Do we understand how stellar winds change stellar fireworks? - - PowerPoint PPT Presentation

NAC – 24.05.2016

Do we understand how stellar winds change stellar fireworks?

Mathieu Renzo

PhD @ API Collaborators: S. E. de Mink, C. D. Ott, S. N. Shore, E. Zapartas,

• Y. G¨
• tberg, C. Neijssel

1 / 16

Outline

Importance of Massive Stars...

• ... and their mass loss

Stellar Winds

• Outline of the Theory
• Treatment in Evolutionary Codes

Preliminary Results

• Final Masses
• Impact on the core structure

Conclusions

2 / 16

Why are Massive Stars Important?

MZAMS 8 − 10 M⊙

• Nucleosynthesis
• Chemical Evolution
• f Galaxies
• Effects on Star Formation
• Re-ionization Epoch
• Observations of

Farthest Galaxies

• Catastrophic Events

3 / 16

Mass Loss – Why does it Matter... ... for the environment of the stars?

• Pollution of the InterStellar Medium (ISM)
• Tailoring of the CircumStellar Material (CSM)
• Effects on the Star Formation

... for the stellar structure?

• Evolutionary Timescales
• Appearance & Classification (e.g. WR)
• Light Curve and Explosion Spectrum
• Final Fate (BH, NS or WD?)

4 / 16

Possible Mass Loss Mechanisms

Radiative Driving

⇐

Stellar Winds Dynamical Instabilities

⇐

LBVs, Pulsations, Super-Eddington Winds, Centrifugal Disk Shedding, Binary interactions

⇐

Roche Lobe OverFlows (RLOF)

Figure: η Carinae.

5 / 16

Outline

Importance of Massive Stars...

• ... and their mass loss

Stellar Winds

• Outline of the Theory
• Treatment in Evolutionary Codes

Preliminary Results

• Final Masses
• Impact on the core structure

Conclusions

6 / 16

Radiatively Driven Winds in One Slide

∆p = h

c (νi cos(θi) − νf cos(θ f ))

fcl

def

= ρ2

ρ2=1 ⇒Inhomogeneities⇒ ˙

M=4πr2ρv(r) Problems: High Non-Linearity and Clumpiness:

7 / 16

Radiatively Driven Winds in One Slide

∆p = h

c (νi cos(θi) − νf cos(θ f ))

fcl

def

= ρ2

ρ2=1 ⇒Inhomogeneities⇒ ˙

M=4πr2ρv(r) Problems: High Non-Linearity and Clumpiness:

Risk: Possible Overestimation of the Wind Mass Loss Rate

7 / 16

Mass Loss in

Figure: From Smith 2014, ARA&A, 52, 487S

(Semi–)Empirical parametric models. Uncertainties encapsulated in efficiency factor: ˙ M(L, Teff, Z, R, M, ...)

⇐

η ˙ M(L, Teff, Z, R, M, ...)

η is a free parameter:

η ∈ [0, +∞)

8 / 16

Different dM/dt algorithms with

Grid of Z⊙ ≃ 0.019, non-rotating stellar models:

• Initial mass:

MZAMS = {15, 20, 25, 30, 35} M⊙;

• Efficiency:

η ≡

• fcl = {1, 1

3, 1 10} ;

• Different combinations of wind mass loss rates for

“hot” (Teff ≥ 15 [kK]), “cool” (Teff < 15 [kK]) and WR stars: Kudritzki et al. ’89; Vink et al. ’00, ’01; Van Loon et al. ’05; Nieuwenhuijzen et al. ’90; De Jager et al. ’88; Nugis & Lamers ’00; Hamann et al. ’98.

9 / 16

Outline

Importance of Massive Stars...

• ... and their mass loss

Stellar Winds

• Outline of the Theory
• Treatment in Evolutionary Codes

Preliminary Results

• Final Masses
• Impact on the core structure

Conclusions

10 / 16

Results 1: Impact on the Final Mass

15 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 M/MZAMS 20 25 MZAMS [M⊙] 30 35

VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VdJH VNJH KdJH KNJH

Legend:

• η = 0.1

x η = 0.33 + η = 1.0

• η → largest

uncertainty

• Dust driven

(vL) → very high mass loss for lower MZAMS.

11 / 16

Results 1: Impact on the Final Mass

15 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 M/MZAMS 20 25 MZAMS [M⊙] 30 35

VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VvLNL KvLNL KNJNL KdJNL VNJNL VdJNL VdJH VNJH KdJH KNJH

Impossible to map: Mf ≡ Mf(MZAMS)

Legend:

• η = 0.1

x η = 0.33 + η = 1.0

• η → largest

uncertainty

• Dust driven

(vL) → very high mass loss for lower MZAMS.

11 / 16

Computing Advanced Burning Stages

• Initially small effect ⇒ Nzones 20 000;
• Complex nuclear burning ⇒ Niso 200;

SurfSara’s Cartesius Computer.

12 / 16

Results 2: Core Structure

Compactness Parameter: ξ2.5(t)

def

=

2.5/M⊙ R(M)/1000 km

• “Large” ξ2.5 ⇒ harder to explode ⇒ BH formation
• “Small” ξ2.5 ⇒ easier to explode ⇒ NS formation

13 / 16

Results 2: ξ2.5 @ Oxygen Depletion

(Reduced grid) Legend:

• η = 0.1

x η = 0.33 + η = 1.0 Post O burning evolution ⇐ Core contraction ⇐ Amplification of the differences.

14 / 16

Outline

Importance of Massive Stars...

• ... and their mass loss

Stellar Winds

• Outline of the Theory
• Treatment in Evolutionary Codes

Preliminary Results

• Final Masses
• Impact on the core structure

Conclusions

15 / 16

Conclusions:

• η has a larger influence on the final mass than the wind algorithm;
• Early (“hot phase”) mass loss influences the further evolution;
• Uncertainties in stellar winds prevent to go back in time and infer

MZAMS of observed evolved stars;

• Different algorithmic representations of stellar winds ⇒ Qualitatively

different evolutionary tracks, and predicted final fate; (...Cartesius still crunching numbers for post-O burning evolution...)

Thank you!

16 / 16

Outline

Backup slides

17 / 16

Supernova Taxonomy

Back

18 / 16

Roche Lobe OverFlow

Back

Mass Transfer in Binaries

19 / 16

Wind Oservational Diagnostics

Back

• P Cygni line profiles
• Optical and near UV lines (e.g. Hα)
• Radio and IR continuum excess
• IR spectrum of molecules (e.g. CO)
• Maser lines (for low density winds)

Assumptions commonly needed:

• Velocity structure: v(r) ≃
• 1 −

r R∗

β with β ≃ 1

• Chemical composition and ionization fraction
• Spherical symmetry:

˙ M = 4πr2ρv(r)

• Steadiness and (often) homogeneity

˙ M derived from fit of (a few) spectral lines. No theoretical guaranties coefficients are constant.

20 / 16

Wolf-Rayet Stars

Observational Definition: Based on spectral features indicating a Strong Wind:

• Hydrogen Depletion (= Lack of Hydrogen)
• Broad Emission Lines
• Steep Velocity Gradients

Sub-categories: WN,WC,WO,WNL, etc. Computational Definition ( ):

• Xs < 0.4

Impossible to distinguish sub-categories without spectra!

Back

21 / 16

Why Impulsive Mass Loss?

Observational Evidence:

• LBVs
• Progenitors of H-poor core

collapse SNe (∼ 30%)

• Dense CSM for Type IIn SNe

Theory: Dynamical Events ⇒ not ready

• Pulsational Instabilities
• Roche Lobe Overflow

in binaries

• Catastrophic Eruption(s)

∆Mwind ≪ ∆Mimpulsive (?)

22 / 16

The Stripping Process

3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 log10(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) mSGB hMR MCE M = 15M⊙, Z = Z⊙ unstripped

Remove mass in steps of 1M⊙, max{∆Mimpulsive} = 7M⊙.

Red dot: Teff = 104 [K]; Yellow Triangle: R ≥ Rmax/2 = 375R⊙; Cyan Diamond: Maximum Extent Convective Envelope.

23 / 16

Chosen Stripping Points

3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 log10(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) mSGB hMR MCE M = 15M⊙, Z = Z⊙ unstripped

t(MCE) − t(mSGB) ≃ 104 [yr] ≪ 14.13 × 106 [yr]

24 / 16

Stripped series on the HR diagram

3.6 3.8 4.0 4.2 4.4 4.2 4.4 4.6 4.8 5.0 5.2 log10(L/L⊙) mSGB 3.6 3.8 4.0 4.2 4.4 log10(Teff/[K]) hMR 3.6 3.8 4.0 4.2 4.4 MCE

Evolutionary tracks depend only on ∆Mimpulsive

25 / 16

Evolution toward Higher Teff

3.55 3.60 3.65 3.70 3.75 log10(Teff/[K]) 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) A B C D E F G unstripped MCE 7M⊙ MCE 7M⊙, η = 0

Impulsive + wind mass loss drives blueward evolution

26 / 16

pre-SN Stripped Structures

1 2 3 4 5 6 7 8 9 10 11 12 M [M⊙]

• 10
• 8
• 6
• 4
• 2

2 4 6 8 10 log10(ρ/[g cm−3]) He core CO core Si core unstripped mSGB 1M⊙ mSGB 2M⊙ mSGB 3M⊙ mSGB 4M⊙ mSGB 5M⊙ mSGB 6M⊙ mSGB 7M⊙ hMR 1M⊙ hMR 2M⊙ hMR 3M⊙ hMR 7M⊙ hMR 5M⊙ MCE 1M⊙ MCE 2M⊙ MCE 3M⊙ MCE 4M⊙ MCE 5M⊙ MCE 6M⊙ MCE 7M⊙ 2.0 2.5 3.0 3.5 2 3 4 5 6

27 / 16

Light Curves from Stripped Models

50 100 42 43

Comparison of three progenitor grids

1 M⊙ stripped 2 M⊙ stripped 3 M⊙ stripped 4 M⊙ stripped 5 M⊙ stripped 6 M⊙ stripped 7 M⊙ stripped

Time [days] log10 L [erg s−1]

mSGB hMR MCE Figure: Morozova et al. – ApJ,814,63M

28 / 16

nearly super-Eddington Regime

LEdd

def

= 4πGM(R)c κ(r) , dPgas dr = dPrad dr LEdd Lrad − 1

• 5.0

5.5 6.0 6.5 7.0 log10(T/[K]) 0.5 1.0 1.5 2.0 2.5 κ [cm2 g−1] OPAL: X = 0.7, log(ρ/T63) = −5 Z=0.02 Z=0.01 Z=0.004 Z=0.001 Z=0.0001

MZAMS 20M⊙ ⇒ insufficient FMLT

conv

MLT++:

∇T − ∇ad → α∇ f∇(∇T − ∇ad)

α∇ ≡ α∇(β, ΓEdd), f∇ ≪ 1

• r/R⊙

log (ρ)

70 M⊙, Teff = 5000 K

a)

−10.1 −10.0 −9.9 log

• P

gas

• b)

2.31 2.38 log (P)

c)

2.7 3.0 3.3 S/ (N

AkB) d)

1000 1100 1200 1300 60 80

Figure: From Paxton et al. 2013, ApJS, 208, 5p

29 / 16

Evolution of a Massive Star in one Slide

3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 4.6 log10(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) M =15M⊙, Z = Z⊙ MS ∆tMS ∼ 1.3 · 108 yr OC ∆tOC ∼ 7.9 · 105 yr SGB ∆tSGB ∼ 1.8 · 105 yr R S G ∆ t

R S G

∼ 1 . 2 · 1

7

y r Vink et al., de Jager et al.

30 / 16

Evolution of a Massive Star in one Slide

3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 4.6 log10(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) M =15M⊙, Z = Z⊙ MS ∆tMS ∼ 1.3 · 108 yr OC ∆tOC ∼ 7.9 · 105 yr SGB ∆tSGB ∼ 1.8 · 105 yr R S G ∆ t

R S G

∼ 1 . 2 · 1

7

y r Vink et al., de Jager et al.

30 / 16

Evolution of a Massive Star in one Slide

3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 4.6 log10(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) M =15M⊙, Z = Z⊙ MS ∆tMS ∼ 1.3 · 108 yr OC ∆tOC ∼ 7.9 · 105 yr SGB ∆tSGB ∼ 1.8 · 105 yr R S G ∆ t

R S G

∼ 1 . 2 · 1

7

y r Vink et al., de Jager et al.

30 / 16

Evolution of a Massive Star in one Slide

3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 4.6 log10(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log10(L/L⊙) M =15M⊙, Z = Z⊙ MS ∆tMS ∼ 1.3 · 108 yr OC ∆tOC ∼ 7.9 · 105 yr SGB ∆tSGB ∼ 1.8 · 105 yr R S G ∆ t

R S G

∼ 1 . 2 · 1

7

y r Vink et al., de Jager et al.

30 / 16

P Cygni Line Profiles

Back

• Blue shifted Absorption

Component

• Red shifted Emission

Component

• Broadening from scattering

into the line of sight ˙ M = 4πρv(r) Assuming: Chemical composition Velocity Structure the fit of the line profile gives ρ

Figure: 34 Cyg or P Cygni, first star to show the eponymous profile.

31 / 16

End of the hot evolutionary phase

Vink et al. only: Tjump ∼ 25 [kK] ⇒ Fe3+ → Fe2+

32 / 16

M(t) for MZAMS = 20M⊙ with with

1 2 3 4 5 6 7 8 9 10 t [Myr] 8 9 10 11 12 13 14 15 16 17 18 19 20 M [M⊙] TAMS MZAMS = 20M⊙

Vink et al., de Jager et al. Kudritzki et al., Nieuwenhuijzen et al. Kudritzki et al., de Jager et al. Vink et al., Nieuwenhuijzen et al. Kudritzki et al., van Loon et al. Vink et al., van Loon et al. η = 1.0 η = 0.33 η = 0.1

33 / 16

M(t) for MZAMS = 25M⊙ with with

1 2 3 4 5 6 7 8 t [Myr] 12 14 16 18 20 22 24 M [M⊙] TAMS MZAMS = 25M⊙

Vink et al., de Jager et al. Kudritzki et al., de Jager et al., Hamman et al. Kudritzki et al., Nieuwenhuijzen et al. Kudritzki et al., de Jager et al. Vink et al., Nieuwenhuijzen et al. Kudritzki et al., Nieuwenhuijzen et al., Hamman et al. Kudritzki et al., van Loon et al. Vink et al., van Loon et al. η = 1.0 η = 0.33 η = 0.1

34 / 16

M(t) for MZAMS = 30M⊙ with with

1 2 3 4 5 6 t [Myr] 14 16 18 20 22 24 26 28 30 M [M⊙] TAMS MZAMS = 30M⊙

Vink et al., de Jager et al. Kudritzki et al., de Jager et al., Hamman et al. Kudritzki et al., Nieuwenhuijzen et al. Kudritzki et al., de Jager et al. Vink et al., Nieuwenhuijzen et al. Kudritzki et al., Nieuwenhuijzen et al., Hamman et al. Kudritzki et al., van Loon et al. Vink et al., van Loon et al. η = 1.0 η = 0.33 η = 0.1

35 / 16