Sunday, June 14, 2009

Shrinking Betelgeuse

This is news I can't let uncommented, after all my recent posts about interferometry: In a recent Astrophysical Journal Letter, a team of astronomers lead by Charles H. Townes, Nobel laureate for his contribution to the development of the laser, report "A Systematic Change with Time in the Size of Betelgeuse".

Betelgeuse, or α Orionis, is the bright red star in the shoulder of the constellation Orion. In 1921, Albert Michelson and Francis Pease succeeded in measuring its diameter as about 50 milliarcseconds (mas), using an interferometer mounted onto the main telescope of Mount Wilson observatory. This was the first time the diameter of a star could be measured. 50 milliarcseconds is a tiny angle, corresponding roughly to the apparent size of an object 100 metres big on the Moon. But given the distance to Betelgeuse, this means the star, when put in the place of the Sun, would fill the solar system up to the orbit of Jupiter.

In the meantime, Betelgeuse could be resolved directly using the Hubble space telescope, and its distance measured to be 200 parsecs (650 light years), though there remains an uncertainty of about 25 percent. And it has been the object of many interferometric studies.

But given the diffuseness of the outer limits of this Red Supergiant star, which on average has a density less than 1/10000 the density of air, measurements of its diameter depend on the wavelength of observation, as the thin outer atmosphere has different transparency for different wavelengths. Here is a comparison of the shape of Betelgeuse, as reconstructed from interferometric measurements taken within two weeks at three different wavelengths in the infrared:

Source: Surface imaging of Betelgeuse with the Cambridge Optical Aperture Synthesis Telescope (COAST) and the William Herschel Telescope (WHT).

The two images on the right were reconstructed using data from the Cambridge Optical Aperture Synthesis Telescope (COAST), and the image on the left with data from the William Herschel Telescope (WHT). Each image has a side length corresponding to 100 mas, and one sees big differences in the diameter of the star, and of its surface features, depending of the wavelength at which it is observed.

This fact makes it difficult to compare directly angular diameters of Betelgeuse obtained with different instruments at different wavelengths. Now, the letter of Charles Townes' team reports on data collected with one instrument at one wavelength over the last 15 years. Using the Infrared Spatial Interferometer (ISI) on Mount Wilson, not far from where Michelson and Pease did their pioneering work, they observe at a wavelength of around 11000 nanometres, and write:

The diameters measured by the ISI are not particularly inconsistent with previous measurements, are rather accurate, and show clearly that the star has systematically decreased in size by about 15% over the past 15 years.

(Click for larger view – data from Townes et al., APJ Lett 697 (2009) L127-L128)

The graph shows their data (note that the zero-line is suppressed, so the change appears bigger than it actually is), together with a best fit by a parabola in grey, and the Michelson & Pease result in orange. Michelson and Pease estimated that the actual diametre of Betelgeuse could be larger by 10%-15% because of systematic errors in their measurement method. I have added the line in magenta, which shows, for comparison, the radius of the orbit of Jupiter, using the latest Hipparcos/VLA parallax distance to Betelgeuse to convert angles in actual distances.

Betelgeuse has been known to be a star with variability, but it seems that this change in size is odd and unexpected, even more so as its luminosity has been roughly constant. It's fascinating to me that such a change can be measured. As the APJ Letter concludes:

ISI measurements over the last 15 years clearly show a systematic change in the diameter of α Ori. This change may or may not be periodic; if it is, the period is likely rather long, perhaps a few decades. [...] It should be valuable to continue accurate measurements of α Ori's size and other characteristics in order to understand the dynamics involved in this striking change, and to have systematic long-term measurements of similar stars.


Phil Warnell said...

Hi Stefan,

It is interesting that this star could be actually shrinking so rapidl,y since I’ve read elsewhere it is expected to go supernova at some point. Many are speculating it will do so within the next thousand years with some saying it might have happened already with only the speed of light being the delay in us knowing about it. If it truly has shrunk 15% since the first measurements, I would wonder if estimates have been done in relation to this expectation.

It has been said that it would outshine the moon when it does and although since it’s not polar aligned with us not seen asto pose to be a catastrophic (environmental) threat . None the less, that doesn’t take into account how sensitive modern electronics are to even radiation bursts of a far less greater magnitude, such as our own sun puts out at times of elevated activity. Perhaps then it would be prudent to look into further hardening of devices where this could present itself as a danger.



Odysseus said...

Yeah, Betelguese is going to blow up! :D

Seriously, this is very intriguing. Even if the change was 'only' periodic, that would be a great confirmation of the theory of massive star development (they are supposed to periodically grow and shrink in their terminal years, aren't they?). Just a minor typo in your text: I guess they used 1100nm light rather then 11000nm.

Arun said...

Evidence of faster-than-light travel - Weight Watchers' commercials have obviously reached Betelguese and are having an effect - it is reducing its girth.

Arun said...

Hi Phil,

Another source of global warming just when we need it :)

Will a Betelguese supernova kill our space telescopes, or otherwise put limitations on space exploration?


Uncle Al said...

1) Betelgeleuse goes Type II supernova (locally) 21 December 2012. Both LIGO wiggle then Kip Thorne goes to Sweden. Super-Kamiokande gets a 10 second neutrno pulse. Positron decay in the supernova cloud finally arrives to blast everything electronic in orbit.

2) God's pre-production prototype for a water molecule. He should clean up afterwards.

3) Somebody's ballast resistor is decaying.

Neil' said...

Around 1970, I saw a cover on Analog magazine of a giant star which was red, white, and blue! IIRC that color trend went from surface to core. An article (not story, but maybe with tie-in) with title perhaps "The red, white, and blue giant" (sorry, no Google hits) said some stars might be like that - a hotter core that shines through much of the cooler, outer atmosphere. Cores are hotter anyway, sure, but it was unexpected that they would show through that well.

I don't know how well the idea held up. But different size patches for different wavelengths, like with Betelgeuse, imply a radial color dependency. It seems more like shorter is bigger and thus bluer instead of redder, but the patches are variegated so I'm not sure.

BTW, isn't the peachy, Mars-like color of Betelgeuse redder than a simple BB color temperature? We look at light sources that are about the same 3,500 K, and they don't seem quite that orange. Maybe H-alpha lines, etc? But REM that our atmosphere, even looking up and not the horizon, makes things a bit yellower than in space. Hence, the real color of the Sun is said to be greenish white. Any color reports from astronauts?

stefan said...

Hi Phil, Odysseus

it is interesting that this star could be actually shrinking so rapidly

Good question... I am not sure what to make of this size change, both in amplitude and time scale – more comments by someone with more expertise in stellar astrophysics than I have is welcome!

Actually, I've just tried to find out more about the changes going on in cepheid variables, and came across this paper: The Angular Size of the Cepheid ℓCarinae: A Comparison of the Interferometric and Surface Brightness Techniques.

This yields an interesting comparison: The star changes size by about 15%, from 2.7 mas to 3.2 mas and back to 2.7 mas - see figure 3 of the paper - within one period of only 35.5 days! Now, ℓCarinae is twice as far away than Betelgeuse, which means that its actual diameter at maximum is about 20% that of Betelgeuse. That's still big enough to engulf the Earth orbit! And it shrinks towards halfway to Venus orbit within two weeks or so, and then inflates again.

Compared to that, this change in size of Betelgeuse over 15 years is quite slow, even though the actual size of Betelgeuse is bigger. On the other hand, and of course, cepheids also vary in brightness in the same rhythm, while the brightness of Betelgeuse seems to have remained constant...

Quite puzzling...

Seriously, this is very intriguing. Even if the change was 'only' periodic, that would be a great confirmation of the theory of massive star development (they are supposed to periodically grow and shrink in their terminal years, aren't they?)

Ah, thanks for that remark. I don't know anything about the late-stage evolution of such giant stars. What is the time scale of such pulsations?

Concerning "going supernova", while this option sounds very dramatic, and exciting, I didn't mention it because, well, I thought that there are probably many other options. Actually, would a supernova of such a star really kick off with a shrinking over a time scale of ten years?

I guess they used 1100nm light rather then 11000nm.

It's indeed 11.15 µm they are using. I don't completely understand the technique of combination of the signals, it's seems not just adding the amplitudes from different telescopes. In the paper, they write: "The ISI operates at wavelengths near 11 µm, using heterodyne detection with CO2 lasers as local oscillators. This converts the infrared signal into a microwave frequency signal".

Cheers, Stefan

stefan said...

Now, ℓCarinae is twice as far away than Betelgeuse, which means that its actual diameter at maximum is about 20% that of Betelgeuse.

Oops, I did miss a factor of two somehow ... ℓCarinae is just about 10% as big in diameter as Betelgeuse. This then mens that it "eats" Mercury, but not Venus.. Sorry for the confusion.

stefan said...

Hi Neil'

this "The red, white, and blue giant", that sounds strange indeed. But how could one measure this "shell structure" in the 1970s? Perhaps the spectrum was a mixture of two blackbody components, and the star definitely not a binary?

If I understood that correctly, that Betelgeuse looks smaller in longer wavelengths is because the photo at longer wavelengths is less disturbed by reflecting clouds of dust in the outer atmosphere of the star.

About colors, I have always trouble to clearly discriminate colours of stars. OK, Betelgeuse is somewhat reddish, but that's it..

Cheers, Stefan

Neil' said...

Stefan, the article was a theoretical prediction. I don't think anyone then was yet empirically studying surface variations on the stars - ?

REM regardless of why one color image is larger than the other, superposing them leaves rings of differing colors. That reminds me, about my prediction of color fringes in diffraction disks imaged even by perfect optical systems: the red Airy disk is larger than the green which is larger than the blue, etc.

Phil Warnell said...

Hi Stefan,

Yes all very interesting. Two of several possibilities being this is a normal oscillation of such a star or perhaps it’s the final contraction before going supernova. The truth is there has never been an opportunity to study a star just prior to it going supernova, so the scenarios are based on models considering what we currently understand.

In looking around it seems even Red China has its eye on this red giant . Not that there is any reason to believe this is eminent, yet in the article I pointed to Charles Townes, out of Berkeley remarked, "measurements showed ... over 15 years, it has decreased in size about 15 percent, changing smoothly, but faster as the years progressed,". One would think if this where an oscillation or rebound effect the retraction speed would be slowing rather then accelerating. However as Townes also reminded more will be learned with what's revealed in the next few years.

One side of me says bring it on, for I don’t believe a Supernova event has happened so close in proximity to earth in all of recorded history and thus would certainly be a unique thing to witness. Then again since there are so many unknowns in such circumstances one is left to consider perhaps this is not something one would like to happen within one’s own lifetime.

For me this once againd serves to both demonstrate and remind how spectacular, powerful and awesome simply natural events can sometimes be and how concerns like those regarding the LHC or even climate change are rendered to be almost trivial in comparison. The fact is if the orientation of this star was so that it was polar aligned with our system ( which it’s not). it going supernova would render both of these concerns almost meaningless. It therefore appears that once again Goldilocks has come to the rescue, well we hope she has;-)



Arun said...

The Local Bubble

Anonymous said...


I recently read a book titled CHILLING STARS - 2nd edition. In it there is a discussion on how cosmic rays influence Earth's climate. Increased cosmic rays means increased global cooling due to increased cloud formation that would reflect the suns energy.

A supernova(<20 solar masses)/hypernova event would accelerate the interstellar gamma rays towards earth - increasing Earth's "apparent" cosmic ray load. Would it trigger another Ice Age?

Read the book and check out their conclusions - they track FE60 ions in the geological record - seeing the 1054 Crab Nebula footprint for example.

Anonymous said...


Climate change modeler's say that a 1% change in cloud cover is the difference between global cooling or global warming.