The unique cataclysmic variable BD Pavonis
The variable star BD Pavonis (RA 18h43m12s, Dec -57°30'44.2") entered the annals of history via the Annals of the Harvard College Observatory, when it was discovered on photographic plates taken at Boyden Observatory, Bloemfontein (Shapley et al. 1940).
The photographs were taken with the 10-inch Metcalf triplet, the 24-inch Bruce doublet and the 60-inch reflector, and were studied and archived at Harvard Observatory under the hawkish eye of Harlow Shapley. Assisted by two ladies, Emily H. Boyce and Constance D. Boyd, the emulsion-coated glass plates were painstakingly searched for variable stars.
The credit goes to Ms Boyd for finding the "new" star in Pavo, on plates taken with the Metcalf. A plate taken on 1934 September 08 showed the star at magnitude 12.4, whereas on an earlier plate it was fainter than the emulsion could record (dimmer than 15th magnitude). A plate taken on September 15 showed it had faded to 13.2 mag, and by September 28 it was again fainter than could be recorded. With four positive observations in hand, she then went though 308 other photographic plates to confirm that the star wasn't shown on any of them. It was announced in 1940 as a nova, Nova Pavo 1934.
It's classification as a nova continued through the 1970s, and it was included in Wyckoff & Wehinger's (1978) finder-chart catalogue of old novae.
Then in 1981, Barwig and Schoembs acquired a spectrum and carried out photometry of the star, at the European Southern Observatory's observing station at La Silla in Chile.
From the light curve they obtained a period of 0.1793015 days for the time between deepest minima. They also noted that "flickering, a characteristic property of cataclysmic variables, is found with timescales of minutes in the individual light curves."
They later published a thorough analysis of the star suggesting that it's classification be changed (Barwig & Schoembs 1983). Besides the rapid flickering observed in the light curve, analysis of its spectrum showed features similar to those found in dwarf novae. They considered it to be a dwarf nova, with an exceptionally bright (evolved) companion, undergoing rare outbursts and long-term spectroscopic variations.
In April 1995, Zwitter & Munari (1996) again studied BD Pav, now classified as a cataclysmic variable, with the ESO's 1.5-m telescope in Chile, confirming the photometric results obtained by Barwig and Schoembs.
The current understanding is that BD Pavonis is a dwarf nova as well as an eclipsing variable. In the latest (2004) version of the Combined General Catalogue of Variable Stars (Samus et al. 2004a) it is classified as a "UG:+EW/WD" type, the only known example of its kind on record. It is one of only four known deeply-eclipsing dwarf nova with orbital periods greater than 3 hours (Gansicke 2000).
Before we look at Monard's recent observations, let's take a moment to decipher that letter-spaghetti code "UG:+EW/WD" by asking, just what are cataclysmic variables?
Also called "explosive" or "novalike" stars, these variables show outbursts caused by thermonuclear burst processes in their surface layers. (In the case of supernovae, these explosions occur deep in their interiors). The majority of cataclysmic variables (CVs) are close binary systems and their components strongly influence each other's evolution (Samus et al. 2004b).
CV's come in the following flavours: novae (fast, slow, very slow, and recurrent types), supernovae (types I and II), dwarf novae (U Geminorum-type, SS Cygni-type and SU Ursae Majoris-type), Z Camelopardalis-type stars, and Z Andromedae (symbiotic) stars.
Dwarf novae (labelled "UG", for U Geminorum-type variables) are close binary systems consisting of a dwarf or subgiant star that has expanded, and a white dwarf (that's the "WD" in the spaghetti) surrounded by an accretion disk. Orbital periods vary from just over an hour to half a day.
The brightness of the system varies only slightly, and in some cases, rapidly, but from time to time the brightness increases rapidly by several magnitudes and, after an interval of from several days to a month or more, returns to the original state. Intervals between two consecutive outbursts for a given star may vary greatly. These systems are frequently sources of X-ray emission.
Some of these systems – as is the case with BD Pavonis – also undergo eclipses (hence the "EW" in the spaghetti code). According to the characteristics of the light curve, dwarf novae may be subdivided into three types: SS Cyg, SU UMa, and Z Cam.
It's commonly accepted that BD Pav is an SS Cygni sub-type; such stars are characterized by an increase in brightness by two to six magnitudes in one to two days, returning to their original brightness over several days; their activity cycles range from ten to several thousand days.
On the other hand, SU UMa-type variables show two types of outbursts: normal and supermaxima. Normal, short outbursts are similar to those of SS Cygni types, while supermaxima are brighter by two magnitudes, are more than five times longer and happen less frequently. During supermaxima the light curves show superposed periodic oscillations (superhumps), with periods close to the orbital period.
Here's where the plot begins to thicken. It was reported on 2006 August 29 that BD Pav was again in outburst. Berto Monard has been monitoring the star, and he thinks its classification as an SS Cygni type may not be correct. He writes:
"Looking at my light curves of the early phase of the present long outburst, it might not necessarily be [an SS Cygni type]. I made observations over four nights now; the one of last night was undisturbed by clouds ... and shows the least modulation in-between eclipses, except for the secondary eclipse."
The light curve above shows differential photometry on BD Pav (top curve), showing its brightness change over a 7.5 hour period. The lower curve shows the magnitude of the (constant) check star (R = 12.1 mag), three arcminutes south of the variable.
"BD Pav is an eclipsing system," he adds, "since we see it from side-on. The eclipses are fixed events during the orbit and are a perfect means to derive the orbital period of the system. I'd like to do further photometry on this system to make sure it is not of the SU UMa-type. The thing to look for is the development of superhumps. These are modulations of around 0.5 mag with an orbital period within 5% of that of the orbital period and possibly shifting."
So, join me in holding thumbs – for clear skies and superhumps.
I asked Berto to discuss the lightcurve he generated, and he provided some extra feedback too:
"The sharp eclipses are obvious and the interval between the minima determines the orbital period of the system. This primary eclipse refers to the eclipse of the hot white dwarf by the cool companion star (which donates matter to it). Halfway during those eclipses is visible the secondary eclipse where the white dwarf and the accretion disk are getting in front (partly obscuring) the companion star.
"Taichi Kato from VSNET, who also received my data, comments:
"We've received 4-nights data of BD Pav from Berto Monard. The light curves show eclipses with a period of 0.1792(1) days, in good agreement with a previous estimation (e.g., Barwig & Schoembs 1983). Other modulations are visible in the light curves, which are under investigation.
These variations outside primary eclipses look like to be the combination of secondary minima of the irradiated secondary and ellipsoidal variations, rather than superhump-like variations (the identification would require model fitting). Unfiltered CCD photometry of high-inclination long-orbiral period dward novae sometimes show this kind of orbital variation.
Nevertheless, the system clearly show prominent eclipses even during a bright outburst. The signature of the hot spot looks like to be weak (overall profile very different from that of U Gem).
"That describes it well. Actual superhumps are therefore not there and this system is in all likelyhood a ugSS [SS Cygni] dwarf nova undergoing a long outburst."
nothing more to see. please move along.