Philip J' Humphrey UC Irvine

1
The evolving LMXB populations of early-type
galaxies
Philip J. Humphrey (UC Irvine)
David A. Buote (UC Irvine), Andreas Zezas, Ewan
O'Sullivan (SAO)
2
LMXBs in early-type galaxies

• Chandra has allowed LMXBs to be found
  Characterized in early-type galaxies (eg Fabbiano
  2006)
• Provides a rich LMXB population to assess
  questions of formation and evolution.

NGC 1399
3
Long-term LMXB evolution
8 Gyr
ASCA
White, Sarazin Kulkarni (2002)
4
LMXB-GC connection
ASCA
Chandra
White, Sarazin Kulkarni (2002)
Irwin 2005
5
Our Galaxy Sample

• Sample of 19 early-type galaxies from Chandra
  archive
• Added 4 new datasets with recent mergers
  (Schweizer Seitzer 1992).
• Sample contains 6 of the 7 most disturbed in
  SS92.
• Stellar ages from SS92 or Lick index fitting.
• Central WFPC2 pointings for all galaxies to study
  GCs.
• Detailed study of subset presented in Humphrey
  Buote (astro-ph/0612058)
• Numbers of LMXB, GC inferred by LF fits.
• For shallow data, LMXB LX from spectroscopy.

6
Spectroscopic Lx vs XLF Lx
7
Evolution in the XLF?

• Does the XLF shape change with age?
• If sources spend most of their active lives at
  similar Lx, the XLF will evolve (Wu 2001).
• The XLF can evolve a knee which changes with
  age.
• Note This cannot explain the XLF break unless
  lifetime 6 Gyr

8
An evolving field population?
9
Our LMXB-GC relation
Omits very young systems
10
LMXB luminosity evolution with age
11
A simple evolutionary model

• In lieu of a detailed population synthesis model,
  simple approach based on White Ghosh (1998)
• Post-supernova binaries (PSNBs) evolve over
  timescale tp (1Gyr) into LMXBs, which have
  lifetimes tL.
• PSNBs form at a rate proportional to SFR.

• Assume 2 bursts of star formation a Hubble time
  ago and at the time of the last merger.
• Use SS92's estimate of fraction of stars born in
  the merger.

12
Normalizing the model

• Problem we actually measure LX, which depends on
  number of sources and duty cycle, beaming factor,
  etc...
• How do we normalize the model to relate it to
  SFR?

• Solution Use the Milky Way!
• Assume constant SFR history for MW disk SFR2
  Msun yr-1
• From LX,excess(MW) we can normalize the model,
  for a given tp and tL.
• We further assume tp1 Gyr (following White
  Ghosh 1998).

13
Amount of field formation
14
Results
15
Conclusions

• Sample of 22 early-type galaxies, 6 post-merger.
• Shape of the XLF insensitive to population age.
• Accounting for production in GCs, number of LMXBs
  per unit light does not evolve with age.
• Consistent with a short (Gyr) LMXB lifetime.
• This implies negligible field LMXB production.
• Supported by specific frequency of LMXBs being
  proportional to specific luminosity of GCs.