Galactic and Extragalactic star formation

M.Walmsley (Arcetri Observatory)

Simplified View of Extragalactic Star Formation

• Based upon :

– Assuming local IMF is valid

everywhere

– Assuming for the SF efficiency some

version of the Schmidt LawAMAZING IF TRUE !

What does galactic star formation tell us?

• Star formation occurs in molecular clouds

• More precisely, it occurs within the dense parts of molecular clouds

This raises the question of whether star formation occurs because the gas is dense or because it is molecular orboth

Galactic Extragalactic connection

• That there is a connection is clear from the various correlations found between star

formation and molecular line luminosity (most recently from Wu et al)

• The HCN luminosity tracks the IR luminosity with the same relationship for galactic and extragalactic SFR

HCN luminosity

In galactic clouds, L(HCN) is thought to be

roughly given by:

L(HCN) = ∫Cex n(H2) n(HCN) dV

This involves the HCN abundance and hence the Wu et al. result suggests similar chemistry in the extragalactic clouds as in galactic.

So the Wu et al. result seems to imply that extragalacticstarbursts are similar to galactic starbursts but muchlarger

Do abundances differ in galactic and extragalactic clouds?

• Yes ! But there are some rough correspondences as one sees comparing NGC253 to galactic clouds (S.Martin et al.)

Simplified view of IMF

• Field Star IMF is within errors same as that inferred for ONC (Orion Nebula Cluster) and other nearby star forming regions

• It has a power law (Salpeter) down to about 0.5-1 M(Sun) with most mass in solar mass stars but most luminosity at high M

• Evidence for deviations from standard IMF in some Gal. Center clusters

One possible explanation of the IMF

• It reflects the mass distribution of the cloud fragments or cores in the molecular cloud

• The “typical” mass of around 1 MO then reflects the Jeans Mass (very T dependent)

M(JEANS) ~ T3/2 n-1/2

The origin of the Initial Mass Function

(see also Testi & Sargent 1998; Motte et al. 2001)

Submm continuum surveys of nearby protoclusters suggest that the mass distribution of pre-stellar condensations mimics the form of the stellar IMF

NGC2068 protocluster at 850 m

Motte et al. 2001

Condensations mass spectrum in Oph

The IMF is at least partly determined by fragmentation at the pre-stellar stage.

Consequences for extragalactic SF

• If fragmentation is fundamental in determining the IMF, the Jeans Mass and hence the temperature may determine the critical turn-over mass

• This could cause the IMF in galactic nuclei to be more biased towards high mass ???

• Temperatures in Galactic Center clouds are high

The Schmidt Law

• The Schmidt Law for the star formation rate (SFR) has many forms:

SFR = d∑/dt ~∑p with p=1-2

Alternatively : d∑/dt ~∑/t(SF)

Where ∑ is col.density and t(SF) is timescalefor star formation

Galactic timescale for Star Formation tSF

• One might naturally think it was the free-fall time at the mean density of molecular clouds

• But as pointed out in the 70s by Zuckerman and Evans, real galactic SF Rate is lower (tSF=109 yr) than from free fall time (tff roughly 106 years)

• This has given rise to two classes of theories:– “slow”: including “ambipolar diffusion” modulated theories.– “inefficient”: turbulence, HII regions and winds

Star Form. Rate in Galactic Dense Clumps

• From Plume et al (1997) the SF rate in galactic clouds corresponds to a timescale of 107 to 108 years - but tff is 105 yr

tff/tSF as in GMCs

But it cannot be too inefficient

• Some cluster masses are 10 percent of maximum GMC masses (Blitz et al., Clark)

Both “slow” and “ineffcient” SF may be needed

Maybe better to write the Schmidt law:

d∑/dt = ∑/ tSF

where tSF = ß tff

/ ~ 0.01

Conclusions

• Extragalactic star formation may well be just galactic writ large

• But we do not understand what determines the efficiencies and timescales

• Of course the IMF might be playing tricks

Stars form in spiral arms

M33 Spitzer ImageFrom Verley et al.

What theory can say about Schmidt Laws

• Mainly that Nature can conspire to make t(SF) equal to orbital time scale

(Tan, McKee and others)

For example cloud collision rate depends on shear which in turn depends on orbital parameters ??