The Impact of Multiplicity on The Impact of Multiplicity on Planet FormationPlanet Formation

Adam KrausAdam KrausHawaii-IfAHawaii-IfA

Michael Ireland (Sydney), Michael Ireland (Sydney), Frantz Martinache (Subaru), Lynne Hillenbrand (Caltech)Frantz Martinache (Subaru), Lynne Hillenbrand (Caltech)

Beta Pic

CoKu Tau/4

LkHa 330

Canonical Model of Star FormationCanonical Model of Star Formation

The canonical model of star formation

(Shu et al. 1987)

Protoplanetary Disks as Sites of Protoplanetary Disks as Sites of Ongoing Planet FormationOngoing Planet Formation

CoKu Tau/4 (Model; Quillen et al. 2004)

Transitional DisksProtoplanetary disks have large cleared gaps – perhaps signposts of planet formation?

LkHa 330 (Submm image; Brown et al. 2008)

The Timescale for Planet FormationThe Timescale for Planet Formation

Disk Frequency as a Function of Age

Haisch et al. (2001)

The Disk dispersal timescale (~5 Myr) sets the timescale for possible planet formation.

Our SurveyOur Survey

Most of our survey is being conducted with Keck using a new technique called aperture-mask interferometry. We could detect equal-brightness companions at ~1/4 λ/D (15 mas) and achieve the best sensitivity (ΔK~6) at ~λ/D (50 mas).

We’ve observed ~100 members each for Taurus (T~2 Myr) and Upper Sco (T~5 Myr), spanning masses of 0.5-2.0 Msun.

Aperture Mask InterferometryAperture Mask InterferometryPlacing an aperture mask in the pupil plane turn the single mirror into a sparse interferometric array.

Interferometric analysis techniques (i.e. closure phases) can be used to filter almost all noise from turbulence and AO errors. This yields very stable performance.

Our plan is to pursue a new planet-search strategy. We trade lower sensitivity (103 versus 106) for superior resolution (~λ/D versus 10λ/D), and then we observe distant young stars where substellar companions are much brighter.

Survey ResultsSurvey Results

We found many new companions, but few with ΔK>3 and none with ΔK>4; a few might be substellar. Our limits extend to ΔK~6 at Keck.

(Keck limits extend to ~8-15 MJup for Upper Sco members or ~6-13 MJup for Taurus members.)

Red circles: Masking detectionsCrosses: Known binaries

Dashed lines: 99.5% Detection Limits

The Disk around CoKu Tau/4The Disk around CoKu Tau/4SED modeling suggests that the inner 10-15 AU of the disk around CoKu Tau/4 have been cleared; this has been attributed this to active planet formation.

We actually found that CoKu Tau/4 an ~8 AU binary; the disk truncation is therefore a natural result of tidal truncation and not planet formation.

Troubling Question: Many of the other hosts for “transitional” disks haven’t been thoroughly surveyed for multiplicity. How many are actually circumbinary disks?

Circumbinary “transitional” disks?Circumbinary “transitional” disks?

Detection limits (K-band multiplicity survey; solid lines) for GM Aur, UX Tau, LkCa15, and LkHa330 from Keck. Next step is a deep L-band survey of these targets to search for planetary-mass companions (dotted lines).

Circumbinary “transitional” disks?Circumbinary “transitional” disks?

20% of all solar-type stars appear to have 2-40 AU 20% of all solar-type stars appear to have 2-40 AU binary companions, so why don’t we see ~20 “fake” binary companions, so why don’t we see ~20 “fake” transitional disks in places like Taurus?transitional disks in places like Taurus?

Many of the disk-free stars in Taurus are turning out to Many of the disk-free stars in Taurus are turning out to be close binaries! be close binaries! (Hubble 4, LkCa4+5+21, V827 Tau, almost anything (Hubble 4, LkCa4+5+21, V827 Tau, almost anything with CIDA in its name)with CIDA in its name)

This isn’t a new idea (Ghez et al. 1997; White & Ghez 2001; Cieza This isn’t a new idea (Ghez et al. 1997; White & Ghez 2001; Cieza et al. 2008), but our survey is the first with enough resolution and et al. 2008), but our survey is the first with enough resolution and sensitivity to study most of this separation range.sensitivity to study most of this separation range.

But wait a minute…But wait a minute…

Binaries and DisksBinaries and Disks

The disk fraction as a function of separation for Taurus solar-type stars in binary systems. The disk location (primary, secondary, or circumbinary) hasn’t been determined for most systems).

Disk Lifetime in Close Binary SystemsDisk Lifetime in Close Binary Systems

The Multiplicity CorrectionThe Multiplicity CorrectionIt seems that many of the disk-free stars at 1-2 Myr are binaries; if we remove them, the disk fraction of genuinely single stars goes up by ~15-20%.

Under this picture, most circumstellar disks last at least 2-3 Myr; most circumbinary disks, on the other hand, seem to disperse promptly.

(Disk fractions from Hillenbrand 2005).

Under this picture, most circumstellar disks last at least 1-2 Myr; most circumbinary disks, on the other hand, seem to disperse promptly.

Numerous surveys have found that even circumstellar disks are mostly gone by 5 Myr, though. (Carpenter et al. 2006)

Disk LifetimesDisk Lifetimes

(Figure from Alexander & Armitage 2009)

Implications for Planet FormationImplications for Planet FormationMost single stars have >1-2 Myr within which to form planets (but not much more than 5).

Close binary systems have <1 Myr. (This doesn’t bode well for planet searches around close binaries…)

Note: There are variable outcomes. Judging by the SEDs, DP Tau has a full disk, CoKu Tau/4 has a circumbinary disk, and Hubble 4 has no disk. Why?

Some binaries have long-lived disks; most seem to be in binary systems with equal mass components and/or circular orbits. (CoKuTau/4, ST 34, GG Tau, HD 98800)

SummarySummary If disk gaps are cleared by companions, then most are

planets. Some are close binaries, but they seem to be Some are close binaries, but they seem to be a small minoritya small minority

The majority of close (<40 AU) binaries lose their The majority of close (<40 AU) binaries lose their protoplanetary disks extremely quickly (<0.5 Myr). protoplanetary disks extremely quickly (<0.5 Myr). Planets in close binary systems should be rare. (alpha Cen, Sirius, Procyon, 61 Cyg…)(alpha Cen, Sirius, Procyon, 61 Cyg…)

This prompt dispersal of circumThis prompt dispersal of circumbinarybinary disks removes disks removes the need for prompt dispersal of circumthe need for prompt dispersal of circumstellarstellar disks. disks. Most single stars and wide binaries have ~3 Myr to form planets (so why do only ~10% have gas giants?)