The impact of metallicity on the demographics of ULXs

Transcription

1 1,2,3 Michela Mapelli 1 INAF, Padova Observatory MERAC prize FIRB fellow The impact of metallicity on the demographics of ULXs Collaborators: Mario Spera, Nicola Giacobbo, Elena Ambrosi, Alessandro Bressan, Emanuele Ripamonti, Monica Colpi, Anna Wolter ULXs and their environments, Strasbourg, June

2 OUTLINE 1. ULX ENVIRONMENT: SFR METALLICITY OFFSET 2. ULX INTERPRETATION: MASSIVE STELLAR BLACK HOLES STAR CLUSTER DYNAMICS INTERMEDIATE-MASS BLACK HOLES 3. Conclusions

3 1. ULX environment: STAR FORMATION In late-type galaxies, ULXs strongly correlate with SFR MM ) ~1-2 ULXs/(Msun yr ~1-2 ULXs/(Msun yr-1) Grimm+2003; Ranalli+ 2003; Gilfanov+ 2004; Kaaret & Alonso-Herrero 2008; Swartz+ 2011; Mineo late-type nearby galaxies with good SFR and X-ray

4 1. ULX environment: METALLICITY SOME HISTORY: HoII X-1 in low Z (~0.1 Z ) nebula (Pakull & Mirioni 2002) NGC1313 X-2 in ~0.2 Z nebula (Zampieri+ 2004, Liu+ 2007) NGC4559 X-7 in low Z ( Z ) region (Soria et al. 2005) ULXs more frequent in low Z (Swartz et al. 2008) Frequency of ULXs normalized to SFR depends on Z (MM+ 2009, 2010, 2011) ULXs more frequent in blue compact dwarf galaxies (Kaaret et al. 2011, Kaaret & Feng 2013, Prestwich et al. 2013, Brorby et al. 2014)..and other studies for single sources... NO NODEFINITIVE DEFINITIVEEVIDENCE EVIDENCE

5 1. ULX environment: METALLICITY From a sample of 66 nearby late-type galaxies with 'good' X-ray data, SFR and metallicity estimate Z<0.2 Z Z>0.2 Z MM Z from Te + Pilyugin & Thuan 2005 for giant spiral galaxies Z measured at 0.7 R25 (average distance of a ULX from centre, background subtracted, Liu, Bregman & Irwin 2006)

6 ULX environment: METALLICITY 1. ULX environment: METALLICITY Number of ULXs per host galaxy normalized to SFR Slope!=0 significant at 96% with F-test MM

7 ULX environment: METALLICITY 1. ULX environment: METALLICITY THE EXTREMELY METAL POOR GALAXIES: I Zw18 and SBS Slope!=0 significant at 96% with F-test MM

8 1. ULX environment: METALLICITY 25 nearby (d 50 Mpc) XMPs (O/H) + 12 < 7.65(~0.05 Zsun) by Prestwich et al low-z galaxies +25 XMPs only 5 Interm.-Z galaxies 17 high-z galaxies + comparison sample from SINGS: Spitzer Infrared Nearby Galaxy Survey 1) high-z 12+log(O/H) ) intermediate-z 8.00<12+log(O/H) ) low-z 12+log(O/H) 8.00 Z from Pilyugin & Thuan 2005 with integrated spectra Prestwich et al. 2013

9 1. ULX environment: METALLICITY And the fundamental plane of Lx SFR metallicity by Brorby et al Brorby et al See Matthew's talk on Monday

10 1. ULX environment: CLOSE TO YOUNG CLUSTERS 1arcsec= 107 pc Accurate astrometry with precision <0.5 arcsec The Antennae, Poutanen (Sergei's talk yesterday afternoon) See also Zezas+ 2002, Zezas & Fabbiano 2002, Kaaret+ 2004, Rangelov+ 2012, Berghea+ 2013

11 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES WHY A ULX-SFR-METALLICITY CONNECTION? 1. BHs are more massive at LOW METALLICITY + some SMALL SUPEREDDINGTON FACTOR (~2-5) e.g. MM et al. 2009, 2010, 2011a, 2011b, 2013; Zampieri & Roberts 2009; MM & Zampieri 2014; Spera, MM & Bressan 2015; MM LOW METALLICITY ENHANCES FORMATION OF HMXBs + ULXs are SUPEREDDINGTON (~10) HMXBs e.g. Dray 2006; Linden+ 2010

12 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES WHY MORE MASSIVE BLACK HOLES AT LOW METALLICITY? 1. STELLAR WINDS: Massive stars (>30 Msun) might lose >50% mass by winds Stellar wind models underwent major upgrade in last ~10 yr (Vink+ 2001, 2005; Bressan+ 2012; Tang, Bressan+ 2014; Chen, Bressan+ 2015) Mass loss depends on metallicity 2. SUPERNOVA: Direct collapse: if final mass of star >30-40 Msun there is no supernova and most star mass becomes BH (Fryer+ 1999, 2001; Heger+ 2003; MM+ 2009; Belczynski+ 2010; Fryer+ 2012; MM+ 2013; Spera, MM, Bressan 2015; Spera, Giacobbo, MM 2016) = METAL-POOR STARS PRODUCE MORE MASSIVE BH

13 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES PARSEC stellar evolution (Tang, Bressan+ 2014; Chen, Bressan+ 2015) + delayed SN model (Fryer+ 2012) New SEVN code (Stellar Evolution for N-body) MM+ 2009; Spera, MM & Bressan 2015; Spera, Giacobbo & MM 2016 WITHOUT VERY MASSIVE STARS WITHOUT PAIR INSTABILITY SN (PISN)

14 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES PARSEC stellar evolution (Tang, Bressan+ 2014; Chen, Bressan+ 2015) + delayed SN model (Fryer+ 2012) New SEVN code (Stellar Evolution for N-body) MM+ 2009; Spera, MM & Bressan 2015; Spera, Giacobbo & MM 2016 WITH VERY MASSIVE STARS WITHOUT PAIR INSTABILITY SN (PISN)

15 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES PARSEC stellar evolution (Tang, Bressan+ 2014; Chen, Bressan+ 2015) + delayed SN model (Fryer+ 2012) New SEVN code (Stellar Evolution for N-body) MM+ 2009; Spera, MM & Bressan 2015; Spera, Giacobbo & MM 2016 WITH VERY MASSIVE STARS WITH PAIR INSTABILITY SN (PISN)

16 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES IF WE ASSUME THAT ALL ULXs ARE HMXBs powered by massive stellar BHs, we obtain the model shown by the RED line Basic assumptions: Number of BHs SFR, Mass of BHs Z0.8 MM+ 2010, 2011

17 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES DO MASSIVE BLACK HOLES EXIST? GW DETECTION!!! LIGO laborarory GW150914: merger of two BHs 36 (+5,-4) Msun 29 (+4,-4) Msun observed by LIGO Abbott+2016 GW shows that 1. BH-BH binaries exist 2. they can merge in a Hubble time 3. massive stellar BHs exist i.e. stellar BHs with mass >25 Msun

18 2. ULX interpretation: MASSIVE STELLAR BLACK HOLES DO MASSIVE BLACK HOLES EXIST? GW DETECTION!!! PARSEC stellar evolution (Tang, Bressan+ 2014; Chen, Bressan+ 2015) GW delayed SN model (Fryer+ 2012) GW Spera, MM & Bressan 2015 used as fig.1 by Abbott paper on Astrophysical implications of LIGO detection

19 2. ULX interpretation: STAR CLUSTER DYNAMICS BUT 1. population-synthesis models (e.g. Linden+ 2010) show that binary evolution quenches formation of massive stellar BHs 2. how do we explain the correlation between ULXs and star clusters? POSSIBLE ANSWER: STAR CLUSTER DYNAMICS

20 2. ULX interpretation: STAR CLUSTER DYNAMICs LARGE FRACTION (~80%) OF STARS FORM IN YOUNG STAR CLUSTERS (YSCs, Lada & Lada 2003), especially MASSIVE STARS DENSE YSCs are YOUNG (<100 Myr), RELATIVELY MASSIVE ( M ), FORM IN LOCAL UNIVERSE, and ARE DYNAMICALLY ACTIVE Embedded cluster (RCW 38) VLT Arches NICMOS HST

21 2. ULX interpretation: STAR CLUSTER DYNAMICs DYNAMICAL EXCHANGES: 3-body encounters (and especially EXCHANGES) ENHANCE THE FORMATION OF BH-BINARIES Exchanges enhance the formation of MASSIVE BH-binaries because favour the formation of MASSIVE binaries > 90% BH binaries in star clusters form via exchange (Ziosi, MM+ 2014)

22 2. ULX interpretation: STAR CLUSTER DYNAMICs TO STUDY DYNAMICS WE NEED N-BODY SIMULATIONS STARLAB (Portegies Zwart+2001): - accurate N-Body integration of SC dynamics - stellar evolution at solar Z; each particle is a star with evolving radius, luminosity, temperature OUR UPGRADE OF STARLAB (MM+ 2013): - metallicity dependence of stellar evolution - metallicity dependent stellar winds for MS and WR - metallicity-dependent recipes for SN and BH mass SEVN: OUR NEW POPULATION SYNTHESIS CODE for N-body simulations (Spera, MM+ 2015; Spera, Giacobbo, MM 2016): - stellar evolution with PARSEC tracks (Bressan+ 2012; Chen+ 2014; Tang+ 2014) - metallicity-dependent recipes for SN and collapse (Fryer+ 2012; O'Connor & Ott 2011; Ertl et al. 2016)

23 2. ULX interpretation: STAR CLUSTER DYNAMICs 600 YOUNG SCs with Z=0.01, 0.1 and 1 Z - rvirial = 1 pc - total mass ~ 3500 Msun per SC - primordial binaries (~10%) - Kroupa IMF (Kroupa 2001) - RUN for 100 Myr -cfr Orion Nebula Cluster ** details in MM+2013 ** part of ONC (infrared, NICMOS)

24 2. ULX interpretation: STAR CLUSTER DYNAMICs EXCHANGES FAVOUR HIGH-MASS BHs in RLO systems exchanged non-exchanged RLO BH-binaries ALL METALLICITIES exchanged systems start RLO LATER exchanged systems contain more massive BHs MM & Zampieri 2014

25 2. ULX interpretation: STAR CLUSTER DYNAMICs RLO systems at LOW/HIGH METALLICITY exchanged TIME since YSC formation Z=Zsun non-exchanged Z=0.1 Zsun MM & Zampieri 2014

26 2. ULX interpretation: STAR CLUSTER DYNAMICs RLO systems at LOW/HIGH METALLICITY exchanged BH mass Z=Zsun non-exchanged Z=0.1 Zsun MASSIVE STELLAR BHs >25 Msun MM & Zampieri 2014

27 2. ULX interpretation: STAR CLUSTER DYNAMICs RLO systems at LOW/HIGH METALLICITY exchanged donor mass Z=Zsun non-exchanged Z=0.1 Zsun MM & Zampieri 2014

28 2. ULX interpretation: STAR CLUSTER DYNAMICs OPTICAL LUMINOSITY and COLOUR of the SIMULATED RLO SYSTEMS: come from - donor star from N-body simulations - X-ray reprocessing of disc and donor star from code by Patruno & Zampieri (2008, 2010) We produce optical luminosity and colours of simulated RLO systems in B, V Johnson filters, Vegamag, as observed at 5 Mpc distance See talk by Elena Ambrosi MM & Zampieri 2014

29 2. ULX interpretation: STAR CLUSTER DYNAMICs primordial binaries exchanged binaries exchanged binaries are redder because lower mass donor stars low-mass BHs MSBHs RLO-MSBHs are a sub-group of exchanged binaries

30 2. ULX interpretation: STAR CLUSTER DYNAMICs HST counterparts from Gladstone et al. 5Mpc: := F555W, F435W more reliable := F606W, F435W (F606W shifted to match F555W)

31 2. ULX interpretation: STAR CLUSTER DYNAMICs HST counterparts from Gladstone et al. 5Mpc: M81 X-6, HoIX X-1, NGC1313 X-1, NGC 1313 X-2, IC-342 X-1, M83 XMM1, NGC 2403 X-1, NGC 5204 X-1, NGC3034 ULX5 MM & Zampieri 2014

32 2. ULX interpretation: STAR CLUSTER DYNAMICs ULX star cluster offset: effect of dynamical ejections? Kaaret et al (bright X-ray binaries in M82, NGC1569, NGC5253) Berghea, PhD Thesis, 2009 Berghea et al (ULXs in nearby galaxies) Poutanen et al (bright X-ray binaries in the Antennae)

33 2. ULX interpretation: STAR CLUSTER DYNAMICs ULX star cluster offset: effect of dynamical ejections? Kaaret et al (bright X-ray binaries in M82, Kaaret et al (bright NGC1569, NGC5253) X-ray binaries in M82, MM etngc1569, al. PhD 2011Thesis, NGC5253) Berghea, 2009 simulated BH 2013 binaries in Berghea et al. Berghea, YSC nearby NOThesis, stellar 2009 (ULXs in PhD galaxies) evolution Berghea et al (ULXs in nearby galaxies) Poutanen et al (bright X-ray binaries in the Poutanen et al Antennae) (bright X-ray binaries in the Antennae) MM et al simulated BH binaries in YSC NO stellar evolution

34 2. ULX interpretation: STAR CLUSTER DYNAMICs ULX star cluster offset: effect of dynamical ejections? Kaaret et al (bright X-ray binaries in M82, NGC1569, NGC5253) Berghea, PhD Thesis, 2009 Berghea et al (ULXs in nearby galaxies) Poutanen et al (bright X-ray binaries in the Antennae) MM et al simulated BH binaries in YSC NO stellar evolution MM et al simulated RLO binaries in YSC with star evolution

35 2. ULX interpretation: Intermediate-mass black holes THE RUNAWAY COLLISION SCENARIO: Mass segregation fast in young star clusters: Massive stars segregate to the centre where collide with each other? Massive super-star forms and possibly collapses to IMBH What is the final mass of the collision product? Colgate 1967; Sanders 1970; Portegies Zwart+ 1999, 2002, 2004; Gurkan+ 2004; Freitag+ 2006; Giersz+ 2015, and many many others

36 2. ULX interpretation: Intermediate-mass black holes MM 2016: First N-body simulations of runaway collisions with stellar winds and failed SNae * Max. mass up to 500 Msun * NO IMBHs at solar metallicity * 1/10 BHs in the IMBH regime at Z = Zsun MM

37 2. ULX interpretation: Intermediate-mass black holes MM 2016: First N-body simulations of runaway collisions with stellar winds and failed SNae * MOST MASSIVE BHs in these simulations END UP IN BH - BH BINARIES in ~ 10 Myr (no stable X-ray sources) 4 BH-BH at Z = 0.01 Zsun 1 BH-NS at Z = 0.01 Zsun 2 BH-BH at Z = 0.1 Zsun 2 BH-BH at Z = 1 Zsun ALL DOUBLE COMPACT OBJECTS PERIOD from few hours to few yr MM

38 3. CONCLUSIONS: 1- ULXs have strong correlation with SFR, strong correlation with star clusters, and maybe anti-correlate with metallicity (Z) (MM+ 2010, 2011) 2- models predict existence of MORE MASSIVE stellar BHs (>20 Msun) at low Z (<0.5 Zsun) than high Z (MM+ 2009, 2013; Spera+ 2015) 3- the DYNAMICAL ENHANCEMENT of massive BH binaries in star clusters can explain correlation of ULXs with star clusters (MM & Zampieri 2014) 4- the DYNAMICAL EJECTION of BH binaries from star clusters can explain the OFFSET between ULXs and star clusters (MM+ 2011, 2013) 5- ULXs born from dynamically formed binaries are more massive and form later than primordial binaries (MM & Zampieri 2014) 6- RUNAWAY COLLISION in star clusters might produce even IMBHs, but only at low metallicity (MM 2016) THANK YOU