[PPT] - Systematic Study of Mass Loss in the Evolution of Massive Stars PowerPoint Presentation

SLIDE 1

Tesi di Laurea Magistrale

Universit` a di Pisa - Dipartimento di Fisica 03/06/2015

Systematic Study of Mass Loss in the Evolution

f Massive Stars

Mathieu Renzo

advisors:

Prof. S. N. Shore, Prof. C. D. Ott

1 / 27

SLIDE 2

Outline

Introduction

Importance of Massive Stars
How do they lose mass?

Stellar Winds

Outline of the Theory
Methods
Results: Amplitude of the Uncertainty
Results: Blue Loops in 15M⊙ models

Impulsive Mass Loss Events

Motivations for This Study
Methods
Results: Wind + Impulsive Mass Loss
Results: pre-SN Stripped Structures

Conclusions

2 / 27

SLIDE 3

Outline

Introduction

Importance of Massive Stars
How do they lose mass?

Stellar Winds

Outline of the Theory
Methods
Results: Amplitude of the Uncertainty
Results: Blue Loops in 15M⊙ models

Impulsive Mass Loss Events

Motivations for This Study
Methods
Results: Wind + Impulsive Mass Loss
Results: pre-SN Stripped Structures

Conclusions

3 / 27

SLIDE 4

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

4 / 27

SLIDE 5

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
Final Fate (BH, NS or WD?)
Light Curve and Explosion Spectrum
Appearance: CSM and Wind Features (e.g. WR)
Role in the Solution of the RSG Problem ?

5 / 27

SLIDE 6

Possible Mass Loss Mechanisms

Radiative Driving

⇐

Stellar Winds Dynamical Instabilities

⇐

LBVs, Impulsive Mass Loss, Pulsations, Super-Eddington Winds Binary interactions

⇐

Roche Lobe OverFlows (RLOF)

Figure: η Carinae.

6 / 27

SLIDE 7

Outline

Introduction

Importance of Massive Stars
How do they lose mass?

Stellar Winds

Outline of the Theory
Methods
Results: Amplitude of the Uncertainty
Results: Blue Loops in 15M⊙ models

Impulsive Mass Loss Events

Motivations for This Study
Methods
Results: Wind + Impulsive Mass Loss
Results: pre-SN Stripped Structures

Conclusions

7 / 27

SLIDE 8

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:

8 / 27

SLIDE 9

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

8 / 27

SLIDE 10

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, +∞)

9 / 27

SLIDE 11

Different dM/dt algorithms with

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

Initial mass:

MZAMS = {15, 20, 25, 30} 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.

10 / 27

SLIDE 12

Results: Relative Final Mass

15 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 MO depl/MZAMS 20 25 MZAMS [M⊙] 30 VdJNL VNJNL VNJH VdJH VvLH VvLNL KdJNL KNJNL KNJH KdJH KvLH KvLNL VdJNL VNJNL VNJH VdJH VvLH VvLNL KdJNL KNJNL KNJH KdJH KvLH KvLNL

Diamonds ⇔ η = 1.0, Squares ⇔ η = 0.33, Circles ⇔ η = 0.1.

11 / 27

SLIDE 13

M(t) for MZAMS = 15M⊙ with

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

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

;

SLIDE 14

M(t) for MZAMS = 15M⊙ with

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

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

;

Only η = 1.0

12 / 27

SLIDE 15

Comparison of Hot Wind Algorithms

Example: MZAMS = 15M⊙ evolutionary tracks

3.4 3.6 3.8 4.0 4.2 4.4 4.6 log(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log(L/L⊙) Kudritzki et al., de Jager et al. η = 1.0 η = 0.33 η = 0.1 3.4 3.6 3.8 4.0 4.2 4.4 4.6 log(Teff/[K]) 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 log(L/L⊙) Vink et al., de Jager et al. η = 1.0 η = 0.33 η = 0.1

⇒ Early (“hot”) wind influences subsequent evolution

13 / 27

SLIDE 16

Why Blue Loops? 1/2

Blue loop ⇔ Large He-core
Convection mixes H down,

determining MHe

µ is higher in He-rich regions

−10 −8 −6 −4 −2 2 15M⊙, Vink et al., Teff = 15000 K η = 1.0 η = 0.1 η = 0.33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 M [M⊙] −10 −8 −6 −4 −2 2 log10(ρ/[g cm−3]) 15M⊙, Kudritzki et al., Teff = 15000 K η = 1.0 η = 0.1 η = 0.33 4.3 4.4 4.5 4.6 4.7 4.8 0.0 0.2 0.4 0.6 4.3 4.4 4.5 4.6 4.7 4.8 0.0 0.2 0.4 0.6 14 / 27

SLIDE 17

Why Blue Loops? 2/2

Blue loop starts when H-burning

shell reaches the edge of the He core

Lower µ and higher X ⇒

Variations of εnuc

Envelope responds on its

thermal timescale

⇐

if η < 1 ⇒ He core edge

too deep for Blue Loops

Vink et al. rate yields larger

cores allowing for Blue Loops

15 / 27

SLIDE 18

Why not Blue Loops?

5 6 7 8 9 10 11 12 13 14 M [M⊙] −8 −6 −4 −2 2 log10(ρ [g cm−1]) εnuc ≥ 104 [erg g−1 s−1] Hot wind: Vink et al., η = 1.0 MZAMS = 15M⊙ age ≃ 13.3 [Myr] de Jager et al. van Loon et al. Nieuwenhuijzen et al.

;

Density profiles at the onset of Blue Loops

SLIDE 19

Why not Blue Loops?

5 6 7 8 9 10 11 12 13 14 M [M⊙] −8 −6 −4 −2 2 log10(ρ [g cm−1]) εnuc ≥ 104 [erg g−1 s−1] Hot wind: Vink et al., η = 1.0 MZAMS = 15M⊙ age ≃ 13.3 [Myr] de Jager et al. van Loon et al. Nieuwenhuijzen et al.

;

Ideal gas EOS: Pgas =

ρ µmpkbT

16 / 27

SLIDE 20

Summary

Results of the Comparison of Wind Algorithms:

η has a larger influence on the final mass than the

wind algorithm;

Early (“hot phase”) mass loss influences the further

evolution;

˙

M is more uncertain when it is higher (RSG phase);

Different algorithmic representations of stellar winds

⇒ Qualitatively different evolutionary tracks;

Small number (8) of WR stars, none with η < 1 ⇒

Other mass loss mechanism(s) to form WR?

17 / 27

SLIDE 21

Outline

Introduction

Importance of Massive Stars
How do they lose mass?

Stellar Winds

Outline of the Theory
Methods
Results: Amplitude of the Uncertainty
Results: Blue Loops in 15M⊙ models

Impulsive Mass Loss Events

Motivations for This Study
Methods
Results: Wind + Impulsive Mass Loss
Results: pre-SN Stripped Structures

Conclusions

18 / 27

SLIDE 22

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 (?)

19 / 27

SLIDE 23

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.

20 / 27

SLIDE 24

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]

21 / 27

SLIDE 25

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

22 / 27

SLIDE 26

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

23 / 27

SLIDE 27

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

24 / 27

SLIDE 28

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. – eprint arXiv:1505.06746

25 / 27

SLIDE 29

Outline

Introduction

Importance of Massive Stars
How do they lose mass?

Stellar Winds

Outline of the Theory
Methods
Results: Amplitude of the Uncertainty
Results: Blue Loops in 15M⊙ models

Impulsive Mass Loss Events

Motivations for This Study
Methods
Results: Wind + Impulsive Mass Loss
Results: pre-SN Stripped Structures

Conclusions

26 / 27

SLIDE 30

Conclusions

Large systematic uncertainties in massive star mass

loss rates

Different algorithms ⇒ Qualitatively different

evolutionary tracks

Uncertainty increases at higher MZAMS and η
Combined impulsive + wind mass loss drives

blueward evolution

Does impulsive mass loss have an effect on the

“Explodability” of the star? Thank you for your attention.

27 / 27

SLIDE 31

Figure Credits

Roughly in order of appearance. Some figure where modified. Figure not listed are from myself. Click for original link.

30 Doradus (Tarantula Nebula)
Observative HR
Crab Nebula
Orion
Reionization Epoch
Bubble Nebula
SN1987A
CCSN entropy rendering
SN observations
η Car
Betelgeuse
Mass Loss Rate plot
AG car
Type Ib SN
WR 124
WR spectra
P Cygni line profile:S. N. Shore

“Astrophysical Hydrodynamics”, Wiley-VCH, 2007.

P Cygni (34 Cyg)

28 / 27

SLIDE 32

Supernova Taxonomy

Back

29 / 27

SLIDE 33

Roche Lobe OverFlow

Back

Mass Transfer in Binaries

30 / 27

SLIDE 34

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

31 / 27

SLIDE 35

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.

32 / 27

SLIDE 36

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

33 / 27

SLIDE 37

Evolution of a Massive Star in one Slide

Back

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.

34 / 27

SLIDE 38

Evolution of a Massive Star in one Slide

Back

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.

34 / 27

SLIDE 39

Evolution of a Massive Star in one Slide

Back

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.

34 / 27

SLIDE 40

Evolution of a Massive Star in one Slide

Back

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.

34 / 27

SLIDE 41

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.

35 / 27

SLIDE 42

R(t) for 15M⊙ Models during Blue Loops

12.9 13.0 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 13.9 14.0 14.1 14.2 t [Myr] 100 200 300 400 500 600 700 800 900 1000 R [R⊙] s15VdJNL s15KNJNL s15KdJNL s15VNJNL s15KvLNL s15VvLNL η = 1.0 η = 0.33 η = 0.1

36 / 27

SLIDE 43

End of the hot evolutionary phase

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

37 / 27

SLIDE 44

Stellar counts

1 2 3 # stars [arbitrary units] 4.9 5.0 5.1 5.2 5.3 5.4 log10(L/L⊙) Vink et al., MZAMS = 25M⊙ η = 1.0 η = 0.33 η = 0.1 1 2 3 # stars [arbitrary units] 4.9 5.0 5.1 5.2 5.3 5.4 log10(L/L⊙) Kudritzki et al., MZAMS = 25M⊙ η = 1.0 η = 0.33 η = 0.1 4.35 4.40 4.45 4.50 4.55 4.60 log10(Teff/[K]) 0.5 1.0 # stars [arbitrary units] Vink et al., MZAMS = 25M⊙ η = 1.0 η = 0.33 η = 0.1 4.35 4.40 4.45 4.50 4.55 4.60 log10(Teff/[K]) 0.5 1.0 # stars [arbitrary units] Kudritzki et al., MZAMS = 25M⊙ η = 1.0 η = 0.33 η = 0.1

Cannot be compared to

clusters or single populations

Higher η ⇒ lower M ⇒

slower evolution

Different cut-offs in L and

Teff

Kudritzki et al. rate with

η = 1.0 produces a loop in the HR diagram tracks, resulting in the

ver-population shown.

38 / 27

SLIDE 45

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

39 / 27

SLIDE 46

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

40 / 27

SLIDE 47

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

41 / 27