Hard Xray Emitting White Dwarfs in Symbiotic Stars

Hard X-ray Emitting White Dwarfs in Symbiotic Stars: a Progress Report Koji Mukai (NASA/GSFC/CRESST & UMBC) Koji. Mukai@nasa. gov Jamie Kennea (PSU), Juan Luna (Cf. A), Jeno Sokoloski (Columbia) X-ray Univserse 2008 Granada, Spain 2008 May 29 X-ray Universe 2008

Symbiotic Stars A Symbiotic Star is binary containing a red giant with emission lines. In many cases, there is a white dwarf accreting from the giant, producing a blue continuum that ionizes the wind and other circumstellar matter. Some Symbiotic Stars are also recurrent novae: once every several decades or so, they undergo thermonuclear runaway. 2008 May 29 This talk concentrates on those symbiotic stars with a white dwarf companion (others, e. g. , GX 1+4, have a neutron star companion) X-ray Universe 2008 2

CVs and Symbiotic Stars Cataclysmic Variables (CVs) • Are semi-detached binaries with a white dwarf accretor • Have a K-M ~main sequence mass donor • Always accrete via Roche-lobe overflow • Have a typical orbital period of hours • Are usually dominated by accretion luminosity in the visible band • Have been well studied in the Xrays (see the rest of this session) 2008 May 29 Symbiotic Stars • Also have a white dwarf accretor • Have M giant mass donors • Are usually thought to be wind accretors (but question of high accretion rate) • Have a typical orbital period of years • Are usually dominated by the photospheric emission of the M giant in the visible band • Have not been well studied in the X-rays X-ray Universe 2008 3

ROSAT View: Muerset et al. ROSAT observations of symbiotic stars (Muerset et al. 1997) revealed several distinct types of X-ray emission. • Supersoft (e. g. , RR Tel) - detectable below 0. 5 ke. V, photospheric emission from nuclear burning white dwarfs • “Beta-type” (e. g. , AG Peg) - peaks suggests optically thin thermal emission with k. T~1 ke. V, interpreted as due to colliding winds • Others (e. g. , GX 1+4) - harder emission, in this case from accretion onto a neutron star 2008 May 29 X-ray Universe 2008 4

BAT detections of the Gang of 4 4 symbiotic stars have now been detected in the Swift/BAT survey (2 were also detected with INTEGRAL) Neither the supersoft or the “beta type” emission seen with ROSAT can explain these BAT detections So what is going on? 2008 May 29 X-ray Universe 2008 5

The curious case of CH Cyg, in addition to the “beta type” emission, showed a much harder component in ROSAT data (which, in fact, had been known from earlier satellites) 2008 May 29 X-ray Universe 2008 6

ASCA Spectrum of CH Cyg 2008 May 29 X-ray Universe 2008 7

ASCA Spectrum of CH Cyg Ezuka & Ishida (1998) found that the ASCA spectrum of CH Cyg to be • Clearly separated into two spectral components • Above about E~2 ke. V, there is a highly absorbed, hard component. – There is a clear detection of Fe K complex, indicating thin thermal origin – A single k. T (~5 ke. V) fit would work – But the effective bandpass was narrow (E~2 -10 ke. V) so only weak constraints can be placed on models – Reminescent of 2 -10 ke. V X-rays seen in CVs, indicating accretion onto the white dwarf as the probable origin • Below about E~2 ke. V, there is an unabsorbed, soft component – Several emission lines (Ne, Mg, Si) are detected, again indicating thin thermal origin – Origin of this component has since been a subject of lively debate 2008 May 29 X-ray Universe 2008 8

Recent & Ongoing Projects In the rest of this talk, we will present selected item from: • Kennea et al. (2008) summarizes works done with Swift – Analysis of BAT light curves and spectra – Follow-up using pointed XRT observations for source identification and joint BAT-XRT spectral fits – Hours~days timescale spectral variability • Additional works: – Suzaku observation of CD -57 3057 aka SS 73 -17 (Smith et al. 2008), Chandra and Suzaku observation of the same (approved) – Chandra DDT observation of RT Cru (Luna & Sokoloski 2007) – Suzaku observation of CH Cyg (Mukai et al. 2007), RT Cru, & T Cr. B (Luna et al. , work in progress) – RXTE/PCA monitoring of T Cr. B and CH Cyg – Chandra TOO observation of CH Cyg (approved) 2008 May 29 X-ray Universe 2008 9

BAT 9 -month Light Curve All 4 symibotics are variable. RT Cru appears to have been much brighter in hard X-rays in 2003 (INTEGRAL; Chernyakova et al 2005) CH Cyg went into a low state in the spring/summer of 2005 2008 May 29 X-ray Universe 2008 10

Two Types of Low States in CH Cyg From 2005 summer to present, CH Cyg appears to have been in a second type of low state: the soft X-ray probably is still bright while the hard X-rays appear to be low 2008 May 29 X-ray Universe 2008 During the 2001 March Chandra/HETG observation, both the soft and the hard component was weak (which in turn allowed detection of spatially extended Xray emission from the jet; Galloway & Sokoloski 2004) 11

Clues from the Fe K lines Both in high state (ASCA) and in 2001 low state (Chandra), the Fe complex is dominated by the He-like (6. 7 ke. V) line -the luminosity reflects accretion rate In Suzaku data, the 6. 4 ke. V line dominates: the direct line-of-sight is blocked and we only see the scattered component 2008 May 29 X-ray Universe 2008 12

RXTE Spectra of CH Cyg Currently CH Cyg is very faint for RXTE/PCA, but untillvery recently, it appears to have been dominated by the 6. 4 ke. V line. This may be changing (still a factor of ~10 lower than in ASCA data) 2008 May 29 X-ray Universe 2008 13

BAT+XRT Spectra of T Cr. B 2008 May 29 X-ray Universe 2008 14

BAT+XRT Joint Fits • Swift XRT has problems with CH Cyg because it is extremely bright in the optical • The other three can be fitted with a bremsstrahlung continuum plus a single Gaussian near 6. 5 ke. V, with a strong partial covering absorber. – Note BAT spectrum is average over 9 months, XRT data are from relatively short, pointed observations – More complex models are not warranted given the quality of the data – k. T~37 ke. V for RT Cru, 17. 2 ke. V for T Cr. B, and 17 ke. V for CD 57 3057 – BAT only fit gives k. T~5 ke. V for CH Cyg 2008 May 29 X-ray Universe 2008 15

Spectral Variability of RT Cru Swift/XRT spectra of RT Cru at three flux levels - the hard part is more or less steady while the soft part strongly varies 2008 May 29 X-ray Universe 2008 16

XRT Count Rate vs. Hardness 2008 May 29 X-ray Universe 2008 17

The Origin of Variability • All three symbiotic stars - RT Cru and CD-57 3057 definitely and probably also T Cr. B - show strong partial covering absorber • All three (again, not so definitely for T Cr. B) become softer when brighter • We do not see the beta-type emission in these three • We do not know why all 4 hard X-ray bright symibotics show so strong an intrinsic absorber. – This is not a general characteristic of symbiotic stars. If so, we would never see any supersoft component from symbiotic stars • Absorber must be close to the hard X-ray source, i. e. the white dwarf (remember these binaries have AU-type separation!) Circumbinary absorber shouldn’t be able to vary on hour timesclaes. 2008 May 29 X-ray Universe 2008 18

The Lack of Coherent Periods Although these symbiotic stars are all variable on short time scales, they do not show coherent periods • in X-rays or • in the optical Chandra light curves of RT Cru The lack of coherent periods strongly suggest the accreting objects in the hard X-ray symbiotic stars as a class are not magnetic white dwarfs 2008 May 29 X-ray Universe 2008 19

Spectra of the Hard Component • The hard X-rays are optical thin, thermal emission from the accreting plasma, as also seen in magnetic and non-magnetic CVs • These symbiotic stars have harder spectra than non-magnetic CVs • Yet they are unlikely to have a magnetic white dwarf • Non-magnetic CVs can have high temperatures if the accreting white dwarf is massive - SS Cyg at ~1. 1 Msun is the hardest among dwarf novae • By scaling from SS Cyg in outburst (ask me for details later), we estimate ~1. 35 Msun for T Cr. B and CD-57 3057, and ~1. 4 Msun for RT Cru • This agrees with the independent estimates for T Cr. B, which is also a recurrent nova. • Maybe the hard X-ray survey is a good way to find near Chandrasekhar mass white dwarfs in symbiotic stars 2008 May 29 X-ray Universe 2008 20

Spectroscopy of the hard component Suzaku XIS+HXD/PIN spectra of T Cr. B 2008 May 29 X-ray Universe 2008 21

Spectroscopy of the hard component Suzaku spectrum of RT Cru 2008 May 29 X-ray Universe 2008 22

Fe Kalpha Complex XIS clearly resolves the Fe K complex into three lines 2008 May 29 X-ray Universe 2008 23

Conclusions so far and Questions • Swift/BAT survey has identified 4 symbiotic stars as bright hard X-ray (>10 ke. V) sources. – Why these 4, and not others? RS Oph also is a recurrent nova with similarly massive white dwarf • The hard X-rays are thin thermal emissions, i. e. , from plasma accreting onto the white dwarf – Are our inference for very massive white dwarfs correct? • They all have strong intrinsic (partial covering and variable) absorbers – Where exactly is the absorber? It’s unlikely to be geometric if all hard Xray bright symbiotic stars have such an absorber • They are variable on timescales of hours to years, including two types of low states for CH Cyg • How many others like these are there? – Current BAT sensitivity is about ~2 x 10 -11, so pointed XMM-Newton (etc. ) observations or the e-ROSITA survey can detect many more 2008 May 29 X-ray Universe 2008 24