Triggered star formation, HII regions and Spitzer bubbles Mark - - PowerPoint PPT Presentation

Triggered star formation, HII regions and Spitzer bubbles

Mark Thompson

Outline of the lectures

1.Observational surveys for infrared bubbles

• Spitzer bubbles & the Milky Way Project

2.Theory of bubble formation & triggered star formation

• HII regions & wind-blown bubbles
• Collect & Collapse and Radiative-Driven Implosion

3.The star-forming environment of bubbles

• Sequential star formation
• Statistical studies

Theoretical considerations

HII regions & wind-blown bubbles Triggered star formation processes

• Radiative driven implosion
• Collect & Collapse
• Cloud Collisions
• Turbulence

Simulations of HII regions & triggered star formation

HII regions

Massive stars (> 8 M⊙ or type <B3) generate sufficient amounts of short wavelength photons to continuously ionise the surrounding gas. Ionisation potential of Hydrogen is 13.6 eV, corresponding to ~912 Å Ionised regions are known as HII regions Within HII region continuous recombination of p + e gives rise to bremsstrahlung or free-free emission and recombination lines

Classic textbook derivations: Lyman Physical processes in the interstellar medium Dyson & Williams The Physics of the ISM

Strömgren spheres

Mean free path of UV photons (λ < 912 Å) is negligible compared to size

• f an HII region (Strömgren 1939)

So HII regions have sharp boundaries - bounded by an ionisation front Radius of the initial ionisation front is called the Strömgren radius

Qo - ionising photon rate

nH - hydrogen number density β2 - n=2 recombination coefficient Assuming constant density homogenous surrounding medium

Strömgren spheres are overpressured

HII regions are overpressured with respect to the surrounding gas If dominant heating is via thermal electrons in the HII region: → Once initial Stromgren radius is ionised, the HII region will then expand over time → Drives a photoionisation shock into the surrounding gas → Preceding the ionisation front is a photon dominated region (PDR)

ci - sound speed of ionised gas (~10 km/s)

Again, assuming constant density homogenous surrounding medium

Ionisation bounded vs Density bounded

Ionisation bounded: HII region surrounded by neutral gas & its extent depends upon how much gas has been ionised Density bounded: star(s) have ionised all the surrounding neutral gas & the extent of the HII region depends on the surrounding gas In real life things are more complex Molecular clouds have complex clumpy & filamentary structure HII regions can be density bounded in some directions and ionisation bounded in others...

Ionisation bounded vs Density bounded

Trifid Nebula (M20)

Pillars & globules

IC 1396

52 18:26:50 48 46 44 42 25:30 12:26:00 30 27:00 30 Right Ascension (J2000) Declination (J2000)

Blister HII regions & champagne flow

When an HII region reaches the edge of a molecular cloud, ionised gas can escape quickly (ci ~ 10 km/s) Pressure reduces & expansion into rest of cloud slows to ~ few km/s Escape of ionised gas known as champagne flow

The complication of winds

For early O stars and OB giants the picture is complicated by strong stellar mass loss in the form of stellar winds Winds inject momentum and shocks into the gas 4 phases:

• 1. Free expansion at the wind velocity
• 2. Adiabatic expansion phase
• 3. “Snowplough” phase, where swept up gas collapses into a shell
• 4. Dissipation phase, where shell dissipates into surroundings

Snowplough phase lasts most of the lifetime of the star (Castor et al 1975, Weaver et al 1977)

Anatomy of a wind-blown bubble

“Perhaps the most interesting implication of our theory is the likelihood of molecules in the outer part of the shell... It will be most interesting to consider molecular radio emission mechanisms in such a shell”

From Povich (2012), after Weaver et al (1977) Initial analytic solutions in Castor et al (1975) & Weaver et al (1977)

Examples

N49 N49 - single 06.5 ionising star M17 - cluster of 12 O stars, including O4 V Blue - Chandra soft X-ray

The relevance for star formation

Star formation controlled by balance between gravity & thermal pressure (+ turbulence, magnetic fields...) For an isothermal spherical homogeneous cloud of constant mass the relationship between external pressure & radius is: Obtained by solving equation of virial equilibrium in the presence of an external pressure pext

Induced or triggered star formation

No matter how stable the cloud, if the external pressure is raised high enough it will collapse. → Photoionisation shocks, wind-driven shocks, supernovae shocks...

Curve gives possible equilibrium states as function of pext For high pext - no equilibrium possible For low pext - two possible equilibria (A & B): A - unstable equilibrium B - stable equilibrium

Possible mechanisms

• 1. Formation of small scale instabilities

(Jeans instability - low-mass SF)

• 2. Formation of large scale instabilities or Collect

& Collapse (e.g. Elmegreen & Lada 1977)

• 3. Ionisation of turbulent structure

(e.g. Gritschneder et al 2009)

• 4. Radiative driven Implosion (RDI) of existing

dense clumps (Bertoldi 1989, Lefloch & Lazareff 1994) Possible, not exclusive mechanisms May dominate more under certain conditions or even operate together: e.g. dense clumps formed by Collect & Collapse then eroded by RDI Schematic from Deharveng et al 2010

Collect & Collapse

Originally put forward by Elmegreen & Lada (1977) to explain propagation

• f OB associations

Massive stars drive ionisation & shock fronts into nearby cloud. Swept up layer becomes gravitationally unstable and collapses. Collapse timescale ~ 1-3 Myr for typical cloud densities

Collect & Collapse

Investigated in more numerical detail by Whitworth (1994) For photoionisation shocks: Similar results obtained for stellar wind shocks Note: weak dependence on ionising photon rate & fragments are typically

• f higher mass (forming more massive stars?)

Radiative-driven implosion

First description in Bertoldi (1989), Bertoldi & McKee (1990) Numerical work by Lefloch & Lazareff (1994, 1995) Ionisation of pre-existing cloud structure Under most typical conditions a D-critical shock is set up & propagates into cloud - bounded by an ionising boundary layer D-critical shock indicated by pressure balance between ionised & neutral gas Shock compresses the cloud and causes collapse Mass loss of photoevaporative flow from cloud surface yields a “Rocket effect”

Radiative-driven implosion

Evolution of cloud t=0.04 Myr to 0.2 Myr from Lefloch & Lazareff (1994) Key parameters in model: ratio of ionising gas density to cloud density Fraction of UV photons absorbed in ionised boundary layer

RDI simulations

Miao et al (2009) illustrate dramatic rise in number density Prediction of RDI is that stars form at the centre of the cloud head within ~1 Myr

RDI results in more massive stars?

Motoyama (2007) numerical models of RDI RDI positively affects accretion rate High accretion rates required to form massive stars Triggered star formation may have a skewed IMF?

Ionisation of turbulent structure

So far have ignored turbulence as a factor But turbulence common in SFR and also predicted by Collect & Collapse model Recent SPH simulations incorporating ionisation (Gritschneder et al 2010) Highly supersonic turbulence forms pillars (similar to the Eagle Nebula) Density peaks resist ionisation longer than low density material

Ionisation as a flashlight?

Does the ionisation just reveal the underlying cloud structure? Underlying dense gas structure

• nly marginally affected by

ionisation in these models Difficult to tell which stars may have been triggered & which were just revealed... (More in Lecture 3)

No feedback vs wind feedback

Simulations courtesy James Dale

Ionisation in unbound cloud

Simulations courtesy James Dale

Ionisation of fractal clouds

Walch et al (2012) follow the evolution of clouds with fractal structure

Ionisation of fractal clouds

Strong evolution in density PDF with ionisation feedback Gas at large N correlated with star formation rate (Lada et al 2010) Gas at low N dispersed from cloud.

The efgect on star formation

In simulations it is easy to see the effect of feedback on SF Turn off feedback in the simulation and see how many stars form Dale et al (2011): effect of ionising feedback ~ 30% over cloud lifetime Walch et al (2012): ionising feedback results in rapid star formation at t < 1 Myr, but blows the cloud apart at t > 1 Myr Look at observational constraints in Lecture 3

Summary

HII regions are overpressured & expand Winds & ionisation fronts drive shocks into ISM and sweep up material Collect & Collapse: swept up shells become self-gravitating Radiative Driven Implosion: cloud clumps induced into collapse Turbulence & fractal structure important for pillar formation Easy to see which stars are triggered in simulations! Coming up: observational study of star formation around bubbles...