Searching for Substellar Objects

1
Searching for Substellar Objects

• Michael McElwain
• Advisor Dr. James Larkin
• Science in collaboration with Dr. Adam Burgasser

2
Overview

• Introduction to Brown Dwarfs
• Lick Wide-Field T Dwarf Search
• 2 field T dwarf discoveries
• First substellar subdwarf discovered
• Digital Filtering to Search for Substellar
  Companions in the Halo of Nearby Stars
• Advantage of using OSIRIS data cubes
• Quicklook v2.0

3
Brown Dwarf Formation Evolution

• Brown dwarfs have masses below the hydrogen
  burning minimum mass (HBMM)

log(L) Ls
Teff K
log(age) Gyrs
log(age) Gyrs
Burrows et al. 1997
4
Searches for Brown Dwarfs

• Companions to nearby stars
• Direct Imaging
• Infrared speckle imaging
• Coronographic imaging
• Radial velocities
• Adaptive Optics Imaging
• Young clusters
• Field Searches
• 2MASS

5
T dwarfs

• Spectral features dominated by CH4, H2O, CIA H2,
  and K I absorption features.
• T lt 1350 K
• MJ 14-16

T dwarf Near Infrared Spectra
6
Why Search for Additional Field T Dwarfs?

• Only 39 T dwarfs known
• Improved search parameters
• Larger samples needed to lower uncertainties in
  substellar properties
• substellar statistics in the Solar Neighborhood
• substellar mass function
• Discover unique objects
• Discover cooler substellar objects
• Characterize substellar properties

7
Sample Coverage Selection Techniques

• Sample Selection
• d gt -20
• 15 lt b lt 88
• J lt 16
• J-H lt 0.3
• H-Ks lt 0
• No optical counterpart within 5 of the 2MASS
  coordinates on the USNO-A2.0 catalog
• 2MASS database is flagged to ignore detections of
  known minor planets

L
M
T
8
Selection Techniques

• 267,646 candidates pass the initial selection
  criterion.
• Visual examination of DSS images for faint
  optical counterparts to candidates
• Roughly 99.5 of candidates are removed in the
  visual examination

9
Reimaging Campaign

• Gemini Infrared Camera (Lick Observatory)
• Uncataloged minor planets remain
• Reimage field at J K to confirm the presence of
  a candidate
• Roughly 25 of the remaining candidates are
  removed in this manner
• Palomar 60 CCD Camera
• Identify faint background stars
• r-J lt 6 that can pass initial criteria
• Roughly 80 of the remaining candidates are
  removed in this manner

10
Spectroscopy

• Gemini Infrared Camera
• Simultaneous observations at J/K or H/K
• Low resolution (?/?? ? 500)
• CH4 H2O absorption detection a decisive test
  for identifying T dwarfs
• Comparison spectra were taken for known M, L, and
  T dwarfs

11
J Comparison Spectra
M3V

• T dwarfs are recognized by the strong CH4
  absorption and increased H2O absorption.

M6V
M8V
L1V
sdL
T2V
T5.5V
12
Results
2MASS 05328246 SdL

• 13 spectra of candidates
• 2MASS 0516-0445 (mid T)
• J-K-0.5
• 2MASS 15032525 (T 5.5V)
• J13.9, third brightest T dwarf known
• D 8pc
• 2MASS 05328246 SdL
• First Substellar Subdwarf
• Strong CIA H2 absorption in K band

2MASS 0516-0445 (mid T)
2MASS 15032525 (T 5.5V)
13
Substellar Companions

• Multiple systems occur in roughly 60 of
  solar-type systems (Duquennoy Mayor 1991)
• Rate of multiplicity decreases for lower mass
  stars
• Brown Dwarf Desert to solar type stars reported
  by radial velocity measurements a lt 4AU (Marcy
  Benitez 1989)
• 35 of field M dwarfs, a 3-30AU (Fischer
  Marcy 1992, Henry McCarthy 1993, Reid Gizis
  1997)
• 20 of field L dwarfs, a lt 15 AU (Koerner et al.
  1999, Reid et al. 2001, Leggett et al. 2001,
  Close et al. 2002)
• 20 of field T dwarfs, a lt 3AU (Burgasser et al.
  2003)

14
Using OSIRIS to Search for Substellar Companions

• High angular resolution
• Keck telescope
• Keck AO system
• Moderate Spectral Resolution
• Obtain simultaneous spatial and spectral
  information
• If unresolved, use spectral information to search
  for companions

15
Digital Filtering of OSIRIS Data Cubes

• Many reasons to apply digital filters to OSIRIS
  data cubes
• Suppress OH contamination
• Make a K or Ks image from the broadband K
• Simulate JHK filter transmissions from other
  instruments or telescopes
• Search for substellar companions in the halos of
  nearby stars

16
Weighted Digital Filter for Substellar Companions

• When looking for a companion in the halo of the
  host star, the spectra will constructively
  interfere within a particular spatial element in
  the OSIRIS data cube.

L1V
(L1V/ G8V)-1
?
G8V
17
Quicklook v2.0

• Comprehensive 3 dimensional image analysis tool
  written in IDL conforming to Keck coding
  standards
• Object Oriented Program, capable of managing
  multiple windows
• Typical image manipulation functions as well as
  more specific image analysis tools

18
Quicklook v2.0 Plots

• Ability to take cuts of the data cube in multiple
  orientations.
• Easily customized plot parameters
• Can set QL2 to remember plot parameters.

19
Applying Digital Filters with QL2

• QL v2 is supplied with a tool to apply digital
  filters.
• 2 column data containing wavelength and
  multiplication factor are read.
• Data is sampled onto the OSIRIS wavelength grid
  and displayed in a plot window.
• Can apply and remove filters from the image in
  the image window gui.

20
Simulation of the Keck AO Point Spread Function

• Wavelength coverage 2.0-2.3 µm and a sampling
  rate of 0.002 µm.
• ro of 0.3 at 0.5 µm
• Strehl ratio of 0.6
• Plate scale of 0.02 "/pix
• 2 second integration time

Problem!
Bruce MacIntosh, LLNL
21
Digital Filter Applied to the Host and Target
Simulated with the Simulated PSF
Before filter
After filter

• Simulated PSF produces an inaccurate
  representation of the halo for a 15 minute
  exposure
• Resolved components increased contrast by a
  factor of 10.

22
Conclusions

• Lick Wide Field T Dwarf Search identified two T
  dwarfs, one substellar subdwarf
• OSIRIS is an ideal instrument to search for
  substellar companions
• Simultaneous spatial and spectral information
• Advantage of using OSIRIS data cubes
• Quicklook v2.0 is a comprehensive 3 dimensional
  analysis software

23
T dwarfs

• 39 Known T dwarfs
• 2MASS (19)
• Sloan (11)
• Deep Fields (3)
• Stellar/Substellar Companions (6)

24
2 Micron All Sky Survey (2MASS)

• Made possible by the introduction of large-scale,
  sensitive infrared arrays
• 1.3 m telescopes on Mt. Hopkins, AZ and CTIO,
  Chile.
• 256?256 HgCdTe array with 8.5 FOV
• Simultaneous observations in J, H, Ks

25
Spectral Reduction

• Spectra were reduced using the REDSPEC software
• Developed at UCLA by Prato, Kim, McLean for
  NIRSPEC data
• Reduction procedure
• Spatial rectification
• Spectra rectification
• Subtract dithered frames to remove background
  emission
• Divide by flat-dark frame
• Divide target spectra by a calibrator spectra to
  correct for telluric absorption
• Paschen and Brackett hydrogen lines are removed
  from the calibrator spectra by linear
  interpolation.
• Multiply by a blackbody consistent to the
  calibrator spectra

26
H Comparison Spectra

• Continuum spectrum in M dwarfs, late types
  exhibiting H2O absorption
• L dwarfs marked by increased H2O absorption.
• Again, T dwarfs are recognized by their strong
  CH4 absorption and increased H2O absorption.

27
K Comparison Spectra

• Continuum spectrum in M dwarfs, exhibiting some
  CO absorption.
• L dwarfs have increased CO absorption
• T dwarfs are marked by their suppressed flux due
  to CIA H2. CH4 absorption is present in this band
  too.

28
Next Infrared Sky Survey?

• HgCdTe and SiAs 1024? 1024 arrays to survey at
  3.5, 4.6, 12 and 23 microns
• 2.2 pixel scale
• Sensitive to brown dwarfs, even cooled to
  temperatures less than 200 K!
• Principal Investigator, Dr. Edward L. Wright

29
OSIRIS Overview

• OH-Suppressing Infra-Red Imaging Spectrograph
• z, J, H, K bands
• Spectral resolution of 3900
• 16?64 broad band (1700 spectral channels) or
  64?64 narrow band mode (400 spectral channels)
• 0.02, 0.035, 0.5, and 0.10 pixel scales
• FOV from 1.32?1.28 to 6.4?6.4
• Combined with Keck AO, OSIRIS will be the most
  sensitive spectrograph to date

30
Construction of the Weighted Digital Filter

• Linearly interpolate the normalized flux of the
  target and host onto the OSIRIS wavelength grid.
• Calculate the weight to assign for each channel,
  ti/ hi.
• Construct a zero mean filter by subtracting 1.