Brian Warner left Oxford for the U.S.A. about 33 years ago as a young Coude spectroscopist with a background in measuring element abundances on the Sun. He little suspected at the time that he would become a leading figure in developing the current theoretical picture of cataclysmic variables, viz. interacting binary systems in which mass transfer is taking place, in many cases via an accretion disk, onto a white dwarf primary star. The development of high-speed photometry in the 1960s and 70s, notably by pulse counting, was instrumental in providing the observational support for this theoretical advance. In the early 1970s Brian was also at the forefront in the discovery of superhumps during superoutbursts of SU UMa-type dwarf novae. His book on Cataclysmic Variable Stars (CUP, 1995) is the most comprehensive and authoritative review of the subject currently available.
There were about 100 participants (corresponding to the capacity of the lecture theatre) at the symposium, which was held in honour of Brian's 60th birthday, not primarily to review past developments but rather to address the current state of knowledge in the field of Cataclysmic Variables (CVs) with emphasis on active research areas and major outstanding problems. Each day there were several invited half-hour talks plus numerous short talks, with poster sessions in the morning and afternoon breaks. I was allowed to give a short talk on amateur contributions to CV research.
The main topic areas covered in the symposium were (after a general Introductory Session):
|Accretion Disks||Outburst Mechanisms||Superhumps|
|Magnetic Systems||Novae||Cataclysmic Variable Evolution|
|White Dwarf Binaries||SW Sex Systems||Supersoft X-ray Sources|
|Low States in Cataclysmic Variables||Properties of Cataclysmic Variables||What we still don't know about Cataclysmic Variables|
Below I summarise highlights from the talks, not comprehensively, but mainly according to what I could appreciate and understand best, and what seems likely to be of most interest to readers of the VSS Circular. Thus there is no attempt to cover detailed discussions of magnetic systems or of CV evolution, nor of some of the more esoteric theoretical discussions in other areas. The official Symposium Proceedings, which are intended to cover fully the subjects addressed, are due to be published later this year by Elsevier.
With respect to CVs themselves, Brian mentioned the diagram in which magnetic moment is plotted against accretion rate; polars appear at the top and full disk-bearing systems at the bottom, the intermediate region being populated by polaroids (e.g. U Gem, which is suspected of being in synchronous rotation, implying that the spin of the primary is locked to the orbital motion). A major problem is the cause of the angular momentum loss from binary systems - gravitational radiation is an inadequate explanation. Another major problem concerns the observed 2-3 hour gap in the orbital period distribution of CVs; it is uncertain whether this can be explained by a change in magnetic torque when the secondary star (usually a cool main-sequence star) becomes convective. The nova-like VY Scl stars (e.g. MV Lyr), which typically have orbital periods of 3-4 hours, are generally bright (a "high" state), but can suddenly fade (into a "low" state); the problem here is why they remain in a low state for significant periods of time.
With respect to the physics of accretion disks, Brian noted that the classical theory of boundary layers (where the disk material reaches the primary star) fails because the temperatures it predicts are too low; realistic disk models are probably beyond reach at present. Outburst mechanisms are not fully understood either; some intermediate polars (e.g. V1223 Sgr) show very short outbursts which are not due to the standard dwarf nova disk instability. Tilted accretion disks have been invoked to explain phenomena such as negative superhumps (superhumps with a period shorter than the orbital period); do these also occur in low-mass X-ray binaries? Another recently-observed phenomenon is that of non-radial oscillations in white dwarf primaries (e.g. in the dwarf nova GW Lib), akin to those observed in ZZ Cet stars. This requires further investigation. Finally Brian noted that the meeting held in Cambridge in 1975 had been seminal for CV research, notably in the area of common envelope evolution, which refers to the evolution of a CV system following a nova eruption. The material ejected in the nova eruption forms an extended envelope which exerts frictional drag on the secondary star. The envelope material is thus ejected from the system in an outflow concentrated near the orbital plane which removes angular momentum as well as mass from the system. This phase in the life cycle of CV systems is an area of current research interest.
Keith Horne (St. Andrews) spoke on the subject of "More Surprises from CVs". He first noted the significance of CVs in the more general astrophysical context, namely that studies of CV accretion disks provide useful insights into other systems (e.g. forming stars, active galactic nuclei), and that the problem of angular momentum transport in CVs is generic rather than specific to CVs. He noted that with respect to dwarf nova outbursts, the disk instability (DI) model makes specific predictions for the radial distributions of temperature and surface density. These can be tested by eclipse mapping (e.g. as in Z Cha). However the observed distributions are too hot and too thin compared with the predictions of the DI model. He suggested that chromospheric emission might play a role in producing the observed distributions. He also noted that IP Peg in decline from outburst differs markedly from the predictions of the DI model; there is no sign of the predicted cooling fronts.
Keith noted that for nova-like systems (with high mass accretion rates) one would expect gaussian (bell-shaped) eclipse light curves, but in fact SW Sex systems show more V-shaped eclipses (e.g. UX UMa, RW Tri), implying flatter temperature profiles than expected. These systems (e.g. BT Mon) also show anomalous single-peaked trailed spectra, known as the "SW Sex syndrome".
As for dwarf novae in quiescence, multi-wavelength eclipse observations of OY Car have allowed extraction of the white dwarf spectrum, but it looks unlike any "standard" white dwarf! The best fit to the observed spectrum is provided by assuming simple absorption by gas in the outer disk.
IP Peg has provided other surprises too. Emission line spectroscopy (presumably in quiescence) shows double-peaked lines which imply supersonic velocities in the disk. In outburst observations suggest that two spiral shock waves develop in the disk. These are predicted by dynamical analysis of particle motions, and the implied presence of gas out of the orbital plane may affect the interpretation of observations in the outburst state.
One of the most remarkable phenomena noted by Keith is the "magnetic propellor" effect exhibited, for example, by AE Aqr. Here the white dwarf spin period is 33 seconds, while the orbital period is 9.88 hours. Many freak phenomena are observed in this system, notably the 33 second oscillations which have now been under observation for 15 years, and the very large flares which are seen from time to time. The spin is slowing down, and it is believed that this is associated with an exit stream of material being flung outwards from the system along magnetic field lines anchored in the white dwarf; there is no accretion disk in AE Aqr. Keith speculated whether such a mechanism could work for SW Sex stars, with the magnetic field rooted in the accretion disk - in my view this cannot work because angular momentum exchange with accreting material will quickly eject the field-retaining material from the system.
The first talk under this heading was by Koji Mukai (Goddard Space Flight Center), on "X-ray Observations of non-Magnetic CVs in the ASCA Era and Beyond". He noted that ASCA had a much wider spectral range (0.4-10 keV in energy terms) than Einstein or ROSAT, and that the SIS instrument had a 2% spectral resolution. He then noted that soft X-ray/EUV is often seen in dwarf novae, SS Cyg and U Gem being the best-studied examples. He suggested that all BD (bright disk) CVs (dwarf novae in outburst, nova-like systems) probably have optically thick disks at temperatures of order 10,000 K, while QD (quiescent disk) systems have optically thin disks but a compact hard X-ray source; thus for example OY Car does not show soft X-ray eclipses during superoutbursts, but shows sharp X-ray eclipses (as does HT Cas) in quiescence. It is not clear how an optically thick boundary layer produces hard X-rays, as for example in the BD system V603 Aql.
John Hameury then spoke on "Disk and Secondary Irradiation in Dwarf Novae". He noted that the Shakura-Sunyaev disk viscosity parameter alpha cannot be constant in transient systems; really it is just a parameterisation of our ignorance of angular momentum transport in accretion disks! However the standard model in which alpha and the inner and outer disk radii are allowed to vary over the outburst cycle reproduces dwarf nova outbursts reasonably well. On the other hand it does not predict UV/X-ray delays in soft X-ray transients, nor the superoutbursts in SU UMa systems or in U Gem. The tidal instability theory introduced by Osaki can predict superhumps and the outburst pattern typical of SU UMa stars, but not the U Gem superoutburst of 1985 or the short supercycle of DI UMa. The speaker suggested that disk illumination and irradiation of the secondary star might play important roles, the former in explaining re-brightenings after superoutbursts and the latter increased accretion rates, and that these effects, together with inner and outer disk radius variations, should be included in future modelling of dwarf novae.
In a subsequent talk Erik Kuulkers (Oxford) addressed the subject "WZ Sge Stars, TOADs and SXTs: Close Encounters of the Same Kind". He considered these phenomena (SXTs are Soft X-ray Transients) in terms of five basic characteristics:
Steve Howell (Arizona) gave a talk entitled "TOADs have WARTs", explaining that WARTs are Weird Anomalous Red Things, or very low-mass, low-temperature secondary stars. He discussed the possible evolution of CVs in terms of orbital period, and then gave details of a K-band spectrum of the polar system ST LMi (which is below the period gap). The spectrum contained absorption bands due to water, implying a low surface temperature; he speculated whether there might be a dark region on the secondary star due to a star-spot group. He mentioned that HST optical and infra-red data on WZ Sge indicated a surface temperature of below 1700 K on the secondary star.
David Buckley, in speaking on "The Power in Intermediate Polars", noted that these systems show periodicities in the optical and X-ray regions, generally at the orbital period and at the spin period of the primary; thus YY Dra has an orbital period of 3.91 hours and a spin period of either 529 seconds or 550 seconds. In most IPs the spin period dominates the power spectrum. In some cases (e.g. TX Col) the power spectra change with time. In some systems the optical polarisation is modulated at the spin period of the primary. David noted that only two deeply eclipsing IPs were known (XY Ari, DQ Her); more would be very welcome!
Mike Shara (Space Telescope Science Institute) then spoke on "The Recovery and Characterisation of Old Magellanic Novae". In fact he covered work on globular clusters, both Magellanic Clouds and the galaxy M81 using archive plates and modern searches. Photographs of the globular cluster M80 (where the nova T Sco erupted) showed how incredibly difficult it is to locate the remnants of novae in such crowded fields. However the remnant of T Sco has now been identified by very careful astrometry, and the tally from 17 searches covering 47 Tuc and M80 is now 15-17 CVs, including two dwarf novae. As for the Large and Small Magellanic Clouds, there were 7 certain recoveries of novae (all but one in the LMC), with a further 4 or 5 possible recoveries. This is about half the expected total. The work on M81 was done with the help of archive plates only recently released to Mike by Allan Sandage. This resulted in the recovery of 24 novae in M81, mainly in the disk (spiral arms) of the galaxy, implying that such novae are relatively massive Population I systems. This is a very similar conclusion to that recently obtained by Ciardullo et al in a CCD study of novae in M31.
The light curve of one of the M81 novae studied by Mike enabled a distance measurement based on the MMRD (Maximum Magnitude/Rate of Decline) relationship for classical novae, which yielded a distance modulus of 27.8 (3.63 Mpc or 11.8 million light years). This is within 0.1 of the modulus obtained from observations of Cepheids in M81! However Mike said he would only be truly convinced that this agreement was not accidental if something like 20 such nova distance estimates were available.
In a short talk in the Nova session Thomas Zwitter (Lubliana, Slovenia) reviewed the photometric and spectroscopic data on the recurrent nova U Sco, which had been detected in outburst by Patrick Schmeer in February 1999. All three recent outbursts (1979, 1987, 1999) were photometrically very similar, but spectroscopically the 1999 outburst was very different from the 1987 one. In the recent outburst the width of the H-alpha line had decreased linearly with time as a result of thinning of the ejecta, and the white dwarf became visible again 22 days after maximum light. The profile of the H-alpha line at this point strongly suggested a high-velocity bipolar outflow of material from the white dwarf. There was some consensus that outbursts of U Sco may have been missed - its outburst interval may be shorter than the minimum of 8 years recorded to date. It is an eclipsing system with an orbital period of 1.3 days.
Knox Long then discussed "The White Dwarfs in CVs". He noted that it is often difficult to separate out the white dwarf contribution to the ultraviolet spectra of dwarf novae such as U Gem, VW Hyi and WZ Sge. The spectral quality obtained from IUE and HUT were inadequate for good results on white dwarfs, but the data from GHRS is much better. As a result the key parameters of some systems are now better determined than ever before; thus for U Gem the orbital period is 15285 seconds, the primary mass 1.04+/-0.04 solar masses, and the K1 velocity (the line-of-sight velocity amplitude due to orbital motion) 107.1+/-2.1 km/s. The main uncertainty in this and other dwarf novae is the systemic velocity (i.e. its motion through space). Figures for white dwarf temperatures were quoted for WZ Sge (14,800 K) and for RX And (35,000 K before an outburst, 36,000 K afterwards), but data on SS Cyg was more difficult to obtain using GHRS.
Coel Hellier (Keele) addressed the subject "The SW Sex Stars - Observations and Models". He noted that these systems have orbital periods in the range 3-4 hours, and that several are eclipsing. Examples are PX And, which has distorted emission-line wings in its spectrum, which can be modelled by stream overflow across the disk - there is a bright region extending well into the disk (V795 Her and LS Peg are similar). V1315 Aql shows noticeable absorption at phase 0.5; modelling by stream absorption in the outer/mid-disk region gives a reasonable fit to the observed spectra. Coel noted that the dwarf nova EX Hya behaves like a SW Sex star when in outburst; its eclipses can be modelled by a simple stream without a disk.
Paul Groot then spoke on "The Spectroscopy of SW Sex". If the out-of-eclipse spectrum of SW Sex in its excited ('high') state is compared with a standard emission-line spectrum, an extra blue continuum is seen corresponding to the disk hot spot. All of the absorption dips can be matched by the spectrum of a B0 star. The hot spot region is optically thick with a temperature of about 25,000 K. Trailed spectra of several lines show a second component due to the irradiated side of the secondary star. The speaker's concluding comment was to the effect that adequate eclipse mapping of this system has to be 3-dimensional, with an optically thick disk rim, though this would have to be done very carefully. [Comment: This seems to be related to the problem of mapping the disk in outbursting dwarf novae, notably IP Peg, which also requires a thick disk rim.]
Boris Gaensicke (Goettingen, Germany) gave on observational update in his talk "Towards a New Model for Supersoft X-ray Binaries (SSXBs)". In this context "supersoft" means that 90% of the photons from a source are below 0.5 keV, with an effective temperature of 10^5 - 10^6 K. Boris said that there were 11 optically identified SSXBs with known orbital periods, including T Pyx, V Sge and QR And. The nature of the central object is difficult to determine because of interstellar absorption. SSXBs are characterised by bipolar jets of up to 4,000 km/s from the central object, which is probably a white dwarf. In some supersoft X-ray sources there is no evidence that a secondary star is present at all. He noted that there are problems with current models of SSXBs, and is working on a "bootstrapping" model which involves evolution from short towards long orbital periods (contrary to the general course of CV evolution), although it is not yet clear how such a model will work. He posed the question of whether the classical SSXBs (QR And etc) are evolved CVs.
F V Hessman spoke on "The Symptoms and Origins of CV Low States". Low states are characterised by:
For those interested in the eclipsing dwarf nova EM Cyg, Rachel North (Southampton) displayed a poster showing trailed high-resolution spectra from which it was clear that the light from the system contains a contribution from a third star of late spectral type (and hence similar to the dwarf nova secondary star). This finding resolves long standing confusion over the masses of the binary components. It is not clear whether the third star is physically bound to the dwarf nova.
A poster by Steve Howell drew attention to the Faint Sky Variability Survey (FSVS), an ongoing photometric survey towards moderate to high galactic latitudes. It is intended to detect and/or place constraints on the number of faint CVs in the Galaxy. Theoretically most CVs should be faint with short orbital periods and have very low mass transfer rates. Such systems are strongly discriminated against in conventional colour-based surveys. The FSVS limit is a V magnitude of 24 and can detect V-band variabilty on time scales down to about 15 minutes. Watch this space!