MULTIPLICITY IN EARLY STELLAR EVOLUTION

B. Reipurth (University of Hawaii, Institute for Astronomy, Hilo, United States), A.P. Boss (Carnegie Institution, Department of Terrestrial Magnetism, United States), C.J. Clarke (University of Cambridge, Institute of Astronomy, United Kingdom), S.P. Goodwin (University of Sheffield, Department of Physics and Astronomy, United Kingdom), L.F. Rodriguez (Universidad Nacional Autonoma de Mexico, Mexico), K.G. Stassun (Vanderbilt University, Department of Physics and Astronomy, United States), A. Tokovinin (Cerro Tololo Inter-American Observatory, Chile), H. Zinnecker (NASA Ames Research Center, SOFIA Science Center, Germany)

Observations from optical to centimeter wavelengths have demonstrated that multiple systems of two or more bodies is the norm at all stellar evolutionary stages. Multiple systems are widely agreed to result from the collapse and fragmentation of cloud cores, despite the inhibiting influence of magnetic fields. Surveys of Class 0 protostars with mm interferometers have revealed a very high multiplicity frequency of about 2/3, even though there are observational difficulties in resolving close protobinaries, thus supporting the possibility that all stars could be born in multiple systems. Near-infrared adaptive optics observations of Class I protostars show a lower binary frequency relative to the Class 0 phase, a declining trend that continues through the Class II/III stages to the field population. This loss of companions is a natural consequence of dynamical interplay in small multiple systems, leading to ejection of members. We discuss observational consequences of this dynamical evolution, and its influence on circumstellar disks, and we review the evolution of circumbinary disks and their role in defining binary mass ratios. Special attention is paid to eclipsing PMS binaries, which allow for observational tests of evolutionary models of early stellar evolution. Many stars are born in clusters and small groups, and we discuss how interactions in dense stellar environments can significantly alter the distribution of binary separations through dissolution of wider binaries. The binaries and multiples we find in the field are the survivors of these various destructive processes, and we provide a detailed overview of the multiplicity statistics of the field, which form a boundary condition for all models of binary evolution. Finally we discuss various formation mechanisms for massive binaries, and the origin of massive trapezia and their role in the dynamical evolution of clusters.

Intro

Most stars are single (M-dwarves), but most were born multiple

• Many new binary surveys completed
• simulations produce binary statistics (too) easily
• new focus on high order multiples, ultra-wide systems

New Surveys

Duquenny & Mayor 1991 was the old standard Now Raghavan is.

• # of triples doubled
• Binary fraction increases towards higher masses

O-stars:

• close binaries very common
• Sana 2012: 1/4 of O-stars should merge
• many others should interact

M-stars:

• "early" and "late" M-stars (large dynamic range in masses)
• smooth trend towards more equal mass ratios towards lower-mass objects
• separation larger for M than brown dwarf

Multiplicity fraction: # of stars not single Companion Star Fraction: # of companion stars per system

• Chen 2013: companion fraction decreases with age from Class 0 (high) to I to main sequence
• Many caveats on class 0 companions (e.g. Maury et al 2010 - not clear they are multi-stellar systems)

Class 0 Non-hierarchical triple (Rodriguez, Reipurth 2013)

• very unstable
• [what is the line-of-sight separation?]
• end state is a close binary and an ejection (e.g. Reipurth 2000)

Class I - Connelly 2008

• Decline in binary fraction with spectral index
• probably losing companions as a function of time
• lose companions at orbital velocity
• 9 of 9 binaries have a companion near enough to be ejected by this mechanism
• early days of the field

Simulations

Reproducing binary fractions is apparently easy

• Bate 2009: multiplicity by mass = perfect match * surprising since there's not much physics * Maybe it's just not that sensitive to physics * metallicity, feedback seem not to change binary stats
• Maybe we're just lucky, though?
• Gobs of simulations papers cited (Bate, offner, Price, Hennebell, Machida, Burzle, Joos, Seifreid, Boss, Commercon...)
• Don't ignore magnetic fields even though clouds are supercritical (there's plenty of B-field energy)

Do B-fields cause problems?

• Hennebelle & Teyssier 2008: braking so efficient that prevents formation of disk * therefore prevents binary fragmentation * braking drives mass ratio down by reducing angular momentum and accreting onto primary star
• B-field alignment may be the problem
• turbulence reduces efficiency of B-braking (see Talk 5)

Simulations are not well-converged

What happens if you add MRI?

• Shi 2012: more vigorous accretion onto binary
• adds to uncertainty

Gravitational Dynamics

Interaction between binary and its environment

• Are binaries processed by cluster?
• Can they constrain environment?
• Need good N-body simulations

• King 2012 - complete only for Taurus, Cham I, Oph, IC 348, ONC, with range of number densities

  • limited range 62-620 AU
  • no apparent different in multiple fraction (~20%)

• binaries in clusters maybe excess at small separations... which is not expected?

Exotica

Ultrawide tail of Raghavan distribution

• separations > natal cloud cores
• Dhital looking for proper motion pairs

Surprising mode of creating wide binaries in clusters

• Kouwenhoven 2010, Moeckel & Bate, Moeckel & Clarke
• temporarily bound pairs many many times
• if cluster expanding, "transient" bound to neighbor -> actually bound over long timescale

Statistics of Higher Order Multiples

Tokovinin et al in prep... (initial results)

• multiple systems fill all phase space that is stable
• not consistent with pure N-body decay of initially non-hierarchical systems
• pure n-body -> high eccentricity, and most often the ejected fellow is small * observationally, this doesn't quite work... * to get away from stability boundary, need gas (Clarke & Pringle 1991)
• Reipurth & Mikkola 2010, 2012 -> populate binary and triple diagrams
• PP7 will compare data to these simulations
• Within triples: * Easier to make high mass ratio binary pairs at early times (1 Myr) * low-mass ratio binary pairs late (100 Myr)

Questions

• Q: N-body systems required to form wide binaries... not really, not the only way. In low-density environments, born that way. Even though you can make wide binaries in clusters, it's not the only mechanism. I am concerned with N-body codes - you can create whatever you wish with N-body.
• A: There are other dynamical routes. You CAN invoke 3-body interaction. Otherwise you need a 10^5 AU core rotating with beta=1
• Q Kastner: Circumbinary system mislabeled. But, illustrates a ton of interesting things. Hierarchical multiple, surrounded by massive circumstellar disk. 10 Myr (if part of beta pic), far-companion that is also a double.
• Q Kastner: Gas-rich - how gas rich must it be?
• A: Orbital evolution, a small amount of mass can go a long way. For formation, obviously must be comparable to mass of star.
• Q Susanne ?: Binary fraction decreases with age. Why?
• A: Reconfiguration of unstable multiples. Don't have much time from Class 0 to Class I. 10% in surviving triples, many more form in triples
• Q Tobin: Binary separations are different in ONC and Taurus. Is that still current or is that an obsolete result?
• A: No wide proper motion pairs in ONC. Maybe still true?
• Q: AK Sco - xmm newton, periastron passage -> sudden outburst. See my poster.
• Q: Mike Myers: Change in q distribution as a function of primary mass? We found q consistent except for BD distribution.
• A: Not monotonic. O-stars have preference for q=1, so do low-mass stars. If purely dynamic, don't want both.
• Q MM: in Science paper, they chose flat q....