Ultra-Compact Binaries

1. Ultra-Compact Binaries Gavin Ramsay (MSSL-UCL) Pasi Hakala (Turku), Danny Steeghs (CfA), Tom Marsh (Warwick), Gijs Nelemans (Nijmegen), Paul Groot (Nijmegen), Kinwah Wu (MSSL), Ralf Napiwotzki (Herts), Mark Cropper (MSSL)

2. Outline: • What are Ultra-Compact Binaries? • How are they formed? • What is their astrophysical significance? • What is their observational properties? • Emphasis on RX J0806+15 and RX J1914+24 • What is their space density? • Outline the strategy & status of RATS project. • First results and future strategy.

3. What is an accreting binary? A binary system where components are so close they exhange mass I am going to discuss binaries where the compact and more massive star is a white dwarf. The evolution of binary depends on orbital period, the masses of both components and how evolved the secondary is.

4. What are `Ultra-Compact’ Binaries? aka `AM CVn’ stars Gansicke (2005) Accreting binaries with white dwarf primaries and main sequence secondaries have binary orbital periods greater than 80 mins. For shorter period systems the secondary must have degenerate or semi-degenerate. eg white dwarf - white dwarf binaries.

5. How compact are these binaries? Cataclysmic Variable with Porb~2hrs UCB with Porb~5min RX J0806+15 UCB with Porb~10min RX J1914+24

6. How are ultra-compact binaries formed? 3 main formation mechanisms: • Double white dwarf binary loses angular momentum via gravitational radiation - period gets shorter and shorter and starts mass transfer at Porb~2 min. Over time orbital period increases and mass transfer rate decreases. • Mass transfer takes place in white dwarf - helium star binary. At point where helium star becomes semi-degenerate the period starts to increase and mass transfer rate decreases. • From Cataclysmic Variables (CVs) with evolved secondaries. At point where secondary becomes a He-core, period increases and mass transfer rate decreases.

7. Nelemans (2005) In principle if you know a systems orbital period, the rate of change of its orbital period and its abundance, you can located it on the above plot and derive the relative importance of each mechanism.

8. The Astrophysical Signficance of UCBs I: Their space density is a sensitive test of binary and population synthesis models. Nelemans, Yungelson, Portegies Zwart (2004) The predicted number of systems brighter than V=20mag and X-ray flux greater than 10^-13 erg/cm^2/s (0.1-2.4keV)

9. The Astrophysical Significance of UCBs: II UCBs are predicted to be the very known sources that LISA is expected to detect. LISA => Gravitational Wave Observatory due for launch 2015+ Three spacecraft transmitting laser beams. By interferometric techniques they can measure the separation between the test masses. As gravitational waves pass by the distances between them will change. Aim to get accuracy of 1/10 size of an atom!

10. Optical Properties of UCBs: I The most obvious observational signature is their optical spectrum - lack of hydrogen lines. Spectrum of typical UCB (Anderson et al 2006) Spectrum of typical hydrogen accreting binary (de Martino et al 2006)

11. Optical Properties of UCBs: II Optical Photometric Properties Can be split into systems in low state, high state and those which undergo outbursts. Systems in outburst and in a high state show charateristic modulations in their optical light curve. HP Lib(Porb=18.4min) Power spectrum shows peaks corresponding to beat periods between orbital period and spin period of the white dwarf. Patterson et al (2002)

12. XMM-Newton observations of UCBs: 3 medium resolution X-ray detectors ~0.1-10keV 2 high resolution X-ray detectors ~0.3-2.4keV Optical/UV telescope - filters range from 2200A-6000A

13. The XMM-Newton view of UCBs: I XMM-Newton observations show they have wide range of UV modulation characteristics and no evidence of any periods in X-ray data. A surprise! Ramsay et al (2005)

14. The XMM-Newton view of UCBs: II Ramsay et al (2005) These spectra are be modelled using multi-temperature thermal models with highly non-solar abundances - with typically large amount of nitrogen required to get good fits.

15. The XMM-Newton view of UCBs: III Prior to XMM-Newton it was thought most of the accretion luminosity would be emitted in X-rays. 10^30 erg/s Ramsay et al (2006) HST parallax programme allowed accurate distances for many systems. We find that most of accretion luminosity emitted in UV and is in good agreement with predictions. This is first time that these predictions have been verified by observations. Suggests that UV band a good place to discover new systems.

16. The candidate systems RX J0806+15 and RX J1914+24 RX J1914+24 - V407 Vul (569sec) RX J0806+15 (321sec) Faint Phase Israel et al (2002) Bright Phase Ramsay et al (2002)

17. XMM-Newton spectra: RX J0806+15 RX J1914+24 Ramsay et al (2006) Soft blackbody, kT~60eV Lx~1e32 erg/s for d=500pc Thermal plasma model T~0.2keV with highly non-solar abundances plus edge at 0.83keV. Lx~1e33 erg/s for d=1kpc

18. Characterising their orbital evolution RX J0806+15 ROSAT 1994-1995 VLT+NOT 2001-2002 NOT 2003 INT+NOT 2003-2004 Hakala, Ramsay & Byckling (2004)

19. RX J0806+15 Hakala, Ramsay & Byckling (2004) Spinning up at a rate of 1.9x10-11 s/s

20. Spin up in RX J0806+15 (II): Spinning up at a constant rate of 1.9x10^-11 s/s.

21. RX J1914+24 Ramsay et al (2005) Spinning up at a constant rate of 3.2e-12 s/s.

22. More periods in RX J1914+24? Previously known that the long term X-ray flux varied (Ramsay et al 2000) Now find evidence for power at 556 and 585 sec at some epochs. Not clear if this is due to beat between the dominant period at 569 sec and a longer period or secular variations (Ramsay et al 2006).

23. Optical Spectra - very different! RX J1914+24 RX J0806+15 Israel et al (2002) Weak Helium lines Steeghs et al (2006) • Looks like a G star. • Radial velocity limits rule out period < 14 hrs • A triple system? • A chance line-of-sight?

24. Radio (6cm) observations of UCBs VLA observations: RX J1914+24 => no detection, 75µm/beam RX J0806+15 => 5sig detection 0.1mJy ATCA observations: ES Cet (10.3min) => no detection, 24µm/beam Further observations of RX J0806+15 planned. Significance of these observations will be explained shortly!

25. Models put forward to account for these observations An intermediate polar:(Norton, Haswell & Wynn (2004) - The periods are the spin period of the white dwarf so the spin up is not a problem. Problems – lack of strong optical emission lines. Would expect the secondary to show up. Could be a double degenerate IP. Regarded as unlikely. A double degenerate polar:(Cropper et al 1998). - A strongly magnetic accreting white dwarf. Its spin period is locked with the binary orbital period. Problems – lack of strong optical emission lines, polarisation and hard X-rays. Can’t be excluded but regarded as unlikely.

26. Direct impact model:Marsh & Steeghs (2002), Ramsay et al (2002) A double degenerate system where the accretion stream impacts the accreting white dwarf directly. Would expect optical emission lines. In M&S interpretation the optical emission is from cooling X-ray tail. Works in relatively narrow parameter range. Unipolar inductor or electric star model:Wu et al (2002) A double degenerate system in which a non-magnetic white dwarf transverses the magnetic field of a magnetic white dwarf causing large currents to be driven causing heating of the white dwarf. Strongly circularly polarised radio emission predicted.

27. More on unipolar-inductor model: Sounds far fetched by it has been seen on Jupiter! Can see tracks where satellitesmagnetic field lines enter atmosphere of Jupiter.

28. Status of RX J0806+15 and RX J1914+24 The weak detection of radio emission in RX J0806 is consistent with UI but the non detection in RX J1914 appears to rule out UI in this system. This is a similar conclusion to that reached by Willes, Wu & Ramsay (2006) and Dall’Osso, Israel & Stella (2006) who take the observed spin-up rates and use the UI model to predict their luminosities and lifetimes. Conclusion: RX J0806+15 candidate for UI system. RX J1914+24 an odd triple system?

29. The search for new UCBs Test whether theoretical models overpredicts their space density or whether there are many more systems which await discovery. UCBs, especially with Porb < 25 min, show prominent variations in the optical. Exisiting wide field surveys such as the Faint Variability Survey could have missed them because of their relatively coarse sampling. The RApid Transient Survey (RATS) takes 30 sec exposure of sky for 2-3 hrs, giving time resolution of a few mins. Our pilot survey took place using the INT and the WFC in Nov 2003. 12 fields were observed including one which included RX J0806+15.

30. INT Wide Field Camera Four 2x4k EEV CCDs ~0.5x0.5deg field. White light observations, made near zenith

31. Field selection: • No stars brighter V~10 • Galactic Latitude 15-30deg • Close to zenith to prevent differential extinction. • Use optimization code to select fields.

32. Expected objects: • Ultra-compact binary systems • ZZ Ceti systems • Pulsating sdB stars • Cataclysmic variables • Precataclysmic variables • Delta Scuti systems • W UMa systems • Algols • Flare stars • Extrasolar planets • Eclipsing systems • Asteroids • New phenomena

33. Detection strategy: • Perform aperture photometry on all sources relative to comparison star. • Run Discrete Fourier Transform on all light curves. • Select variable sources based on significance parameter. • For `interesting’ sources, followup spectroscopy required to indentify their nature.

34. RX J0806+15 - test field

35. RATS I - results • 46 new variables discovered. • Most of these were W UMa or contact binaries. • 3 eclipsing systems • 2 Flare stars • 4 objects with periods less than 70 mins. • 16 minor planets found - 5 not previously known. • Results in Ramsay & Hakala (2005).

36. Ramsay & Hakala (2005)

37. Followup observations of the 4 sources with Mod<70min Mod Amp=25mmag Period ~40min Mod Amp=50mmag Period ~374 sec Period ~66min Mod Amp=40mmag Period ~60min Mod Amp=0.1mag Ramsay, Napiwotzki, Hakala & Lehto (2006)

38. RAT J0455+1305 - A high amplitude sdB star Spectral Fits: T~29200K Log g~5.2 Cool end of EC 14026 instability strip. consistent with high amplitude modulation. Ramsay, Napiwotzki, Hakala & Lehto (2006)

39. 3 New SX Phe - dwarf Delta Scuti stars Spectral Fits: T~7200-7500K This together with their modulation amplitudes, and periods & infrared colours show they are SX Phe stars. Only 20 `field’ SX Phe stars currently known. RATS good strategy for discovering rare pulsators. Ramsay, Napiwotzki, Hakala & Lehto (2006)

40. Further data from INT and ESO 2.2m telescopes: INT data for another 10 fields. Analysis still in progress but more candidate SX Phe stars found. ESO analysis complete: 20 fields, most with 20<b<40. Over 200 new RATS discovered. Candidate SX Phe from INT2 data. SAAO 1.9m spectra of brighter objects A handful of fields have b<10, and 2 had fields containing the large globular clusters M4 and M22. Software improved to account for field crowding.

41. Observations of M4 and M22. Did not sample cluster core, but did sample stars well within their tidal radius. M4: 24 new variable objects M22: 74 new variable objects Many contact binaries, but also shorter period variables and many eclipsing systems. AAOmega spectra of these fields. We have discovered a handful of accreting binary systems - but a hydrogen accreting rather than helium accreting. New accreting source in M4

42. Summary and Status of RATS project: • We have discovered more than 300 new variable objects. • A small number of these are rare pulsating stars. RATS is therefore a good strategy for discovering them. • Followup spectroscopy has been obtained using AAT & SAAO for brighter targets. • Up to 10 eclipsing accreting binaries have been discovered. • Currently no ultra-compact binaries have been discovered. • Early indications imply that models over estimate their number. • However, since models suggest they should be concentrated in the galactic plane, future fields need to be biased towards Galactic plane.