CMFGEN - GREAT-ESF Stellar Atmospheres in the Gaia Era ...

Model**in**g **the** w**in**d and photosphere of massive

stars with **the** radiative transfer code **CMFGEN**

Image credit: N. Smith (UCB), NASA/ESA

Jose Groh (Max-Planck-Institute for Radioastronomy, Germany)

Acknowledgments: John Hillier (U Pittsburgh, USA)

Outl**in**e

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

Jose Groh - Model**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

Outl**in**e

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

2. Spectroscopic analysis of hot stars with **CMFGEN**: diagnostics

Jose Groh - Model**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

Outl**in**e

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

2. Spectroscopic analysis of hot stars with **CMFGEN**: diagnostics

OB stars

Jose Groh - Model**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

Outl**in**e

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

2. Spectroscopic analysis of hot stars with **CMFGEN**: diagnostics

OB stars

Lum**in**ous Blue Variables (LBV)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

Outl**in**e

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

2. Spectroscopic analysis of hot stars with **CMFGEN**: diagnostics

OB stars

Lum**in**ous Blue Variables (LBV)

3. Spectroscopic analysis of hot stars with **CMFGEN**: example

O stars with weak w**in**ds

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. Why do we study massive stars?

Cosmic eng**in**es

Massive stars are key constituents of **the** Universe:

✦ ioniz**in**g photons;

✦ energy;

✦ some chemical elements.

M33 galaxy

(size= 10 000 pc)

Image credit: R. Nemiroff & J. Bonnell

First stars and reionization

(size= 50 000 000 pc)

Car**in**a nebula

(size = 15 pc)

Image credit: N. Smith, NASA/ESA credit: Friedich+ 2011

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The need for spectroscopic model**in**g

Enhance **the** amount of **in**formation obta**in**ed from observations

From observations only, one generally measures:

Photometry: magnitudes, colors

NGC 6101

(Marconi+ 2001)

Spectroscopy: EW, l**in**e ratios, radial

velocities, spectral energy distribut.

5

H gamma

AG Car

A-hypergiant with a high mass-loss rate

(Groh+ 2011 **in** prep.)

(Stahl et al. 1982, Stahl et al. 2001, Groh et al. 2009, **in** prep.)

Normalized Flux

4

3

2

Fe II + Ti II

He I

1

4400 4500 4600 4700 4800

Wavelength (Å)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The need for spectroscopic model**in**g

Enhance **the** amount of **in**formation obta**in**ed from observations

From observations only, one generally measures:

Photometry: magnitudes, colors

NGC 6101

(Marconi+ 2001)

Spectroscopy: EW, l**in**e ratios, radial

velocities, spectral energy distribut.

5

H gamma

AG Car

A-hypergiant with a high mass-loss rate

(Groh+ 2011 **in** prep.)

(Stahl et al. 1982, Stahl et al. 2001, Groh et al. 2009, **in** prep.)

Normalized Flux

4

3

2

Fe II + Ti II

He I

1

4400 4500 4600 4700 4800

Wavelength (Å)

But we want:

L, Teff, log g, abundances, ioniz**in**g fluxes,

Mdot, v**in**f !

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

General **in**fo and purposes

Developed and ma**in**ta**in**ed by John Hillier (University of Pittsburgh, USA);

ma**in** reference: Hillier & Miller 1998, ApJ 496, 407.

Code and atomic data are publicly available at:

http://kookaburra.phyast.pitt.edu/hillier/web/**CMFGEN**.htm

Purpose is to provide:

✦ accurate stellar parameters and abundances → stellar evolu,on

✦ accurate EUV ( λ < 912 Å) radiation fields → nebular photoioniza,on

✦ fundamental data for **the** study of starbursts → star forma,on

✦ better understand**in**g of **the** hydrodynamics of w**in**ds → radia,on hydrod.

✦ testbed of approximate methods that can be used **in** more complex

geometries and **in** **in**homogeneous media → mul,-‐dimensional Rad. Transf.

✦ distances and diagnostics of **the** progenitor of Type II SNe.

→ extragalac,c studies, cosmology

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Code Suite

**CMFGEN** → atmospheric and w**in**d structure, T, popula,ons

CMF_FLUX → spectral computa,on

DISPGEN → display package for T, popula,ons, tes,ng, etc

SPEC_PLT → spectral plots, comparison, etc

MAIN_LTE → compute Rosseland opacity for a given composi,on, atomic data

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Atomic data (I)

Opacity project

Seaton 1987, Hummer+ 1993

Bob Kurucz

Bell & Kurucz 1995; http://kurucz.harvard.edu

Keith Butler (Munich)

Sultana Nahar & Anil Prandham

http://www.astronomy.ohio-state.edu/~nahar/nahar_radiativeatomicdata/**in**dex.html

NIST

http://nist.gov/pml/data/asd.cfm (energy levels, f values)

http://nist.gov/pml/pubs/atspec/**in**dex.cfm (**in**tro to atom.sp.)

Gary Ferland / Verner

charge exchange rates, ground state photoionization cross-sections.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Atomic data (II)

Atomic data is stored **in** ASCII files **in** unique directory:

Example: NIII

/home/jgroh/atomic/NIT/III/24mar07

niiicol.dat → collisional cross sec,on

niiiosc_rev.dat → oscillator strengths/energy levels

niii_auto → autoioniza,on rates

f_to_s_ls → superlevels designa,on

phot_sm_0_A.dat → ground state photoioniza,on cross sec,on

phot_sm_0_B.dat → excited state photoioniza,on cross sec,on

Data format is unique to **CMFGEN** -- conversion from published data

**in**tro **CMFGEN** is time consum**in**g (QUESTION A1 from Discussion I yesterday)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Atomic data (II)

Atomic data is stored **in** ASCII files **in** unique directory:

Example: NIII

/home/jgroh/atomic/NIT/III/24mar07

niiicol.dat → collisional cross sec,on

niiiosc_rev.dat → oscillator strengths/energy levels

niii_auto → autoioniza,on rates

f_to_s_ls → superlevels designa,on

phot_sm_0_A.dat → ground state photoioniza,on cross sec,on

phot_sm_0_B.dat → excited state photoioniza,on cross sec,on

Data format is unique to **CMFGEN** -- conversion from published data

**in**tro **CMFGEN** is time consum**in**g (QUESTION A1 from Discussion I yesterday)

Do not assume **the** atomic data for a certa**in**

non-standard spectral feature is correct!

see talk

N. Przybilla

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

The problem to be solved

1) Model atmospheric construction: ρ, T, Ni, Ne

✦ generally **in**sensitive to atomic details: gf values, **in**tr**in**sic l**in**e profiles,

coherent/**in**coherent electron scaPer**in**g;

✦ **in**clude all species;

✦ **in**clude all l**in**es, even those with bad gf values.

2) Precise spectroscopic analysis

✦ use fixed model atmosphere;

✦ details matter: **in**dividual l**in**es constra**in** T eff , log g, and abundances; accurate gf

values (usually) essen,al.

Unfortunately, H, He, CNO, and Fe are strongly coupled to **the**

atmospheric structure.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Transfer options

1) Plane parallel (if you still want to use it)

✦ Formal and moment equations

2) Plane parallel co-mov**in**g frame

✦ Monotonic vertical velocity field

✦ Radiation computed **in** frame mov**in**g with **the** gas: opaci,es and emissivity

are isotropic.

✦ Formal and moment equations

✦ Zero order **in** v/c (only reta**in** v/c term that multiply δ/δv terms)

3) Spherical co-mov**in**g frame

✦ radial monotonic velocity field

✦ Radiation computed **in** frame mov**in**g with **the** gas: opaci,es and emissivity

are isotropic.

✦ Formal and moment equations

✦ Zero order **in** v/c (only reta**in** v/c term that multiply δ/δv terms)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Transfer options

4) Spherical co-mov**in**g frame relativistic

✦ Formal and moment equations

✦ fully relativistic but stationary monotonic flows

5) Spherical time-dependent co-mov**in**g frame

✦ First order v/c terms

✦ only implemented for Hubble flows (as **in** SNe)

✦ Moment equation only

6) Spherical time-dependent co-mov**in**g frame fully relativistic

✦ monotonic velocity field

✦ Moment equation only

✦ UNDERGOING TESTS!

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

L**in**earization of rate equations

Consider a function of 1 variable. In general **the** equation can be written as:

f(x)=0

For some estimate x1, f(x1) ≠ 0. We can compute a correction Δx us**in**g a

Taylor expansion. To first order, we require:

0=f(x1)+ Δx f’(x1) (‘ denotes derivative)

This gives

Δx=-f(x1)/f’(x1)

Thus:

x2 = x1 - (-Δx)

This can be easily extended to more than one variable:

note: B = BA, F=STEQ **in** **CMFGEN**

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative

1. The radiative transfer code **CMFGEN**

Construct**in**g

T2*:')43,*7$'5($R%$F+')1I$

**the** BA matrix

!"#$%&'()*+,-(&.$/($'+0($1*'2$+3324*'$'5($3246&1*7$/1'5$*(17582)1*7$9(6'5:$

!;=?@A%B#$$+66)2+35#C$;5($71-(:$+$8&230D;%?@A%B$:('$2E$:1F4&'+*(24:$

(G4+,2*$H$(+35$8&230$F+')1I$1:$A;CA;C$J2&-(9$8.$F+')1I$-():12*$2E$'5($;K@L%J$

+&72)1'5FC$=*$'51:$+66)2+35$/($5+-($A>CA;$:1F4&'+*(24:$(G4+,2*:$'2$:2&-($DDDM$$

NOPQPPP$3246&(9$:1F4&'+*(24:$(G4+,2*:C$

ARS$$>=?%@A%BS$ $$ $K175()$:'+81&1'.$/5(*$32))(3,2*:$+)($&+)7(C$

$;=%?@A%BS$ $T2*-()7(:$E+:'()C$

;5($:2&4,2*$1:$3+))1(9$24'$4:1*7$B%U%TV$)24,*(:C$R4'$1*$82'5$3+:($/($92$"$')130:S$

!W#$X($$+9Y4:'$'5($(&(F(*':$2E$'5($R%$F+')1I$:2$'5+'$/($:2&-($E2)$'5($E)+3,2*+&$

32))(3,2*:$!1C(CQ$ZA[A#C$%RJ@B\;]B^$]JJ]A;=%B$H$1'$5(&6:$'2$6)(-(*'$:1*74&+)$

F+')13(:C$

!"#$X($+66&.$B%U%TV$)24,*(:$'2$6)(32*91,2*$'5($:1F4&'+*(24:$(G4+,2*:C$

X1'524'$!W#$'51:$92(:$*2'$/2)0C$

J1*3($/($+)($:2&-1*7$E2)$$ZA[AQ$/($92*$*2'$*((9$'5($32))(3,2*$'2$8($2E$5175$6)(31:12*C$

;5($32))(3,2*:$$:1F6&.$*((9$'2$8($)(+:2*+8&($:2$'5+'$/($:,&&$7('$32*-()7(*3($

!6()5+6:$+'$'5($(I6(*:($2E$F2)($1'()+,2*:#C$

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative

B,/3)405'6-67/'

transfer code **CMFGEN**

Iteration cycle

!"#$%&'()*'+',-./'01'2,/3)405'6-67/*'888'/)6('6-67/'.3092:/*''5/;'/*4'&2>')5:',(/':/5*2,-'?*0

1. The radiative transfer code **CMFGEN**

Advantages

✦ spherical-symmetry;

✦ non-LTE l**in**e formation;

✦ quasi-hydrostatic structure until just below **the** sonic po**in**t;

✦ simultaneous treatment of **the** atmosphere and w**in**d;

✦ can be applied to hot stars **in** different evolutionary classes;

✦ w**in**d (micro) clump**in**g;

✦ x-rays **in** **the** photosphere and w**in**d;

✦ full l**in**e blanket**in**g due to metals with Z up to 30.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Effects of L**in**e Blanket**in**g on Energy Distribution

Effects of l**in**e blanket**in**g on **the** spectral energy distribution

Key:

H, He;

H, He, CNO

H, He,

CNO, Fe,

etc

(smoo**the**d)

HD93250

L = 1.3 x 10 6 L ! , T eff = 45,700 K, R = 18.3 R ! , log g = 4.0

M = 5.6 x 10 -7 M ! /yr , V ! = 3000 km s -1

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. The radiative transfer code **CMFGEN**

Cons

✦ momentum equation of **the** w**in**d is not solved; velocity law has to be

assumed a priori (but can be constra**in**ed from observed l**in**e profiles).

(Mart**in**s+ 2005)

✦ computational time can be large compared to o**the**r codes such as

FASTWIND (Puls+ 05; Sergio Simon Diaz & Alex de Koter’s talks);

typical **CMFGEN** run takes 10-20 hours on 1 CPU; decreases by ~60%

with OPENMP (4 cores).

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

2. Spectroscopic analysis of hot stars with **CMFGEN**: diagnostics

OB stars

Lum**in**ous Blue Variables (LBV)

3. Spectroscopic analysis of hot stars with **CMFGEN**: example

O stars with weak w**in**ds

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Evolution of an 85 M star (no rotation)

O-type Lum**in**ous Blue Variable WR SN

LBV

O-type

WR

(evol. tracks from Meynet & Maeder 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: Teff

(Mart**in**s 2011)

He I

He II

Ionization balance:

l**in**es of different ionization

stages from **the** same

chemical element.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: Teff

(Mart**in**s 2011)

Ionization balance:

l**in**es of different ionization

stages from **the** same

chemical element.

AG Car**in**ae (LBV)

Same pr**in**ciple for OB, LBVs,

WRs, and o**the**r emission l**in**e

stars.

(Groh+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: log g

(Mart**in**s 2011)

pressure broaden**in**g:

strong l**in**es of H not affected by **the** w**in**d.

NGC 346 MPG 12 (O9.5 V)

(Bouret+ 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: log g

(Mart**in**s 2011)

pressure broaden**in**g:

strong l**in**es of H not affected by **the** w**in**d.

AG Car**in**ae (LBV)

log g generally cannot be determ**in**ed for LBVs and WRs.

(Groh+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: log g

(Mart**in**s 2011)

pressure broaden**in**g:

strong l**in**es of H not affected by **the** w**in**d.

W243 (LBV): log g ~0.65

(Ritchie+ 2009)

log g generally cannot be determ**in**ed for LBVs and WRs (unless has low

Mdot ).

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: lum**in**osity

Fit of **the** SED; **the** larger **the** spectral region **the** better; may obta**in**

redden**in**g parameters towards **the** star, i.e. E(B-V) and Av.

Hen 3-759 (O8 Iaf): log L=5.9 ± 0.2

(Crow**the**r & Evans 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: lum**in**osity

Fit of **the** SED; **the** larger **the** spectral region **the** better; may obta**in**

redden**in**g parameters towards **the** star, i.e. E(B-V) and Av.

AG Car (LBV): absolute error 0.08 dex **in** log L

log L=6.17 log L=6.17 log L=6.17 log L=6.17

log L=6.17 log L=6.17 log L=6.17 log L=6.17

log L=6.17 log L=6.17 log L=6.17 log L=6.00

log L=6.00 log L=6.00 log L=6.00 log L=6.04

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

(Groh+ 2009)

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: lum**in**osity

Fit of **the** SED; **the** larger **the** spectral region **the** better; may obta**in**

redden**in**g parameters towards **the** star, i.e. E(B-V) and Av.

R136a1 (WN6h): log L=6.94 ± 0.09

corresponds to M = 265 Msun ! (accord**in**g to evolutionary models)

log λ [Angs]

(Crow**the**r+ 2010)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: Mdot

(Mart**in**s 2011)

Mdot is determ**in**ed from **the** strength of UV resonance l**in**es and/or from a

strong recomb**in**ation l**in**e such as Halpha.

]

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: Mdot

(Mart**in**s 2011)

Mdot is determ**in**ed from **the** strength of UV resonance l**in**es or from a

strong recomb**in**ation l**in**e such as Halpha

Zeta Oph (O9.5 Vnn)

C IV

λ [Angs]

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: Mdot

(Mart**in**s 2011)

Mdot is determ**in**ed from **the** strength of UV resonance l**in**es or from a

strong recomb**in**ation l**in**e such as Halpha

Eta Car**in**ae (LBV)

C IV

λ [Angs]

(Hillier+ 2001)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: w**in**d (micro)clump**in**g

(Mart**in**s 2011)

Constra**in**ed from non-saturated UV l**in**e profiles (**in** stars with weak w**in**ds)

and/or electron scatter**in**g w**in**gs of strong recomb**in**ation l**in**es (e.g. H alpha).

HD 190429A (O4 If+)

λ [Angs]

(Bouret+ 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: w**in**d (micro)clump**in**g

(Mart**in**s 2011)

Constra**in**ed from non-saturated UV l**in**e profiles (**in** stars with weak w**in**ds)

and/or electron scatter**in**g w**in**gs of strong recomb**in**ation l**in**es (e.g. H alpha).

AG Car**in**ae (LBV)

λ [Angs]

λ [Angs]

(Groh+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: w**in**d term**in**al velocity

(Mart**in**s 2011)

V**in**f is very well constra**in**ed from saturated P Cygni l**in**e profiles **in** **the** UV;

lower limit can be obta**in**ed from broaden**in**g of strong recomb**in**ation l**in**es.

HD 93204 (O5 V((f))): v**in**f = 2900 km/s

λ [Angs] (Mart**in**s+ 2005)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: w**in**d term**in**al velocity

(Mart**in**s 2011)

V**in**f is very well constra**in**ed from saturated P Cygni l**in**e profiles **in** **the** UV;

lower limit can be obta**in**ed from broaden**in**g of strong recomb**in**ation l**in**es.

AG Car**in**ae (LBV): v**in**f = 105 km/s

λ [Angs]

(Groh+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: chemical abundances

(Mart**in**s 2011)

In pr**in**ciple, use l**in**es from a given species that are less affected by model

details; best to use l**in**es from two or more ionization stages.

AV 83 (O7 Iaf+) :N/H=2.2e-6 by number

(Hillier+ 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics: chemical abundances

(Mart**in**s 2011)

In pr**in**ciple, use l**in**es from a given species that are less affected by model

details; best to use l**in**es from two or more ionization stages.

AV 83 (O7 Iaf+) :N/H=2.2e-6 by number

(Hillier+ 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

2. Spectroscopic analysis of hot stars with **CMFGEN**

Ma**in** diagnostics

(Mart**in**s 2011)

Indicative; has to be adapted to a particular class of objects and dataset

available.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

1. Introduction to **CMFGEN**: radiative transfer; pros and cons

2. Spectroscopic analysis of hot stars with **CMFGEN**: diagnostics

OB stars

Lum**in**ous Blue Variables (LBV)

3. Spectroscopic analysis of hot stars with **CMFGEN**: example

O stars with weak w**in**ds

see also poster by L. Mahy

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of hot stars with **CMFGEN**

Evolution of an 85 M star (no rotation)

O-type Lum**in**ous Blue Variable WR SN

Assum**in**g a standard mass-loss

rate prescription (V**in**k+01)

LBV

O-type

WR

(evol. tracks from Meynet & Maeder 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of hot stars with **CMFGEN**

Evolution of an 85 M star (no rotation)

O-type Lum**in**ous Blue Variable WR SN

Assum**in**g a standard mass-loss

rate prescription (V**in**k+01)

LBV

Mass-loss rate as a function of

time ultimately determ**in**es **the**

fate of a massive star (Chiosi &

Maeder 1986).

O-type

WR

(evol. tracks from Meynet & Maeder 2003)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of hot stars with **CMFGEN**

Evolution of an 85 M star (no rotation)

O-type Lum**in**ous Blue Variable WR SN

Assum**in**g a standard mass-loss

rate prescription (V**in**k+01)

LBV

Mass-loss rate as a function of

time ultimately determ**in**es **the**

fate of a massive star (Chiosi &

Maeder 1986).

O-type

WR

(evol. tracks from Meynet & Maeder 2003)

When is **the** star go**in**g to explode, how, and after los**in**g how much mass?

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: atomic model

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: Teff

Show essentially a photospheric optical and UV spectra.

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

Teff diagnostics:

He I λ4471/He II λ4542 and He I λ4713/He II λ4686 →Teff = 34 ± 1 kK

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

(Marcol**in**o+ 2009)

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: Teff

Show essentially a photospheric optical and UV spectra.

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

He I

He II

Teff diagnostics:

He I λ4471/He II λ4542 and He I λ4713/He II λ4686 →Teff = 34 ± 1 kK

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

(Marcol**in**o+ 2009)

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: Teff

Show essentially a photospheric optical and UV spectra.

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

He I

He II

He II

He I

Teff diagnostics:

He I λ4471/He II λ4542 and He I λ4713/He II λ4686 →Teff = 34 ± 1 kK

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

(Marcol**in**o+ 2009)

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: log g

Show essentially a photospheric optical and UV spectra.

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

log g diagnostics:

broaden**in**g of w**in**gs of Hdelta → log g = 4.0 ± 0.1

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: log g

Show essentially a photospheric optical and UV spectra.

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

Hdelta

Hgamma

log g diagnostics:

broaden**in**g of w**in**gs of Hdelta → log g = 4.0 ± 0.1

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of O stars with **CMFGEN**

O dwarfs: Teff and log g

Show essentially a photospheric spectrum. Few spectral l**in**es are affected

by **the** w**in**d → mass-‐loss rate diagnos,cs

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

Halpha

Mdot diagnostics:

H alpha, C IV 1548-1551 → log Mdot = -9.35 ± 0.7

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of O stars with **CMFGEN**

Weak w**in**ds of O dwarfs?

Show essentially a photospheric spectrum. Few spectral l**in**es are affected

by **the** w**in**d: C IV 1548-‐1551, Halpha → mass-‐loss rate diagnos,cs

green l**in**e = ultraviolet observations of HD 216898

black l**in**e = **CMFGEN** model

Model:

log L = 4.73 ± 0.25

P V

P V

Fe IV forest

C III

Teff = 34 ± 1 kK

log g = 4.0 ± 0.1

log Mdot = -9.35 ± 0.7

v∞ = 1700 ± 500 km/s

Si IV Si IV

Fe IV forest

C IV

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

3. Spectroscopic analysis of O stars with **CMFGEN**

Weak w**in**ds of O dwarfs?

Problem: mass-loss rate found from several studies us**in**g **CMFGEN** is

much lower than what is predicted by **the**oretical models (V**in**k et al. 2001).

Key for **the** plot:

stars: C IV, H alpha w/ **CMFGEN** (Marcol**in**o+ 09)

open triangles: UV l**in**es, Halpha w. **CMFGEN** (Mart**in**s+ 05)

filled symbols:Halpha w/ FASTWIND (Mokiem+ 07)

Do **the** standard radiative transfer

models provide reliable massloss

rates?

Role of magnetic fields?

Rotation? X-rays?

Possible role of macroclumps

(optically-thick) **in** **the** w**in**d: this

could **in**crease **the** derived Mdot

(Sundqvist+ 2011).

(Marcol**in**o+ 2009)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

**CMFGEN**: conclud**in**g remarks

State-of-**the**-art **in** 1D spectroscopic analysis of massive stars:

spherically-symmetric, non-LTE, simultaneous treatment of

photosphere and w**in**d, full l**in**e blanket**in**g (large database of atomic

data, metals up to Z=30), w**in**d clump**in**g, X-rays.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

**CMFGEN**: conclud**in**g remarks

State-of-**the**-art **in** 1D spectroscopic analysis of massive stars:

spherically-symmetric, non-LTE, simultaneous treatment of

photosphere and w**in**d, full l**in**e blanket**in**g (large database of atomic

data, metals up to Z=30), w**in**d clump**in**g, X-rays.

Ma**in** limitation: momentum equation of **the** w**in**d is not solved;

velocity law has to be assumed a priori (but can be constra**in**ed from

l**in**e profiles).

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

**CMFGEN**: conclud**in**g remarks

State-of-**the**-art **in** 1D spectroscopic analysis of massive stars:

spherically-symmetric, non-LTE, simultaneous treatment of

photosphere and w**in**d, full l**in**e blanket**in**g (large database of atomic

data, metals up to Z=30), w**in**d clump**in**g, X-rays.

Ma**in** limitation: momentum equation of **the** w**in**d is not solved;

velocity law has to be assumed a priori (but can be constra**in**ed from

l**in**e profiles).

Multi-wavelength analyses of hot massive stars **in** different evolutionary

stages: OB, LBVs, WRs, SN.

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011

**CMFGEN**: conclud**in**g remarks

Cont**in**uous update of **CMFGEN**: time-dependent w**in**ds (Groh & V**in**k

2011), application to SNe (Dessart & Hillier 2011), multi-dimensional

studies via separate 2D code (Busche & Hillier 2005, Groh+ 2006, 2009)

(Groh & V**in**k 2011)

**in**g **the** w**in**d and photosphere of massive stars with **CMFGEN** **GREAT**-**ESF** workshop, Brussels, 24 June 2011