A Universal Stellar IMF? A critical look at variations in ...conference.astro.ufl.edu/STARSTOGALAXIES/science_final/talks/covey... · A critical look at variations in Galactic environments

A Universal Stellar IMF? A critical look at variations in Galactic environments

Kevin R. CoveyFrom Stars To Galaxies

4/8/2010

Nate Bastian (Exeter) Michael Meyer (ETH Zurich)

Annual Reviews of Astronomy & Astrophysics, 2010

(in press; arxiv:1001.2965)

A UNIVERSAL STELLAR IMF? A CRITICAL LOOK AT VARIATIONS:

EXTREME ENVIRONMENTS

Nate Bastian (Exeter) Kevin Covey (Cornell), Michael Meyer (ETH Zurich)

Annual Reviews of Astronomy & Astrophysics (2010, volume 48, arXiv:1001.2965)

Environments Examined: - Extended Solar Neighborhood- ‘normal’ young clusters (Taurus->ONC)- super star clusters- Galactic Center

- Globular Clusters- Nearby Galaxies- high-z universe

Do observations provide unambiguous evidence for systematic IMF variations?

Do observations provide unambiguous evidence for systematic IMF variations?

For Theory: McKee & Ostriker 2007; Bate, Clark, Hartmann, Krumholz, Kunz, Mouschovias, Myers,

Hsu (DP3), Urban (DP2)

Environments Examined: - Extended Solar Neighborhood- ‘normal’ young clusters (Taurus->ONC)- super star clusters- Galactic Center

- Globular Clusters- Nearby Galaxies- high-z universeNot Yet.

Alpha/Gamma plot

Rising

Falling

Flat

Low Mass High Mass

From Stars to Galaxies, Gainesville, Florida, 7th-10th April 2010

Star Cluster Formation and the Origin of Stellar Properties

Matthew R. Bate (University of Exeter)

I discuss the results of recent numerical simulations of star cluster formation that simultaneously resolve

large (parsec) scales and small (AU) scales. The large dynamic range allows large numbers of stars and

brown dwarfs to be formed while at the same time resolving binary and multiple systems and circumstellar

discs so that the properties of multiple stellar systems can be determined. This allows the statistical

properties of the stellar systems to be compared with observations. I find that hydrodynamical simulations

that use sink particles to model the stars and brown dwarfs can reproduce many of the observed

properties and trends of stellar systems (e.g. the increasing multiplicity as a function of primary mass;

Bate 2009a). This implies that many stellar properties may originate primarily from dissipative N-body

dynamics. However, these simulations do display some significant discrepancies with observations. In

particular, they over-produce brown dwarfs relative to stars. Adding radiative transfer into the simulations

can correct this deficiency because heating of the gas nearby existing protostars inhibits the

fragmentation of massive discs and nearby gaseous filaments (Bate 2009b). This increases the

characteristic stellar mass (see the figure). Furthermore, while calculations that exclude radiative

feedback result in a characteristic stellar mass that scales linearly with the mean Jeans mass in the

progenitor cloud (Bate & Bonnell 2005), the inclusion of radiative feedback significantly weakens or even

removes this dependence. This may help to explain why the stellar initial mass function is not found to

vary substantially from region to region within our Galaxy. Finally, adding magnetic fields into the

simulations acts to decrease the star formation rate. The combination of strong magnetic fields (plasma

beta < 1) and radiative feedback results in low star formation rates of less than 10% per free-fall time,

similar to those rates that are inferred from observations (Price & Bate 2009).

Left: Distribution of stellar masses from the hydrodynamical star cluster formation calculation of Bate

(2009a) at 1.136 initial cloud free-fall times compared to the Salpeter, Kroupa (2001) and Chabrier (2003)

mass functions. Right: Distribution of stellar masses from an identical calculation at the same time, but

performed using radiation hydrodynamics. The characteristic stellar mass has clearly increased.

References:

Bate, M.R., Bonnell, I.A., 2005, MNRAS, 356, 1201

Bate, M.R. 2009a, MNRAS, 392, 590

Bate, M.R. 2009b, MNRAS, 392, 1363

Chabrier, G. 2003, PASP, 115, 763

Kroupa P. 2001, MNRAS, 322, 231

Price, D.J., Bate, M.R. 2009, MNRAS, 398, 33

63

Forms

Low Mass High Mass

Bate 2009

Field MFs

Rising

Falling

Flat

0.01 0.10 1.00 10.00 100.00Stellar Mass [Msun]

-2

-1

0

1

2

!

Chabrier (2005)Kroupa (2002)

SalpeterAssociations

ClustersField

Star Forming

Low Mass High Mass

See: Rana & Basu 1992; Maciel & Rocha-Pinto 1998; Reid, Gizis & Hawley 2002; Schroder & Pagel 2003; Martini & Osmer 1998; Gould et al. 1997; Reid & Gizis 1997; Zheng et al. 2001; Reyle & Robin 2001;

Schultheis et al. 2006; Robin et al. 2007; Vallinari et al. 2006; Deacon, Nelemans & Hambly 2008; Covey et al. 2008; Bochanski et al. 2010; Reid et al. 1999; Allen et al. 2005; Metchev et al. 2008; Pinfield et al. 2008

Field’s Achilles Heels:IMF/SFH degeneracy;

Mixed stellar pop

IMF Universality? 13

Figure 3: The derived present day mass function of a sample of young star-forming regions (§ 2.3),open clusters spanning a large age range (§ 2.2), and old globular clusters (§ 4.2.1) from thecompilation of de Marchi, Parsesce & Portegies Zwart (2010). Additionally, we show the inferredfield star IMF (§ 2.1). The dashed lines represent “tapered power-law” fits to the data (Eqn. 6).The arrows show the characteristic mass of each fit (mp), the dotted line indicates the meancharacteristic mass of the clusters in each panel, and the shaded region shows the standard deviationof the characteristic masses in that panel (the field star IMF is not included in the calculation of themean/standard deviation). The observations are consistent with a single underlying IMF, althoughthe scatter at and below the stellar/sub-stellar boundary clearly calls for further study. The shiftof the globular clusters characteristic mass to higher masses is expected from considerations ofdynamical evolution.

represent a clear violation of the hypothesis of a spatially and temporally invariant IMF, even ifonly over relatively small spatial scales.

The second explanation is that mass segregation is simply the result of a cluster’s dynamicalevolution, where more massive stars sink to the center of the cluster in about a relaxation time dueto energy equipartition (e.g. Binney & Tremaine 1987). Mass segregation in young clusters is oftenregarded as evidence for primordial mass segregation since their ages are smaller than their currentrelaxation timescale, and thus there has not been su!cient time for dynamical mass segregation totake hold. The Orion Nebula Cluster provides a useful example: it has been known for nearly acentury that the cluster’s most massive stars (the Trapezium stars) sit in the cluster’s center, andBonnell & Davies (1998; see also Hillenbrand & Hartmann 1998) concluded this mass segregationwas primordial, a relic of the cluster’s initial state. Several recent theoretical studies, however,suggest that the timescale for dynamical mass segregation may be shorter than previously thought.Many young clusters are currently expanding (Mackey & Gilmore 2003; Bastian et al. 2008), andpossibly formed through the merger of several smaller sub-clusters. Both of these e"ects suggestthat stars in clusters may have inhabited smaller, denser environments in the past, with corre-spondingly shorter relaxation timescales and faster dynamically evolution into a mass segregated

Cluster MFs

Bastian, Covey & Meyer (via de Marchi)

(see also Wright, DP12; Povich, DP8; Offner, DP1; Sadavoy, BP4)

Cluster MFs

Rising

Falling

Flat

Orion Nebular Cluster

0.01 0.10 1.00 10.00 100.00Stellar Mass [Msun]

-2

-1

0

1

2

!

Chabrier (2005)Kroupa (2002)

SalpeterAssociations

ClustersField

Star Forming

Low Mass High Mass

EXTREME STAR FORMATION IN THE GALAXY

Figer et al.

2 * 104 Msol

Stolte et al. 2005

turn-over at 6-7 Msol

Kim et al. 2006

=0.91.1

EXTREME STAR FORMATION IN THE GALAXY

Arches, NGC 3603 (Pang, DP9; Rochau, DP10), Westerlund 1 (Kudryatseva, DP11) and the ONC are consistent with having the same global IMF through resolved star counts.

All show some sign of mass-segregation

Not necessarily primordial - as mass segregation can happen < 1 Myr and may be a product of cluster formation

Galactic center (<3pc) still an open question

YOUNG MASSIVE CLUSTERS

Larsen & Ritchler 2004

J. Hibbard

Maraston et al. 2004Bastian et al. 2006

5 * 105 Msol

4 * 105 Msol

8 * 107 Msol

1.6 * 107 Msol

M83

NGC 7252

YOUNG MASSIVE CLUSTERS (>20 MYR)

M82-F

Bastian et al. 2007

bottom heavy

top heavy

Maraston 2005 SSP models

Conclusions

Kroupa/Chabrier MF seen in vast majority of field/‘normal’ cluster MFs.

Brown dwarf MFs differ as func. of age: evol. models?

some outliers deserve more attention (Taurus, NGC 2323, Hyades/Praesepe, etc.), but no systematic variations seen.

IMF in ‘extreme’ environments appears to be the same as seen locally

(<1) < ρ [Msol pc-3] < (>105)

SFR surface densities > 5 x 104 Msol kpc-2 yr-1

Orders of magnitude above even the most extreme galactic starbursts in the nearby or distant universe