Supersoft X-ray Sources in M31 in Be binaries? An astronomical game of

1
Supersoft X-ray Sources in M31 in Be binaries?
An astronomical game of Guess Who

• Thomas Nelson and Marina Orio
• INAF-Padova and University of Wisconsin

2
Supersoft X-ray Sources

• First discovered by Long et al. (1981) in
  Magellanic clouds with Einstein, became class
  with ROSAT
• Characterized by very soft emission - no flux at
  energies gt1 keV
• Fit with blackbody models of 105 lt Teff lt106 K,
  and 1036 lt Lx lt 1038 erg/s
• A large fraction must be WD systems
• Unerstanding whether they are the single
  degenerate progenitors of type Ia SN is crucial

CAL 83
CAL 87
3
Type Ia supernovae
SN1994D in NGC 4526 (HST) Sne Ia are often in
young populations! RateSFR Favorite single
degenerate models are very old (recurrent
novae, symbiotics)
4
Optical counterparts of Galactic and Magellanic
Cloud sources

• Proximity of 2(4) Galactic and 14 MC sources
  allows for study of optical counterpart
• SSS can be foreground objects, SNRs and PN, but
  most are WD binaries classical novae in outburst
  (1), symbiotics, and other WD in binary
  systems burning accreted hydrogen in a shell
• Trying to narrow down the phenomenological
  definition in X-ray range makes things even more
  confusing (e.g. RS Oph vs S And)
• The nature of some sources (e.g. MR Vel, Cal 87)
  is still uncertain.
• Sources can be persistent, recurrent or transient
  in soft X-rays. SMC mostly persistent.
• 75 of SSS contain a white dwarf - possible SN
  Ia progenitors?

CAL 87 finding chart (Pakull et al., 1998)
CAL 87 V band lightcurve (Callanan et al., 1989)
5
M31 - the largest SSS population
ROSAT PSPC survey (Supper et al., 1997)
XMM-Newton EPIC survey (Pietsch et al., 2005,
Orio 2006)
Chandra ACIS-S (Di Stefano et al., 2004)
6
M31 SSS population properties
Satellite No. of sources Notes
ROSAT 46 many sources transient and never detected again
Chandra 16 sources mainly in bulge
XMM-Newton 18 includes several novae, also re-detected several Chandra sources

• Most of the sources are transient (especially
  ROSAT!)
• A few are recurrent, and have been detected at
  different epochs with different instruments
• Larger distance to M31, and increased crowding in
  the galaxy make optical counterparts hard to
  identify
• gt33 of M31 SSS have been identified with novae

7
A UV and optical counterpart search

• Search for counterparts to the M31 SSS population
  using GALEX, WIYN and the images of the Local
  Group Survey (Massey et al., 2006)
• If (and only if) the SSS hosts a shell burning
  white dwarf, it should be detected as a UV source
  with GALEX, and a very blue star (U-B, B-V
  negative) in the Massey survey
• Both the GALEX and LGS images suffer from source
  confusion in the bulge of M31 no counterparts
  within central circle of radius 5
• This leaves 60 sources which could be detected
  with GALEX
• The majority of the sources are ROSAT objects -
  only 4 Chandra and 12 XMM sources lie outside of
  the bulge

8
The Nearby Galaxy Survey with GALEX
Thilker et al. (2005)

• 50 cm telescope
• Simultaneous imaging in 2 bands
• FUV 1350-1750 A
• NUV 1750-2800 A
• Average exposure time 2700 s

9
The Local Group Survey
Massey et al. (2006)

• UBVRI photometry of 370,000 stars in M31
• 1 photometry at UBVRI 21
• lt10 photometry at UBVRI 23
• Coverage not as extensive as GALEX, but covers
  all but 5 of the 60 SSS
• Follow up with WIYN images

10
FUV
RX J0039.74030
RX J0043.34118
NUV
11
A surprising result!

• We find that only 9 SSS have a UV counterpart in
  the X-ray error circle and they are all are
  inside, or within 30 of an OB association in
  M31!
• In addition, several of the sources not detected
  with GALEX also lie within or near an OB
  association, including one Chandra source.
• So we find a number of SSS that appear to be
  associated with young stellar population.
• We have proof in a few cases that only one UV
  object among them is a B star that has colors
  consistent with a binary WD system SSS.
• This is in contradiction with most accepted
  models of accreting white dwarf SSS, which have a
  long delay time and so should be associated with
  OLD populations.

12
Is the proximity to OB associations a coincidence?

• Assume that SSS really are an old population
  phenomenon
• Therefore, should be distributed randomly over
  the ROSAT survey area

total area of ROSAT survey 23,400 arcmin2
total area of OB associations in M31 1293
arcmin2 probability of chance alignment 6 We
find 20 of all ROSAT sources are inside an OB
association! This is not just a coincidence!
13
So what are these sources?

• Detections could be spurious - ROSAT count rates
  are very low an effect of spread of soft tails
  of many stellar winds?
• Could be supernova remnants. Young supernova
  remnants can be quite soft, anomalous SNR in
  cavities and bubbles.
• Some kind of new LMXB? Ot LMXB in new state?
• Something else?

14
One possibility WD-Be binaries

• Be stars are rapidly rotating stars, which may be
  formed as a result of massive star binary
  evolution
• Raguzova (2001) predicts that 70 of all Be stars
  formed as a result of binary evolution should
  have a WD companion
• WD accretes wind from the Be star and begins to
  burn H in a shell
• Could this be what we are seeing?
• Kahabka et al. (2006) report the XMM detection of
  a new supersoft source, XMMU J052016.0-692505 in
  the LMC, coincident with Be star in the LMC

15
An exciting new XMM-Newton source! A clue to
this puzzle?
We found a new SSS in the XMM-Newton archive,
better error box
0.15 - 10 keV
0.6 - 10 keV
16
FUV
U
One star inside error circle V 21.827 B-V
-0.33 U-B -0.705 V-R 0.151 Colors are
consistent with a B star in M31!
NUV
17
Blackbody model with
T 5 - 6 x 105 K
EPIC-pn spectrum. Lx in range 1037 - 1038 erg/s
18
Summary

• We have carried out a search for optical and UV
  counterparts of SSS in M31
• Of 60 sources which could be detected with GALEX,
  we find only 9 have UV counterparts
• These sources all lie inside or near OB
  associations in M31, and it is likely the UV
  sources are single, extremely hot stars in these
  young populations.
• Conventional SSS should not be found in young
  populations - these sources are therefore likely
  to be different, and if they are WD binaries,
  they may account for prompt component of SNe Ia.