Could the cause of the dimming NOT be because of eclispsing material?

Hi Tony,Many people have studied this most mysterious star system. There are a great many papers published on it and lots of theories. While new ideas are always welcome I suggest doing some research and see what the current thinking is and why. Epsilon Aurigae is a star system called an eclipsing binary even though only the F star has been seen. No spectrum of a second star. But something huge is coming between the F star and the Earth and takes around 2 years to pass by. With each new set of observations more is learned and the puzzle gets closer to being solved. Bob Stencel's (Dr. Bob) recent interferometry work at CHARA on Mount Wilson in California may provide the best evidence yet as to what the eclipsing body is. If you wish to learn more about epsilon Aurigae I suggest you check the Campaign web site at http://www.hposoft.com/Campaign09.htmlWe will be publishing the next Campaign Newsletter (#15) in a couple of weeks. All the Newsletters are available as downloadable pdfs.JeffCounting PhotonsHopkins Phoenix ObservatoryPhoenix, Arizona USAphxjeff@hposoft.com

Tonyome,Good questions. As part of my Ph.D. dissertation on eps Aur, I've been reading through the literature and as Jeff points out there have been several explanations for the eclipse. Although I've read only a fraction of the literature (there are over 420 papers on eps Aur), I've seen some fairly esoteric solutions. For instance, Ludendorff proposed a "swarm of meteors" was causing the eclipse. Kuiper, Struve and Stromgren (some big names in astrophysics) proposed that the eclipse was caused when a hot star went behind a giant "infrared star."Like yourself, I too was unconvinced about the current interpretation of the system and your question affords me an opportunity to review some literature and cement my understanding of the system. But I've got to agree with you, the current proposed solution seems too elaborate, doesn't it? Because nobody, including myself, has read all of the literature (found a paper Dr. Bob hadn't read the other day, so I think I'm safe saying this), I'll try to summarize a few facts about the system and offer the broader interpretations. After that I'll wrap stuff together into the current theory/model of eps Aur. Please let me know if you would like additional elaboration on any of these items.So here are a few facts about the system with reasonable (i.e. party-line) interpretations.1) Epsilon Aurigae exhibits out-of-eclipse light variations of ~ 0.2 mag.This means that eps is a variable star. The easiest explanation for this is that the star pulsates, just like Cephied variables.2) The out-of-eclipse variations are non-periodic.Most Cephied variables follow a well-defined period-luminosity relationship so several attempts have been made to find the period of the pulsation. I've actually participated in this work as well and found prominent periods that appear to shift, something Cephieds don't do too much. Using data from Boyd and Hopkins, we have very good coverage of the V-band since the 1983 eclipse. If one were to analyse the data as a whole, there appears to be a peak around ~100 days, but if you break the 27 years of data into two segments, this peak shifts. A strong statistical analysis was done by Gary Henson as part of his Ph.D. thesis, and he concluded that the star shows no consistent period. 3) Polarization measurements show variable out-of-eclipse variations.Gary Henson's thesis was actually on analysing polarization data taken after the 1983 eclipse. What he found is that there is evidence (but no strong conclusions) that the F-star may be a non-radial pulsator. I talked briefly about this in my talk, basically it means that the star doesn't pulsate uniformly in a radial direction, but instead develops beat patterns (see http://www.univie.ac.at/tops/dsn/texts/nonradialpuls.html for a quick overview), possibly in the l=2 or l=3 modes.3) Once every ~27 years there is a significant (i.e. ~0.8 mag) drop in light that lasts ~15 months.If this were traditional pulsation, we would expect the event to be periodic on an ~15-month timescale or some (small) integer multiple of ~15 months. So we're then left with a problem, if it isn't "normal" pulsation what is it? A simple solution would be to have something, like an exo-planet orbiting the F-star. Several planets have been discovered observed dimming in the light-curve of the parent star, or by spectroscopic shifts.4) The in-eclipse light curve is exceptionally flat bottomed and sharp-cornered.The problem with transiting planets is that they cause smooth changes in the parent star's light. Some of the planets move by so quickly that they don't really have a flat bottom to the eclipse (see http://brucegary.net/book_EOA/WebEOA/ExoplanetObservingAmateurs.html), or for those that stick around a while, the ingress and egress light curves are smooth (this isn't a good example, but the interpolated curve displays the principle I'm getting at: http://www.ursa.fi/sirius/HD209458/HD209458_eng.html). 5) The light curve dictates that we find something that can eliminate ~50% of the star's light for two years.So the object needs to be big.6) Spectroscopic orbit solutions imply that there is a second equally massive object in the system, but spectroscopic observations show *almost* no signatures of a companion.Here's the zinger. There is no evidence for a companion object if you look in the normal places (i.e. visible light). In the infrared we see an odd emission line that should be there if there is only an F-supergiant. We also see an ~600 K blackbody in the mid-infrared which was reported as infrared excess during the 1983 eclipse. UV spectroscopy (which can only be done from very high in the atmosphere or in space) also shows evidence for something being there, but it's difficult to pin down. (There is a paper in-the-works that explicitly says what it is... the results were made public during the Adler conference, but the authors didn't post their presentations online because the results have yet to be published so I'll refrain from mentioning that information out of professional courtesy).The thing is that for something equally massive to the F-star, we would expect to see more spectroscopy than what we do see. If this system were a classic binary system, the companion object should be equally luminous to the F-star.7) Spectroscopic observations imply something rotating by as the eclipse progresses.As the eclipse begins, some spectroscopic lines show additional red-shifted absorption, which becomes blue-shifted after eclipse. The lines in the F-star are broadened by the F-star's own rotation, but the odd absorption characteristic that I mentioned shows up only during eclipse.----So we are left with the question. What could be equally massive to the F-star, obscure 50% of the light from the F-star, have *almost* no spectroscopic signature (but be ~600 K), and go rotating by as the eclipse progresses? Su-Shu Huang figured out a method of fitting the light-curve alone, thereby explaining the flat-bottomed, sharp cornered nature of the eclipse. He proposed an opaque object with a (projected) rectangular appearance passing in front of the star. This model was subsequently revised to fit the mass constraints and observed spectroscopic rotation curves to become a dark disk-like object. To explain why we can't see the contents of the disk, it was proposed that instead of a single star inside of the disk (akin to proto-planetary disks), that there be two smaller stars whose total mass nearly equals the F-star.The disk model, as far fetched as it may seem, appears to fit the observed data.As for your question about Hubble observing the star, it can't and wouldn't help. First the "can't part": eps Aur is too bright. It would saturate Hubble's cameras nearly immediately. Dr. Stencel looked into using Hubble's UV cameras for doing spectral work back in the 80's and 90's, but found that the exposure times required to keep from saturating were so short it would have been difficult to organize and subsequently justify. Now for the "wouldn't help" part. Due to recent interferometric observations, we know the angular diameter to be about 2.3 milliarcseconds. Hubble's resolution is limited to about 7 milliarseconds (if I recall correctly). So even to Hubble, eps Aur would appear as a single pixel.There is hope though. Interferometry has the highest available resolution with effective mirror diameters in the hundred of meter range. Dr. Stencel and I just observed at CHARA about two weeks ago and obtained data. We have additional observations scheduled for December and we intend to re-apply for observing time next season. With CHARA we can observe objects as small as 0.5 milliarcseconds using the longest baselines. Well, I hope this answers your questions. There are still several issues I didn't even begin to discuss and would be glad to type about them if someone brings them up.Brian

That is all very interesting Brian. When my Astronomy club friends and I were discussing this last week, they were saying that there was no spectroscopic signature of a companion, but from what you said, that isn't entirely true. It does show that something is there, but very difficult to detect. Very strange.One thing I'm a little puzzled about with Hubble, and perhaps you can explain, I remember seeing news reports of Hubble viewing Jupiter, Uranus, and Neptune, and recently, wasn't Hubble even focused on the moon when they crashed that spacecraft into some craters looking for water? These are far brighter than EA, so I don't get it. When Hubble focused in on the "Pillars of Creation" for example, it could distinguish things invisible from earth. It's a shame it can't do the same with EA.

Tonyome,The spectroscopic "no signature" thing is really a quick way of summarizing the situation during certain times. Most of the lines that I know about are absorption lines that don't show up until the eclipse happens. Some of the lines, like those from carbon monoxide, don't show up until mid-eclipse implying some form of asymmetry.You make a good point with the Hubble observations. If I recall the problem was with wanting to use the UV camera with such short exposures, but I'll have to check with Dr. Stencel who considered the proposal back in the day. Perhaps I don't have the entire story behind that one.

Hi Brian,I think the main problem with Hubble is it is hard to justify observing epsilon Aurigae with it. As you noted, interferometry provides much better resolution so there is really very little Hubble could do that ground based photometry and spectroscopy cannot do. It would be nice to get photometric observations during the time of year it is difficult for ground based observer (very high air mass). Some think epsilon Aurigae goes behind the Sun during the summer, but it does not. It is above the plane of the solar system and could be observed year around. It may be more politics than actual problems for Hubble.Jeff

Hi!I guess like many, I'm by now a bit confused as to what is positively known and what is accepted theory (because it fits observations, but cannot be directly deduced from it).Spectroscopic evidence. Let me summarize ans check if I got this correctly:I can think of three different kinds of spectroscopic evidence concerning the secondary:a) traces of emissions from the secondary ==> weak evidence b) traces of absorption the primary's light by secondary during eclipse ==> weak evidence c) "wobble" of radial velocity of primary indicating presence of a secondary in the first place ==> strong evidence for itSo we can assume for a fact that there IS a secondary which must be almost invisible. But if or how this secondary plays a role in eclipsing the primary is not known with equal certainty, right? It's convenient to construct a primary that can eclipse the primary and "eclipse itself", tho.As for the mass of the primary: if read about papers discussing a low-mass and a high-mass model, so I wonder how certain the characterization of the primary as an F-supergiant is. Will the CHARA measurements be able to close this question once and for all?Cheers Heinz

It has been pointed out that care must be used when talking about primary and secondary objects. Because the eclipsing body is 10 times larger than the F Star it could be considered the primary. To be clear it is better to talk about the F star as one object and the eclipsing body as the other object and not use primary and secondary terminology.Jeff

Bikeman,I understand the confusion. You correctly summarized the spectroscopic evidence:1) There is absorption during the eclipse that shifts from blue to red.2) Few (one that I know of) emission lines both in/out of eclipse. In a "normal" F-supergiant, the line, Brackett Alpha at 4.05 um, would be an absorption line, but in eps Aur it's an emission line. See Backman 1985 PASP for further details.3) There entire star wobbles, causing spectroscopic shifts in the F-star's spectrum. This is fairly well defined, but the out-of-eclipse variations cause some noise which causes the orbit to be hard to fit to the photometric mid-eclipse date.So we can safely assume that there is something tugging at the F-star, therefore there must be a two components (note, the masses are almost 1:1 so I refrain from using "primary" and "secondary" to refer to the system).The current party-line interpretation is that the F-star is eclipsed by a disk-like object (a disk for the reasons I mentioned above).As for the CHARA measurements, I can't really comment too much about them because we're going to try to publish the results in a prestigious journal. I can say that previous interferometric measurments (i.e. those from NPOI and PTI) indicate an ~2.27 milliarcsecond (mas) F-star. At the Hipparcos distance of 625 (+/- 150?) parsecs, either the high-mass or low-mass model could fit.

I was wrong, Hubble has observed eps Aur at least once:http://adsabs.harvard.edu/abs/1999PASP..111..829SI haven't read the paper, but it's in my queue (which presently has 60 papers).

That is interesting about Hubble - I wish they could get a visual photo of it mid eclipse. I wonder if Hubble could detect the small stars that are theorized to be at the heart of the disk that are invisible from telescopes on earth.

Unfortunately the resolution of Hubble is too low at ~7 milliarcseconds [mas] per pixel. Interferometers have measured the angular diameter of the F-star to be ~2.2 mas so the entire F-star would appear as a single pixel on Hubble (if it were even observable without saturating the new, more sensitive camera).Fortunately, Dr. Stencel and I have recently observed eps Aur with the CHARA interferometer (resolution 0.5 mas) and have a great result. We have additional observing time in December (weather and road conditions permitting) and hope to get a second "snapshot" of the system. Note, interferometers don't take pictures or images. Image must be reconstructed from the data with several assumptions made along the way that could change the outcome of the reconstructed image. Think of it as figuring out what the image is in a 1000 piece puzzle with only 100 of the pieces and the knowledge of exactly where those pieces would belong.We won't be able to share our result with the CS community until after it has been published in a journal (publications get us funding, funding pays the bills...), but I can tell you that it is absolutely amazing!