like this field blue stragglers and related mass transfer issues george preston, eso, santiago, 2012

like this

FIELD BLUE STRAGGLERS AND

RELATED MASS TRANSFER ISSUES

George Preston, ESO, Santiago, 2012

or this

We stand on the shoulders of giants who pioneered stellar structure and evolution in interacting binaries

my giants

G. P. Kuiper

O. StruveZ. Kopal

R. Kippenhahn

B. Paczynski

F. Hoyle

H. Bondi

W. H. McCrea

L. Lucy

my giants

G. P. Kuiper

O. StruveZ. Kopal

R. Kippenhahn

B. Paczynski

F. Hoyle

H. Bondi

W. H. McCrea

L. Lucy

mentor

mentor

friendfriend

G. P. Kuiper

O. StruveZ. Kopal

R. Kippenhahn

B. Paczynski

F. Hoyle

H. Bondi

W. H. McCrea

DetachedSemi-detached

Contact

gas streamsBeta Lyrae

Bondi-Hoyle accretion

Hoyle-Bondi accretion

The BS mass-transfer

model

CasesAB&C

Roche lobe overflow

Mass transfer

The Algol paradox

& more L. Lucy

The mostcamera-shy

giant known

my giants

M3

M3, where it all started

In a dense stellar environment

Allan Sandage PhD thesis

unlike my FBS domain

.

.

mostlyempty space

.

.

and here

like here

etc.

While reading papers on the subject of my talk, I was surprised to learn …

how often first-class astronomers ignore each other’s work!

Thus,Sandage (AJ 1953) first identified blue stragglers unambiguosly in M3

McCrea invented an explanation for Sandage’s blue stragglers 11 years later (MNRAS 1964) with no reference to Sandage.

Böhm-Vitense (ApJ 1980) confirmed operation of McCrea’s process 16 years later in the ζ Capricorni system with no reference to eitherSandage or McCrea.

McClure (ApJ 1984) brilliantly generalized Böhm-Vitense’s result 4 years later with no reference to Sandage or McCrea or Böhm-Vitense.

how often first-class astronomers ignore each other’s work!

Thus,Sandage (AJ 1953) first identified blue stragglers unambiguosly in M3

McCrea invented an explanation for Sandage’s blue stragglers 11 years later (MNRAS 1964) with no reference to Sandage.

Böhm-Vitense (ApJ 1980) confirmed operation of McCrea’s process 16 years later in the ζ Capricorni system with no reference to Sandage or McCrea.

McClure (ApJ 1984) brilliantly generalized Böhm-Vitense’s result 4 years later with no reference to Sandage or McCrea or Böhm-Vitense.

Three decades of ignoring!

WTF!

Finally, in 1989 Peter Leonard set the stage for this conferenceAJ 98, 217

HOW TO IDENTIFY FBS

Of necessity, in the field we first identified metal-poor FBS by colorimetry.

Preston et al. 1994

BM

P st

ars

HK Survey: Beers et al. 1985, 1992

MS [Fe/H] = 0

MS [Fe/H = 1

Metal poor stars near GC turnoff

BHB

W. W. Morgan would have calledthe BMP stars a“natural group”.

Of necessity, in the field we first identified metal-poor FBS by colorimetry.

Any photometricsystem with a

uv filter will work.

RHB

Preston et al. 1994

BM

P st

ars

HK Survey: Beers et al. 1985, 1992

MS [Fe/H] = 0

MS [Fe/H = 1

Metal poor stars near GC turnoff

BHB

W. W. Morgan would have calledthe BMP stars a“natural group”.

Of necessity, in the field we first identified metal-poor FBS by colorimetry.

Any photometricsystem with a

uv filter will work.

RHB

FBS area subset of BMP

In BMP domain isochrones with a wide range

of ages and metallicities overlap in a tangled mess.

Preston & Sneden 2000

Isochrones of various [Fe/H] values and ages overlap in a 2-color diagram of the BS domain.

RYI IsochronesGreen et al. (1987)

main sequence isochrones

subgiant isochrones

Turnoffs for: [Fe/H] = 2.2 ages 3 7,10 Gy

Hence, “straggle”

The cool (red) edge of the BS domain in any stellar system is defined by stars

that are not members of the domain

STRANGE DEFINITION

Mandushev, Fahlman, Richer 1997, AJ

MOST FBS ARE A SUBSET OF A LARGER FAMILY OF MAIN SEQUENCE MASS TRANSFER BINARIES

Mandushev, Fahlman, Richer 1997, AJ

Use of MSTO coloras a boundary

obscures this reality

12 Gy isochrone

surely FBS

CS 22949-008pri. & sec.

CASE IN POINTMetal-poor carbon stars below MSTO in hierarchical triples

Masseron et al. 2012, ApJ, 751:14

mass transfer sequence

CS 22964-161

= CEMP, literature

Various tools have been devised to isolate field blue stragglers.Pier (1983) pioneered the identification of FBSs.

FBS

BHBD(0.2)

0.2

Various tools have been devised to isolate field blue stragglers.Sersic (b,c) Balmer parameters do this particularly well.

They do not require knowledge of photometric colors.

FBS

BHB

Clewley et al. 2002

Various tools have been devised to isolate field blue stragglers.Sersic (b,c) Balmer parameters do this particularly well.

They do not require knowledge of photometric colors.

FBS

BHB

Clewley et al. 2002

An application of Sersic parameters to Sloan data in the distant halo

(n = 4985)

(n = many)

Xue et al. 2011

Sarajedini 1993, ASP Conf. Ser.

Globular clusters provide luminosity calibrationfor Galactic structure applications, e.g. halo (R)

MV) ~ 0.5 mag/star(distance) ~ 3 %/100 stars

Sarajedini 1993, ASP Conf. Ser.

Globular clusters provide luminosity calibrationfor Galactic structure applications, e.g. halo (R)

GROUP PROPERTIES OF FBS

It is easy to find binaries among FBS candidates

RV-constant stars Binary stars

A child can do it.

JD - 2400000

A high % of FBS are members of spectroscopic binaries

Preston & Sneden 2000

4 km/s

20 km/s

vertical scales are not uniform

Very high

Preston & Sneden 2000, AJ

A high % of FBS are members of spectroscopic binaries

But a low % of FBS are in double-lined spectroscopic binaries

Very low

≤

The one DLSB, CS 22873-139, included here is contested by Spite et al. 2000, A&A, 360, 1077

Preston & Sneden 2000, AJ

FBS have longish orbital periods & a high % of small orbital eccentricities reminiscent of their carbon-star cousins

deficit of short periods excess of low

eccentricitiesat P > 100 d

Normal MS binaries disk = xhalo = o

BMPblue metal-poor

C, s-process rich

Preston & Sneden 2000, AJ

and their mass functions suggest companions of lower-than-normal mass (like maybe white dwarfs?).

f1 = K13 P / (2 π G) = M2

3 (sin3 i) / (M1 + M2)2

FBS Ba, CH cousins

Preston & Sneden 2000, AJ

All of the preceding:High binary fractionDeficit of short periodsLow orbital eccentricitiesSmall mass functionsNo visible secondaries *And Thank You, Erika, for Capricorni

tell us that FBS are a species sui generis

* Hierarchical triples excepted

white dwarfflux red wing of L

in white dwarf

All of the preceding:High binary fractionDeficit of short periodsLow orbital eccentricitiesSmall mass functionsFew (no?) visible secondariesAnd Thank You, Erika, for Capricorni

tell us that FBS are a species sui generis

My perspective:Wide binary disruption is main reason

for the specific frequency deficitIn GCs relative to the Galactic field

4.0

FBS

FBS

Specific frequency of FBS appears to be the upper bound of a sequence defined by OCs (DeMarchi et al 2006) and GCs (Piotto et al 2004)

This is a lo

garithm

This isn’t

Something like “concentration” must be what matters

Preston & Sneden 2000, AJ

visualbinaries

c.p.m.binaries{

radial velocitybinaries P<4000 d

This 13% of radial velocity binaries with P 5 d merge in less than a Hubble time (Vilhu, 1982, A&A, 109, 17). Hence, the deficit of short period binaries in the field.

These more frequent wide binaries are largely disrupted in GCs. Hence, the relatively low specific frequency of blue stragglers in globular clusters.

This interpretation follows fromthe presumption that the

Duquennoy-Mayor (1991, A&A)period distribution is universal.

Ockham’s Razor, etc

William of Ockham1248-1307

DUSTING

How do you hide it?

DUST: A thin layer that you notice on tables

just before your guests arrive

In the house use this

DUST: A thin layer that you notice on tables

just before your guests arrive

In stars redistribute by thermohaline mixing.

DUST: A thin layer that you notice on tables

just before your guests arrive

Reconciliation of theory and observation

Observational facts/issues that can be improved:

Numbers, orbital parameters of MS and post-MS C and Ba starsDetection criteriaDetection threshholds, bias, completenessObserver persistence (RV)

Theory:

Mass transfer by winds Including stuff like Davies-Pringle (1980) paradox

AGB theory and practice from Busso & Gallino andConvective envelopesGravitational settling and stabilizing molecular gradientsDilution with and without thermohaline mixing

those

those

all

other Italians

those

all

Do all MS CEMP survive the RGB ascent?

INNOCENT QUESTION:

Is the observed density ratio (RGB)/(MS) OK?

MS CEMP(RGB)

CEMP(MS)

(kpc-3) = space densityn = nuclear time scale

relative volumes searched (in apparent magnitude limited surveys)

RGB

IF LRGB /LMS ~ 40

Volume ratio ~ 250 andn(RGB)/n(MS) ~ 102

THENVol* n ~ 2.5

those

those

all

Stars with “no-dilution” solutions

dilution = log MCE/MAGBacc

x

dilutions from Bisterzo et al. 2012, MNRAS, 422, 849

All the stars with “no-dilution” solutionslie near MSTO; ~ half are FBS)

MSTO

RGB(CEMP)/MS(CEMP) ~ 2.5 in Bisterzo et al. sample. I was a bit surprised.

dilu

tio

n

pure AGB

We have very little information about wind accretion at large distances from an AGB donor.

[s]

Regression below is anchored by two long-period SLSBswith primaries originally classified as ordinary GK giants!

Griffin 1985

G8 III

K0 III

[s]

Regression below is anchored by two long-period SLSBswith primaries originally classified as ordinary GK giants!

Griffin 1985

G8 III

K0 III

Want more candidates? See McWilliam 1990, ApJS, 74

23/671= 4% of GK giantsV(mag) 6.0, [s] > 0.2 Expect 800 more in HD (V<11)

[s]

A bunch of calculations for MS stars with various

ages and [Fe/H]

Pinsonneault, DePoy, & Coffee 2001, ApJ

FBS convection zone rapidly disappears at Teff > 6500 K

He, Sr, Ba, sink in M2 prior to AGB evolution of M1.

Negative –gradient inhibits thermohaline mixingWhat happens when Ba-rich long-P FBS ascends RGB?

Will we recognize it, or will it disappear?

7000

-3.5

-4.0

Humoris still alive in

the USA!

FBS live here

Jorissen & Van Eck 2001

F-G binaries ofDuquennoy & Mayor 1991

VB

CPM

The period distribution of Ba stars declines at P > 2000 d and dies at P ~ 5000 d, but the period distribution of MS binaries peaks at P > 10000 d.

Extant data provide a clue.

Jorissen & Van Eck 2001

F-G binaries ofDuquennoy & Mayor 1991

VB

CPM

The period distribution of Ba stars declines at P > 2000 d and dies at P ~ 5000 d, but the period distribution of MS binaries peaks at P > 10000 d.

Extant data provide a clue.

Orbital speeds at P = 10000 days are easily measured by modern spectrographs.

An unavoidable conclusion: winds “run out of gas” at some long P. Which?

2 3 4 5 6

log P(days)

0

5

10

15

20

K1 (k

m/s

)

m1 = 0.8, m2=0.5

K13P = constant

Newton’s laws are very tolerant of long orbital periods.

VB CPMRV

THOSE YOUNG A-TYPE STARS ABOVE THE PLANE:

ANOTHER INCONVENIENT TRUTH

This all began with Perry and Rogers, Harding & Sadler

Perry 1969 Rogers, Harding & Sadler 1981

I circled and boxeddata for

metal-rich A starsthat stray far from

Galactic plane

AVG = 5250 AVG = -0.23

This all began with Perry and Rogers, Harding & Sadler

Perry 1969 Rogers, Harding & Sadler 1981

I circled and boxeddata for

metal-rich A starsthat stray far from

Galactic plane

AVG = 5250 AVG = -0.23

Santos et al. 2009, A&AThe average [Fe/H] for these A-Stars 5 kpc from the Galactic plane is only slightly lower than average value near the sun.

Density distributions of early A-type stars far from the Galactic planehave a two-slope signature

Perry 1969 Lance 1988

north south

Perry 1969 Lance 1988

The Lance (1988) stars below the Galactic plane are hotter than the BMP stars studied by Preston & Sneden (2000).

north south

Perry 1969 Lance 1988

isochrone fort = 5.0E+8 y

They are apparently quite youngIf they are blue stragglers, what do they straggle behind?

Perry 1969 Lance 1988

isochrone fort = 5.0E+8 y

They are apparently quite youngIf they are blue stragglers, what do they straggle behind?

Lance hypothesis: star formation during recent interaction

of a metal-poor, gas-rich satellite with the disk of the Milky Way.

CHEMICAL ABUNDANCES ABOVE AND BELOW

THE MSTO

Deep mixing accompanying merger, collision, mass transferburns 7Li to 4He ash low Li is a straggler signature.

BUT confirmation requires a big telescope for most FBS.

Data from Thorburn 1994, ApJ, 421

data from P&S 2000, AJ

Galactic FBS

dSph red giants

halo stars

Compare dSph red giants to Galactic stars – Venn et al. 2004, AJ

thick disk

thin disk

What we might learn from the -elements.

What we might learn from the -elements.

data from P&S 2000, AJ

Galactic FBS

dSph red giants

halo stars

Compare dSph red giants to Galactic stars – Venn et al. 2004, AJthick disk

thin disk

RV-constant BMP stars: Intermediate-Age captures

from a dSph satellite like Carina?

Stetson et al. 2011, ESO Messenger

13 Gy

8 Gy

CH

The “non-variable” spectrum is the

mean of three stars

The FBS binary CS 29497-030 illustrates how the carbon signature responds to an increase in Teff

C I

Enter the AGB: Carbon and n-capture elements

G-band becomes inconspicuous

Atomic C becomesbetter signature

Sneden, Preston, & Cowan 2003, ApJ

Note similarities of Fe-peak lines

versus disparities of the Ba and Pb lines

The “non-variable” spectrum is the mean

of three stars

CS 29497-030 = FBS (B-V)0 = 0.30

CS 29497-030 = FBS (B-V)0 = 0.30

ELABORATION

Sneden, Preston, & Cowan 2003, ApJ

Lead-rich binary CS 29497-030

The “non-variable” spectrum is the mean of

three stars

Question:Is this a FBS?(B-V)0 = +0.30(U-B)0 = - 0.14[Fe/H] = - 2.6

CS 29497-030 = FBS (B-V)0 = 0.30

CS 29497-030 = FBS (B-V)0 = 0.30

MSTO color vs [Fe/H]

Every FBS has its own red color boundary

Lead-rich binary CS 29497-030

Answer: Yes

One solution to the red color boundary problem for FBS:Use [Fe/H] to assign MSTO boundary star-by-star

Question:Is this a FBS?(B-V)0 = +0.40(U-B)0 = - 0.14[Fe/H] = - 2.2

Answer: No

MSTO color vs [Fe/H]

neutron/seed ratio decreases

X = IA pop = BS pop

x

Orbit dimensions, RGB, AGB mass transferFBS binaries ( ): n-capture normal abundances (low [C/N])

mass transfer only during RGB evolution

FBS binaries ( ): n-capture over-abundances (low [C/N])mass transfer during RGB & AGB evolution

FBS single (X): n-capture normal abundances (high [C/N])

FBS PULSATORS

Pulsation like this would produce chaos in FBS binary RV curves,so we know that such pulsators are rare among FBS binaries.

30 km

/s

1 hour

ORBIT

PULSATION

ΔV ~ 0.1 mag.

Shift due toorbital motion

Preston & Landolt 1999 AJ

30 km

/s

1 hour

Conclusion: Such pulsation is rare among metal-poor FBS(like 1/42 = 2 or 3 %) although most FBS lie in instability strip.

ORBIT

PULSATION

20 km/s

Preston & Landolt 1999 AJ

HADS confined to narrow band in the instability strip.McNamara (1997), Petersen & Dalsgaard (1999)Kepler data confirm (Balona & Dziembowski 2011)Pulsator fraction: ~ 0.5 in central band < 0.5 elsewhere in instability strip

Sct instability strip

McN

amar

a’s

field

sta

rs

Poretti et al. 2008, ApJ

RR Lyraes fill their instability strip, but SX Phe stars in Fornax II probably are confined to McNamara’s HADS strip.

RR Lyraes fill their instability strip, but SX Phe stars in Fornax II probably are confined to McNamara’s HADS strip.

Poretti et al. 2008, ApJ

RR Lyr

SX Phe

STARS THAT DON’T FIT

2500 3000 3500 4000 4500 5000 5500

Julian Date

-60

-50

-40

-30

-20

-10

0

10

20

30

RV

(km

/s)

TY Gru 2003 - 2010

C & s-process rich stars that don’t fitC & s-process rich stars that don’t fit

1 km/s

20 years

2500 3000 3500 4000 4500 5000 5500

Julian Date

-60

-50

-40

-30

-20

-10

0

10

20

30

RV

(km

/s)

TY Gru 2003 - 2010

C & s-process rich stars that don’t fitC & s-process rich stars that don’t fit

After so many years of this– NADA!

1 km/s

2500 3000 3500 4000 4500 5000 5500

Julian Date

-60

-50

-40

-30

-20

-10

0

10

20

30

RV

(km

/s)

TY Gru 2003 - 2010

C & s-process rich stars that don’t fitC & s-process rich stars that don’t fit

After so many years of this– NADA!

1 km/s

Tengo 82 anos – no mas para mi. I quit!