Evidence for systematic IMF variability "ESO@50 - the first 50 years of ESO" September 3 - 7, 2012 ESO, Garching Pavel Kroupa Argelander Institute for Astronomy (AIfA) University of Bonn Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 1 The observationally derived IMF of stars places firm constraints on the cosmological matter cycle. Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 2 Kroupa et al. 2012 (150 page IMF review) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 3 Good working hypothesis: The IMF is universal. It is the same, independent of the physical conditions of star formation. But see The IMF canonical / standard / universal Unmeasurability two-part power-law IMF : Theorem −α ξ(m) ∝ m i α1 =1.3 logdN/dlog(m) α2 =2 . 3 (Massey-Salpeter) α3,Massey =2.3 M stars G stars O stars 0 log(m) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 4 The low-mass system end Brown Dwarfs vs Stars ? Pavel Kroupa: Sternwarte, University of Bonn Monday, 3 September 2012 5 Hennebelle 2012 Cloud fragmentation predicts too few BDs Padoan & Nordlund 2002 Significant deficit of theoretical BDs and IMF is not a log-normal Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 6 Hennebelle 2012 Cloud fragmentation predicts too few BDs "Our model reproduces Padoan & Nordlund 2002 well . an initial mass function that is i) very close to the Chabrier "given the success of the IMF" present model in predicting the observed shape of the stellar IMF" recent IMF work sociologically driven Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 7 What we know from observation : Brown dwarf desert (nearly) only star - star binaries Binary fraction among stars in MW > 50 % (100 % in dynamically young systems, 50 % in dynamically evolved systems, e.g. open clusters, Galactic field) m Approx. flat mass-ratio distribution for 0.2 < primary < few M ! BD - BD binary fraction 15 % ≈ What this implies : Note: this is purely a result inferred of counting # stars/m stellar IMF "observed" system IMF BDs stars m Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 8 (Thies et al. 2010, ApJ) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 9 Thies, Kroupa, Goodwin, Basu & Vorobyov 2012 Stamatellos, Whitworth 2010 Pavel Kroupa: Sternwarte, University of Bonn Monday, 3 September 2012 10 Thies et al. 2010 log # 18 DRAGON computations : 58 BD & VLMS sinks formed +0.3 Sternwarte, University of Bonn αBD =0.1 0.4 − convergence of theory and observation (correctly interpreted !). Pavel Kroupa: Monday, 3 September 2012 11 Thies et al. 2010 log # 18 DRAGON computations : 58 BD & VLMS sinks formed +0.3 Sternwarte, University of Bonn αBD =0.1 0.4 − convergence of theory and observation (correctly interpreted !). Pavel Kroupa: Monday, 3 September 2012 12 Thies et al. 2010 log # 18 DRAGON computations : 58 BD & VLMS sinks formed +0.3 Sternwarte, University of Bonn αBD =0.1 0.4 − convergence of theory and observation (correctly interpreted !). Pavel Kroupa: Monday, 3 September 2012 13 Thies et al. 2010 log # 18 DRAGON computations : 58 BD & VLMS sinks formed +0.3 Sternwarte, University of Bonn αBD =0.1 0.4 − convergence of theory and observation (correctly interpreted !). Pavel Kroupa: Monday, 3 September 2012 14 Kirk & Myers 2012 Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 15 Luhman et al. 2009 Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 16 new universal / canonical discontinuous three-part power-law IMF : −α ξ(m) ∝ m i α1 =1.3 ( the result of logdN/dlog(m) one correlated α =2.3 2 star-formation event (CSFE) - α0 ≈ 0.3 star cluster ! ) 0 ≈ α3,Massey =2.3 No log-normal form ! (sorry Chabrier!) BDs M stars G stars O stars 0 log(m) discontinuity: Thies & Kroupa (2007, 2008) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 17 Ingo Thies (AIfA, Bonn) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 18 ESO data ! e.g. : Bouvier et al., 2008, "Brown dwarfs and very low mass stars in the Hyades cluster: a dynamically evolved mass function", EMMI Moraux et al. 2007, "The lower mass function of the young open cluster Blanco 1 : from 30 Mjup to 3 Msun", FORS2, SOFI Nürnberger & Petr-Gotzens 2002, "Infrared observations of NGC 3603. I. New constraints on cluster radius and Ks-band luminosity function", ISAAC Muench et al. 2002, "The Luminosity and Mass Function of the Trapezium Cluster: From B Stars to the Deuterium-burning Limit", SOFI Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 19 Bottom-heavy IMFs in metal-rich environments / in E galaxies ? Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 20 With increasing metallicity, SF may be producing increasingly "bottom heavy" IMFs : α 1.3 + 0.5[Fe/H]; m<0.7M ! ≈ Kroupa 2001, 2002 Find long-sought cooling flow population of low-mass stars using gravity-sensitive spectral lines : Kroupa & Gilmore 1994 With increasing E-galaxy mass, IMFs in E galaxies indeed seem to become increasingly "bottom heavy". α =3.41 + 2.78[Fe/H] 3.79[Fe/H]2;0.1 < m/M < 100 − ! Cenarro et al. 2003 see also van Dokkum & Conroy 2011 Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 21 ESO data ! Saglia et al. 2002, "The Puzzlingly Small Ca II Triplet Absorption in Elliptical Galaxies", EMMI Abstract: " . (2) The steepening of the IMF at low masses required to lower the CaT* and CaT indices to the observed values is incompatible with the measured FeH index at 9916 Å and the dynamical mass-to-light ratios of elliptical galaxies. " ? Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 22 IMF variation : the canonical IMF Pavel Kroupa: Sternwarte, University of Bonn Monday, 3 September 2012 23 canonical / standard / universal Remember : two-part power-law stellar IMF : −α ξ(m) ∝ m i α1 =1.3 logdN/dlog(m) α2 =2.3 (Salpeter) α3,Massey =2.3 M stars G stars O stars mmax 0 log(m) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 24 Variation of the . So : canonical / standard / universal IMF ? dlogN/dlog(m) 0 stars log(m) 0.5 M e.g. 150 M ! ! Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 25 Variation of the . So : canonical / standard / universal IMF ? R136 logdN/dlog(m) −α ξ(m) ∝ m i ONC mmax 150 or 300 M ≈ ! Taurus mmax 50 M mmax 2 M ≈ ! M stars G stars ≈ ! O stars 0 stars log(m) 0.5 M e.g. 150 M ! ! Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 26 The m max ( M ecl ) relation Weidner & Kroupa 2005, 2006; Weidner et al. 2010 mmax = 300 M (Crowther,∗ Schnurr et" al. 2010) physical maximum stellar mass ? mmax,∗ ≈ 150 M" (Weidner & Kroupa 2004; Figer 2005; Oey & Clarke 2005, Koen 2006; Maiz Appelaniz et al. 2007) mmax∗ 1= ξ(m) dm !mmax mmax Mecl = mξ(m) dm !ml ESO data ! Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 27 Top-heavy IMF Globular Clusters ? (ancient star bursts) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 28 A sample of 20 Galactic GCs (de Marchi, Paresce & Pulone 2007) with solid global MF measurements from One of the most deep HST or VLT data. important recent MF papers ! ESO data ! canonical (universal) IMF normal low-mass star population expectation too few low-mass stars low-concentration clusters ought to be dynamically less evolved Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 29 Nbody models of binary rich initially mass segregated clusters with redisual gas expulsion after birth (Marks, Kroupa & Baumgardt 2008) gas expulsion stellar loss independent loose mostly of mass low-mass stars Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 30 Cluster reaction to sudden gas removal : (movie by Baumgardt) Baumgardt & Kroupa 2007, Bastian & Goodwin . 31 Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 31 Nbody models of binary rich initially mass segregated clusters with redisual gas expulsion after birth (Marks, Kroupa & Baumgardt 2008) −α [ ξ(m) ∝ m ] Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 32 Nbody models of binary rich initially mass segregated clusters with redisual gas expulsion after birth (Marks, Kroupa & Baumgardt 2008) −α [ ξ(m) ∝ m ] ✓for residual gas expulsion + mass segregated clusters (de Marchi, Paresce & Pulone 2007) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 33 Top-heavy IMF in extreme-density environments : 3.5 Galactic globular clusters Marks et al. 2012 6352 UCD sim (SFE 1.0,HE 1.00) UCD sim (SFE 0.4,HE 1.00) 3 6496 UCD sim (SFE 0.4,HE 0.03) 2.5 6809 288 standard IMF slope 6838 2 3 ! 6712 1.5 α3 = fn(Mecl) top-heavy slope 6656 6218 6397 1 2298 6254 7099 6752 0.5 5272 5139 144 6341 Arches (2-3 Myr, d =25pc) 6121 0 GC NGC 6303 (1-3 Myr, dGC=8kpc) 0.01 0.1 1 10 100 1000 6 -3 Cloud density [10 Msun pc ] Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 34 Michael Marks (AIfA, Bonn) 35 Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 35 Top-heavy IMF in UCDs ? the M/L ratio (ultra-compact dwarf galaxies) Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 36 Properties of Ultra Compact Dwarf galaxies (UCDs) UCDs occur UCD mostly in galaxy clusters dE Image by M. Hilker Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 37 From close distance, a UCD probably looks similar to this: ω Cen Image from ESO Pavel Kroupa: AIfA, University of Bonn Monday, 3 September 2012 38 M/L vs mass : (Dabringhausen et al.