Planets around evolved stellar systems Tom Marsh Department of Physics, University of Warwick Tom Marsh, Department of Physics, University of Warwick Slide 1 / 35
Tom Marsh, Department of Physics, University of Warwick Slide 2 / 35 Outline 1. White dwarfs 2. White dwarf binary stars 3. ULTRACAM 4. Eclipsing white dwarfs 5. Timing measurements 6. A role for variable star observers? 7. Conclusions
Tom Marsh, Department of Physics, University of Warwick Slide 3 / 35 White Dwarfs The most common end-point of stellar evolution; 95% of all stars > 1 M become white dwarfs. A typical white dwarf is comparable in size to Earth, but with a mass similar to the Sun, they have mean densities ρ 1 tonne per cc. There are around 1000 million white dwarfs in our Galaxy. NGC 6543 (Cat s Eye), with nascent white dwarf (HST)
Tom Marsh, Department of Physics, University of Warwick Slide 4 / 35 White dwarfs & Type Ia Supernovae Exploding white dwarfs are thought to make the Type Ia supernovae (2011 Physics Nobel Prize dark energy ): 1. Add mass to a white dwarf 2. As mass nears 1.4 M, the white dwarf shrinks and compresses. 3. Carbon/Oxygen material undergoes runaway fusion... Khoklov et al (1997)
Tom Marsh, Department of Physics, University of Warwick Slide 4 / 35 White dwarfs & Type Ia Supernovae Exploding white dwarfs are thought to make the Type Ia supernovae (2011 Physics Nobel Prize dark energy ): 1. Add mass to a white dwarf 2. As mass nears 1.4 M, the white dwarf shrinks and compresses. 3. Carbon/Oxygen material undergoes runaway fusion... Khoklov et al (1997)
Tom Marsh, Department of Physics, University of Warwick Slide 5 / 35 Tycho s supernova was a Type Ia Remnant of an exploded white dwarf in X-rays (Tycho s SN of 1572, now about 30 light-years across). Light-echo from Tycho s SN shows that it was a Type Ia
Tom Marsh, Department of Physics, University of Warwick Slide 6 / 35 White Dwarfs in Binary Stars Modern astronomical surveys are producing large numbers of white dwarfs in binary systems. So many, that even the exclusive class of eclipsing systems is growing rapidly Number of eclipsers known vs time
Tom Marsh, Department of Physics, University of Warwick Slide 7 / 35 CVs accreting white dwarfs Cataclysmic variables are a long-lived phase of evolution in which mass transfer is driven by weak angular momentum loss. They give us the chance to measure the build-up or otherwise of mass on white dwarfs.
Tom Marsh, Department of Physics, University of Warwick Slide 8 / 35 Detached WD+MS systems Simple systems with spectra which are a combination of white dwarfs and low-mass main-sequence stars. Can be detected through simultaneous flux excess in ultraviolet and red filters. Around 2000 now known, mostly from the Sloan Digital Sky Survey. Pyrzas et al (2009), spectra of 4 eclipsing WD/dM systems.
Tom Marsh, Department of Physics, University of Warwick Slide 9 / 35 ULTRACAM White dwarfs are small, with orbital speeds at their surfaces of 4000 km s 1. They must be observed fast to resolve variations, typically 1 to 30 sec. In May 2002 we commissioned a new high-speed camera to facilitate observations of white dwarf binary stars. ULTRACAM mounted on the 4.2m WHT in La Palma
Tom Marsh, Department of Physics, University of Warwick Slide 10 / 35 ULTRACAM To reduce light-losses due to filters, in ULTRACAM we split the light up into three bands, UV, green and red, before imaging. The detectors used are frame transfer CCDs in which one exposure is taken while the previous one is being read out. Can take 100s of frames/second with high efficiency. Timestamps are taken from the GPS. ULTRACAM design
Tom Marsh, Department of Physics, University of Warwick Slide 11 / 35 ULTRACAM ULTRACAM and two astronomers ULTRACAM at the 8.2 m VLT, Cerro Paranal, Chile
Tom Marsh, Department of Physics, University of Warwick Slide 12 / 35 White dwarf / main-sequence eclipsers Exquisite measurements are possible on large telescopes (the VLT in this case). [Magnitude: V = 17.0, eclipse length: 10 mins, exposure time: 1.5 sec.] Parsons et al (2010a), GK Vir, a hot WD + 0.1 M M-dwarf eclipser.
Tom Marsh, Department of Physics, University of Warwick Slide 13 / 35 White dwarf + very low mass star Very dim companions are outshone by the white dwarf at all wavelengths. Rise in between eclipses caused by heating of cool star by the white dwarf. The dip at the top of the lightcurve occurs as the white dwarf transits the heated face. T W = 35, 300 K, M W = 0.51 M, M C = 0.09 M. Parsons et al (2011), CSS 03170, P = 94 mins
NN Serpentis In the best cases, complete solution of M W, R W, M R, R R and i is possible. e.g NN Ser M W = 0.535 ± 0.013 M R W = 0.0211 ± 0.002 R M R = 0.111 ± 0.004 M R R = 0.149 ± 0.002 R i = 89.6 ± 0.2. Parsons et al (2010b), NN Ser. Tom Marsh, Department of Physics, University of Warwick Slide 14 / 35
Tom Marsh, Department of Physics, University of Warwick Slide 15 / 35 Testing Equations of State These are good enough to test models of both white dwarfs and very low mass stars (and, potentially, brown dwarfs), themselves dependent upon the behaviour of high density matter. Parsons et al (2010b)
Tom Marsh, Department of Physics, University of Warwick Slide 16 / 35 One motivation for high-speed work is to study the angular momentum loss that drives binary evolution. A steady rate of period change Ṗ alters the times of eclipse by a quadratic function of time t = 1 Ṗ 2 P t2. Timing studies In practice, rather erratic variations always seem to be the rule. Period changes in WD/dM system QS Vir, O Donoghue et al (2003)
Tom Marsh, Department of Physics, University of Warwick Slide 17 / 35 Applegate s Mechanism Applegate (1992) suggested that such variations are driven by variations in the shape of the MS star driven by solar type cycles. No angular momentum is lost in the process; Applegate s mechanism is not a driver of evolution. Applegate s mechanism does require energy however. Variations in shape alter the gravitational attraction between the stars
Tom Marsh, Department of Physics, University of Warwick Slide 18 / 35 Violating Applegate We have found that some systems show too large a period change for the MS star to have supplied the energy needed. Here the eclipses in QS Vir are seen to arrive 200 sec earlier than expected from O Donoghue et al. s data. It seems certain that there is a third body orbiting the binary, probably a brown dwarf ULTRACAM Last 3 O Donoghue times ULTRACAM QS Vir, Parsons et al (2010)
Tom Marsh, Department of Physics, University of Warwick Slide 19 / 35 Third bodies from timing Unseen object Unseen object different time Eclipse arrival time delayed or advanced
Tom Marsh, Department of Physics, University of Warwick Slide 20 / 35 Planets! NN Ser is the best measured system to date. Its timing variation in NN Ser can be fit to within very small errors with two planets, M c sin i = 6.9 M J, M d sin i = 2.2 M J in 15.5 and 7.7 yr orbits. Two planet fit to NN Ser times, Beuermann et al (2010)
Tom Marsh, Department of Physics, University of Warwick Slide 21 / 35 Planets around a CV Here, M c sin i = 6.3 M J, M d sin i = 7.7 M J, with periods of 16 and 5.3 years. System here is accreting and so may be subject to variability-induced scatter. In optimum cases such as NN Ser, timing is sensitive to planets of a few Earth masses in long period orbits. Two planet fit to UZ For times, Potter et al (2011)
Tom Marsh, Department of Physics, University of Warwick Slide 22 / 35 Liverpool Telescope Project Even well-measured systems can be hard to pin down, and well-spaced sampling is needed. Start using robotic Liverpool Telescope data to extend coverage. Some possible orbits of NN Ser given the data in hand.
Tom Marsh, Department of Physics, University of Warwick Slide 23 / 35 First LT results (started Feb 2011) Upper panel: an eclipse of NN Ser; lower panel: QS Vir times. More than 30 eclipses of different systems measured.
Tom Marsh, Department of Physics, University of Warwick Slide 24 / 35 All systems show timing anomalies... RR Caelum RXJ2130.6+4710
Tom Marsh, Department of Physics, University of Warwick Slide 25 / 35 Are the planets really there? Orbital stability can place strong constraints upon multiple-planet systems. In NN Ser, the planets seems to be close to a 2:1 resonance. New data has significantly shrunk the allowable parameter space, but is still consistent with the 2:1 resonance.
Perhaps not always... Tom Marsh, Department of Physics, University of Warwick Slide 26 / 35
Tom Marsh, Department of Physics, University of Warwick Slide 27 / 35 Test with double eclipsing white dwarfs Pairs of eclipsing of white dwarfs should be much less vulnerable to stellar-induced timing noise than white dwarf/main-sequence binaries. Although the chances are low, the first such system, NLTT 11748, was discovered last year. 0.2 M + 0.7 M pair of white dwarfs in a 5.6-hour, at 0.1 from being exactly edge-on. The eclipses are total but only 6% and 3% deep.
Tom Marsh, Department of Physics, University of Warwick Slide 28 / 35 A second eclipsing DWD This year we discovered a second with 40% and 10% deep eclipses using Liverpool Telescope and Gemini data. Remarkably 4 are now known. If these systems show planets, it will be a firm indication of their reality. Watch this space!
Tom Marsh, Department of Physics, University of Warwick Slide 29 / 35 Origin of the circum-binary planets Systems like NN Ser are only a solar radius or so apart, with orbital periods of 10 hours or less. However prior to the formation of the white dwarf they were 1.5 AU apart, raising problems with stability of the planets. Two possibilities: 1. The planets predated the white dwarf and spiralled in as the white dwarf lost its envelope 2. The planets formed out of the material lost by the white dwarf.
Tom Marsh, Department of Physics, University of Warwick Slide 30 / 35 Calling Variable Star Observers In several cases we are limited by coverage with large gaps. Some systems have variations of order a minute or more. There are a few systems V471 Tau, QS Vir, DE CVn, RXJ2130.6+4710 and (for Australians) RR Cae within reach of small telescopes. Please contact me at tom.marsh@warwick.ac.uk if you are interested.
Tom Marsh, Department of Physics, University of Warwick Slide 31 / 35... and a couple of longshots... White dwarfs could be totally eclipsed by any planets they host. Two interesting possibilities: G24-9 aka V1412 Aql: V = 15.75 white dwarf with two reports of 2 mag. eclipses. Arno Landolt requested an AAVSO campaign in Feb 2009. No reports of success, but very interesting to continue this. EG 131: V = 12.3, a white dwarf but reported as varying by ±0.35 mag.
Tom Marsh, Department of Physics, University of Warwick Slide 32 / 35 V471 Tau with ULTRACAM V471 Tau (V = 9.8) is bright, but lacks nearby comparison stars (ULTRACAM FOV = 5 to 6 ).
Tom Marsh, Department of Physics, University of Warwick Slide 33 / 35 V471 Tau with ULTRACAM V471 Tau (V = 9.8) is bright, but lacks nearby comparison stars (ULTRACAM FOV = 5 to 6 ). The eclipse is much stronger in the UV than in red light, so with ULTRACAM we can use V471 Tau as its own comparison!
Tom Marsh, Department of Physics, University of Warwick Slide 34 / 35 V471 Tau with ULTRACAM V471 Tau (V = 9.8) is bright, but lacks nearby comparison stars (ULTRACAM FOV = 5 to 6 ). The eclipse is much stronger in the UV than in red light, so with ULTRACAM we can use V471 Tau as its own comparison!
Tom Marsh, Department of Physics, University of Warwick Slide 35 / 35 Conclusions Large surveys are hugely increasing the number of eclipsing white dwarf binaries enabling precision parameter studies. These require high time resolution and large apertures to exploit fully. The same studies give precise eclipse times. Almost all eclipsing detached systems show timing anomalies that can be interpreted in terms of planets. It remains to be seen whether this will stand the test of time.