On MOND and dark matter in ultra compact dwarf galaxies

On MOND and dark matter in ultra compact dwarf galaxies Jörg Dabringhausen Argelander-Institut für Astronomie, University of Bonn Pavel Kroupa Michael Hilker, Michael Fellhauer, Steffen Mieske, Michael Marks,...

Properties of ultra compact dwarf galaxies (UCDs) luminosities between 10 7 and 10 solar luminosities 6

Properties of ultra compact dwarf galaxies (UCDs) luminosities between 10 7 and 10 solar luminosities Half-light radii of less than 50 pc 6

Properties of ultra compact dwarf galaxies (UCDs) luminosities between 10 7 and 10 solar luminosities Half-light radii of less than 50 pc 6 high densities

Properties of ultra compact dwarf galaxies (UCDs) luminosities between 10 7 UCD and 10 solar luminosities 6 Half-light radii of less than 50 pc de high densities Image by M. Hilker

Properties of ultra compact dwarf galaxies (UCDs) luminosities between 10 7 UCD and 10 solar luminosities 6 Half-light radii of less than 50 pc de high densities intermediate to high ages Image by M. Hilker

From close distance, a UCD probably looks similar to this: Image from ESO

From close distance, a UCD probably looks similar to this: ω Cen Image from ESO

The M/L ratios of UCDs The M/L ratios of UCDs are of particular interest: GCs UCDs (Data from Mieske et al.2008)

The M/L ratios of UCDs The M/L ratios of UCDs are of particular interest: GCs UCDs A part of the difference between ʻclassicalʼ globular clusters and UCDs can be explained by the evolution of ʻclassicalʼ globular clusters......but not all of it. (Data from Mieske et al.2008)

The M/L ratios of UCDs Compare the M/L ratio of UCDs with predictions from models for stellar populations.

The M/L ratios of UCDs Compare the M/L ratio of UCDs with predictions from models for stellar populations. For this, an IMF has to be assumed:

The M/L ratios of UCDs Compare the M/L ratio of UCDs with predictions from models for stellar populations. For this, an IMF has to be assumed: The canonical IMF

The canonical IMF: α=1.3 α=2.3

The canonical IMF: α=1.3 α =2.3 This IMF is apparently invariant in open clusters (Kroupa 2001), which makes it a good initial assumption for UCDs as well.

The M/L ratios of UCDs predicted M/L ratio

The M/L ratios of UCDs 32 UCDs predicted M/L ratio 14 UCDs

The most likely M/L ratios of UCDs exceed the prediction of even very old simple stellar population models (with the canonical IMF)! (Mieske & Kroupa 2008) (Dabringhausen, Hilker & Kroupa 2008)

MOND in UCDs?

MOND in UCDs? GCs UCDs average radii of UCDs increase with their mass! (Data from Mieske et al.2008)

MOND in UCDs? half-mass radius: 33pc Image from Haghi et al. (2009)

MOND in UCDs? half-mass radius: 33pc UCDs Image from Haghi et al. (2009)

MOND in UCDs? half-mass radius: 33pc UCDs MONDian and Newtonian mass estimaes are almost the same for UCDs Image from Haghi et al. (2009)

MOND in UCDs? half-mass radius: 33pc UCDs MONDian and Newtonian mass estimaes are almost the same for UCDs MOND does not solve the problems of the elevated M/L ratios in UCDs Image from Haghi et al. (2009)

Dark matter in UCDs?

Dark matter in UCDs? This cannot be non-baryonic cold dark matter - UCDs are to compact to contain an appreciable amount of it (Murray 2009).

Dark matter in UCDs? This cannot be non-baryonic cold dark matter - UCDs are to compact to contain an appreciable amount of it (Murray 2009). So, could the IMF vary after all? - This would allow for more baryonic (almost) dark matter in some systems. Motivation: UCDs may be associated with the most extreme star bursts (not observed locally). Thus, we may observe an invariant IMF in resolved systems, but it needs not be invariant everywhere.

A possible explanation: A top-heavy IMF

A possible explanation: A top-heavy IMF top-heavy IMF: for old populations, high M/L ratio through over-abundance of stellar remnants

Top-heavy IMF in UCDs Find an IMF so that stellar remnants can explain the M/L ratio of UCDs (Dabringhausen, Kroupa & Baumgardt 2009)

Would UCDs with a top-heavy IMF survive their early evolution?

Would UCDs with a top-heavy IMF survive their early evolution? Perform N-Body simulations of UCDs with mass-loss through gas explusion and stellar evolution

Would UCDs with a top-heavy IMF survive their early evolution? Perform N-Body simulations of UCDs with mass-loss through gas explusion and stellar evolution UCDs can also form with top-heavy IMFs, but this implies extreme initial conditions for them. (Dabringhausen, Fellhauer & Kroupa 2010)

Initial parameters thereby implied for UCDs small present-dayucd massive present-day UCD 1.5 1.9 2.3 2.3 10 100 6 Mass [10 Solar masses ] (Dabringhausen, Fellhauer & Kroupa 2010) 1000

Consider a UCD with some 10 solar masses today 7

Consider a UCD with some 10 solar masses today 7 Initially, it may have had: A mass of some 10 solar masses... 8

Consider a UCD with some 10 solar masses today 7 Initially, it may have had: A mass of some 10 solar masses... 8...but a half-mass radius of only a few pc! (expansion through mass-loss!)

Consider a UCD with some 10 solar masses today 7 Initially, it may have had: A mass of some 10 solar masses... 8...but a half-mass radius of only a few pc! (expansion through mass-loss!) A population of 10 O-stars... 6

Another clue to top-heavy IMFs: Abundance of neutron stars Compared to one with the canonical IMF, a stellar system with a top-heavy IMF should have many neutron stars.

Another clue to top-heavy IMFs: Abundance of neutron stars Compared to one with the canonical IMF, a stellar system with a top-heavy IMF should have many neutron stars. Thus, it can have many binary systems where a neutron star accretes matter from a close companion star, so called low-mass X-ray binaries (LMXBs).

Low-mass X-ray binaries LMXBs make neutron stars visible as bright X-ray sources.

Low-mass X-ray binaries LMXBs make neutron stars visible as bright X-ray sources. The creation of LMXBs is driven by encouters involving stars and neutron stars - such encounters can make binaries close enough for accretion from the star to the neutron star.

LMXBs in globular clusters and UCDs in Virgo The encounter rate is given as Γ n s n ns r 3 c σ. GCs constant radii UCDs radii vary with mass Γ for non-changing stellar mass function

LMXBs in globular clusters and UCDs in Virgo The encounter rate is given as Γ n s n ns r 3 c σ.

LMXBs in globular clusters and UCDs in Virgo The encounter rate is given as Γ n s n ns r 3 c σ.?

LMXBs in globular clusters and UCDs in Virgo The encounter rate is given as Γ n s n ns r 3 c σ.? Is there an IMF, such that the probability for an LMXB in a UCD can be proportional to Γ?

LMXBs in globular clusters and UCDs in Virgo The encounter rate is given as Γ n s n ns r 3 c σ.? Is there an IMF, such that the probability for an LMXB in a UCD can be proportional to Γ? A changing IMF implies n s n ns changes in, and σ.

There is such an IMF:

There is such an IMF: (Dabringhausen, Marks & Kroupa, in preparation)

There is such an IMF: result similar to the one that came from the M/L ratios (Dabringhausen, Marks & Kroupa, in preparation)

LMXBs in globular clusters and UCDs in Virgo?

LMXBs in globular clusters and UCDs in Virgo top-heavy IMF

Conclusions

Conclusions UCDs generally have a higher mass-to-light ratios than expected, if they are a pure stellar population that formed with the canonical IMF.

Conclusions UCDs generally have a higher mass-to-light ratios than expected, if they are a pure stellar population that formed with the canonical IMF. A top-heavy IMF is a possible explanation for this finding. A top-heavy IMF does not contradict the survival of UCDs until today...but their initial conditions must have been very extreme.