Transcript Chandra

The Impact of Chandra on M31
and M33 X-ray Studies
Ben Williams (UW)
X-rays from Nearby Galaxies
September 5, 2007
Collaborators
Michael Garcia (CfA)
Jose Galache (CfA)
Albert Kong (MIT)
Sachi Naik (ISAS)
Robin Barnard (Open)
Jeff McClintock (CfA)
Joseph Gelfand (CfA)
Lorant Sjouwerman (NRAO)
Frank Primini (CfA)
Julian Osborne (Leicester)
Ulrich Kolb (Open)
Stephen Murray (CfA)
Rosanne Di Stefano (CfA)
Paul Callanan (Cork)
Paul Plucinsky (CfA)
Terry Gaetz (CfA)
Manami Sasaki (CfA)
Knox Long (STScI)
Tsevi Mazeh (TAU)
Avi Shporer (TAU)
Wolfgang Pietsch (MPI)
Frank Haberl (MPI)
Parviz Ghavamian (JHU)
Jack Hughes (Rutgers)
Frank Winkler (Midd)
William Blair (JHU)
Richard Edgar (CfA)
Robert Kirshner (CfA)
Miguel de Avillez (Vienna)
Dieter Breitschwerdt (Vienna)
Ralph Tϋllmann (CfA)
Randall Smith (JHU)
Thomas Pannuti (Morehead)
David Helfand (Columbia)
Luciana Bianchi (JHU)
Jonathan Grindlay (CfA)
Kip Kuntz (JHU)
David Thilker (JHU)
Motivation: Why M31 and M33
• Interpretation of distant galaxies, where less detail is available.
• Most nearby examples of early and late type spiral galaxies.
• Probe endpoint of stellar evolution, which drives galaxy evolution.
• Additional samples for any class of X-ray sources, ULX, more
galactic nuclei.
• Disks not edge on.
• Sources at a common distance.
Outline
I. Nuclei
Chandra’s abilities to resolve the crowded nucleus.
II. SNRs
Resolved Multiwavelength Studies (X-ray, optical,
radio).
III. Transients
Well-sampled long-term light curves and counterpart
candidates.
IV. Conclusions
Nuclei
• M31 and M33 offer 2 of the closest spiral
galaxy nuclei other than the Milky Way
• Chandra can potentially resolve the weak
nuclear emission.
• The best places to study under-luminous
nuclei.
Nuclei
Bondi radius resolved by Chandra (Garcia et al. 2005).
X-ray
Optical (X-ray contours)
Chandra-HRC image of M31 nucleus and optical overlaid with X-ray contours
Correcting for small extraction radius and
absorption (assuming Γ=1.7, NH=7X1020
cm-2), gives 9X1035 erg/s (0.3-7 keV).
Nuclei
Variability on timescales of < Months.
Lightcurve
Nucleus
Nuclei
• Like Milky Way, low luminosity for mass and gas density
• Accretion rate is similar to star formation in nuclear cluster
Lx/LBondi (max luminosity
from observed gas density in
a classical accretion disk)
vs.
Lx/LBondi
RBondi (radius of accretion
influence) for several SMBHs.
M31
MWG
RBondi (arcsec)
•For M31, LBondi = 5X1041 erg/s, assuming ne =
0.1 cm-3 (Dosaj et al. 2002) (~1 M๏ /yr)
•Observed accretion (2x10-6 M๏/yr) ~mass and
age of nuclear cluster (10-6-10-5 M๏/yr [Chang
et al. 2007])
Nuclei
M33 has a ULX within 0.6′′ of the nucleus (Dubus & Rutledge 2002).
Not associated with the nucleus (Gebhardt et al. 2001; Long et al. 2002)
M33 Nucleus in U-band
(Dubus et al. 1999)
Optical Nucleus dominated by
star cluster (Long et al. 2002)
(LX (0.3-8 keV) = 1.5 x 1039 erg/s)
BH mass <1500 M๏
Close-up of M33 Nucleus
in Chandra-HRC image.
Yellow circle is 0.6′′
positional error with
respect to the nucleus.
Nuclei
The ULX is variable and dwarfs any contribution from the nucleus
to the X-ray emission by a factor of at least 14 (Dubus et al. 2004).
Lightcurve
Radial Profile
Spectral variability, UV spectrum (40 Myr old starburst, Long et al.
2002) and lack of optical counterpart all consistent with a black hole
binary, perhaps an HMXB like LMC X-3 (La Parola et al. 2003; Foschini et al. 2004).
Nuclei
The Chandra ACIS Survey of M33 (ChASeM33) data may be able to
separate the extended nuclear emission from the ULX.
Model PSF
ACIS Image
Pileup makes central ~2.5” difficult to interpret!
Nuclei
Only diffuse emission is seen outside of 2.5”. Perhaps
more detailed analysis and simulations will show more.
Nuclei
• M31 is only resolved with Chandra, and shows faint
emission from the nucleus, of similar type to Milky Way.
• M33 nucleus is yet to be detected behind the ULX,
even with Chandra resolution.
Supernova Remnants (SNRs)
• Key for learning current SN
rate, SN feedback effects, and
shock physics.
• Chandra resolves X-ray
counterparts to known SNRs and
discovers new SNRs from their
resolved X-ray appearance.
• Resolved SNR studies constrain
ages and shock properties.
http://chandra.harvard.edu/photo/2005/casa
Chandra Discovers and Resolves M31 SNRs
1
2
• Two resolved discoveries
• Intermediate age
• Low density ISM
kT=0.17, t~17 kyr, n=0.2
kT=0.3, t~9 kyr, n=0.3
Williams et al. 2004
Chandra Discovers and Resolves M31 SNRs
1
2
X-ray contours
Radio image
Kong et al. 2003
Radio discoveries confirmed by resolved Chandra imaging
Williams et al. 2004
SNR r3-63: The Brightest X-ray SNR
• X-ray bright
• Old
• Low density ISM
Williams et al. 2005
Kong et al. 2002b; Williams et al. 2005a
Radius = 21 pc
LX(0.3-7)~1037 erg/s
kT ~ 0.16 keV
0.1<n0<0.3 cm-3
Age ~ 20 kyr
(Kong et al. 2002b;
Williams et al. 2005a)
X-ray
SNRs: M33
ChASeM33: X-ray remnants
appear to trace the southern
spiral structure (Plucinsky et al.
2007).
Ha
Green circles mark SNRs.
Great correspondence:
Younger SNRs?
High ISM density?
IR 3.6 μm
9.3′ X 15.6′
(Also see work by Ghavamian et al. and Pietsch et al)
SNRs
The brightest one (SNR21; Gordon et al. 1998) has a wellsampled X-ray spectrum and an interesting morphology
(Gaetz et al. 2007).
Narrow-band optical
image of the H II
region NGC592. The
strong [S II] emission
of the SNR makes it
appear green in this
image.
Red=Ha, Green=[S II], Blue=[O III]
SNR21
Smoothed Chandra
Deconvolved Chandra
Narrow-band optical
(continuum subtracted)
Narrow-band optical
To H II Region Core
High-mass progenitor?
Relatively Young
Hot
High ISM density
LX(0.25-4.5) = 1037 erg/s
Vshock ~ 600 km/s
Radius = 10 pc
kT ~ 0.46 keV
n ~ 2 cm-3
Age ~ 7 kyr
Log[E0(ergs)] ~ 51.25
Gaetz et al. 2007
M33 SNRs: More to come
• Strong emission lines suggest it is O-rich (J. Hughes et al., in prep)
• We are detecting the ejecta directly!
• Very young; a gem for studying heavy element creation and distribution.
ACIS-I Spectrum
Initial spectral
fitting suggests
high O/Fe,
Ne/Fe, and
Mg/Fe ratios,
comparable to
the known O-rich
SMC SNR
1E 0102-7219
SNRs
M31 and M33 both have SNRs well-resolved by Chandra.
M31 ISM appears to be of generally low density.
Many optical SNRs in M33 are strong X-ray emitters.
(Young? More dense ISM?)
X-ray Transients (XRTs)
• Chandra’s resolution and pointing capabilities (rotating
solar panels) both key for this research.
• Large fraction of Galactic LMXBs that exhibit bright
(>1037 erg/s) transient outbursts have been shown to
contain black holes (McClintock & Remillard 2004).
• There are only ~20 such objects in the Galaxy, and
each is vital for studies of accretion physics in the strong
gravity regime. Finding more is worthwhile.
(In addition to our work, lots of great work by, e.g. Trudolyubov et al.; Pietsch et al.)
M31 Transient Population
8 Years of Chandra Observations (1999-2007)
ACIS-I
More than 50 transients!
HRC
The Transient Population: Movie
ACIS-I
central 5’ X 5’
4-12 ks images
>2e36 erg/s
Red = 0.3-1 keV
Green = 1-2 keV
Blue = 2-7 keV
The Transient Population: Movie
ACIS-I
central 5’ X 5’
4-12 ks images
>2e36 erg/s
Red = 0.3-1 keV
Green = 1-2 keV
Blue = 2-7 keV
Luminosity
The Transient Population: Lightcurves
Stars: HRC
Full Squares: ACIS-I
Open Squares: ACIS-S
Circles: XMM
Time
Williams et al. 2006b
Williams et al. 2005
Similar Global Properties to Galactic transients
Distribution of Decay Times
Distribution of Peak
Luminosities
Williams et al. 2006b
log (d) = 1.5 +/- 0.5 (Galactic is 1.2 +/- 0.4; Chen et al.
1997)
Most sources 1037<LX<1038 (Galactic sample has mean of
LX=0.2 LEdd; Chen et al. 1997)
Cumulative number with lower duty cycle
Similar Global Properties to Galactic transients
• The best light curves allow duty
cycle constraints.
• About half are constrained to
<0.1, and nearly all are <0.2.
• Constrains NS/BH ratio!
Fundamental stellar evolution/IMH
parameter.
Williams et al. 2006b
Duty Cycle
About 1/100 sources > 2*1036 erg/s is transient.
If transients are black holes, and others are
neutron stars, then NS/BH1*(dc/0.01).
Optical Luminosity and Orbital Period
Optical Luminosity
•Orbital Period of Galactic
LMXBs is correlated with
Optical and X-ray luminosity
(van Paradijs & McClintock
1994). Blue points show
more recent transients
during outburst.
• Possible Explanation:
larger disks, which may be
associated with longer
period binaries, glow
brighter in optical.
LX * Orbital Period2/3
The first XRT follow-up with ACS was 16’ off-axis
Linear decay indicates large disk
Flux
X-rays
on
X-rays
off
Time
• If this is the counterpart,
Lx=1.4E38 erg/s , MV ~ -0.3,
P ~ 1 day.
Williams et al. 2005b
Optical
on
Optical
off
Flux
A more typical follow-up was only 4’ off-axis
Optical
on
Optical
off
X-rays
on
X-rays
off
Time
•
•
•
•
0.2” error circle
Counterpart faded, but is blended with another star
If star to NE in blend is the counterpart, MV ~ -1.2, Lx ~ 6E37 erg/s, P~8 days
If counterpart is a third unresolved source, MV > -1.2 and P < 8 days
(Williams et al. 2005c)
The Search continues…
• Thirteen transients followed up so far, with luminosities
ranging from 1036 – 1038 erg/s (Williams et al.; Galache et al.)
• Follow-up optical continues, but more difficult with WFPC2..
• Periods still range from 1-23 days, with most being <5
days, similar to the distribution of Galactic XRTs.
• Period distribution is a fundamental constraint for models of
binary formation and evolution.
X-ray Transients in M33
• Smaller galaxy, no large bulge, therefore no old and
dense population of stars.
• Transients may not be dominated by LMXBs.
• ChASeM33 scheduling not optimal for transient search.
• Most of the galaxy has been observed more than once.
Relative Flux
Seven Good Candidates Found
Time
Active and Quiescent X-ray Images
Different Exposure Times!
Low-count spectra are of a wide variety
Hard
Soft
Softer
The hardest spectra have photon indices of <1, possible HMXBs.
Different from M31, where all of the spectra have been soft (Γ>1.5).
The soft spectra have photon indices of ~1.5-2.0, possible LMXBs or AGNs.
Softer spectrum is blackbody with 0.04<kT<0.5 keV, possible BHB.
Low-count spectra are of a wide variety
Flux
Black: XMM-PN
02-Aug-2000
Resid
Red: XMM-PN
04-Aug-2000
0.35
keV
Softest!
1.0
Green: Chandra
ACIS-S 30-Aug2000
Supersoft source with kT = 50 eV
(agrees with Grimm et al. 2007)
(See work on periodicities in SSSs in M31 by Trudolyubov et al. 2007, King et al. 2002
Optical counterparts
XRT-1
XRT-3
XRT-4
Soft Transient – Red variable Star
Hard Transients (HMXBs) – Blue Variable Stars
Infrared counterparts from Spitzer
XRT-7
XRT-2,4,5
Stern et al. 2005
One transient has the IR colors of an
AGN; the rest are consistent with stars.
Transients in M33
7 (6 new) transient candidates found in M33
3 could be LMXBs in different states
2 are likely HMXBs
1 is a supersoft source
1 is likely an AGN
These transients appear to be generally fainter and
more spectrally diverse than those we often see in
M31.
Still more ChASeM33 data to be analyzed.
Conclusions: Comparing Galaxies
Property
M31
Milky Way
M33
Type
Metal-rich,
luminous early
spiral
Moderately metalrich intermediate
spiral
Metal-poor low
luminosity, late spiral
Nucleus
Very faint,
massive black
hole
Very faint, lowermass black hole
Dominated by ULX,
little or no massive
black hole.
Transients
(LMXBs)
Soft spectra, short Soft spectra, short Small sample, faint,
periods, 30 day
periods, 30 day
and diverse
decay times.
decay times.
classification.
SNRs
Low ISM density
X-ray bright are
intermediate age
Distance
uncertainties;
extinction; difficult
comparison
Good
correspondence
between X-ray and
optical; Young?
XLF
Relatively steep,
dominated by old
population (Kong et
Both HMXB and
LMXB populations
Dominated by young
population HMXBs
al, Voss & Gilfanov,
Trudolyubov et al., etc.).
(e.g. Grimm et al. 2002)
(Grimm et al. 2005)