The Delay Time The Delay Time Distribution of Type Ia Distribution of Type Ia

Supernovae: Constraints Supernovae: Constraints on Progenitorson Progenitors

Chris Pritchet (U. Victoria), Mark Sullivan (Oxford), Chris Pritchet (U. Victoria), Mark Sullivan (Oxford), Damien LeBorgne (IAP), Matt Taylor (PUC Chile), Damien LeBorgne (IAP), Matt Taylor (PUC Chile),

+ SNLS Collaboration+ SNLS Collaboration

UWO Sep 2009 3Mt Wash Feb 2009Mt Wash Feb 2009

brightness1010 Lsun,

Std candle

progenitor 1 or 2 white dwarfs:

mechanism mass transfer or merger

progenitor age ~108-10 yr:

evolution with z Little:

Remnant (ns/bh) no

Metals ejected Fe, Ni …

109 Lsun

Not std candlemassive star

core collapse:

~107 yr

(1+z) 2-4 :

yes

O, Ne, Si …

or

SNe Ia CC SNe

brightness1010 Lsun,

Std candle

progenitor 1 or 2 white dwarfs:

mechanism mass transfer or merger

progenitor age ~108-10 yr:

evolution with z Little:

Remnant (ns/bh) no

Metals ejected Fe, Ni …

UWO Sep 2009 4Mt Wash Feb 2009Mt Wash Feb 2009

or

SNe Ia

energy release COFeno H in spectrumlight curve shapepresence in old stellar pops

progenitor mechanism – 2 broad classes

5

SN Ia Progenitors - 2 Broad Classes

Single Degenerate - white dwarf + 2ndary evol. (M ~ 1.4 Msun at explosion)

Double Degenerate - 2 white dwarfs (Mtot >= 1.4 Msun at explosion)

Key point: white dwarf maximum mass M = 1.4 Msun (Chandrasekhar mass)

brightness1010 Lsun,

Std candle

progenitor 1 or 2 white dwarfs:

mechanism mass transfer or merger

progenitor age ~108-10 yr:

evolution with z Little:

Remnant (ns/bh) no

Metals ejected Fe, Ni …

UWO Sep 2009 6Mt Wash Feb 2009Mt Wash Feb 2009

or

SNe Ia

energy release COFeno H in spectrumlight curve shapepresence in old stellar pops

progenitor mechanism – 2 broad classes

7

Type Ia SNe as Standard Candles

Bright - seen to cosmological distances

Max brightness makes an excellent standard candle - ±6% distance errors

Standard candle seems to have a physical basis

SNeIa are “well-understood” - thermonuclear disruptions of C+O white dwarfs - std physics

Systematics – possibly, but ample opportunity to study with potentially hundreds of objects

But …

explanation of stretch – L relation explanation of colour – L relation nature of scatter in L after calibration nature of progenitor

Delay time distribution

• DTD(t) = rate of supernovae as a function of time from a burst of star formation

• SNR(t) = SFR(t) ★ DTD(t)

log t

SFR(t)

DTD(t)SNe/yr/1010 M

Importance of DTD(t)

• potential to discriminate among progenitor models

Greggio 2005

DTD History

• pre-1990 – “prevailing wisdom” was that all SN Ia were old because they occur in E/S0 galaxies

• by 2004 – SNe Ia have higher rates in young galaxies – both young and old progenitors

Recent DTD Determinations

Totani et al 2008:Subaru/XMM survey65 variable objectsages from SED fitting

• from age/SFH estimates of SN host and field galaxies (SN age ~ galaxy age)

Recent DTD Determinations• from age/SFH estimates of SN host and field

galaxies

MaozMaoz et al 2010:LOSS survey82 SNeIa SFH from SDSS

Supernova Legacy Survey (SNLS)

• 2003-2008, 4 deg2, ugriz, 4d samples, CFHT 3.6m+MegaCam

• spec types and z (VLT, Gemini, Keck) - 370 SNe Ia (0.2<z<1)

DTD from SNLS

• completeness estimate and weight for each supernova

• host galaxy age for each supernova …• assumes host age = SN progenitor age

• … and an age for all other objects too• gives total available mass at a given age

z distribution and completeness

• Perrett et al 2011

SNIa*SNIa

SN weightingwi =(1+ zi )⋅

1ε( field,ti ,zi ,si ,ci )

⋅1Δti

⋅1

Nseason

Perrett et al 2011

length of eachobserving season

SNe / year(all fields,rest-frame)

# of observing seasons

Pegase/zpeg ages and redshifts

• mass, SFR, age, z for different evol scenarios

DTD Calculation

• Use only SNe with hosts in magnitude-limited catalogue

• assumes that SN DTD does not depend on host galaxy mass

• In each time bin of DTD t1t2, sum wi values for SNe with t1<ti<t2; normalize by host mass in time bin:

Dt1→ t2=

wit1→ t2

∑Mt1→ t2

2 different M(t) methods

• 0.2 < z < 0.75, 4 SNLS fields (3.6 deg2)• dashed=SFR(z), solid=zpeg SED fits

log t

log M(t)log

M

Hopkins and Beacom 2006

SFR(z)

DTD

• other z ranges give the same result

DTD from 2 different M(t) methods

• 0.2 < z < 0.75, red=SFR(z), black=obs

Comparison with Totani et al 2008

Mannucci

Totani

t-1

Power-law fit

t-1.35

Two power laws

t-0.7

t-3

cutoff real

Comparison with DD

solid – Mennekens et al 2010dotted – Ruiter et al 2009dashed – Yungelson and Livio 2000

Comparison with SD

solid – Mennekens et al 2010dotted – Ruiter et al 2009dashed – Hachisu et al 1999dash dot – Han and Podsiadlowski 2004

Further corrections

• Have assumed that TSN=<Thost>. Not necessarily true

• iterative approach to correct statistically

• correction for dead stars • slope steeper by ~0.1

• effects of bursts• effects of catastrophic errors in M or age

Making a standard candle Supernova light curve stretch s

csMMOBB βα −−+= )1(

aka Phillips relation

Stretch dependence of DTD

• not due to age systematics• two types of progenitors?? or …

Conclusions

• SNIa DTD may be more complex than a simple ~ 1/t power-law

• match to DD population synthesis models• pop syn needs further work

• s<1 and >1 show differences in DTD below 109 yr – different progenitors? or PDF of ages?