Common Envelope Evolution through Planetary Nebula Eyes

1
Common Envelope Evolutionthrough Planetary
Nebula Eyes Orsola De Marco American Museum of
Natural History
Collaborators H.E. Bond, M. Moe, M.-M. Mac Low,
E. Sandquist, F. Herwig, R. Taam
Merging binaries. Simulations UKAFF
2
Outline

• The common envelope (CE) interaction and its
  progeny populations.
• The CSPN RV survey can PN be by and large a CE
  phenomenon?
• CE simulations
• the determination of the CE efficiency parameter
  (?? and  other parameters.
• post-CE populations (sdB and CSPN) to constrain
  CE simulations.
• Further/future simulations common envelope
  mergers.

3
Common Envelope A twice-in-a-lifetime
opportunity
R
AGB R 500-1500 Ro
R
RGB R 100-300 Ro
4
Common envelope
Unstable Roche Lobe overflow
5
Depending on the efficiency of the energy
transfer from the companion to the CE (?), one
might get
A short-period binary, or a merged star
6
The existence of a CE phase is inferred by the
presence of evolved close binaries CVs, Type Ia
SN, LMXB, post-RGB sdB binaries, and... Binary
CSPN, with P lt 3-5 yr
If you are interested in the PN for
themselves How many post-CE PN? How are they
different from single or merged CSPN?
If you are interested in the CE interaction Can
we use the post-CE CSPN to constrain our
understanding of the CE interaction?
7
PN RV survey10/11 RV variables (De Marco et al.
2004 H. Bonds talk) Most CSPN in close
binaries !?
8
Could our finding be right and most/all PN
derive from a CE?Rephrasing the question Could
most/all galactic PN derive from main sequence
binaries that enter and survive an AGB CE?
of stars in the Galaxy 2.1 x 1011 (Total mass
Kulessa Lynden-Bell 1992 IMF Kroupa et al.
1993) Primaries w/ lifetime shorter than age of
Galaxy 7.6 (based on 9 Gyr del Peloso et al.
2005 corresponding to M gt 1.03 Mo Schaller et
al. 1992 Bressan et al. 1993) Percentage of
stars w/ companion 60 (Duquennoy Mayor 1 ie
of binary systems0.375x2.1x1011) Binaries w/
100 Ro lt a lt 500 Ro 12 (I.e. that enter CE on
the AGB Duquennoy Mayor) M1/M2 gt 0.2 73
(Duquennoy Mayor 1991 I.e. secondary ejects
the envelope even for low ?) Mean age of
primaries 1.15 Gyr (Schaller et al. 1992
Bressan et al. 1993 for mean mass of 2.03Mo
from Kroupa
IMF with limits 1.03Mo and 10Mo) PN visibility
time 20,000 yr (ESO catalogue) of post-CE
binary CSPN 9100 (OK within a factor of 5)
of PN in the Galaxy (actual) 3000 (ESO
catalogue Parker Phillips 1998) of PN in
the Galaxy (estimated) 7200 /- 1800 (Peimbert
1990) Despite uncertainty, 9100 post-CE
binaries, is commensurate with the galactic
PN, lending circumstantial support to the RV
survey.
9
Some population syntheses (e.g., Han et al.
1995)predict only 20 of all PN in
close-binaries.Is this inconsistent with our
earlier accounting?

• Population syntheses count the fraction of all
  binary stars that enter and survive a CE. They
  do not count the absolute numbers.
• When plugging the star numbers into those
  simulations the 20 will result in absolute
  number of PN close to our estimate.
• In passing if all stars that ascend the AGB
  make a PN, too many PN are predicted in the
  Galaxy (in the tens of thousands).

10
If the majority of CSPN are post-CE
binaries Where are the single (or merged)
post-AGB stars?

• The total of PN in the galaxy might be lt10,000
  rather than the often-quoted 20,000 (3000
  known).
• Single stars in the post-AGB-to-pre-WD phase
  might have an invisible PN (see Subag Soker
  (submitted)).
• Can we quantify the population of naked
  post-AGB stars via their integrated UV flux
  in external galaxies?

11
The CE phase plays a fundamental role in CVs,
Type Ia SN, LMXB, post-RGB sdB stars in close
binaries, and close-binary CSPN.
Despite past work, out theoretical understanding
of the CE interaction is still rudimentary. In
particular what is ?? We know it is not
constant, but a function of stellar and system
parameters. Without knowing ?, population
synthesis models cannot predict/explain period
distributions and other characteristics of,e.g.,
CVs, Type Ia SN progenitors.
Past work in common envelope theory Ostriker
1975, Paczynski 1976 (proposal)eg, Rasio Livio
1996 (analytical)eg, Taam Sandquist 2000
(numerical)
Past work in common envelope observations e.g.
Hillwig et al. 2002, Drake Sarna 2003Sarna et
al. 1995, Bleach et al. 2000
12
The determination of ? De Marco et al. 2003 in
prep.
Code Burkert Bodenheimer 1993Method
Sandquist et al. 1998

• 3D nested grid hydro code.
• Self gravity only (no B fields).
• Primary calculated via 1D code (Herwig), and
  mapped into the cartesian grid.
• Companion and AGB star core are point masses,
  separation 3 AU, P3 yr
• Max resolution in inner grid 1.75x1011cm
  cf. primary radius 1013cm, core radius
  108cm, companion radius lt1010cm

13
4 common envelope tests
1 Main Sequence Mass 1.5 Mo
Bottom of the AGBTop of the AGB - Thermal Pulse
10
14
Reminder depending on the efficiency of the CE,
the outcome can be
A short-period binary, or a merged star
The efficiency is measured by ?? ?EBin /
?Eg Hence
??? 1 is more likely to result in a close
binary. ??ltlt 1 is more likely to result in a
merger.
15
Resultsthe outcome is a very sensitive function
of initial parameters, including the
evolutionary state of the primary.
? highly variable, while population studies
assume it is constant!
16
The period distribution of post-CE populations,
is a sensitive function of ?.
Period distribution of WDMS post-CE systems from
thetheoretical population synthesis models of
Han et al. 1995 (Fig 4)
17
Theoretical period distributions using the new ?
values can be compared to the period
distributions of post-CE populations such as
or post-AGB CSPN
Post-RGB sdB stars binaries
?
Maxted et al. 2001 Morales-Rueda et al. 2003
De Marco et al. (2004) and work in progress.
This calibration makes population simulations
more reliable to understand theaction of
magnetic breaking or gravitational wave radiation
in tightening binariesleading to the onset of
phenomena like CV behaviour or type Ia SN.
18
CE outcome is a sensitive function of the exact
evolutionary status of the primary.
A 0.1-Mo companion has little effect on
a bottom-of-the-AGB star, but is devastating
for a top-of-the-AGB one
Bottom-AGB
with 68 of the envelope lost in 10 yr and a
resulting binary. The mass lost (unbound mass on
the grid) has a bipolar configuration
(PN morphology?)
Top-AGB TP10
Orbital plane
Perpendicular plane
19
We will also address (code FLASH Fryxell et al.
2000)
1) What happens to the companion in the final
phase of the spiral-in? useful in (i)
can low mass companions eject the envelope?
(formation of CVs with BD
companions
Politano 2004) (ii) can a
planet change into a more massive
object by accreting (e.g., Siess Livio
1999)? 2) What happens when companions merge
with the primarys core? useful in (i) Blue
stragglers (Saffer et al. 2000)
(ii) R Coronae Borealis stars (Clayton
1996) (iii) Wolf-Rayet central
stars (De Marco Soker 2002)
(iv) SN Type Ia (Langer et al. 2000)
(v) Other types of SN??? (suggestion
by E.F. Brown)
20
Summary

• CSPN might be predominantly in close period
  binaries.
• If so, the of PN in the Galaxy might be better
  explained, than if single stars readily make
  PN.
• CE calculations assist population syntheses that
  predict the characteristics of binary classes
  (CV, SN Type Ia). Binary CSPN population used
  to constrain models.
• New generation of simulations is underway, to
  understand accreting secondaries and mergers.

21
Thank you! Please send questions
to orsola_at_amnh.org