Limits on Ia progentors from pre-explosion imaging :

Lessons from core collapse progenitor studies

S. J. SmarttAstrophysics Research CentreQueen’s University Belfast

2

Detected SN rates in Local Volume

19980101-20080630 : 139 SNe discovered in galaxies with Vvir<2000 kms-1 (13.2 SNe yr-1)

LOSS : controlled, within ~60Mpc.

Smartt et al., 2009

LOSS : Li et al. (2010)

M101

Core collapse :~26% SN-HST image coincidence rate

Type Ia :~38% SN-HST prediscovery coincidence rate

Detection of progenitors

• Within the volume limited, 10.5 yr survey for progenitors : three “gold” events

• SN2008bk, SN2005cs, SN2003gd

• Red star identified coincident with all three.

• Typical magnitudes :– Mv ~ -4.5 ; MI ~ -6.5

• Discovery papers : Van Dyk et al. 03, Smartt et al. 04,

Maund et al. 05, Li et al. 06, Mattila et al. 08.

WFPC2 F300W, F606W, F814W HRC F435W, F555W, F814W

Figures from Smartt 2009 ARAA

Elias de la Rosa & Van Dyk

The disappearance of 2003gdSN2003gd:V=25.8 ± 0.15 V–I=2.5 ± 0.2Smartt et al. 04, Van Dyk et al. 03

Aug 2003

Maund & Smartt (2009)

Four confirmed cases of disappearance : SN1987A , SN1993J, SN2003gd, SN2005gl (Gal-Yam & Leonard 2009)

Sep 2008

Aug 2002 May 2002

Mass estimates from stellar evolutionary tracks

Red points : Milky Way red supergiants (Levesque et al. 2005) STARS stellar evolutionary tracksSN progenitors : SN2003gd (black), SN2005cs (blue box)

Maund et al. (2005), Li et al. (2006)

Smartt, ARAA 2009

Other examples: no detection

• SN1999gi in NGC3184,

• HST U+V pre-explosion

• D=11Mpc (Leonard et al. 2002)

• M 12 M

• SN2001du in NGC1365

• HST UVI pre-explosion

• D=17Mpc (Cepheid Key P.)

• M 15 M

Smartt et al. 2001

Smartt et al. 02, Van Dyk et al 02

Attempts to find Ia progenitors

• Mannucci & Maoz 08 : NGC1316 radio galaxy in Fornax hosting 06dd, 06mr

• D = 19 Mpc ; μ = 31.4 ; Av(foreground) = 0.06m

• HST-ACS prediscovery imaging

9

Filter Exptime 3σ mag

06dd & 06mr MB,V,I +AB,V,i

F435W 1240 26 -5.8

F555W 6980 26 -5.5

F814W 4860 25.4 -6.1

Quantitative constraints ?• Stellar evolutionary tracks from

Girardi et al. 2000, Salasnich et al. 2000

• Rules out : – stars >6M at tip of AGB

– Young post-AGB stars >4M

– Post RSG stars >9M

• At d=5 Mpc ; μ = 28.5

• Limit would become MI ≈ -3

• Even for MV ≈ -2 then – main-sequence stars of 6-7Mwould go undetected

• Difficult to rule out intermediate mass progenitor systems

10

• Mannucci & Maoz 08

8 further type Ia progenitors• Nelemans et al. 08 (2007sr) ; Voss & Nelemans 08 (2007on)• MSci project in 2006 – (Neale Gibson, QUB) • HST imaging of pre-explosion sites for 6 Ia SNe. • No coincindent objects, upper limits below • 2004ab – closest, but in galactic bulge, high SB, no meaningful limit

11

03cg 03gs 03hv 04w 05df 04ab 07sr 07on

F450W -6.2

F475W -3.9 -4.5

F555W -6.2 -6.3 -5.5

F606W -7.1 -5.1

F814W -8.3 -6.8 -7.5 -4.5

F850LP -5.5

D(Mpc) 17.3 17.9 21.3 17.3 24 14.1 12-25 20

Hopeless ?

• SN2002ic-like with AGB companion ?

• Han & Podsiadlowski 06 : super-soft channel, mass-loss results in fainter companion

– Ia systems with AGB companions (~ 1%)

– 1 Type Ia per decade within 10Mpc

– ~40% chance of HST– ~50% chance of very deep HST

• Go back after to search for remaining companion ?

12

Stellar population ageing

• Transient NGC300-OT (μ=26.5), nature uncertain, deep HST pre-explosion images. Progenitor visible in Spitzer

• Age of surrounding stellar population ?• Isochrone fit to 50pc region• Gogarten et al. 09 : “stars that formed in a single burst between 8 and 13 Myr ago,

with 70% confidence.” 13

Stellar populations around SNRs

• Badenes et al. 09 ; spatially resolved SFH of LMC compared with SNRs

• Statistically powerful method of dating a progenitor population

14

Statistical – not deterministic

• Valid way of adding value to progenitor studies e.g. Van Dyk et al. 1979C

• Not sensible to apply to single objects for deterministic age

• Coeval ? Small numbers of stars. How to pick a radius ?

• Type Ia within 28Mpc with HST :– ~14 with pre-explosion imaging– Further ~15-20 with more imaging (Li,

Filippenko, Van Dyk snapshot surveys)– Some subset should have homogeneous

filter/depth combinations– Comparisons with II-P and Ibc stellar

populations ?

15

Van Dyk et al. 1999

X-ray detections ?SN2007on : Voss & Nelemans 2008

• SN Ia in NGC1404 (d = 20pc)

• Astrometric calibration of Chandra image with 2MASS and DSS sources : 0.9 ±1.4”

• Source detection (14.1 ± 4.6 counts), 4σ

• L ≈ 3 x 1037 ergs-1 (at 20Mpc)

Roelofs et al. (2008)

• More accurate alignment with WFI (pre-explosion), EFOSC (post-explosion) to give 7 sources in the X-ray

• Offset : 1.15 ± 0.27”

16

Chandra

HST (MB,I > -4.5)

Swift

Not coincident, but source gone

• Bootstrap calibration : 1.15 ± 0.27”• But source is gone – looks convincing (80-99% CL that source is fainter,

depending on chosen aperture)• Was it flux from the progenitor ?

17

The disappearance of 2003gdSN2003gd:V=25.8 ± 0.15 V–I=2.5 ± 0.2Smartt et al. 04, Van Dyk et al. 03

Aug 2003

Star A in HST : Differential astrometry:

r = 13 ± 33 mas

But r = 147± 71 mas in Gemini image (10 point sources).

We assumed blend ? But whole object has disappeared.

RMS likely underestimated

Sep 2008

Aug 2002 May 2002

X-ray source :progenitor or background AGN ?

• Astrometry only : probability of relation small (~1%), but equal to prob of chance alignment (~1%)

• Mis-alignment could just underestimate of uncertainties

• Significant the source is close and gone

• Deeper Chandra – useful but can’t enhance the significance of it being the progenitor

• Deeper HST - faint source near the X-ray position could very likely be AGN : 1224s in F814,

I ~ 26.5

19

Summary • Direct pre-explosion images in the optical.

- Deepest limits : MB,V,I ≥ -4

- Not restrictive : not deep enough to rule out any viable models

• Stellar population analysis

- More promising. Data probably exists, or can be obtained

- Needs careful methodology

- Large effort to do it properly - not worth doing badly

• X-ray searches (high priority)

- More work on 2007on (HST and Chandra)

- Could target Chandra for nearby galaxy catalogue

- Check every possible target

€

→

SS X-ray source comparison

• Nelemans et al. 2009• SN2007sr :L Grey

boxes: upper limits for 2007sr

• Bottom : blue and red bars SN2007on proposed “progenitor” detection

21

MW,LMC,SMC, SS sources

M101, M51, M83, N4967

2007on

• X-ray spectrum – not clearly a SS X-ray source

22

8m AO imaging - new approach

Gemini and VLT diffraction limited K-band AO images0.08” and ~0.02” pixels = well sampled PSFTypical ~15 - 20 stars identified in common between KS and V or I bandDifferential astrometry ~ 20 milliarcseconds RMS

Gemini + Altair :Crockett et al. 07

VLT + NACO :Mattila et al. 08

Lessons Learned

24

• 10 years of searching – not as easy as first thought

• 93 CCSNe within 28Mpc : ~32 with good pre-explosion images

• 4 high significance, unambiguous detections. 3 questionable ones,

plus 3 on unresolved host clusters

• 5-10% yield (but large number of upper limits restrictive)

• High resolution images (HST or 8m AO) crucial

The future :

• Extend to another 10-20 years (any takers….!)

• Focus on the 10Mpc volume : HST MCTP for all high SFR

galaxies (800 orbits). Huge legacy science (D. Calzetti)

• Guaranteed ~15 CCSNe in 10 years. With full mosaic, deep

WF3/ACS of the galaxies, discovery potential high

Summary of II-P progenitors : 10.5yr search

Smartt et al. 2009,MNRAS : used Cambridge STARS code, homogeneous analysis, consistent luminosity and mass estimates

Does a Salpeter/Scalo IMF fit ?

• Solid : Salpeter IMF

maximum mass of

16.5M

• Dashed : Salpeter

IMF, maximum mass

of 30M

• Lower mass limit : 7-

8M (WD limits :

Williams talk)

Summary

Wolf Rayet stars : not Ibc progenitors?

28

• LMC (or M31) WR magnitude distributions ~5-10% probability we have had

no detections by chance

• SN2008ax : detection of WNL progenitor of a IIb (Crockett et al. 08)

SN2008ax detection MV = -7.4 From Crockett 2009

(PhD Thesis), See also Van Dyk et al. 03Maund & Smartt 05,Maund et al. 05Gal-Yam et al. 05

SN2008ax : IIb + WNL progenitor ?

29Crockett et al. 2008, Pastorello et al. 2008

Comparison of codes

Maximum likelihood approach

•mmin : is better measured with the

detections only. Unconstrained IMF if limits used.

• mmax : calculated using both

detections and limits : mmin = 8 +1

-1.5 M

mmax = 16.5 1.5 M