other models for type ia supernovae pair instability ...woosley/ay220-13/lecture18.13.pdf · pair...

Lecture 18

Other Models for Type Ia Supernovae

Pair Instability Supernovae,

Pulsational Pair Supernovae,

and Population III

He

0.05 - 0.2 M

C,O

0.7 - 1.1 M

M= 2 - 6 x 10−8 M

y−1

SUB-CHANDRASEKHAR MASS MODELS A critical mass of He accretes from a companion. The helium ignites and detonates. This may set off a secondary detonation of the carbon

Nomoto (1980, 1982) Woosley et al (1980, 1986, 1994, 2011)

Moll and Woosley (2013)

Single point detonation. Detonation ignites in the helium layer, propagates around the white dwarf

and collides at the anti-pode. The collision induces a

detonation in the carbon- oxygen core.

Note similarity to

Chicago “GCD” model for Chandrasekhar mass

white dwarfs

Study of asynchronous multiple ignion points by Moll and Woosley (2013). All models studied detonated the CO core provided the helium itself detonated.

Advantages

•  Should be frequent event. Donor could provide H or He. Symbiotic variable, AM Canum Venaticorum variable, etc,

•  Detonation physics easier than deflagration physics

•  Makes some useful missing isotopes by explosive helium burning (44Ca).

•  Naturally provides a range of 56Ni masses

•  Light curves can, in some cases, be similar to those observed for common SN Ia

Symbiotic variable - white dwarf in a wide binary with a cool giant. Mass exchange by wind. AM CVn star – white dwarf in a compact binary with another compact star – e.g., a He dwarf. Short period. Mass excahnging. Gravitational radiation source. Product of common envelope evolution

Disadvantages

•  Chiefly spectroscopic. He detonation produces very high velocity material not seen in the spectrum of SN Ia. The ionization state may not be good.

•  Should produce a greater variety of supernovae than are seen.

Neither of these objections are fatal and might be relieved if the mass of the helium layer could be made trivially small. This could work if the white dwarf is hot to begin with and if the accretion rate is high.

Recent relevant studies:

•  Bildsten et al (2007) – an interesting variety of subluminous supernovae might come from He

detonation alone – “point Ia supernovae” . Smaller He shells may detonate than previously thought possible [12C(p,)13N(,p)16O accelerates ignition] •  Fink et al (2010) – essentially any mass He shell

that detonates will detonate the CO core (assumed spherical or cylindrical symmetry)

•  Sim et al (2010); Kromer et al (2010) – carbon cores detonating w/o He shells look like ordinary SN Ia

Two examples of sub-Chandrasekhar mass explosions

original white dwarf core helium shell

He shell only explodes

Entire star explodes

The general class of sub-Chandrasekhar mass models can give a wide variety of transients ranging from very luminous SN Ia to super “novae”.

Some of these look like SN Ia… Model 10HC (hot 1.0 solar mass CO WD accreting at 4 x 10-8 solar masses per year) – peak light spectrum vs observations.

91T was an unusally bright SN Ia

“Hot” WD here means a white dwarf with L = 1 Lsun

Width luminosity relation for models whose CO cores explode. Closes circles are for hot WD accretors.

The effect of the detonating He shell on the spectral evolution (Model 9C -  0.90 solar mass CO core with 0.12 solar masses of He.

This set of calculations shows the effect of helium shells of increasing mass on the peak light spectra. Only those with the lowest helium shell masses look like SN Ia. Where are the others?

He – deflagration model. What happens if the helium does not detonate. (A CO WD remnant is left behind in a hot luminous state.) Note large production of 44Ti, 48Cr, and 52Fe but little 56Ni.

– 60 –

0 2 4 6 8 10 12days since explosion

-10

-11

-12

-13

-14

-15

-16

abso

lute

mag

nitu

de (p

lus

offs

et)

U+1

B

V - 0.5

R - 1.0

I - 1.5

2000 3000 4000 5000 6000 7000 8000 9000wavelength (angstroms)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

rela

tive

flux

SiIICaII CaIISII

2000 3000 4000 5000 6000 7000 8000 9000wavelength (angstroms)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

rela

tive

flux

1 day

2 days

3 days

5 days

Fig. 18.— Broadband light curves (left) and spectral time series (right, marked in time sinceexplosion) of Model 8HBC1, which involved the explosion of a low mass helium shell. The

light curve is dim and evolves quickly; the spectra show notable features of intermediatemass elements moving at velocities near 9,000 km s!1. The observational characteristics ofModels 7B1, 9A1, and 10HCD1 would probably be similar.

Light curve and spectra of Model 8HBC1 – a helium deflagration model. Characteristic speed 7000 km s-1. Note – quite faint. Powered by 48Cr decay.

Possible counterpart SN 2008ha (Foley 2009). Faint, low v, “SN .Ia”

The difference between helium deflagration and detonation for models whose CO core does not explode.

Minimum helium shell mass required to initiate a CO core detonation for hot WD accretors (bottom) and cold WD accretors (top).

Accretion rate vs initial WD mass where various outcomes are expected. Top solid line is for cold WDs. Ordinary Ia’s can only occur between the lower solid line and the dashed line (or get too much 56Ni in the helium shell)

Issues:

•  Does helium really detonate for very low mass helium shells? Limiting ~ 5 x 105 – 106 g cm-3

•  Where (and how often) does helium ignite?

ignition simultaneously on a shell is unphysical [2013 - probably once but not critical]

•  If the helium detonates does it do it with sufficient symmetry to induce a carbon detonation? [yes]

•  What/where are the observational counterparts of (solely) helium shell detonation (or deflagration) or of CO cores that detonate with more than 0.05 solar masses of He on top?

Woosley and Kasen (2011) Moll and Woosley (2013) Fink et al (2010); Sim et al (2010)

Two attempts at detonating the helium layer On a 1.0 solar mass CO dwarf. The top succeeded; the bottom failed. The mass of the helium shell is 0.045 solar masses (Moll and Woosley 2013)

Merging White Dwarfs

WIde variety of outcomes possible. •  If the merger results in “slow” accretion a common

outcome is the production of a neon-oxygen white dwarf. This seems to be the common case unless

the merger results in a detonation. •  Detonation can occur “promptly” in the merger initiated by compression or “delayed” initaed by shear in a merged differentially rotating object.

•  Typically the explosion is surrounded by a debris disk left over from the merger. •  Difficult to achieve detonation if the white dwarfs

have masses less than 0.8 or maybe 0.9 solar masses

Moll, Raskin, Kasen and Woosley, 2013, in prep

1.06 M + 1.06 M

Light curves and spectra in progress. Only the 0.96 plus 0.81 shows promise of looking like an ordinary

SN Ia.

Moll et al. (2013)

Pair Instability Supernovae

!  Helium core mostly convective and radiation a large part of the total pressure. ~ 4/3. Contracts and heats up after helium burning. Ignites carbon burning radiatively !  Above 1 x 109 K, pair neutrinos accelerate evolution. Contraction continues. Pair concentration increases. Energy goes into rest mass of pairs rather than increasing pressure, < 4/3. Contraction accelerates to high speed.. !  Oxygen and (off-center) carbon burn explosively liberating a large amount of energy. At higher mass silicon burns to 56Ni. At still higher masses photo- disintegration is encountered and the collapse continues. !  The star completely, or partially explodes for helium cores up to 133 solar masses. Above that a black hole is formed.

Barkat, Rakavy and Sack (1967) Rakavy, Shaviv and Zinamon (1967)

133 M

> M (helium core) ≥ ~ 35 - 40 M

Helium stars

degeneracy parameterΨ =

Nomoto and Hasimoto (1986)

He Core Main Seq. Mass Supernova Mechanism

2 ≤M ≤35 10≤M ≤85 Fe core collapse to neutron star

or a black hole

35≤M ≤60 85≤M ≤ 140 Pulsational pair instability followed

by Fe core collapse

60≤M ≤133 140≤M ≤260 Pair instability supernova

M ≥133 M ≥260 Black hole

PAIR INSTABILITY SUPERNOVAE (Thermonuclear - two kinds)

Woosley, Blinnikov and Heger (Nature 2007)

Nucleosynthesis and Energetics of Pair Instability Supernovae

Heger and Woosley (2002)

Initial mass 150 M

Initial mass 150 M?

Depends on rotation and dredge up

Initial mass: 250M���

Initial mass 250 M

Big odd-even effect and deficiency of neutron rich isotopes. Star explodes right after helium burning so neutron excess is determined by initial metallicity which is very small.

CNO → 14N α,γ( ) 18

F (e+υ)

18O α,γ( )

22Ne

η ≈ 0.002Z

Z

⎛

⎝⎜⎞

⎠⎟

65 − 130 M

He cores

Production factor of massive Pop III stars – �standard� mixing included

10−95 M

Main sequence mass

Production factor of massive and very massive Pop III stars

10 − 95 M

MS +

65 - 130 M

He cores

Very metal poor stars do not show such large odd-even effects and this is sometimes taken to imply that the number of these very massive stars was small. However, the production of the iron group does not begin until quite heavy masses so this does not preclude some subset of pair instability supernovae from being “first generation stars”.

Lai et al. 2008, ApJ, 681, 1524

28 metal poor stars in the Milky Way Galaxy -4 < [Fe/H] < -2; 13 are < -.26

Cr I and II, non-LTE effects; see also Sobeck et al (2007)

KE = Eo (20/M)Eexp B

mixing 0.1 would have been "normal"

Pair instability supernovae can, in principle, be very bright and very energetic

Some are very bright

IIp Ia

SN 2006gy

Gal-Yam et al (2009, Nature), SN 2007bi may have been (they say “was”) a pair- instability supernova. Best fit requires a helium core with about 100 solar masses and an explosion energy ~1053 erg.

The late time spectrum suggests the presence of 8 to 11 solar masses of 56Ni was made in the explosion (56Fe by the time the observations were made).

Red-shifted light curve of a bright pair-instability SN

But actually a wide variety of outcomes is possible

Blue Supergiants

Red Supergiants

Helium Cores

(Kasen, Woosley and Heger 2011)

56Ni = 10 M

Very bright pair-instability supernovae require high

mass helium cores

Mass Loss in Very Massive Primordial Stars

•  Negligible line-driven winds (mass loss ~ metallicity1/2) (Kudritzki 2002)

•  No opacity-driven pulsations (no metals) •  Continuum-driven winds likely small contribution •  Epsilon mechanism inefficient in metal-free stars

below ~1000 M! (Baraffe, Heger & Woosley 2000) from pulsational analysis we estimate upper limits: –  120 solar masses: < 0.2 % –  300 solar masses: < 3.0 % –  500 solar masses: < 5.0 % –  1000 solar masses: < 12.0 % during central hydrogen burning

•  Red Super Giant pulsations could lead to significant mass loss during helium burning for stars above ~500 M!

Particularly uncertain is the mass loss rate

for luminous blue variable stars. Could some

stars retain enough mass to be

pair SN even today?

The Pistol Star •  Galactic star •  Extremely high mass loss rate •  Initial mass: 150 (?) •  Will die as much less massive

object

The great uncertainty is whether the stars that make them both formed and retained the necessary high helium core mass when the star died.

Eta Carina Thought to be over 100 solar masses Giant eruption in 1843. Supernova-like energy release. 2nd rightest star in the sky. V = -0.8 12 - 20 solar masses of material were ejected in less than a decade. 8000 light years distant. Doubled its brightness in 1998- 1999. Now visble V = 4.7.

Pulsational Pair Instability (should be more common than PSN)

Starting at helium core masses ~35 solar masses, or about 80 solar masses on the main sequence, post carbon-burning stars experience a pulsational instability in nuclear energy generation that comes about because of pair production

80 M

Helium core 35.7 M

Central oxygen depletion

Pulsational instability begins shortly after central oxygen depletion when the star has about one day left

to live (t = 0 here is iron core collapse).

Pulses occur on a

hydrodynamic time scale for the helium and heavy element core (~500 s).

For this mass, there are

no especially violent single pulses before the star

collapses. Nevertheless, there may be mass ejection.

90 M

Helium core 41.3 M

Pulses commence again after central oxygen

depletion, but become more violent. Two strong pulses send shock waves

into the envelope.

Two days later the iron core collapses.

2 days

80 M

90M

Hydrogen envelope

Velocity at iron core collapse This will eject the

envelope at low speed

Chritian Ott (private communication 2011)

The ~40 M

cores of heavy elements

always become black holes.

(ς ≥0.5)

100 M

Helium core 45.8M

The instability has now shifted to oxygen ignition at the center of the star.

The pulses are much more violent and occur at increasing intervals.

The total duration of the pulses however is still

only 5 months. The supernova will be

almost continuous.

6 yr

“First” pulse above is actually several

Core Collapse

100 M

Helium core 45.8 M

The actual light curve would be smooth due

to mixing and light propagation delay times.

Core collapse 4 years

later produced no observable event.

For helium cores in this mass range the total duration of pulses, ~months, is comparable to the duration of the light curve. Late time energy input leads to a very brilliant, long lasting supernova. The appearance would be affected by pre-explosive mass loss.

Train wreck!

Helium Core

Number Pulses

Energy Range

Interval Range

Msun 1051 erg years

48 6 .11- 2.4 .02 - 0.26

51 4 .44 - 3.7 0.09 - 0.9

52 4 .94 - 3.1 .01 - 3.0

54 3 2.1 - 3.2 0.03 - 12

56 3 1.3 - 3.3 .01 - 110

Woosley, Blinnikov, and Heger (Nature, 2007)

For still larger helium cores, the pulses become more violent and the intervals between them longer. Multiple supernovae occur but usually just one of them is very bright. Finally at a helium core of about 60 solar masses the entire star explodes in a single pulse – the traditional “pair-instability” supernova

SN 2006gy

110 M

model

Pulsational Pair Supernovae (PPSN) Five possible outcomes

•  The whole star becomes a black hole (possible for the lightest masses – less than about 80 solar masses with little rotation

•  The envelope is ejected with very low speed producing a long, faint supernova and leaving a black hole ~ 40 solar masses

•  The pulses all finish in less than 1 day (shock crossing time for the envelope). One long supernova not unlike an ordinary SN IIp but no tail and a black hole ~40 solar masses

•  Strong pulses occur weeks to years after the first mass ejection. Collisions of shells produce a very bright supernova and a black hole ~40 solar masses

•  If the time between pulses becomes years, as it does for the heaviest of the PPSN, one may have multiple supernovae

- and a 40 solar mass black hole..

PPSN – Nucleosynthetic Implications

Since anywhere from most to all of the carbon oxygen core collapses to a black hole the nucleosynthesis

is limited to what exists in the hydrogen envelope (H, He, plus CNO from dredge up), and some elements

from the outer helium core (C, O, Ne, Mg). In particular there is no explosive nucleosynthesis

and no iron group elements are made.

This is especially interesting for first generation (Pop III) stars.

Rotationally induced mixing of C and H Mixing and dredge up

Mass loss

Envelope enhanced in CNO (~10% by mass)

140 M

(X

cent−He≈0.33)

Figure from Tumlinson (2007)

Frebel et al (2008)

140 M

Model

produces

18 M

H

55 M

He

3.5 M

C

8.5 M

N

14 M

O

0.5 M

Ne

40 M

Black hole

Qualitative agreement with what is seen in the oldest stars in the galaxy – the ultra-iron-poor stars

Pulsational – Pair Instability Progenitors (Zero metallicity, including rotation and mass loss)

Main Seq

He core (He-burn)

He-core (final)

Outcome

80 40.6 30 CC 90 47.0 34 weak PPSN

100 53.9 39 PPSN 110 61.0 47 PPSN 120 67.4 48 PPSN 130 72.5 49 PPSN 140 79.2 54 PPSN 150 82.7 55 PPSN 160 90.2 59 PPSN 170 93.9 63 Pair SN (no Fe) 180 102.2 65 Pair SN (no Fe)

Pulsational – Pair Instability Progenitors (Solar metallicity, no rotation and reduced mass loss)

Main Seq

He core (He-burn)

He-core (final)

Outcome

80 35.9 35.8 weak PPSN 90 41.0 40 PPSN

100 45.9 43.3 PPSN 110 51.5 46.3 PPSN 120 56.0 49.8 PPSN 130 58.1 52.7 PPSN 140 63.5 57 PPSN

Shock break-out in pair-instabilty supernovae

Kasen, Woosley, and Heger (2011) see also Whalen et al (2013)

Tcolor = 2 – 4 x 105 K Tcolor = 6 – 8 x 105 K

Spectrum of shock break-out ( R = RSG, B = BSG)