sequentially triggered star formation in ob associations thomas preibisch & hans zinnecker...

Sequentially Triggered Star Formation in OB Associations

Thomas Preibisch & Hans ZinneckerMax-Planck-Institute for Radioastronomy, Bonn, Germany

Astrophysical Institute Potsdam,Germany

Upper Scorpius

Upper Centaurus - Lupus

Oph cloud

age = 17 Myr

generation I

age = 5 Myr

generation II

age <= 1 Myr

generation III

ISO

OB associations: (Ambartsumian 1947)

- Unbound stellar groups containing O–B2 stars, Ø ~ 20 ... 50 pc

- Density < 0.1 M pc-3 unstable against galactic tidal forces < 30 Myr old

Blaauw (1964, 1991): Many OB associations consist of distinct sub-groups

with different ages sequential (triggered ?) formation

Ori OB 1

Sco OB 2

Per OB 2

Summary of recent results on OB associations: Protostars and Planets V chapter by Briceno et al. (2006; astro-ph/0602446)

Many observations show star formation near massive stars:

Q: Was the formation of the YSOs triggered, or did YSOs form prior to the arrival of the shock waves ?

A: Determine ages of the YSOs and compare to shock arrival time

OB associations show the result of a recently completed star formation process,

star formation history and initial mass function allow a

quantitative comparison to models

Star formation in irradiated globules:"radiatively driven implosion"

Star formation in swept-up shells:"collect & collapse" model

Trifid NebulaHST, Hester et al. (1999)

O star

OB stars

RCW 79blue: H, red: 8 mZavagno et al. (2006, A&A 446,171)

YSOs

YSOs

Theoretical models for triggering mechanisms in OB associations:

1. Sequentially triggered formation of OB subgroups (Elmegreen & Lada 1977, Lada 1987)

Predictions: - Bimodal star formation: low-mass stars form independently are on average older, show large age spread - IMF variations: younger OB subgroups should have larger fractions of low-mass stars

Age spread among low-mass stars: [8....12] Myr [4....12] Myr [0....12] Myr

IMF variations: defict excess of low-mass stars

star formation terminated


2. Radiation-driven implosion of globules near OB stars (Bertoldi 1989; Lefloch & Lazareff 1994; Kessel-Deynet & Burkert 2003)

Predictions:

- OB stars form first, are older than low mass stars - Age gradients: stars close to the O star are older than those further away

(see also next talk by W.P. Chen)

Hester & Desh (2005)

Ages of the low-mass stars: 7 , 5 , 3 , 1 Myr


3. Supernova shock wave compression of cloud (e.g. Foster & Boss 1996, ApJ 468, 784; Vanhalla & Cameron 1998, ApJ 508, 291)

At suitable distances of ~ 20 ... 100 pc,

where vshock = 20 ... 50 km/sec,

cloud collapse can be triggered by supernova

shock waves

Predictions:

- High- and low-mass stars have same age

- Small age spread (since vshock > 20 km/sec)

- Age difference of ~ 5...10 Myr between subgroups

NEXT: Predictions versus observations

D= 144 pc49 B-stars

Upper Scorpius D = 142 pc 66 B-stars


D = 116 pc 42 B-stars

Lower Centaurus - Crux

10

The nearest OB association: Scorpius - Centaurus (Sco OB2)

Hipparcos revealed B to F starsde Zeeuw et al (1999, AJ 117, 354)

de Bruijne (1999, MNRAS 310, 585)

25 pc

Sco

= 5 MyrUpper Scorpius = 17 Myr


= 16 MyrLower Centaurus - Crux

10

The nearest OB association: Scorpius - Centaurus (Sco OB2)

25 pc

Sco

Ages of the massive stars from MS turnoff

What about the low-mass stars ?de Geus et al. (1989, A&A 216,44)

Mamajek et al.(2002, AJ 124,1670)

Upper Scorpius

10

25 pc

Sco

Huge field star confusion in extended OB associations:

Young low-mass members must be individually identified

by spectroscopy (6708Å Lithium lines)

- X-ray selected candidates: Walter et al (1994, AJ 107,692) Preibisch et al (1998, A&A 333,619)

- Survey with multi-object spectrograph 2dF: Preibisch et al (2002 AJ 124, 404):

250 low-mass members (representative sample)

+ 114 higher-mass stars from Hipparcos:

364 known members

SpT = B0.5 ... M6 , M = 20 M ... 0.1 M

Statistically robust & well defined sample

Individual spectral types and extinctions known:

derive IMF and star formation history from HRD

The stellar population of Upper Scorpius

HR Diagram for Upper Sco

5 Myr isochrone

High- mass and

low-mass stars

are coeval

Spread of isochronal agesPreibisch et al (2002, AJ 124, 404)

Reasons for spread of isochronal ages:

- distance spread: D / D = 30 pc / 145 pc ~ ± 0.25 mag

- unresolved binaries: <~ + 0.75 mag

- photometric variability <~ ± 0.3 mag

HRD consistent with no age spread, < 1-2 Myr

- Mass function is consistent with field star IMF

age of the high-mass stars: 5 Myr diameter: ~ 30 pc

age of the low-mass stars: 5 Myr 1D velocity dispersion: 1.3 km/sec

age spread < 1-2 Myr lateral crossing time ~ 25 Myr

age spread << crossing time

external agent coordinated onset of star formation over the full spatial extent

Implications on the star formation process

ScoCen is surrounded by several H I shells

De Geus (1992, A&A 262, 258)

Wind & supernova driven

expanding superbubble

from UCL crossed Upper Sco

~ 5 Myr ago

De Geus (1992, A&A 262, 258):

Scenario for the star formation history de Geus (1992, A&A 262, 258); Preibisch & Zinnecker (1999, AJ 117, 2381)

Supernova in USco

star formation triggered


cloud fully dispersed

star formation in Oph triggered

Supernova & wind driven shock wave

from UCL crosses USco cloud

Wind & ionizing radiation of the

massive stars disperse the cloud


from USco reaches Oph cloud

Scenario for the star formation history de Geus (1992, A&A 262, 258); Preibisch & Zinnecker (1999, AJ 117, 2381)

Supernova in USco

star formation triggered


cloud fully dispersed

star formation in Oph triggered


from UCL crosses USco cloud

"Any theory of star formation is incomplete without a corresponding theory of cloud formation" (Elmegreen & Lada 1977)

Hartmann et al (2001, ApJ 562,852) and others:

Molecular clouds are short-lived structures,

i.e.do not exist for > 10 Myr without forming stars

and "wait for a trigger"

Wind & ionizing radiation of the

massive stars disperse the cloud


from USco reaches Oph cloud

Rapid formation of molecular clouds and stars

Ballesteros-Paredes et al.(1999, ApJ 527,285); Hartmann et al.(2001 ApJ 562, 852); Clark et al.(2005, MNRAS 359,809)

Large-scale flows in the ISM

accumulate and compress gas

to form transient molecular clouds

Wind & supernova shocks waves create

coherent large-scale velocity fields,

formation of large structures in which

star formation can be triggered nearly

simultaneously

Hartmann et al. (2001 ApJ 562, 852)

Ballesteros-Paredes et al. (1999, ApJ 527,285)

OB star winds in UCL create

expanding superbubble (v ~ 5 km/sec)

Interaction with ISM flows

starts to sweep up clouds

Triggered cloud & star formation in ScoCen T = - 14 Myr

30 pc

UCL

After supernovae in UCL

added energy & momentum to

the expanding superbubble,

the shock wave (~ 30 km/sec)

crossed the USco cloud,

increased pressure triggered

star formation in USco

Triggered cloud & star formation in ScoCen T = - 5 Myr

60 pc

UCL

Triggered cloud & star formation in ScoCen

Lupus I cloudgeneration III

Shock wave from USco superbubble

triggers star formation

in Oph and Lupus I clouds

T = - 1 Myr


generation I

Upper Scorpius

generation II

Ophiuchus

generation III

Observation:

Determination of centroid spacemotions of the groups:

de Bruijne (1999, MNRAS 310, 585)

Upper Sco moves away from

UCL with v ~ 5 (±3) km/sec

Model Predictions I:

Stellar groups triggered in swept-up clouds

move away from the trigger source

Hipparcos proper motions of USco and UCL members

+

Centroid space motion in the rest-frame of UCL

de Zeeuw et al (1999, AJ 117, 354)

Model versus observations

adapted from:Mamajek & Feigelson (2001,in: Young Stars Near Earth,

ASP 244, p. 104; astro-ph/0105290)

Observation:Mamajek & Feigelson (2001)

Several young stellar groups:

- Cha cluster

- TW Hydra Association

- CrA cloud

move away from UCL with v ~ 10 km/sec

were located near the edge ofUCL 12 Myr ago (when SN exploded)

X (pc)

UCL

Y

(pc)

X (pc)

0

50 100 150

-100

-50

0

T = - 12 Myr

UCL

CrA

TW Hya

ChaY

(p

c)

0 50 100 150

-100

-50

0

T = 0 (today)


Stellar groups triggered in swept-up clouds

move away from the trigger source


USco

UCL

S

Observation:

Lupus I cloud located at

interface of USco and

UCL (post-SN, wind-driven)

superbubbles

Model Predictions II:

Elongated star forming clouds form at the

intersection of two expanding flows

Dust extinction map from Dobashi et al. 2005, PASJ 57,S1


How general are these findings?

Several OB associations show subgroup sequences of three generations that provide evidence for triggered formation, e.g.

ScoCen: UCL USco Oph / Lupus I

Cep OB2: NGC 7160 IC 1396 VDB 142

Superbubbles in W3/W4 (Oey et al. 2005, AJ 129, 393)

BUT: Some associations also contain subgroups with similar ages

e.g.: ScoCen: UCL [17 Myr] - LCC [16 Myr]

Supernova-driven

expanding H I shell

v = 22 km/sec

Gorjian et al. (2004,ApJS 154, 275)

Extragalactic example of supernova triggering in OB associations: Hen 206 (LMC)

Triggered formation of several new OB subgroups

OB association NGC 2018: age ~ 10 Myr

80 pc Spitzer image blue: 3.6+4.5 m, cyan: 5.8 m, green: 8.0 m, red: 24 m

optical image

Explanation for subgroups with the same age (e.g., ScoCen: UCL [17 Myr] - LCC [16 Myr] )

How general are these findings?

Results for well investigated OB associations: (Sco OB2, Cep OB2, Ori OB1)

Briceno et al. (2006; Protostars & Planets V chapter; astro-ph/0602446)

- IMF of most OB subgroups consistent with field IMF, no good evidence for IMF variations

- Low- and high mass stars have the same ages, have formed together

- In most regions, age spreads are (much) smaller than the crossing time

consistent with models of large-scale triggering by shock waves

Supernova (+wind) driven shock waves play an important role, but other triggering mechanisms are also at work in some regions:

Cep OB 2 association:

HD 206267 (O6)

IC 1396 (4 Myr)

class 0 / class Iprotostars

NGC 7160(10 Myr)

IRAS 12 m

1

Spitzer 3.6+4.5 m, 5.8+8 m, 24 mReach et al (2004, ApJS 154, 385)

<1 Myr

VDB142

13 pc

Sicilia-Aguilar et al (2004, AJ 128, 805; 2005 AJ 130, 188)

Reach et al (2004, ApJS 154, 385)

VDB 142:Radiation-driven implosion of globule (no supernova triggering!)

The globule will form a small stellar group,but no OB subgroup!

Conclusions:

OB subgroups with well defined age sequences and small internal age spreads

suggest large-scale triggered formation scenarios.

(Supernova/wind driven shock waves)

Expanding bubbles coherent large-scale ISM flows new clouds

Supernova shock waves cloud compression

triggered formation of whole OB subgroups (several 1000 stars).

Other triggering mechanisms (e.g. radiation-driven implosion of globules)

may operate simultaneously, but seem to form only small groups of stars

(i.e. are secondary processes).

Note: Our Sun formed in an OB association !

Supernova shock wave injected short-lived radionucleids (e.g. 26Al).(Cameron & Truran 1977; Hester & Desh 2005)

THE END

Interpretation of the HR Diagram: Age spread or no age spread ??

Perfect world: no errors, no uncertainties

age histogram

simulated HRD

Coeval population of 5 Myr old stars

Interpretation of the HR Diagram: Age spread or no age spread ??

false impression of a large age spread

and an accelerating star formation rate

in an actually perfectly coeval population !

Perfect world: no errors, no uncertainties

Reality: - photometric variability & errors

- unresolved binaries

- spread of individual distancescenter:145 pc

front:130 pc

back:160 pc

Upper Sco

Analysis for Upper Sco: no detectable age spread, = 5 Myr (±1-2 Myr)

age histogram

simulated HRD

Coeval population of 5 Myr old stars

Ori OB 1C members ?(~ 5 Myr)

e.g. Orion Nebula Cluster:

Distributed T Tauri stars: ~ 2 -10 Myr

Trapezium clusterstars: ~ 1 Myr

BN complex, OMC-S protostars: <~ 0.1 Myr

To Earth

simulated side-view:

O'Dell 2001 ARAA 39, 99

Age sequences / spreads and projection effects

15 Myr 5 Myr 1 Myrline-of-sight

no observed age sequence

false impression of

a large age spread

1 ... 15 Myr

The Supergiant Shell Region in IC 2574 Cannon et al. (2005, ApJ 630, L37)Stewart & Walter (2000, AJ 120,1794)

U + V + I

Expanding shell triggers a second generation of OB associations on its rim

N

central OB associationtotal mass ~150 000 M

age: ~ 11 Myr

young OB associations

M ~ 5000 ... 300000 M

ages ~ 1 ... 4 Myr

cavity surrounded byexpanding shell

Ø ~ 800 pc, M ~ 106 M

v ~ 25 km/sec

N

Note:most massive clouds are at bubble intersection

Yamaguchi et al. (2001, ApJ 553, L185)

shell diameter 1.9 kpcv(exp) = 10 – 40 km/sec

center: 400 OB stars ages 9-16 Myr

stars at the rim are < 6 Myr

H image, green contours: CO open circles: > 10 Myr clusters, filled circles: < 10 Myr old clusters

The Superbubble LMC4

Only 11 of 73 EGGs have YSOs

< 100 stars will eventually form in the pillars,

much less than the stellar population of the

exciting OB cluster NGC 6611

HST optical image; Hester et al. (1996)

"Pillars of Creation" in the Eagle Nebula (M16)

VLT near-infrared image; McCaughrean & Andersen (2002)

Detection of evaporating gaseous globules

"EGGs"; sites of triggered star formation ?

Triggered massive star formation in RCW 79Zavagno et al. (2006, A&A 446,171):blue: H, red: Spitzer 8 m

Data consistent with

"collect & collapse model"

fragmentation of the shocked dense

layer around an expanding HII region

Whitworth et al (1994)

central OB cluster,including an O4 star (~60 M)

compact HII region,ionized by an O9 star (~20 M),

contains many class I protostars

Radius of HII region = 6.4 pc

for n ~ 2000 cm-3 dyn = 1.7 Myr

collected layer fragmented ~105 yr ago,

consistent with ages of YSOs

30 Dor – NGC 2070 ( 40 O3 stars, 60 000 M , age = 2-3 Myr )

1'

15 pc

Radiation of the massive stars in the

central cluster triggers star formation

in the nebular filaments

(Rubio et al. 1998)

"Two stage starburst"

216 pc

Ori association Ori (O8) + 60 B-type stars, ~ 6 Myr

Dolan & Mathieu (2001, AJ 121, 2124) Dolan & Mathieu (2002, AJ 123, 387)

- global IMF consistent with field IMF

- low-mass stars concentrate

- near Ori, ages ~ 1 – 6 Myr

no ongoing star formation, age spread probably real

- near the dark clouds B30+B35

continuing star formation, ages ~ 0 - 6 Myr

IRAS 12+24+100 m

star formation begins

Star formation history:

1 Myr ago, a supernova explosion terminated

the star formation process near the center,

but there is still ongoing star formation in the

dark clouds B30 and B35

Ori

B30

B35

200 OB stars,

age ~ 3 Myr

Rcavity = 60 pc

Maiz-Apellaniz et al. (2004, AJ 128,1196)NGC 604 (in M33)

Superbubbles in W3/W4 Oey et al. (2005, AJ 129,393)

IC 1805: 1-3 Myr

W4

W3"230 pc shell"

age ~ 6 – 10 Myr

W3 Main: 105 yr

W3 North: 105 yr

W3 OH: 105 yr

IC 1795: 3-5 Myr

UCL

Upper Sco

Observation:

Initial configuration for Upper Sco:

from proper-motion back-tracing, Blaauw (1991)

elongation as signature of a swept up cloud


swept-up clouds should be elongated

along the direction of the shock front


Ori OB 1 association Briceño et al (2005, AJ 129, 907)

identify 197 low-mass members

of Ori OB 1a and 1b

derived ages:

1a: ~10 Myr, 1b: ~5 Myr

High- and low-mass stars are coeval

1a (~ 10 Myr)

1b (~ 3 Myr)

1c (~ 4 Myr)

1d (= ONC)

Ages of the OB stars from Brown et al (1994, A&A 289, 101)

2

16 pc

sequentially triggered star formation in ob associations thomas preibisch & hans zinnecker...

Documents