Blogs / Aaron Price's blog / New radial velocity measurements w/interesting implications
New radial velocity measurements w/interesting implications
We have a guest blog article from a coauthor of a recent paper published about epsilon Aurigae. The author describes the system overall for new participants, then describes what they discovered with their radial velocity data.
After 113 years of radial velocity measurements, epsilon Aurigae remains as mysterious as ever. 20 years of new radial velocity observations have been combined with historical data to improve our understanding of the orbital mechanics of the system, but nature of the objects themselves remains uncertain.
This eclipsing binary has the longest period on record – a whopping 27 years! There also seems to be no optical emission from the eclipsing object; but a two-year eclipse duration means this dark companion must be several AU wide. Its mass must also be similar to that of the star itself. There are models for the system, the most accepted being that the companion is a thick, dusty disk. The problem is that we don't have very good constraints on the spectral type of the primary. As you can imagine, in order to have a good idea of the size and mass scales involved we need to know what this primary object is!
In our paper appearing in the March 2010 edition of the Astronomical Journal, we use these radial velocity data to improve the existing orbital solution for the system. Whilst doing so we also identify short-term oscillations which may be due to motion in the stellar envelope. We present two orbital solutions; one based purely on radial velocities and one where the period is tied to the photometric eclipses. Interestingly, we see a nine-month difference in the predicted mid-eclipse. Whether the short-term oscillations or perhaps some other noise may be the source of this, we are unsure, but if not then this could imply that the eclipsing object and the gravitating companion are not one and the same.
The full article is now available at http://www.iop.org/EJ/
- Justin Lovegrove is a physics graduate student at the University of Southampton in the U.K and did some work on this paper while at the Harvard-Smithsonian Center for Astrophysics. He is a coauthor of the paper and presented a poster on the topic at the January 2009 meeting of the American Astronomical Society. A preprint of the paper is available here.
So this suggests that the disc could be elliptical right? No suprise thereperhaps as this was suggested by Lambert and Sawyer as a possible reason for the asymetry in the potassium 7699 line radial velocity with time plot. The only problem is this resultappears to be thewrong way round (L&S proposed that the centre of mass was offeset towards the trailing edge of the disc. If I read it rightthis result suggests it iscloser tothe leading edge. Fascinating! Robin
Amazing and puzzling.... Let's assume the system parameters given here: http://www.citizensky.org/sites/default/files/SystemPropertiesTable.png So the disc is supposed to be relatively light-weight. Even if you would concentrate all the mass of the disk right at the ("photometric") leading edge, it will not be enough to shift the center of gravity enough to cause the 9 month gap (mentioned above) between the photometric mid-eclipse (when the central opening of the disc with the B-star(s) is supposed to be in our line of sight) and the time when the RVdata seems to suggest the center of gravity of the disk + Bstar should be in our line of sight. So I think this time gap cannot be explained just by an elliptical disc. Right? CS HBE
Leaving aside the possible central hole (still subject of debate in some quarters, I believe), it might just be possible. 9 months back from 5 Aug 2010 takes us back to 5Nov 09 when the centre of mass would be directly in front of the F star. Thislooks just possible from the photometry, but explaining the radial velocity curve would be a problem. A comet like object might be a better fit but the change in sign of RV from leading to trailing edge seems tofavour a rotating disc. Robin
Or, perhaps, the inclination of the disk's orbit together with the tilt of the disc would be such that a substantial part of the leading half of the disc doesn't even eclipse the F-star at all. If only the trailing part of the disc fully eclipsed the F-star while the leading half passes mostly "above" or "below" the F-star's disc as seen from Earth, this would be more consistent with the new analysis of the F-star RV data, right? Again the central brightening would not fit in as easily as in the conventional model. CS HBE
Not so sure that it would work that way round. Remember that for a disc in aKeplerian orbit, the part of the disc closest to the B star has the highest orbital velocity and so it is this part of the disc which defines the maximum leading and trailing velocities seen in the RV curve. There might be some effect though if the outer parts of the trailing edge missed the star as this would bias the RV measured on the outgoing side towards inner (higher RV) parts of the disc. Take a look at my trendsof the development in different parts of the K I 7699 line to see what has been happenning so far. http://www.threehillsobservatory.co.uk/astro/spectra_40.htm Robin