Review of SNLS results

Blois 2008 21 May 2008 1

Vanina Ruhlmann-KleiderCEA/Irfu/SPPSaclay

Review of SNLS results

1) The Supernova Legacy Survey2) Cosmology results3) SNe Ia : usable for precision

cosmology ?4) Conclusions

Blois 2008 21 May 2008 2

o The survey (2003-08): - four 1°x1° fields - each field observed every

3-4 nights during 6 monthso MegaCam : - 1°x1° CCD imager - 340 millions of pixels - pixel resolution: 0.18“ - four filters (400-1000nm) o Spectroscopy: time allocated

on 8-10 meter class telescopes ~500 confirmed SNe Ia in 5 years with redshifts in 0.1 – 1.2

SNLS at CFHT

04D1dc

Blois 2008 21 May 2008 3

SNLS "rolling search"

large and homogeneous sample of high-z spectroscopically confirmed SNe Ia, measured with

good photometric accuracy

Blois 2008 21 May 2008 4

a SNIa lightcurve

spectroscopy triggered type Ia confirmed, z=0.627

multi-band LC’s :

test compatibility with SNIa model (trained on SNIa lightcurves and

spectra)

i’, r’, z’, g’ filters

apparent B flux : mB*, (B-V) colour c and stretch s

see e.g J.Guy et al., A&A 466 (2007) 11

Blois 2008 21 May 2008 5

Cosmology with SNe Iao SNe Ia are (assumed to be) standard candles :

mB* from B* L(c,s) /4πdL2 with

dL(z,H0,ΩM,ΩΛ,w,..)

B = mB* - MB + (s-1) - c

apparent rest-frame maximum light

magnitude

reference absolute B-

band magnitudelightcurve shape

variability

(B-V) colour variability(intrinsic variation and

extinction)

known luminosity

o Distance estimator :

+ intrinsic dispersion term, int allowed in cosmological fits to account for our lack of knowledge about SNe

Blois 2008 21 May 2008 6

1st year resultsP.Astier et al., A&A 447 (2006) 31

residuals to (0.26,0.74):

older

fain

ter

First year result (71 SNe) consistent with an accelerating Universe

int: 0.13± 0.02Low-z: 0.15±0.02High-z: 0.12±0.02

Blois 2008 21 May 2008 7

if ΩM+ΩΛ=1:

ΩM=0.263±0.043±0.032 (SNLS,w=-1)

w=-1.023±0.090±0.054 (SNLS+BAO)

w=-1

M+=1

Cosmological parameters

wpde

de

Blois 2008 21 May 2008 8

Preliminary 3rd year result

SNLS3 result (~250 SNe) agrees with SNLS1 result, w consistent with -1 within 6% (SNLS+BAO+WMAP5)

Blois 2008 21 May 2008 9

SNe Ia : usable for precision cosmology ?

Distance estimator :

MB, , assumed

to be z-independent

Is that so ?

compare properties of SN sub-samples split by redshift, host activity…

B = mB* - (MB - (s-1) + c)

Blois 2008 21 May 2008 10

resi

dual

s w

ithou

t s-t

erm

resi

dual

s w

ithou

t c-t

erm

mB* - MB + (s-1) - c – 5 log10(dLfit c-1 H0)

‘’brighter-slower relationship’’ ‘’brighter-bluer relationship’’

consistent behaviour for nearby and distant SNe Ia (1st year data)

P.Astier et al., A&A 447 (2006) 31

blue: z<0.15 / black: z>0.15

bri

ghte

r

slower bluer

Blois 2008 21 May 2008 11

stretch-corrected lightcurves

SN Ia rise-times consistent at 1σ

o low-z vs high-z SNe Iao passive/active host galaxies

A.Conley et al., AJ 132 (2006) 1707

73 SNe z<0.9 – 2 years of data

Blois 2008 21 May 2008 12

SN Ia rate is a function of the

host stellar mass and SFR

Passive

hosts

Active hosts

SN

Ia r

ate

per

unit

mass

per

year

Star formation rate per unit mass

SN stretch depends on host activity

M.Sullivan et al., ApJ 648 (2006) 868

SN Ia environment

Brighter, slower SNe occur mostly in star-forming

galaxies(2 years of data)

10 10 10 10 10

Blois 2008 21 May 2008 13

Two-component rate modelisation1

SNRIa(t) = A Mtot + B dM/dt(t)

galactic stellar mass SFR old stars prompt SNe

evolved galaxies star-forming galaxies

1. Fit A and B from SNLS data2. Use fitted A, B values and a model2 for SFR(z) to

predict rates vs redshift3. Compare predicted rate with measurements

1Scannapieco, Bildsten (2005) 2 Cosmic SFH from Hopkins & Beacom (2006)

Blois 2008 21 May 2008 14

Average SN stretch will evolve

with z

Predicted mix of two SN

populations evolves with z

Implication:

Blois 2008 21 May 2008 15

SNLS

D.Howell et al., ApJ 667 (2007) 37

Stay tuned !

stretch evolution with z: predictions agree with data

<stretch> + 8% for z=0.03-1.12

<intrinsic brigthness> +12%

If is the same for old/prompt SNe, this will

not affect cosmology

Need more data to test , values from sub-samples

split by stretch or host type

Blois 2008 21 May 2008 16

Conclusions

o SNLS: efficient high-z SN Ia detection, optimised temporal sampling, very good photometric accuracy.

o Cosmology: ~250 SNe in SNLS3. Constraints on m, w consistent with SNLS1 results, accuracy on w from 9 to 6% (stat).

o Are SNe Ia good cosmological probes ? • Average SN Ia properties evolve with redshift. So far,

no bias in cosmology detected. • Larger samples needed to test if the empirical

luminosity-shape or color relations need to be refined.

Blois 2008 21 May 2008 17

BACK-UP SLIDES

Blois 2008 21 May 2008 18

A possible type Ia explosion mechanism

reproducible intrinsic luminosity : SNe Ia ~ standard candles

Blois 2008 21 May 2008 19

Spectroscopic time allocated on 8-10 meter class telescopes

more 8-10m time than CFHT time !

Blois 2008 21 May 2008 20

September 06:>350 spec confirmed Ia

>500 spec confirmed Ia by end of June 08

More than 1000 Ia in total

SNLS SNIa data sample

Blois 2008 21 May 2008 21

SNLS data analysis

An example of raw image:

entire 4x9 CCD mosaicone exposure of 300secfilter r (<>=620nm)size: 1.4 Gb

2004/10/21

whole survey: 15 Tb

Blois 2008 21 May 2008 22

SN detections

Reference image Current image Subtraction

SN Ia events found by CCD image subtraction:

Many steps: image cleaning and alignment, photometric calibration, match reference and current image qualities…

SN

Blois 2008 21 May 2008 23

if ΩM+ΩΛ=1:

ΩM=0.271±0.021±0.007

w=-1.023±0.090±0.054

DE equation of state

M+=1

(SNLS+BAO)

wpde

de

1de

dde

dt 3H1w

w=1/3 radiationw=0 matterw>-1 quintessencew=-1 cosmological constant

Blois 2008 21 May 2008 24

Systematics, 1st year paper

photometric accuracy

selectionbias

SN Ia model

SN Ib/Ic contamination: 0.5 in the 1st year sample

Blois 2008 21 May 2008 25

Comparison with previous SN results

Knop et al, 2003 (HST+SCP) : 52 SNe,dotted contours

SNLS 1st year:71 SNe, solid countours

SNLS projected end of surveyshaded area

the superior color and time sampling leads to tighter constraints

Blois 2008 21 May 2008 26

Intrinsic colour vs dust extinction

Colour—luminosity relationship

inconsistent with MW-type dust

Best-fit: β~3

MW-dust: β≡RB=4.1

SN Colour (c)

Resid

ual w

ith

ou

t c-c

orr

ecti

on

β=4.1

Preliminary SNLS3

M.Sullivan, STScI, may 2008

Blois 2008 21 May 2008 27

Colour correction required in all host

types

SN Colour (c)

Resid

ual w

ith

ou

t c-

corr

ecti

on

Resid

ual w

ith

ou

t c-

corr

ecti

on

Passive hosts

Star-forming hosts

40 high-z SNe

180 high-z SNe

Large “local” SN surveys Large “local” SN surveys covering a wide wavelength covering a wide wavelength range (inc. near-IR) urgently range (inc. near-IR) urgently needed to disentangle thisneeded to disentangle this

M.Sullivan, STScI, may 2008

Preliminary SNLS3

Either:Passive hosts have dust?An intrinsic relation dominates over dust?

Blois 2008 21 May 2008 28

Determine (luminosity correction), (color

correction), from subsamples split by…

spiral

elliptical

low s

high s

Early days… larger samples will be definitive

Tension?s<1

s>1

Preliminary SNLS3

M.Sullivan, STScI, may 2008

Blois 2008 21 May 2008 29

Spectroscopic corollaries

T.J.Bronder et al., A&A 477 (2008) 717

SiII equivalent width distributions (residuals / low-z

beahviour)□ E/S0 ■ Spirals

Lower EWs in late-type galaxies, whatever the redshift

SNe Ia in late-type (star-forming) galaxies are brighter (more 56Ni)

and thus hotter more ionisation less intermediate

mass elements (Ca, Si).

(87 SN Ia spectra from Gemini – 3 years of data)

Lower EWs in late-type galaxies, whatever the redshift

Blois 2008 21 May 2008 30

SNe Ia as standard candles

o 1990s: a simple temporal stretch correction aligns the light-curves in a sample of nearby SNe Ia

o Study cosmology with distant SNe Ia• The High-Z project (95-97):

10 SNe Ia in 0.3<z<0.97• The SuperNova Cosmology

Project (93-98): 42 SNe Ia in 0.18<z<0.83

days

days

~flux

Blois 2008 21 May 2008 31

SNLS vs 1st generation experiments

Ex: the High-Z experiment:

o Small CCD camera few nights of observation

only 10 high z SNe Ia

poor light-curve sampling

o Only two filters

Blois 2008 21 May 2008 32

the Universe expansion is accelerating

0.01 0.1

1z~ apparent fluxB

= L /4πdL2

withdL(z,H0,ΩM,ΩΛ,w)

fain

ter

furt

her

back

in t

ime

more redshift more expansion

Cosmology from SNe Ia : 1997-98