a certain area of the extremely small, apparent disk of a distant star is enhanced much more than that from other areas of the disk. If so, the stellar light registered by the terrestrial telescope will come mainly from that particular area. From optical terminology, such an event is referred to as a "caustic crossing" .

However, the exact circumstances are difficult and complex to calculate. The light curve during a lensing event depends on the relative motions of the involved objects or, in other words, on exactly how the distorting and magnifying glass (the lensing object), as seen from the Earth, moves across the background object.

In this context, binary lenses are particularly interesting. Not only can they very efficiently enhance the brightness of the source, but there will also be two "caustic crossings" and two associated light maxima. This implies that once the first crossing/maximum has passed, it may be possible to predict when and how the source will be magnified a second time. In that case, the astronomers will have time to prepare for detailed observations at the moment of the second caustic crossing. In particular, this may then include spectroscopic observations that can reveal the structure of the background star.

The May 2000 microlensing event

On 5 May 2000, the EROS group announced an apparently normal microlensing event in a direction a few degrees from the Galactic Centre (ESO Press Photo eso0022). The brightness of the background star was rising and the PLANET team began to monitor it during its regular operations.

About one month later, on 8 June 2000, the MPS team noticed that the event, now designated EROS-BLG-2000-5, was undergoing an unexpected, sudden and significant brightening. PLANET observers immediately turned their full attention to it, monitoring it continuously from five different observing sites located at suitable longitudes around the Earth. The light curve changed dramatically while the source went through a first caustic crossing (ESO Press Photo eso0022). On 10 June 2000, the PLANET team alerted the community that this particular event was indeed due to a multiple lens, thus indicating that another light maximum would follow at the second caustic crossing.

While continuing to monitor the light curve in order to predict the timing of this second event, the PLANET team contacted ESO with an urgent request to carry out a novel set of observations. The astronomers called attention to the unique possibility of performing detailed spectral observations during the second caustic crossing that could provide information about the chemistry of the stellar atmosphere of the magnified star . ESO concurred and within a day, their observing proposal was granted "Director's Discretionary Time" with the FORS1 spectrograph on the 8.2-m VLT ANTU telescope at the appropriate moment.

Some spectra were taken of the background star while it was still magnified, but had not yet made the second caustic crossing. The star was now identified as a cool giant star, located some 25,000 light-years away [3] in the general direction of the Galactic Centre (in the "Galactic Bulge").

Then the team waited. Their predictions indicated that the second caustic crossing might last unusually long, several days rather than a more normal 10-20 hours. The observing plan was therefore changed to ensure that spectra could be taken on four consecutive nights (ESO Press Photo eso0022) during this caustic crossing. The light curve would then first brighten, and then drop dramatically.

During the four nights, the lens would successively magnify different areas of the disk of the cool giant star while "the gravitational magnifying glass slowly moved across it", as seen from the VLT. First it would mostly be the light from the cool limb of the star that would be amplified, then the hotter middle of the disk, and finally the other, also cooler limb.

The VLT observations

On each of the four nights beginning on July 4, 5, 6 and 7, 2000, ESO astronomers at Paranal performed two hours of service observations according to the detailed planning of the microlensing team. Spectra were successfully taken of the giant star with the multi-mode FORS1 instrument at the 8.2-metre VLT ANTU telescope at the moment of the second caustic crossing. The magnitude was about I=13 at the brightness peak, dropping about 2 magnitudes towards the end of the period (ESO Press Photo eso0117).

In a first scientific assessment of these unique spectra, the team concentrated on an absorption line in the red spectral region (the "H-alpha" line) that is produced by hydrogen in the stellar atmosphere. They found a clear change in the strength of this line of the source star during the four nights (ESO Press Photo eso0117). No such variations were seen in the spectra of neighbouring stars that were observed simultaneously, providing a secure check that the observed changes are real.

The astronomers then went on to interpret this change. For this they performed various simulations by means of a computer model of the atmosphere of the cool giant star, applying the expected effects of the lensing and then comparing with the observed spectra. The expected changes in the strength of the H-alpha absorption line during the crossing from two simple simulations are plotted as lines over the observed data in ESO Press Photo eso0117.

The observed changes of the H-alpha line during the caustic crossing agree well with the model calculations . During this event, the microlens magnifies successive areas of the stellar disk particularly strongly. To begin with, the light from the relatively cool, leading limb of the star dominates the registered spectrum - and here the absorption line strength drops slightly, exactly as expected. It then becomes stronger as the hotter areas near the middle of the disk "come into focus" and then again decreases when the cooler trailing limb is strongly magnified. This is the first time that this effect has ever been measured for all phases of a caustic crossing.

More to come

More quantitative predictions of the modeling will now be carried out, refining the geometry of the caustic crossing and involving many more spectral lines. This will allow a sophisticated tomographic analysis of the atmosphere of this star.

For this, the detailed brightness measurements that were collected from over two thousand observations of EROS-BLG-2000-5 by PLANET observers in Tasmania, Western Australia, South Africa, Chile and the United States will be of great help in determining better the exact geometry of the event. In due time, the VLT spectra data will then make it possible to test directly the best models of stellar atmospheres now devised by astronomers.

Observations like these are very important because they allow detailed investigation of a stellar atmosphere other than that of the Sun. It is remarkable that this is based on the "resolution" of the disk of a star over 25000 light-years away, i.e. about 1.6 billion times more distant than our own Sun [4].


[1] Note the recent ESO Press Release 06/01 about the VLT Interferometer. Observations of binary stars that undergo eclipses from time to time also allow indirect studies of the surfaces of the two components; such objects, however, influence each other and cannot be characterized as "normal" stars.

[2] The team (the PLANET collaboration) consists of Michael Albrow , Kailash C. Sahu (Space Telescope Science Institute, Baltimore, MD, USA) Jin H. An (Dept. of Astronomy, Ohio State University, Columbus, OH, USA), Jean-Philippe Beaulieu (Institut d'Astrophysique de Paris, France), John A. R. Caldwell , John W. Menzies , Pierre Vermaak (South African Astronomical Observatory, Cape Town, South Africa), Martin Dominik , Penny D. Sackett (Kapteyn Astronomical Institute, Groningen, The Netherlands) , John Greenhill , Kym Hill , Stephen Kane , Robert Watson (University of Tasmania, Hobart, Tasmania, Australia), Ralph Martin , Andrew Williams (Perth Observatory, Australia), Karen Pollard (Physics Dept., Gettysburg College, PA, USA) and Peter H. Hauschildt (Dept. of Physics and Astronomy & Center for Simulational Physics, University of Georgia, Athens, GA, USA).

[3] The distance to the Sun is 149.6 million kilometres; 25,000 light-years = 240,000,000,000,000,000 kilometres. 1 billion = 1000 million.

[4] The diameter of the cool giant star is approx. 15 million km (about ten times that of the Sun). At the indicated distance, 25,000 light-years, this corresponds to a very small angle, about 10 micro-arcsec. This is equal to the angle subtended by a human hair (diameter 50 microns = 0.05 mm) at a distance of 1000 km.

Plus d'informations

Further detailed information is available at the PLANET website and in a research paper ("H-alpha Equivalent Width Variations across the Face of a Microlensed K Giant in the Galactic Bulge") that appeared in the April 1, 2001 issue of the "Astrophysical Journal" (available on the web at ApJL 550, L173 or astro-ph0011380).


Penny Sackett
Kapteyn Astronomical Institute
Groningen, Netherlands
Tel: +31-50-3634073

John Menzies
South African Astronomical Observatory
Cape Town, South Africa
Tel: +27-21-447-0025

Peter Hauschildt
Department of Physics and Astronomy, University of Georgia
Athens, GA, USA
Tel: +1-706-542-2860

Connect with ESO on social media

A propos du communiqué de presse

Communiqué de presse N°:eso0117
Legacy ID:PR 09/01
Nom:EROS-BLG-2000-5, Gravitational Microlensing, Spectrum
Type:Milky Way : Star
Milky Way : Cosmology : Phenomenon : Lensing
Facility:Very Large Telescope
Science data:2001ApJ...550L.173A


The light-curve of microlensing event EROS-BLG-2000-5
The light-curve of microlensing event EROS-BLG-2000-5
The sky area of EROS-BLG-2000-5
The sky area of EROS-BLG-2000-5
A VLT spectrum of EROS-BLG-2000-5
A VLT spectrum of EROS-BLG-2000-5
H-alpha line strength of EROS-BLG-2000-5
H-alpha line strength of EROS-BLG-2000-5