Dark energy in the Supernova Legacy Survey

1
Dark energy in the Supernova Legacy Survey

• Mark Sullivan (University of Toronto)

http//legacy.astro.utoronto.ca/
2
French Group Reynald Pain (PI), Pierre Astier,
Julien Guy, Nicolas Regnault, Jim Rich, Stephane
Basa, Dominique Fouchez
Toronto Group Ray Carlberg, Mark Sullivan, Andy
Howell, Kathy Perrett, Alex Conley
UK Gemini PI Isobel Hook Justin Bronder,
Richard McMahon, Nic Walton
USA LBL Saul Perlmutter CIT Richard Ellis
Plus Many students and associate members
throughout the world
3
SNLS Vital Statistics

• 5 year (202n) rolling SN survey
• Goal 500 high-z SNe to measure w
• Uses Megacam imager on the CFHT griz every 4
  nights in queue scheduled mode

• Survey running for 3 years
• 300 confirmed zgt0.1 SNe Ia
• Largest single telescope sample
• On track for 500 by survey end

4
Supernova Legacy Survey
Imaging CFHT Legacy Survey Deep program
Spectroscopy Types, redshifts from 8m-class
telescopes
Discoveries
Lightcurves
Gemini N S (120 hr/yr)
VLT (120 hr/yr)
griz every 4 days during dark time
Magellan (15 nights/yr)
Keck (8 nights/yr)
5
Dark Energy in the SNLS
6
First Year Results (Astier et al. 2006)
Assuming flatness, w-1 OM 0.263 0.042
15 of final sample
7
Dark energy SNLS WMAP
Spergel et al. (2006)
HST/GOODSWMAP
SNLSWMAP
8
The third year sample

• Third Year cosmological analysis
• Data collection complete yesterday (end 06A)!
• SN sample 4 times larger
• Improved z data will make the zgt0.8 SNe more
  cosmologically powerful than in Year 1
• Final results should be ready in the Autumn

9
Preview of 3rd year Hubble Diagram (preliminary)
160 SNe Ia to z0.8 50 are still having data
acquired or are still being reduced 70 at zgt0.8
await an improved k-correction template
Sullivan et al. in prep.
10
UV and U-band k-corrections

• At zlt0.8, rest-frame B-V is used to
  colour-correct SNe
• At zgt0.8
• i and z probe rest-frame U and B no V data
• Understanding of UV/U required for colour
  correction to be performed
• Almost no data error in existing templates
  essentially unknown
• Rest-frame UV study at Keck (PI Richard Ellis)

11
SNe Ia show much diversity in the UV Improving
the k-correction spectral template will decrease
systematics from this region at zgt0.8
Ellis, Sullivan et al. in prep.
12
Constraining population evolution
13
Potential Systematics in measuring w

• Photometric zeropoints
• Mismatches to local SNe observations
• Contamination by non-SNe Ia
• Spectroscopy is critical
• K-corrections
• U and near-UV uncertain evolution in UV?
• Extinction
• Grey dust Effective RB Dust evolution
• Redshift evolution in the mix of SNe
• Population drift environment?
• Evolution in SN properties
• Light-curves/Colors/Luminosities

More mundane
More scientifically interesting
14
Potential Systematics in measuring w

• Photometric zeropoints
• Mismatches to local SNe observations
• Contamination by non-SNe Ia
• Spectroscopy is critical
• K-corrections
• U and near-UV uncertain evolution in UV?
• Extinction
• Grey dust Effective RB Dust evolution
• Redshift evolution in the mix of SNe
• Population drift environment?
• Evolution in SN properties
• Light-curves/Colors/Luminosities

Population Evolution
15
White Dwarf
?

• Many competing models for
• Nature of progenitor system the second star
• Single versus double degenerate
• Young versus old progenitor
• Explosion mechanism?
• Mass transfer mechanism?

16
SNLS SN rate as a function of sSFR
Per unit stellar mass, SNe are at least an order
of magnitude more common in star-forming galaxies
SN rate in SNLS passive galaxies
125 Host Galaxies at zlt0.75
Sullivan et al. (2006)
17
AB Model for SN Ia rate
Scannapieco Bildsten (2005) and Mannucci et al.
(2005) proposed a two-component model

• Confirmed by SNLS results
• SNR is linearly proportional to galaxy mass and
  SFR
• SNe Ia will originate from a wide range in
  progenitor age
• Two components? Or one with a wide range in
  delay-time?
• Either way the mix of the two components will
  evolve with redshift

18
Mix will evolve with redshift

• Relative mix evolves strongly with redshift

B component
19
Population evolution stretch and colour

• Distance estimator used
• (how) Do these vary across environment?
• By understanding and calibrating any
  relationships, we can improve the quality of our
  standard candle

s stretch corrects for light-curve shape via a
c B-V colour corrects for extinction (and
intrinsic variation) via ß
20
Stretch and Environment
Sullivan et al. (2006)
Similar trend observed at low-redshift Simplest
inference Older progenitors produce smaller
stretch, fainter SNe Younger progenitors produce
larger stretch, brighter SNe
Stretch ?Fainter/faster SNe
Brighter/slower SNe ?
21
Yet so far the stretch correction seems to
work equally well in all environments

• (Conley et al. 2006, AJ in press)

No evidence for gross differences between
light-curves in passive and active galaxies
22
Colour relationships
Fainter
Combination of Intrinsic brighter-bluer
relationship Extinction
First year sample ß1.6 (Milky Way dust predicts
ß4.1) But stretch correlates with environment
so perhaps the colour correction (ß) should
correlate with stretch
Brighter
SN Colour
23
Colour relationships low stretch
Preferentially located in passive galaxies Less
dust Intrinsic SN relationship only?
24
Colour relationships high stretch
Preferentially located in star-forming
galaxies Extinction much greater Intrinsic SN
relationship PLUS dust? Or just different
intrinsic SN relationship?
Effective ß differs according to environment
25
Low-stretch rms 0.14
High-stretch rms 0.20
Low-stretch SNe show a far smaller scatter on the
Hubble Diagram but, they are rarer (AB!)
26
Summary

• 3rd year analysis challenge is controlling
  systematics such as population drift
• SNe Ia know and care about their environment
• Stretch depends on age of the progenitor
  population
• SNe with narrow light-curves preferentially
  hosted in passive galaxies show less scatter
• Cosmology with sub-samples of SNe improves the
  power of the standard candle

27
Summary

• The SNLS dataset is the most uniform, well
  understood, and statistically powerful SN Ia data
  set currently the best SN dataset to combine
  with BAO or WMAP data to measure w.
• 3rd year analysis will be completed in the Autumn
  watch this space
• The final SNLS data set will be essential for
  constraining systematics and when planning next
  generation projects like the LSST or NASAs JDEM.

28
Host Galaxies of SNLS SNe

• PEGASE2 is used to fit SED templates to the
  optical ugriz data.
• Recent star-formation rate and total stellar
  mass are estimated.
• Host galaxies classified by their specific
  star-formation rate.

Passive
Star-forming
Starbursting
Sullivan et al. (2006)