Transcript X-ray Astronomy and the search for Black Holes

Hard X-ray Emitting White Dwarfs
in Symbiotic Stars: a Progress Report
Koji Mukai (NASA/GSFC/CRESST & UMBC)
[email protected]
Jamie Kennea (PSU), Juan Luna (CfA), Jeno Sokoloski (Columbia)
X-ray Univserse 2008
Granada, Spain
2008 May 29
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
X-rays (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 keV, photospheric emission
from nuclear burning white dwarfs
•“Beta-type” (e.g., AG Peg) - peaks
suggests optically thin thermal emission
with kT~1 keV, 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
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 keV, there is a highly absorbed, hard component.
– There is a clear detection of Fe K complex, indicating thin thermal origin
– A single kT (~5 keV) fit would work
– But the effective bandpass was narrow (E~2-10 keV) so only weak
constraints can be placed on models
– Reminescent of 2-10 keV X-rays seen in CVs, indicating accretion onto
the white dwarf as the probable origin
• Below about E~2 keV, 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
CrB (Luna et al., work in progress)
– RXTE/PCA monitoring of T CrB 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 keV) line
-the luminosity reflects
accretion rate
In Suzaku data, the 6.4
keV 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 keV 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 CrB
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 keV, 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
– kT~37 keV for RT Cru, 17.2 keV for T CrB, and 17 keV for CD57 3057
– BAT only fit gives kT~5 keV 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 CrB - show strong partial covering absorber
• All three (again, not so definitely for T CrB) 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 ~1.1Msun 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 CrB and CD-57 3057, and ~1.4 Msun for
RT Cru
• This agrees with the independent estimates for T CrB, 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 CrB
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 keV) 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 ~2x10-11, so pointed XMM-Newton (etc.)
observations or the e-ROSITA survey can detect many more
2008 May 29
X-ray Universe 2008
24