Understanding How Stars Die

Q: So what are AGB stars and why did you want to study them?

A: AGB stars, or asymptotic giant branch stars, are low- or intermediate-mass stars like our Sun that are at the end of their lives. At this stage, these stars have become red giants — they’re cooling off, creating extended atmospheres, and they’re losing a lot of mass in a dense stellar wind. They also periodically undergo pulses — about every 10 000 to 50 000 years — which blow material off the surface of the star at a much faster rate than normal. This helps create large shells of gas and dust, and eventually, these stars become planetary nebulae.

We have a basic idea of this mass-loss process but we don’t know many details, in particular how this mass loss is initiated close to the surface of the star. So we wanted to find out more. Different types of AGB stars, like carbon-rich or oxygen-rich stars, have different properties. The mass-loss process is theoretically best understood for carbon-rich stars, so we decided to closely study the carbon-rich AGB star R Sculptoris — and our results are the start of a detailed understanding of what happens.

Q: What can studying AGB stars tell us about the Universe?

A: Studying AGB stars is important to understand stellar evolution. The amount of mass lost by the star actually changes its evolution and creates different types of planetary nebulae. Understanding these processes is important because AGB stars are one of the main producers of dust in the Universe, which means they enrich the Universe with elements.

AGB stars are one part of the element-making puzzle because in the process of their death they can produce a vast range of elements — including 50% of elements heavier than iron. These elements are blown into the Universe to make new stars, new planets, new moons...to create the building blocks of everything else. To understand the stellar evolution process, we need to understand how these elements are created and released into the Universe.

Astronomers captured this ghostly image of AGB star R Sculptoris using ESO’s Very Large Telescope Interferometer (VLTI). The image covers a very small section of the sky: approximately 20x20 milliarcseconds. For comparison, Jupiter has an angular size of approximately 40 arcseconds.

Credit: ESO/M. Wittkowski (ESO)

Q: What did you find out?

A: We were looking specifically at the region close to the stellar surface of the AGB star R Sculptoris, which is the region where mass-loss is initiated. We found that R Sculptoris has one dominant bright spot on its stellar disc, two or three times brighter than the other regions. This contrast is very large so we wondered how to explain it. We know that these stars have large convection cells of moving gas on their surface, but these wouldn’t produce such a large contrast. Moreover, for AGB stars we expect that such detailed structure would be obscured by the extended atmosphere and dense stellar wind. Previous radio observations of R Sculptoris with the Atacama Large Millimeter/submillimeter Array (ALMA) showed an interesting spiral structure within the stellar wind much further out from the star, which hints at the presence of a previously unknown companion star, cutting through the dust as it orbits. However, the distance to this companion is too large to cause the structure close to the surface of R Sculptoris that we observed.

We compared our results to atmosphere and wind models, which predict that these convection cells on the photosphere are also related to mass loss and dust formation. This means that large convection cells with low contrast will lead to asymmetric dust formation — we’d get big blobs of dust forming, instead of a spherical dusty shell. This helped us to interpret our results: we realised we were seeing dust two or three stellar radii out from the star’s surface, forming not uniformly but in large clumps. The bright spot we saw is actually a spot where there is little to no dust, and we can look deeper into the stellar surface, where it’s brighter. The remaining parts appear darker because the starlight is blocked by the forming dust. Our “bright” spot is not actually inherently brighter, it’s simply a region that is less obscured by dust!

A: First of all, there aren’t many known carbon-rich AGB stars. We also needed to use a star that has the right size and the right brightness to observe it. As I mentioned, ALMA has previously found a spiral structure around R Sculptoris, which reveals a mass-loss history at much larger distances — the star had already thrown off large amounts of dust and transported it out to large distances.

The spiral gives us a lot of information about the mass loss history, including how the mass was lost, at which rate, and at what velocity. The observations we took using the Very Large Telescope Interferometer (VLTI) are complementary to the ALMA observations. They show us the present state of the star because we can look directly at the part of the star where the dust forms, very close to the stellar surface. That’s too close for ALMA, which can only see the spiral of dust that occurs at distances a dozen of time larger.

I also find the technique we used very exciting because we can use it to see the surface of stars, which until very recently we could only do for our own Sun.

ALMA observations revealed an unexpected spiral structure in the material around the old star R Sculptoris, probably caused by a hidden companion star orbiting the star.

Credit: ALMA (ESO/NAOJ/NRAO)/M. Maercker et al.

Q: Tell us more about the technique you used to make your observations.

A: We used interferometric imaging to look at R Sculptoris. Imagine if you drop a stone into a lake and it makes a pattern of ripples radiating out through the water. Then you drop a second stone into a lake, which will create a second pattern, and the two will interfere at some point. This interference is what we’re interested in.

Optical interferometry is like the double slit experiment that people do in high school physics, where you combine two beams coming from the same light source and look at the interference pattern produced. In interferometry at the Very Large Telescope, we combine light from different telescopes — sometimes up to a hundred metres apart — and it gives us a higher spatial resolution. It’s like observing a star with a 100-metre telescope. The resultant interference gives us information on very small spatial scales. If we combine a lot of these observations, we can actually reconstruct the image of the star.

Interferometry is not an easy technique, but recent advances in observation efficiency and precision, as well as image-reconstruction techniques, allow us now to image stars other than the Sun. We tried several image-reconstruction methods and they all gave the same results, so we are quite confident that the images we reconstructed are correct.

Inside the tunnels housing instruments for the Very Large Telescope Interferometer (VLTI)

Credit: ESO

A: Usually, if you look at a star with a single telescope, it’s so small that it appears to be a single point. From a single telescope, anything smaller than 30 to 60 milliarcseconds can’t be resolved. Astronomers use arcseconds to measure the angular sizes of objects on the sky; the Moon, for example, is around 30 arcminutes or 1800 arcseconds. Most stars are smaller than 30 to 60 milliarcseconds — R Sculptoris, for example, has an angular size of about 10 milliarcseconds, so it’s much smaller than what we could resolve with one telescope alone. But if we use interferometry to combine the observations of multiple telescopes, we can dramatically increase the resolution we can obtain! With the VLTI and its PIONIER instrument, we have observed scales of one or two milliarcseconds, so we can look at the details on the surface of the star.

One of the VLT’s Auxiliary Telescopes (AT) looks up at the stars of the Milky Way

Credit: ESO/José Francisco Salgado (josefrancisco.org)

Q: What’s next in this area of research?

A: My colleagues and I are so excited about these imaging results coming out — not only for R Sculptoris but also for other similar stars, such as red giants and red supergiants, some with less dust around them. We’re getting a lot of new results in a lot of different wavelengths for more stars. We are now at a point where we can obtain resolved images of a variety of stars, so that’s quite exciting.

We’re planning a workshop at ESO to discuss these results and put them all together. It’s a very exciting time at the moment because we can finally produce these images at different wavelengths. The models I mentioned are also progressing thanks to advances in computing power — we can now calculate and simulate the environment around a star, including the dust formation, in three-dimensions! You can see these clumps of dust in the models too, which is relatively new. So it’s a very exciting moment to bring all of these results together and plan the next steps.