Properties of Mass Loss f rom Supernova Progenitors Determined from Radio Observations

Properties of Mass Loss fromSupernova Progenitors Determined

from Radio Observations

Nino Panagia Nino Panagia (STScI, INAF-CT, Supernova Ltd(STScI, INAF-CT, Supernova Ltd))

Mass Loss Return

From Stars to Galaxies

29 March 2012

Digging into the Past of Digging into the Past of Exploding Stars: Exploding Stars:

Radio Observations of Radio Observations of SupernovaeSupernovae

Nino Panagia Nino Panagia (STScI, INAF-CT, Supernova Ltd(STScI, INAF-CT, Supernova Ltd))

Mass Loss Return

From Stars to Galaxies

29 March 2012

29 March 2012 Nino Panagia Radio Supernovae

3

Largely based on work done (over MANY years) in Largely based on work done (over MANY years) in collaboration with:collaboration with:

• Kurt W. Weiler (NRL)Kurt W. Weiler (NRL)

• Dick A. Sramek (NRAO)Dick A. Sramek (NRAO)

• Schuyler D. van Dyk (IPAC) Schuyler D. van Dyk (IPAC)

• Marcos J. Montes (NRL) Marcos J. Montes (NRL)

• Chris Stockdale (Marquette U) Chris Stockdale (Marquette U)

• Stuart Ryder (AAO)Stuart Ryder (AAO)

• Stefan Immler (GSFC)Stefan Immler (GSFC)

• et al…et al…

Radio Supernovae


4

SN classification

Ia: termonuclear explosion of a WD

II-Ib/c: gravitational collapse of a massive star


66

Radio Supernovae (RSNe)

• ~70 RSNe detected in the radio

• ~30 with light curves

• ~20 with extensive light curves

• Many upper limits (~300 )


7

Our most recentreviewon RSNe

ARAA 40, 387, 2002


8

Why RSNe?

The study of radio emission provides valuable

insight into SN shock/CSM interaction

– History of pre-SN evolution– Mass-loss rate and changes therein– Nature of the progenitor


9

What RSNe?

• All RSNe have in common: – Nonthermal synchrotron with high TB

– Decrease in - dependent) absorption with time

– Power law flux density decline after max

– Final approach to optically, thin constant

• Interesting variations:– Clumps or filaments in CSM

– Possible binary companions


10

Radio Supernovae

• “Turn on”, first at high , progressing to lower

• Power-law decline after maximum at each

• Transition from “optically thick” spectral index (where S ~ + ) to an “optically thin” asymptotic value

• Nonthermal emission with very high TB

SN IbSN II-L

SN IIn


11

Standard Model Physical Parameters(Chevalier 1984)

• Red Supergiant (RSG) progenitor• Slow (10 km s-1), dense (10-6-10-4 M yr-1) wind r-2 [ Mdot/(wwind r2)] density profile • CircumStellar Medium (CSM) ionized by SN

UV/X-ray flash• Relativistic electrons & enhanced magnetic field

arise from shock/CSM interaction• Ionized CSM provides initial absorption (f-f)• Synchrotron Self-Absorption (SSA) may play a

role at very early times

}SNII


12

Circumstellar Interactionparameterized radio light curves

[Weiler et al. 2002]

K5, K6 = internal SSA and f-f optical depth on day 1

(

>105 K


13

Circumstellar Interaction:Estimation of progenitor’s

Mass-loss Rate(from f-f absorption)

(Weiler et al. 1986, 1990, 2002, 2007)


14

Type Ia Supernovae


16

VLA Observations of SNIa

• Observed 27 SNIa over

24 years of monitoring

•

Panagia et al 2006

NO detection


17

All SNIaat once

The lowest 2 upper limit is about 3 10-8 M yr-1


18

The most StringentUpperLimits

M < 310-8 M yr-1


20

Radio Constraints on SNIa Progenitors

The theoretical requirement that accretion rates higher than 10-7 M yr-1 are needed to make a WD mass exceed the

Chandrasehkar mass combined with the radio upper limit to the mass loss rates < 310-8 M yr-1 :

– Rules out accretion via stellar wind from a companion (accretion efficiency <30%)

– Allows accretion via Roche lobe overflow if the mass transfer efficiency is higher than 60%

– Allows the “double degenerate merger”


21

Core-collapse Supernovae


22

Type IIL Supernovae


23

Optical/Radio SN1979C • Optical • Radio

Type IIL

My FIRST supernova!!!


24

SN 1979C: The First 10 Years

A periodic modulation?

20 cm

6 cm


25

SN 1979C: A Sinusoidal Fit

Spiral pattern expected for a binary system including 15 and 10 M stars that are orbiting around each other with a period of ~5000 days

A wide binary system


26

SN1979C: Twenty Years of Observations

An increase ofCSM density

About 20,000 yearsbefore explodingthe progenitorejected a discrete shell?

Pulsational instability?

Panagia & Bono 2001


27

SN1980K - RadioAn decrease ofCSM density (/2)

Type IIL


28

Pre-SN Evolution of Massive Stars[Bono & Panagia 2001, and in prep.]

Hydrostatic Evolution + Pulsational Instabilities


29

Type IIn Supernovae


30

SN 1986Jone of the brightest

RSNe

Type IIn


31

SN 1986J

The gradual riseimplies a clumpyabsorbing medium


32

SN1986J (1999 VLBI obs.)And clumpy it is…


33

SN 1987A or

The unusual explosion ofa normal Type II supernova

[~18M progenitor]


34

SN1987A – Optical [Type II pec]


35

SN 1987A – Radio Ring


36

10 years after explosion the ring started lighting up…

More than 20 spots now seen to brighten, due to the collision of the ejecta with the central ring.

Over the next decades, as the entire ring will light up, the Evolutionary history of the star’s mass loss will be revealed

1996

2006


37

SN 1987A Radio Evolution

Early Evolution:- Fast rise- Low luminosity

Low density CSM

Late Evolution- Gradual Rise- High Luminosity

High density medium


38

Optical, X-ray, & Radio light curves of SN1987A ring

optical

radio

X-ray

The fireworks have started!

1990 1995 2000 2005


39

Type IIb Supernovae


40

SN 1993J in M81 (NGC 3031)Type IIb


41

SN 1993J: Radio Observations [Type IIb]


42

Circumstellar InteractionSN shock wave with the pre-SN dense wind

[Chevalier & Fransson 1994; Fransson, Lundqvist, & Chevalier 1996]

~109 K (adiabatic)

~107 K(radiative)

<104 K (cool shell)

Strong X-ray emissionexpected


44

SN 1993J – Radio Light Curves(only SSA= synchrotron self-absorption)


45

SN 1993J – Spectral Index Evolution(SSA only)


46

SN 1993J – Brightness Temperature(SSA only)


47

SN 1993J – Radio Light Curves(only FFA=free-free absorption )


48

SN 1993J – Spectral Index Evolution(FFA only)


49

SN 1993J – Brightness Temperature(FFA only)


50

SN 1993J – Radio Light Curves(SSA + FFA)


51

SN 1993J – Spectral Index Evolution(SSA + FFA)


52

SN 1993J – Brightness Temperature(SSA + FFA)


53

SN 1993J – Late (>day3100) Evolution(negligible SSA & FFA)


54

SN1993J: VLBI Observations

Expansion of SN 1993J from age 5 months to age 31 months

[Marcaide et al. 2000, 2007]


55

SN 1993J in M81 [SNIIb]

Mass loss rate jumped up to 6 10-6 ~8000 years before explosion and decreased down to 5 10-7 M yr-1 prior to explosion

Evidence for “shell ejection” possibly due to pulsational instability


56

SN 1993J – Radio Observations• Radio data are the most detailed set for any SN (except

possibly SN 1987A) in any wavelength range– ~650 measurements from 0.3 -115 GHz (3 mm – 1 m)

• Conclusions– The best overall fit to the ``early'' (day <3100) includes both

non-thermal SSA, and thermal FFA– Both the radio and the X-rays imply CSM~ r-s, with s~1.6 the progenitor mass-loss rate and/or wind speed varied before

explosion

– After day ~3100 the radio and X-ray decline rate steepens significantly and can be described very well as an exponential with an e-folding time of ~1100 days.

• Evidence for shell ejection?

• Evidence for a RSG BSG transition?


57

SN 2001gd [Type IIb]very similar to SN 1993J


58

Mass-loss Rate Evolution for SNIIb 1993J & 2001gd

Evidence for shell ejection

few 103 years before explosion


59

SN 2001ig - A Chronology[Ryder et al. 2004]

• Dec 10.43 2001 UT: discovery by R. Evans in NGC 7424 (SAB(rs)cd, D=11.5 Mpc, = –41º).

• Early spectroscopy (LCO 6.5m, ESO NTT) suggested similarities with SN 1987K and SN 1993J (Type IIb).

• Dec 15 UT: Detected with ATCA at 8.6 GHz.

• May 2002: Detected with ACIS-S/Chandra – L(0.2–10 keV) ~ 1038 erg s-1.

• Oct 2002: Transition to Type Ib/c complete.

Type IIb


60

Radio “light curve”


61

SN 2001ig best “simple” fit


62

Comparison with SNIIb 1993J and 2001gd

10-6


63

Episodic mass-loss?

• Bumps and dips with P ~ 150 days.

• vexp ~ 15,000 km s-1 R = 0.006 pc.

• w = 10 km s-1 t ~ 600 yr.

• t >> stellar pulsation timescales, but perhaps consistent with thermal pulse (C/He flashes) periods in 5–10 M AGB stars.

Deviations from fit


64

Analogy with SN 1979C

Montes et al. 2000

Schwarz &Pringle 1996


65

Pinwheels in the sky

WR104Tuthill et al.

1999

WR98aMonnier et al.

1999


66

SN 2001ig summary

• SN 2001ig is a Type IIb comparable with SN 1993J, but late-time radio light-curve akin to SN 1979C.

• Spectral index evolution changes in CSM density, rather than optical depth.

• Mass-loss variability (~600 yr) likely due to:– red-supergiant progenitor wind modulated by eccentric

orbital motion about massive binary companion?


67

Type Ib/c Supernovae


68

SN 1983N in M83

Prototypical SNIb

the discovery ofa new type of SN

Sramek, Panagia & Weiler 1984


69

Supernova 1994I in M51 the Type Ic “Rosetta Stone”

Weiler et al 2011

Type Ic

Weiler et al 2011


70

20 cm

6 cm

3.6 cm

2 cm

1.3 cm

Early emission is dominated by synchrotron self absorption

(assuming vexp=20000 km s-1)

Supernova 1994I in M51brightness temperature

SN 1994I – Spectral Indices


72

SN 1994I: Emission Modulation


73

Consistent with a B-type companion at a separation of about 310AU

The Caltech/NRAO Radio Survey of SNIc (z < 0.1)

many more SNIc

(Kulkarni et al. 1998, Berger et al 2003, Soderberg et al. 2004a, 2005a, 2006d, +

A dispersion in light-curves- peak time- luminosity- decay rate

Progenitorparameters:- mass- metallicity- binarity

(Weiler 1986; Kulkarni 1998; Berger 2002,2003; Soderberg 2004,2005a,2006b,c,d & in prep)

Rough Numbers:Targeted 170 SNe Ibc (since 1980’s), 2 GRB-SNe• 60% Ic (5% BL=broad lines), 20% Ib, 20% Ib/c• 19 detections - roughly equal between Ib vs Ic, 2 BLs,2 GRB-SNe• detection rate: 30% Ib, 10% Ic, 25% BL, 100% GRB-SNe Similar fraction of massive stars in close binaries (Podsiadlowski)

What fraction of non-detections (esp. BL) are like 2002ap? EVLA

A Search for Engine-Driven SNe Ib/c (d < 150 Mpc)

7529 March 2012

Statistical Properties of CC-RSNe


76

“Our” SNII, SNIIL, SNIIn


77

Supernova Type Δtpeak

[days]Lpeak

[1026 erg s-1Hz-1]Mass-loss-rate[10-6 M yr-1]

1978K II 802 125 152

1981K II 34 2.1 1.5

1982aa II? 476 1270 103

1970G IIL 307 14 68

1979C IIL 556 26 106

1980K IIL 134 12 13

1986J IIn 1210 197 43

1988Z IIn 898 232 114

1995N IIn 400 14 61

2000ft IIn? 320 12 50

2004ip IIn? >100 35 >100

“Our” SNIIb & SNIb/c


78

Supernova Type Δtpeak

[days]Lpeak

[1026 erg s-1Hz-1]Mass-loss-rate[10-6 M yr-1]

for w = 10 km/s

1993J IIb 133 15 0.5-5.9

2001gd IIb 80 38 7-56

2001ig IIb 74 35 22

2008ax IIb ~20 1.2 9

1983N Ib 12 0.14 0.87 (87*)

1984L Ib 11 0.26 0.74 (74*)

1990B Ib 38 5.6 2.7 (270*)

1994I Ib/c 34 15 1.4 (140*)

* w = 1000 km/s


79

Lpeak-tpeak relation for “our” RSNe

2002ap

1987A

Panagia et al. 2012

Weiler et al. 1998

Mass-loss Rates vs tpeak


80


81

The Lpeak-tpeak relation for RSNe

v = const

L t 1.4

1998bw

2003dh

2002ap

1987A

Panagia et al. 2012Weiler et al. 1998


82

SNII Progenitor Mass Loss Rates (from f-f Absorption)

.

Panagia et al. 2012


83

The SN-GRB Connection


84

The SNIc – GRB Connection• Some GRBs are identified with powerful SNIc, e.g.

– GRB980425 = SN 1998bw (z=0.0085)– GRB030329 = SN 2003dh (z=0.1685)

• Some SNIc are – exceptionally luminous (“hypernovae”?), e.g. SN1997ef,– highly polarized (asymmetric ejecta), e.g. SN 2002ap (BUT not a GRB nor a bright RSN)

• Both SNIc and GRBs appear to be associated with star forming regions


85

The historical SN-GRB Connection

SN 1998bw


86

ATCA Observations of SN 1998bwthe first SN-GRB Evidence

http://www.narrabri.atnf.csiro.au/\~mwiering/grb/grb980425/

N.B.: Is a linear-linear plot


87

SN1998bw/GRB980425[Weiler, Panagia & Montes 2001]

Evidence for periodic modulation? binary system? WR progenitor?


88


90

SN1994I

SN2006aj

SN2003lw

SN1997ef

SN2002ap

SN2003dh

SN1998bw

GRB-SN Nomoto 2007


91

GRB 060614

NO detectable SN


92

Light curves of Ic SNe: GRB-SNe, broad-lined SNe, normal SNe

-13.7

060614upper limit


94

`bright’ branch:

Hypernovae

`faint’ branch:

SN1997D-like

Type Ic SNe Nomoto et al. 2003

Faint type II collapse of massive stars with an explosion energy so small that most 56Ni falls back into the BH

GRB 060614 “fall-back” SN? Tominaga et al. 2007


95

SUMMARY - I• SNe classes are distinct in radio emission properties (thus

distinct in CSM environments):– SNe Ia are undetectable at VLA’s limiting sensitivity (so far)– SNe Ib/c turn on and off quickly– SNe II evolve more slowly

• RSNe are sensitive to Mdot/wwind (~ pre-SN mass loss rate)

• RSNe sample the CSM properties of the pre-SN wind density & structure -- unique stellar evolution probe

• Because vwind ~ 10 km/s and vshock ~ 104 km/s, radio observations are a “time machine”


96

SUMMARY - II• SNIa: Very small (<3 x 10-8 Msun yr-1) matter outflow from pre-SNIa

systems

• SNII: – Red Supergiant Winds (but SN 1987A)– Essentially all change their evolution on a timescale of ~ x 104 yrs– Clumpy CSM and/or cylindrically symmetric density distributions– Variable mass loss rates over 104 years time scales– Evidence for pre-SN binary system wind collisions

• SNIb/c:– More tenuous CSM than SNII– Much higher shock speeds– Evolve much faster

SUMMARY - IIIDerived Properties of RSNe

• Type Ia SNe never detected: very low pre-SN mass loss rates (<3 x 10-8 M yr-1)

• Type II SNe -- slowly evolving radio emitters, with:– Flatter spectrum: = > -1.0 (generally)– Slow decay: - = 0.7 - 1.4– Mass loss rates 10-6 - 10-4 M yr-1

– Most how late time deviation from a smooth radio light curve• Type Ib/c -- rapidly evolving radio emitters, with:

– Steep spectrum: = < -1.0 (generally)– Fast decay: - = 1.2 - 1.6– Mass loss rates 10-7 - 10-6 M yr-1

– Associated (rarely) with GRBs• Deviations from this “standard” picture are, of course, the

most interesting


98

SUMMARY - IVComparison of RSNe & rGRBs• Differences with “standard” RSNe

– Interstellar scintillation (ISS)– Cosmological (z ~ 1)– Relativistic effects

• Higher mass-loss rates than normal Type Ib/c (but affected by assumptions)

• Much more radio luminous than normal Type Ib/c (but not if boosted by ~ 10)

• Rare• “Fast-Hard” GRBs still undetected in the radio

8 June 2006 Expansion Rate of the Universe

80

THE END

.

29 March 2012 99Nino Panagia Radio Supernovae

Properties of Mass Loss f rom Supernova Progenitors Determined from Radio Observations

Documents

radio supernovae rsne

radio supernovaeturn

study of radio emission

sn classificationia

mass loss returnfrom

ff optical depth

extensive light

main sequence core h