DYNAMICAL MASSES IN BLACK HOLE XRAY BINARIES Jorge

DYNAMICAL MASSES IN BLACK HOLE X-RAY BINARIES Jorge Casares (IAC)

Outline 1. - Dynamical Masses: techniques, limitations and systematics 2. - Bowen emission: dynamical masses during outburst 3. - Search for dormant BH XTs: IPHAS/UVEX surveys 4. - Conclusions

1. - BHs in XRBs • Mass distribution of BHs has strong impact in SNe models, evolution of massive stars, chemical enrichment • Stellar evolution models predict ~ 108 BH remnants (Brown & Bethe 94). • But only BHs in interacting binaries can be easily detected and weighed accurately. • Best evidence for BH is dynamical i. e. a compact object with Mx > 3 M

The Classic Method: Weighing BHs do not burst nor pulse dynamical information must be extracted from donor star. Three experiments: V 404 Cyg 1) Radial velocity curve q=MC/MX f (M) < MX R=λ/Δλ≥ 1500 required

Weighing BHs 2) Measure Vrot sini e. g. in V 404 Cyg we measure Vsini=40 km/s and then q=0. 06 R=λ/Δλ≥ 5000 required

Weighing BHs 3) Fit ellipsoidal modulation GRO J 1655 -40 (Orosz & Bailyn. 97) Amplitude is strong function of inclination f(M) + q + i complete solution

BHs in HMXBs: Cyg X-1 • Optical counterpart: O 9. 7 Iab (=HD 226868) • Velocity K= 64 km/s and Porb= 5. 6 d. (Webster & Murdin; Bolton 1972) = 0. 25 M Typically O 9. 7 Iab has ≈ 33 M (and then MX ≈ 7 M , i=90º) • But donor likely undermassive by factor ≤ 3 (van den Heuvel & Ostriker 73) • Mass transfer is powered by stellar winds determination of q & i through Vsini and ellipsoidal fits can be biased low

M 33 X-7: First eclipsing BH Eclipsing X-ray source with a 3. 45 d period (Peres et al. 1989) First ellipsoidal fits and radial velocity curve of the O 7 -8 III donor (Orosz et al. 2007) Eclipse+distance constrain parameter space: RC=19. 6± 1 R , f=0. 78± 0. 02 i=74. 6± 1. 0 o Mc=70± 7 M Mx=15. 7± 1. 5 M

BHs in X-ray Transients Low mass donors have no winds and fill Roche lobes A 0620 -00 OUTBURST: Companion 103 fainter than X-ray irradiated disc QUIESCENCE: companion dominates optical flux radial velocity studies

Galactic Black Holes with dynamical evidence Orosz 2003; Charles & Coe 2006 System GRS 1915+105 V 404 Cyg X-1 LMC X-1 M 33 X-7 XTE J 1819 -254 GRO J 1655 -40 BW Cir GX 339 -4 LMC X-3 XTE J 1550 -564 4 U 1543 -475 H 1705 -250 GS 1124 -684 XTE J 1859+226 GS 2000+250 A 0620 -003 XTE J 1650 -500 GRS 1009 -45 GRO J 0422+32 XTE J 1118+480 Porb 33. 5 d: 6. 470 d 5. 600 d 4. 229 d 3. 453 d 2. 817 d 2. 620 d 2. 545 d 1. 754 d 1. 704 d 1. 542 d 1. 125 d 0. 520 d 0. 433 d 0. 382 d : 0. 345 d 0. 3205 d 0. 283 d 0. 212 d 0. 171 d f(M) Spect. Type 9. 5 ± 3. 0 M 6. 08 ± 0. 06 M 0. 244 ± 0. 005 M 0. 14 ± 0. 05 M 0. 46 ± 0. 08 M 2. 74 ± 0. 04 M 2. 73 ± 0. 09 M 5. 75 ± 0. 30 M 5. 8 ± 0. 5 M 2. 3 ± 0. 3 M 7. 65 ± 0. 38 M 0. 25 ± 0. 01 M 4. 65 ± 0. 21 M 3. 01 ± 0. 15 M 7. 4 ± 1. 1 M : 5. 01 ± 0. 12 M 2. 72 ± 0. 06 M 2. 73 ± 0. 56 M 3. 17 ± 0. 12 M 1. 19 ± 0. 02 M 6. 3 ± 0. 2 M Classification Mx K/MIII Transient 14. 0 ± 4. 4 M et al. 2007 P=30. 8 d Neil K 0 IV , , 12 ± 2 M 09. 7 Iab Persistent 10 ± 3 M 07 III , , 10. 9 ± 1. 4 M 07 -8 III , , 15. 7± 1. 5 M B 9 III Transient 7. 1 ± 0. 3 M F 3/5 IV , , 6. 3 ± 0. 3 M G 5 IV >7. 8 M , , -, , >6. 0 M B 3 V Persistent 7. 6 ± 1. 3 M G 8/K 8 IV Transient 9. 1 ± 0. 6 M A 2 V , , 9. 4 ± 1. 1 M K 3/7 V , , 6 ± 2 M K 3/5 V , , 7. 0 ± 0. 6 M Filippenko & Chornock 2001 IAUC 7644 -, , K 3/7 V , , 7. 5 ± 0. 3 M K 4 V , , 6. 6 ± 0. 3 M K 4 V , , >2. 5 M K 7/M 0 V , , 5. 2 ± 0. 6 M M 2 V , , 4 ± 1 M K 5/M 0 V , , 6. 8 ± 0. 4 M Orosz et al. 2009 (LMC X-1), Cantrell et al. 2010 (A 0620 -00); Orsoz et al. 2010 (J 1550 -564)

Late Results on XTE J 1859+226 Orbital period is not 9. 2 hr (Filippenko & Chornock 2001 IAUC 7644) but either 6. 6 or 7. 7 hr 9. 2 hr GTC 10. 4 m spectroscopy 2 x 3 hr blocks in July & Aug 2010 K≈520 km/s f(M)=4 -4. 5 M Corral-Santana et al. 2010 submitted

Mass spectrum of BHs 17 reliable masses of BHs: 4 -16 M Typical errors 20% – – Do BH masses cluster at a particular value? What are the edges of the BH distribution? Limited by errors and poor statistics New strategies required

Systematics in Ellipsoidal Fits 1) Superhumps: modulation a few % longer than orbital caused by precessing disc Typically seen in outburst, when disc exceeds 3: 1 resonance radius (O’Donoghue & Charles 1996) XTE J 1118+480: near quiescence (Zurita et al. 2002)

Systematics in Ellipsoidal Fits 2) Rapid aperiodic variability: ~ 6 hr flares in V 404 Cyg Zurita, Casares & Shahbaz (2003) Pavlenko et al. (1996) All SXTs show flaring activity when observed at high time resolution. flaring originate in the disc whose contribution fades with λ Fits in the IR assumed to be safe

Systematics in Ellipsoidal Fits 2) Rapid aperiodic variability: varies with time for a particular system Cantrell et al. (2008) i=51 ± 1 deg i=37 ± 6 deg Mx=6. 6 ± 0. 3 M Mx=10 ± 2 M i=41 ± 3 deg Mx=11 ± 2 M A 0620 -00 Cantrell et al. (2010) Shahbaz et al. (1994) Gelino et al. (2001) Critical because inclination dominates the mass error budget

The Future: astrometry of donor stars BH BINARY V 404 Cyg GRS J 1915+105 Cyg X-1 GRO J 1655 -40 A 0620 -00 a/D (μas) 61 40 33 18 14 VLTI+GRAVITY(2013): will provide 10 μas for m. K=15 and will resolve V 404 Cyg and GRS 1915+105 Space Interferometry Mission (SIM 2016) High sensitivity : 20 th mag Angular resolution: 3. 5μas Will provide accurate inclinations and masses ≤ 5% calibrate ellipsoidal fits

32 BH Candidates: prime targets for rv studies SWIFT J 1753 -0127: remained in low-hard state since discovery 2005 2007 2009 2010 Evidence for superhump variability with P=3. 24 hr (Zurita et al. 2008) Shortest period BH binary

BH Candidates: SWIFT J 1753. 5 -0127 Outburst Amplitude vs Porb correlation (Shahbaz & Kuulkers 1998) Since P=3. 24 hr then ΔV ≈ 10 and Vquiescent ~ 26 Too faint for dynamical studies

Black Hole Candidates Only a few are brighter than R~23 (limit of 10 m Telescopes) Beyond R~23 we need ELT New strategies to expand sample of dynamical BHs: 1. - Dynamical studies during outburst 2. - Search for new quiescent BHs

1. - Dynamical studies during outburst: NIII/CIII He. II Bowen Emission from the Donor Steeghs & Casares 2002 Detection of sharp high excitation emission lines Most prominent are CIII/NIII at λλ 4630 -40 NIII powered by fluorescence Doppler shift traces orbit of heated companion GX 339 -4 ● Classic Black Hole candidate ● Quiescence in 2000 -01 ● 2002 outburst AAT+NTT Shahbaz et al. 2001 NIII/CIII emission lines VLT

GX 339 -4 Multigausian fit to NIII lines Porb=1. 76 d from He. II velocities Kem=317 ± 10 km/s K 2 f(M) 5. 8 ± 0. 5 M Black Hole!! (Hynes et al. 2003 Ap. J 583 L 95) Determine f (M) in new SXTs during outburst before they fade (and are lost in quiescence)

Echo Tomography ● Time delays between X-ray and optical variability to map reprocessing regions in a binary O(t) = X(t) * T(t- ) depends on geometry and varies with orbital phase • disc geometry • binary separation a • inclination i Credit: K. 0’Brien Most works use broad band filters • mass ratio q=M 2/MX delays consistent with disc reprocessing

Bowen line Echo Tomography : Sco X-1 Narrow filters to amplify reprocessed signal from companion: Bowen+He. II & continuum subtracted lightcurves of lines Muñoz-Darias et al. 2007 W 3 Flaring Branch =0. 52 W 4 =0. 42 W 1 W 2 =0. 60 W 5 =0. 73

2. - Search for new quiescent BHs BH demography: how many? MAXI SWIFT INTEGRAL XTE SAX ASCA CGRO ROSAT GRANAT 32 BH candidates Cen X-2 GINGA 17 dynamical BHs Discovered at a rate ~1. 7/yr since GINGA Extrapolation of SXTs since late 80’s + assuming outburst duty cycle ~ 50 yrs suggest ~ 103 dormant BH SXTs (van den Heuvel 93, Romani 98) BH SXTs known are the tip of iceberg of Galactic population

The IPHAS/UVEX Survey ● ● INT Photometric Surveys of the northern Galactic plane (|b|<5º) in Hα , r’, i’ (IPHAS) and u’, g’, He. I 5876 (UVEX) down to r’=20 (Drew et al. 2005) Covers ~1800 sq. deg with ~5 x 105 Hα emitters in 300 Mill. stellar objects (0. 1%) CV BH XRT

IPHAS/UVEX/2 MASS Cross-match of IPHAS/UVEX /2 MASS with these colour cuts yields ~5000 candidates. Most will still be contaminants so follow-up spectroscopy required for confirmation.

Conclusions ● ● Best way to measure BH masses is through dynamical studies of X-ray binaries SXTs provide many more & cleaner cases than HMXBs: 17 confirmed BHs with masses in the range 4— 14 M (caution with systematics in ellip. mod. and hence i) Tip of iceberg of hidden population of ~103 BH binaries Better statistics & ≤ 10% errors needed to derive constraints to close binary evolution and SNe models.

Conclusions 2 possible avenues to increase the population: ● ● Exploit Bowen Fluorescence to get dynamical information during outburst. Discover new hybernant BHs using appropriate colour cuts in deep surveys such as IPHAS/UVEX.

XTE J 1752 -223 DECAY TO QUIESCENCE: Aug 2010 WHT i>18. 9 OUTBURST: 26 Oct 2009 IMACS i=15 -16 Torres et al. ATel#2263 Russell et al. ATel #2775 Corral-Santana et al. ATel #2818 Faint quiescent counterparts in crowded fields hamper radial velocity studies of donor star

A low mass BH in LS 5039 ? • First γ-ray Binary detected by HESS at Te. V with Lγ=1034 ergs/s • Spacially resolved radio emission Microquasar or pulsar wind? • Radial velocity curve of the 06. 5 V((f)) indicate Porb=3. 906 d (Casares et al. 2005). But the orbital parameters depend on the spectral lines used. HI and He. I may be contaminated by wind emission because show larger scatter and the solution is blueshiftted (as expected by contamination from P-Cygni) We adopted the He. II solution

1) We fitted UNIFIED MODELS (NLTE+Sphericity) to stellar spectrum to yield Teff, log g, M, R and V∞ 2) Donor almost fills Roche lobe at periastron assume is pseudo-synchronized i = 23. 2 -- 26. 8 o Combining 1+2+orbital solution we get a full solution to the masses. The result suggests that LS 5039 might be a black hole with 3 -5 solar masses

BH Candidates TOTAL = 49 BH transients in 44 years, 17 with dynamical evidence

The K-correction Measures the displacement of emission line region wrt center of mass Kem/K 2= 1 - f (1+q) 0 f with f ( ) 1 ( = disc opening angle) Muñoz-Darias et al. 2005 f=0 1 f=1 Real K 2 can be determined with extra information on q and .

XTE/VLT campaign on burster 4 U 1636 -536 June 2007 3 bursts detected at orbital phases 0. 20, 0. 55 and 0. 83 Time lag varies with amount of continuum subraction, approaching prediction for model with Mns=1. 4 Msun, q=0. 3, i=36 -60, α=12 deg χ2 fits to the 3 burst delays as function of continuum subtraction yields i=45 -50 deg and optimum subtraction ~0. 85 Muñoz-Darias, O’Brien et al. , in preparation