43
Observations of Binaries in Globular Clusters Adrienne Cool San Francisco State University

Observations of Binaries in Globular Clusters

Embed Size (px)

DESCRIPTION

Observations of Binaries in Globular Clusters. Adrienne Cool San Francisco State University. Primordial. Observations of ^ Binaries in Globular Clusters. Adrienne Cool San Francisco State University. OUTLINE. Why bother? What do we want to know? How can we find out? - PowerPoint PPT Presentation

Citation preview

Page 1: Observations of Binaries in Globular Clusters

Observations of Binaries in Globular Clusters

Adrienne Cool San Francisco State University

Page 2: Observations of Binaries in Globular Clusters

Observations of ^ Binaries in Globular Clusters

Adrienne Cool San Francisco State University

Primordial

Page 3: Observations of Binaries in Globular Clusters

OUTLINE

• Why bother?

• What do we want to know?

• How can we find out?

• What’s new?

• What’s next?

Page 4: Observations of Binaries in Globular Clusters

Why bother?

• Binary fraction is a fundamental parameter

• Primordial binaries play a key role in cluster dynamics

• Primordial binaries are implicated in the formation of many more exotic populations

Page 5: Observations of Binaries in Globular Clusters

What do we want to know?

• What fraction of globular cluster stars are binaries? • Do clusters have different binary fractions (fb)? Any correlation with particular cluster parameters?

• Are there signs of dynamical evolution of binaries?

• How are the binaries distributed in…• period (Porb)• eccentricity (e)• mass ratio (q m2/m1)• primary mass

Page 6: Observations of Binaries in Globular Clusters

How can we find out?

radial velocity variables

photometric variables

outliers incolor-magnitude diagrams

3 methods used so far…

Pryor / Hut et al. 1992 Kaluzny et al. 1999 Rubenstein & Bailyn 1997

Page 7: Observations of Binaries in Globular Clusters

Method 1: Radial velocity variables

• ground-based spectroscopy• samples of ~30-300 giants• ~2-4 velocities per star• accuracies ~0.6–3 km/s• baselines ~ 1-20 years

Cote et al. 1996 – M22

sensitive to binaries withPorb ~ days – years

fb estimates depend oneccentricity distribution

Page 8: Observations of Binaries in Globular Clusters

cluster MV c bin/total fb Porb reference47 Tuc, M2, M3

M12, M13, M71–9.4 –

–5.6

2.03 –

1.15

6 / 393 ~ 5 % (circ)

~ 12 % (therm)

0.2 – 20 yr Pryor et al.

1989

NGC 3201 –7.5 1.30 2 / 276 6 – 10 % (circ)

15 – 18 % (therm)

0.1 – 10 yr Cote et al.

1994

NGC 5053 –6.7 0.84 6 / 66 ~ 20 – 30% 3 day – 10 yr Yan & Cohen

1996

M71 –5.6 1.15 12 / 121 10 – 30 % (circ)

15 – 40 % (therm)

3 day – 10 yr Barden et al.

1996

M22 –8.5 1.31 0–1 / 109 1+10–1 % (circ)

3+16–3 % (therm)

0.2 – 40 yr Cote et al.

1996

M4 –7.2 1.59 0 – 2 / 33 15 15% @ > 4 rc 2 day – 3 yr Cote & Fischer 1996

Cen –10.3 1.61 ? / 310 3 – 4% 0.5 – 10 yr Mayor et al.

1996

Pal 5 –5.2 0.70 -- / 17 40 20%

(68% conf)

a < 50 A.U. Odenkirchen et al. 2002

searches for radial velocity variables

Page 9: Observations of Binaries in Globular Clusters

Method 2: photometric variables

Albrow et al. 2001 – 47 Tuc

• HST & ground-based imaging• ~ 2000 – 40,000 stars sampled• 250 – 1300 images• baselines ~ 1 week – 1 month

sensitive to binaries withPorb ~ 0.1 day – few days

fb estimates depend on assumedPorb, e, q distributions

Page 10: Observations of Binaries in Globular Clusters

cluster MV c bin/total fb Porb reference

M71 –5.6 1.15 5 / 5300

~ 0.09 %

4 contact

10 – 72 %

Porb < 800 yr

0 – 5 rc

hours - days Yan & Mateo 1994

M5 –8.8 1.83 6 / 3600

~ 0.17 %

5 contact

22 – 39 %

2.5 day – 550 yr

4 – 8 rc

hours - days Yan & Reid 1996

47 Tuc –9.4 2.03 26 / 46,422

~ 0.06 %

11 detached

16 contact

13 6 %

(5 w/Porb> 4 day)

0 – 4 rc

0.3 hr – 15 day Albrow et al. 2001

47 Tuc –9.4 2.03 28 / 43,067

~ 0.07 %

7 detached

21 contact

--

15 rc – 0.6 rt

2 hr – 60 day Weldrake et al. 2004

M4 –7.2 1.59 0 – 1 / 2102

0.05 %

--

@ 6 rc

4 hr – 6 day Ferdman et al. 2004

searches for eclipsing binaries – selected results

Page 11: Observations of Binaries in Globular Clusters

Method 3: Outliers in color-magnitude diagrams

Cool & Bolton 2002 – NGC 6397

• HST & gnd-based imaging• 100s – 1000s of stars• 2 filters is enough• no repeat measure required• high photometric accuracy

sensitive to binaries with… any Porb, e, inclination!

fb estimates depend on assumed q distribution, F(q)

Page 12: Observations of Binaries in Globular Clusters

cluster MV c fb note ref.

M92 –8.2 1.81 < 9 % @ 10 – 30 rc all q Romani & Weinberg 1991

M30 –7.3 2.50

pcc

< 4 % @ 25 – 65 rc all q Romani & Weinberg 1991

NGC 288 –6.7 0.96 ~ 10 % @ 0 – 6 rc lower limit

q 0.7

Bolte

1992

E 3 –2.8 0.75 29 9 % @ 0 – 2 rc

lower limit?

q 0.7 ?

Veronesi et al.

1996

NGC 2808 –9.4 1.77 24 4 % @ 24 rc lower limit

q 0.8 ?

Ferraro et al.

1998

M30 –7.3 2.50

pcc

5 2 % @ 100 rc lower limit

q 0.7-0.8 ?

Alcaino et al.

1998

searches in color-magnitude diagrams – selected early results

Page 13: Observations of Binaries in Globular Clusters

cluster MV c fb note ref.

NGC 6752 –7.7 2.50

pcc

15 – 38 % [99.7% conf.] @ < 1 rc

< 16 % [99.7 % conf.] @ > 1 rc

Monte-Carlo

various F(q)

Rubenstein & Bailyn 1997

NGC 6397 –6.6 2.50

pcc

3 % (q 0.45) @ 0 – 18 rc

5 – 7 % (all q) @ 0 – 18 rc

flat F(q)

extrapolated

Cool & Bolton

2002

NGC 288 –6.7 0.96 10–20 % [8–38% @ 99.9%] @ < 1 rh

0 – 10 % [<10% @ 99.9%] @ 1-2 rh

Monte-Carlo

various F(q)

Bellazzini et al. 2002

Pal 5 –5.2 0.70 9 1 % @ < 1 rc (Poisson errors)

9 8 % @ 2 – 3 rc

lower limit

(high q only)

Koch et al.

2004

Pal 13 –3.7 1.31 30 4 % (Poisson errors) @ 0-18 rc

lower limit

(high q only)

Clark et al. 2004

M4 –7.2 1.59 ~ 2 % @ < 1 rc

~ 1 % @ > 1 rc

lower limit

(high q only)

Richer et al.

2004

M3 –8.9 1.84 6 – 22 % [3–35% @95%] @ < 1 rc

1 – 3 % [0–22% @ 95%] @ 1-2 rc

Monte-Carlo

various F(q)

Zhao & Bailyn

2005

searches in color-magnitude diagrams -- more recent results

Page 14: Observations of Binaries in Globular Clusters

What fraction fb of globular cluster stars are binaries?

Does fb differ among clusters? Range?

Do any clusters have fb = 0% ? 100% ??

Page 15: Observations of Binaries in Globular Clusters

cluster MV c bin/total fb Porb fb (3 decades)

Cen –10.3 1.61 ? / 310 3 – 4% 0.5 – 10 yr 7 – 9 %

M22 –8.5 1.31 0–1 / 109 1+10–1 % (circ)

3+16–3 % (therm)

0.2 – 40 yr 1.3+13–1.3 % (circ)

4+21–4 % (therm)

NGC 3201 –7.5 1.30 2 / 276 6 – 10 % (circ)

15 – 18 % (therm)

0.1 – 10 yr 9 – 15 % (circ)

23 –27 % (therm)

M4 –7.2 1.59 0 – 2 / 33 15 15% @ > 4 rc 2 day – 3 yr 16 16% @ > 4 rc

NGC 5053 –6.7 0.84 6 / 66 ~ 20 – 30% 3 day – 10 yr ~ 20 – 30%

M71 –5.6 1.15 12 / 121 10 – 30 % (circ)

15 – 40 % (therm)

3 day – 10 yr 10 – 30 % (circ)

15 – 40 % (therm)

radial velocity variables – global binary fractions

• typical fb ~ 15% (~5% per decade)

• trend toward lower fb for massive clusters (high , low Pcrit)

disruption of soft binaries? Cote et al. 1996

Page 16: Observations of Binaries in Globular Clusters

Odenkirchen et al. 2002 – Pal 5

• MV = –5.2

• c = 0.70

• single epoch, VLT

• 17 cluster stars

• accuracy ~ 0.15 km/s

broad pedestal under narrow peak = binaries?

fb = 40 20 %

Pal 5

Page 17: Observations of Binaries in Globular Clusters

photometric variables – global binary fractions

cluster MV c bin/total Porb fb Porb assumed ref

47 Tuc –9.4 2.03 26 / 46,422

~ 0.06 %

11 detached

16 contact

0.1 hr –

15 day

13 6 %

(5 w/Porb> 4 day)

0 – 4 rc

2.5 d – 50 yr A01

M5 –8.8 1.83 6 / 3600

~ 0.17 %

5 contact

hours –

days

22 – 39 %

4 – 8 rc

2.5 day – 550 yr YR96

M71 –5.6 1.15 5 / 5300

~ 0.09 %

4 contact

hours –

days

10 – 72 %

0 – 5 rc

< 800 yr YM94

47 Tuc: fb ~ 2 – 5 % per decade for hard binaries

consistent with vrad results

Page 18: Observations of Binaries in Globular Clusters

Albrow et al. 2001 – 47 Tuc

Page 19: Observations of Binaries in Globular Clusters

CMD outliers – binary fractions – “all q” subset

cluster MV c fb note ref.

M3 –8.9 1.84 6 – 22 % [3–35% @ 95%] @ < 1 rc

1 – 3 % [0–22% @ 95%] @ 1-2 rc

Monte-Carlo

various F(q)

Zhao & Bailyn

2005

M92 –8.2 1.81 < 9 % @ 10 – 30 rc max likelihood

Romani & Weinberg 1991

NGC 288 –6.7 0.96 10–20 % [8–38% @ 99.9%] @ < 1 rh

0 – 10 % [<10% @ 99.9%] @ 1-2 rh

Monte-Carlo

various F(q)

Bellazzini et al. 2002

NGC 6752 –7.7 2.50

pcc

15 – 38 % [99.7% conf.] @ < 1 rc

< 16 % [99.7 % conf.] @ > 1 rc

Monte-Carlo

various F(q)

Rubenstein & Bailyn 1997

M30 –7.3 2.50

pcc

< 4 % @ 25 – 65 rc max likelihood

Romani & Weinberg 1991

NGC 6397 –6.6 2.50

pcc

3 % (q 0.45) @ 0 – 18 rc

5 – 7 % (all q) @ 0 – 18 rc

flat F(q)

extrapolated

Cool & Bolton

2002

Page 20: Observations of Binaries in Globular Clusters

Zhao & Bailyn 2005 – M3

Page 21: Observations of Binaries in Globular Clusters

Zhao & Bailyn 2005 – M3

Page 22: Observations of Binaries in Globular Clusters

Zhao & Bailyn 2005 – M3

q = 1

more high q

~ flat

more low q

mass ratiodistribution F(q)

core 1 – 2 rc

Page 23: Observations of Binaries in Globular Clusters

CMD outliers – binary fractions – “all q” subset

cluster MV c fb note ref.

M3 –8.9 1.84 6 – 22 % [3–35% @ 95%] @ < 1 rc

1 – 3 % [0–22% @ 95%] @ 1-2 rc

Monte-Carlo

various F(q)

Zhao & Bailyn 2005

NGC 288 –6.7 0.96 10–20 % [8–38% @ 99.9%] @ < 1 rh

0 – 10 % [<10% @ 99.9%] @ 1-2 rh

Monte-Carlo

various F(q)

Bellazzini et al. 2002

NGC 6397 –6.6 2.50

pcc

3 % (q 0.45) @ 0 – 18 rc

5 – 7 % (all q) @ 0 – 18 rc

flat F(q)

extrapolated

Cool & Bolton 2002

NGC 6752 –7.7 2.50

pcc

15 – 38 % [99.7% conf.] @ < 1 rc

< 16 % [99.7 % conf.] @ > 1 rc

Monte-Carlo

various F(q)

Rubenstein & Bailyn 1997

• all Porb, all e, any inclination… 4 - 5 decades in Porb

why not higher fb ?

• maybe M3 is okay, but NGC 288?

• how come a post-collapse cluster has such a high fb??

Page 24: Observations of Binaries in Globular Clusters

CMD outliers – binary fractions – high q subset

cluster MV c fb note ref.

NGC 2808 –9.4 1.77 24 4 % @ 24 rc q 0.8 ? Ferraro et al. 1998

M30 –7.3 2.50 5 2 % @ 100 rc q 0.7-0.8 ? Alcaino et al. 1998

M4 –7.2 1.59 ~ 1 – 2 % 0 – 4 rc high q Richer et al. 2004

NGC 288 –6.7 0.96 ~ 10 % @ 0 – 6 rc q 0.7 Bolte 1992

Pal 5 –5.2 0.70 ~ 9 % @ 0 – 3 rc high q Koch et al. 2004

Pal 13 –3.7 1.31 30 4 % @ 0 – 18 rc high q Clark et al. 2004

E 3 –2.8 0.75 29 9 % @ 0 – 2 rc q 0.7 ? Veronesi et al. 1996

• 3 cases with fb ~ 25 – 30 % … extrapolate to all q (×3??) 75 – 90 % ?!

• NGC 2808 so high even far outside core ?

• Pal 5 with tidal stripping… why not higher?

• M4… why so low?

Page 25: Observations of Binaries in Globular Clusters

Koch et al. 2004

Odenkirchen et al. 2002

Pal 5 – compare 2 methods

CMDs: fb ~ 9 1% vrad: fb ~ 40 20 %

Page 26: Observations of Binaries in Globular Clusters

Richer et al. 2004

M4 – compare 3 methods

CMDs: fb ~ 1–2 % (high q)

vrad: fb ~ 15 15 % Cote et al. 1996

variables: fb (observed) < 0.05%

(similar to 47 Tuc w/fb ~ 13%)

Ferdman et al. 2004

Page 27: Observations of Binaries in Globular Clusters

Hut et al. 1992

effect of photometric errors on perceived binary fraction in CMDs

fb = 10%

Page 28: Observations of Binaries in Globular Clusters

Primordial binary fraction in globular clusters

• all GCs studied have at least some binaries

• not all GCs have same binary fraction… at present low end: < 5 – 7 % ? (NGC 6397) high end: ~ 30% ~ 90% for all q ?? (E3, Pal 13)

• fb = 100% is not ruled out for some poor clusters

• fb = 0% is possible in outskirts of some clusters

• trend toward higher fb for poorer clusters, with exceptions

Page 29: Observations of Binaries in Globular Clusters

Cool & Bolton 2002 – NGC 6397Clark et al. 2004 – Pal 13

Page 30: Observations of Binaries in Globular Clusters

effect of photometric errors on perceived binary fraction in CMDs

fb = 100%

Hut et al. 1992

Page 31: Observations of Binaries in Globular Clusters

Are there signs of dynamical evolution of binaries?

Page 32: Observations of Binaries in Globular Clusters

Are there signs of dynamical evolution of binaries?

• trends toward lower fb for higher mass clusters consistent with destruction of binaries beyond hard/soft boundary

… or are fb values in loose clusters justenhanced by tidal stripping?

• low fb in NGC 6397 and M30 destruction in collapsed cores?

… but what about NGC 6752??

Page 33: Observations of Binaries in Globular Clusters

Bellazzini et al. 2002 – NGC 288

Page 34: Observations of Binaries in Globular Clusters

Bellazzini et al. 2002 – NGC 288

Page 35: Observations of Binaries in Globular Clusters

Bellazzini et al. 2002 – NGC 288

mass segregation in NGC 288

Page 36: Observations of Binaries in Globular Clusters

Albrow et al. 2001 – 47 Tuc

mass segregation of binaries in 47 Tuc

71 BY Dra stars in 47 Tuc! an untapped resource

Page 37: Observations of Binaries in Globular Clusters

Weldrake et al. 2004 – 47 Tuc

period segregation of eclipsing binaries in 47 Tuc

contact all stars detached

Page 38: Observations of Binaries in Globular Clusters

• segregation of binaries by mass is observed

• 47 Tuc: shorter period binaries are more centrally concentrated than long period binaries

mass effect?? binary hardening?

More signs of dynamical evolution…

Page 39: Observations of Binaries in Globular Clusters

What about binary parameters?

• Radial velocities can give Porb, e and more… long-term tracking of candidates required

• Eclipsing binaries beginning to give information on Porb

• CMDs in principle can give information on q = m1/m2

Page 40: Observations of Binaries in Globular Clusters

Cool & Bolton 2002 – NGC 6397

Page 41: Observations of Binaries in Globular Clusters

Cool & Bolton 2002 – NGC 6397

Page 42: Observations of Binaries in Globular Clusters

What’s next?

• better constraints on binary fraction and distribution

* vrad – need larger samples! Fabry-Perot underway

* eclipsing – large samples are proven to work

* CMDs – exploit highest possible photometric accuracy

– look for MS-WD pairs too?

• contraints on binary parameters? track candidates!

* vrad – already done for some

* CMDs – spectroscopy on MS-MS binaries? – BY Dra stars: more complete sample?? • HST very valuable, especially in crowded cluster cores• ground-based work equally powerful in sparse clusters (e.g. Pal 13) or outskirts (e.g. 47 Tuc)

Page 43: Observations of Binaries in Globular Clusters

How many “primordial” binaries are really primordial??