Cataclysmic variables : a "cannibal" white dwarf as a companion

Among binary stars, made of two stars orbiting around each other, astronomers already know that some, called cataclysmic variables, may be very peculiar. In such systems, one of the two members is not an ordinary star but a very dense star resulting from the death of a brilliant star, a "white dwarf" with the size of a planet like the Earth but with the mass of our sun. A white dwarf is so dense that its gravity is enough to pull out matter from its close-by companion. When this matter falls onto the "cannibal" white dwarf, it releases a lot of energy causing what sometimes looks like a "cataclysm", a sudden increase in the luminosity of the star. 
Starting from 1976, it was realized that among the most energetic of these cataclysmic variables were those containing a magnetic white dwarf, a white dwarf with a magnetic field several million times stronger than most typical stars like the sun. In this case, the matter falling from the companion is forced to follow the lines of the magnetic field and is channelled by this way through a column down to very narrow spots, the magnetic "poles" of the star. For this reason, the magnetic cataclysmic variables (MCVs), are sometimes called "polars". The same mechanism is at work on the Earth when particles from the Sun are forced to follow the much weaker Earth magnetic field lines, ending up to the North and South poles and causing the well known spectacular and colourful auroras.
In the magnetic cataclysmic variables, the energy concentrated in such small regions is so high that the temperature reaches more than ten millions degrees and the aurora light is emitted mostly in X-rays, a very energetic form of light. More than 60 MCVs have already been discovered and they may represent about 10% of all the binary systems containing a white dwarf in a close orbit.

An artistic view of a magnetic cataclysmic variable : matter from a companion star falls onto the poles of the white dwarf along the field lines of the strong magnetic field. (Credit M. Garlick)

An overabundance of nitrogen

The discovery of an anomaly in the element abundances among some MCVs came in 1987 when J.M. Bonnet-Bidaud and M. Mouchet obtained the first ultraviolet observations of the source BY Cam (a MCV located at an approximate distance of 800 light-years in the Camelopardalis-Giraffe constellation) with the NASA/ESA IUE satellite. The spectrum of the source revealed a very strong emission "line", an excess of light at a given energy characteristic of nitrogen. Astronomers consider that these strong lines are produced when the matter falling to the white dwarf is deeply irradiated by the X-ray flux emerging from the poles of the star. Atoms illuminated by the powerful X-ray light are forced to loose some of their electrons, when they recombine or recapture their electrons they emit light at a given energy, different for each atom. The strongest lines for the CNO (carbon, nitrogen, oxygen) elements appear in the ultraviolet and X-ray light. The IUE spectrum was sufficient to determine that the nitrogen line was about ten times stronger than in any other sources while another important carbon line was simultaneously much weaker. But the astronomers had to wait till the advent of a new satellite, FUSE (Far Ultraviolet Spectroscopic Explorer), operating in the far-ultraviolet to be able to reach an important line of the oxygen element and to carefully compute for the first time the balance between the three CNO elements.

The FUSE satellite, one of the first missions observing the far-ultraviolet light and a collaboration between U.S.A., Canada and France (CNES).

The exact measure of the abundances is normally a very difficult task since for each of the many elements, one has to consider all possible levels of excitation. To do the job, the astronomers have made use of a powerful tool called a "plasma code", a complex computer program including thousands of atomic data on each element. The program called "CLOUDY" was developed at the University of Pennsylvania, for a simple homogeneous medium and a very simple geometry. In the context of the present work, it had to be deeply adapted to the complex shape of the falling stream of gas where the density of the gas is strongly variable. The astronomers were then able to reproduce the intensity of the lines in different conditions. First they checked if the abnormal intensities of the lines could be the effect of an unexpected specific irradiation but they were forced to conclude that the only way to produce the lines was by changing significantly the proportion of the different chemical elements in the gas. Compared to the "universal" abundances which are representative of the "normal" matter in the Universe, the amount of nitrogen (N) had to be multiplied by 25, the carbon (C) divided by 8 and the oxygen (O) only slightly lowered by a factor 2. This is the first quantitative measure and the results are totally unexpected. So large differences are rarely observed in stars and this opens new questions on the origin of the CNO elements.

The CNO factory

In the context of the Big Bang model, only very few elements are produced at the beginning of the Universe, mainly hydrogen, helium and a tiny fraction of the lightest elements lithium, beryllium and boron and therefore life could not exist at that stage. It was only in 1957, with the reference work of Burbidge, Burbidge, Fowler and Hoyle, a "bible" of more than hundred pages, known as BBFH and still widely accepted, that it was demonstrated that all the other elements are cooked into stars and later re-injected into space. This is the reason why life can be considered as made of "stardust". The most abundant atoms are the three CNO elements which represent more than 3/4 of all the elements, apart from hydrogen and helium. Their formation is however very different. While carbon and oxygen are produced directly by fusion reactions into stars (carbon is for instance the result of the fusion of two helium), nitrogen is only a secondary product. It is synthetized in a by-product reaction of the fusion of hydrogen into helium called the "CNO cycle", during which part of the existing carbon (C) (and to less extent oxygen O) is transmuted into nitrogen (N). The measure of the abundances in MCVs which shows an overabundance of N together with an underabundance of C (and also O) is totally in line with an origin in a very efficient CNO cycle. The question remains however to find the place where these reactions could have taken place.

The "universal" cosmic abundances (shown in logarithmic scale). Apart from hydrogen and helium, the CNO elements are the most abundant species but their exact origin in the Galaxy are still unclear (click the image to enlarge).

The first hypothesis was to consider the fusion of freshly deposited hydrogen at the surface of the white dwarf. Such reactions are however very violent and end up with an explosion know as a "nova", with a sudden brightening followed by an expulsion of gas, leaving behind a very hot white dwarf, with a temperature of the order of 1000 000 degres. This is probably not the case in BY Cam for which no such record exists and which has a rather "cool" temperature of 22 000 degrees. Moreover, recent observations done with the Hubble Space Telescope have now also shown that the same abundance anomaly is probably found at the surface of several other MCVs such as MN Hya and V1309 Ori. Though a quiet "nova" explosion is still a plausible explanation, it can hardly happen so often as to affect a significant fraction of the MCVs. 
Instead a more general alternate explanation is now emerging. Recent observations with the european X-ray satellite XMM have just shown that a few other very different binary systems also show an overabundance of nitrogen. This is the case for a source called Her X-1, a well known binary system where the compact star is a neutron star as well as another source, XTE J118+480 which is suspected to harbour a black hole. It is therefore more probable that the peculiar CNO abundances are linked to a particular evolution of the companion star. The elements are normally produced well inside the star and shielded by the outer layers. They can be made visible only if they have been efficiently brought from the core to the surface by an unknown mechanism or if the outer layers were totally lost. The two hypothesis are actively studied. As the majority of stars are in binary system, these mechanisms which have been so far neglected  can lead to a significative revision of the origin of CNO in the Galaxy.

Publications

"The CNO problem in magnetic cataclysmic variables"
J.M. Bonnet-Bidaud & M. Mouchet (2003)  in "Magnetic Cataclysmic Variables", IAU Col. 190, Capetown, Eds.: M. Cropper & S. Vrielmann (e-print astro-ph/0302158).

"The surprising Far-UV spectrum of the polar BY Camelopardalis"(*)
M. Mouchet, J.M. Bonnet-Bidaud, E. Roueff, K. Beuermann, D. De Martino, J.M. Desert, R. Ferlet, R. E. Fried, B.T. Gaensicke, S.B. Howell, K. Mukai, D. Porquet, P. Szkody (2003), Astronomy & Astrophysics (in press) (e-print astro-ph/0302312).

"Far-UV FUSE spectra of peculiar magnetic cataclysmic variables"(*)
M. Mouchet, J.M. Bonnet-Bidaud, E. Roueff, M. Abada-Simon, K. Beuermann, D. de Martino, J.M. Desert, R. Ferlet, R. Fried, B. Gaensicke, S. Howell, K.Mukai, D. Porquet, P. Szkody (2003) in "Magnetic Cataclysmic Variables", IAU Col. 190, Capetown, Eds.: M. Cropper & S. Vrielmann (e-print astro-ph/0302319).

(*) These works are a collaboration between : Paris-Meudon Observatory (F), Astrophysical Department CEA-DAPNIA (F),  Astrophysical Institute of Paris (F), Sternwarte Göttingen (D), Osservatorio di Napoli (I), Braedside Observatory (USA), Dept of Astronomy Southampton (GB), Astrophysics Group Tucson (USA), NASA Goddard (USA), Max-Plank Garching (D), Astronomy Dept Seattle (USA)