In Praise of IRAS
formation and other
Virtues of IRAS
• Mining the IRAS catalog 24(!) years after
launch, with colleagues:
• Carl Melis, Joseph Rhee(UCLA)
• Inseok Song (IPAC => Univ. Georgia)
Why bother with IRAS when we
now have Spitzer which is much
more sensitive and has much
better spatial resolution?
With Spitzer, 1000s of stars are pointed
at during its lifetime, but IRAS covered
100 times more sky. This conveys some
advantages; for example, for the very
rare nearby star with very high LIR/Lbol,(=
“tau”) IRAS is the better search tool.
Because IRAS was first it
skimmed a lot of IR-bright cream
Spitzer stellar campaigns primarily focus on
extensive observation of fainter stars in
regions of already known interest, (e.g., the
h and chi Persei clusters [Thayne Currie]),
so that the rare, very infrared-bright field
sources may be missed.
• We have correlated the IRAS catalog with
the Hipparcos and Tycho catalogs, which
are also all-sky surveys.
• In this way rare, lonely (isolated), infrared
bright stars can be found that otherwise
will be missed by Spitzer and sometimes
were missed in earlier studies that
correlated the IRAS catalog with older
The dustiest not-so-young stars
We have been investigating two classes of
extraordinary field stars first detected by IRAS:
Main sequence stars that likely are windows into
formation processes for terrestrial planets.
A (still) growing collection of first ascent giant stars
with properties never previously seen.
The Zodiacal Light
The COBE Sky (8–200 µm)
Scattered Sunlight & Warm Zodiacal Dust
T ≈ 250 K
T ≈ 6000 K
Leinert & Gruen 1990
“Vega-like” debris disks
Particle lifetimes << age of star.
So, particles are “second
generational”, the debris from
larger, unseen objects.
HST ACS planet search
HST Fomalhaut detection -- consistent with sub-mm maps
Hubble Space Telescope
JCMT SCUBA 450 micron map (Wyatt & Dent 2002)
Kuiper Belt vs asteroid belt
• The dust at almost all Vega-like stars is sufficiently
cold to be orbiting with semi-major axes of 50 AU or
more from the central star. Thus, the debris disks are
almost always to be considered (young) analogs of
the Sun’s Kuiper Belt.
• Until 2005, among the 100+ main sequence stars
with far-IR excess, only one example of warm dust
signifying a potential asteroid belt analog had been
reliably established – at the A-type star zeta Lep, of
age ~100 Myr (Jura & Chen). LIR/LBol ~10-4
• Absence of warm dust is true even for stars with ages
as young as tens of Myr. Thus, dust in the terrestrial
region dissipates very quickly.
Two unusually dusty sun-like
HD 23514: Pleiades member
BD+20 307: main sequence
Beautiful new Spitzer data on
Alycia Weinberger, Inseok Song, Eric Becklin
Pleiad HD 23514
Comparison of LIR/LBol in Sun’s zodiacal cloud and
in analogous regions at stars with warm dust
Name Sp. Type LIR/LBol Age
Lep A3 0.65 x 10-4 300
HD 23514 F6V 0.02 100
BD+20 307 G0V 0.04 ???
Crv F2V 5 x 10-4 600
HD 72905 G1.5 1 x 10-4 400
HD 69830 K0V 2 x 10-4 2,000
Zodiacal dust G2V 10-7 4,600
Era of heavy bombardment in
early solar system
• Until ~600 Myr following the formation of the Sun, the
bombardment rate in the early solar system was
sporadically heavier than at present by factors up to
• At BD+20 307, which is ~1,000,000 times dustier
than the present solar system, the current
bombardment rate might be incredibly large!
Zodiacal dust properties
• In the solar system, the typical zodiacal dust
particle is 30-100 microns in size. As a
consequence of the Poynting-Robertson (PR)
effect, such particles will spiral into the Sun in a
time (~105 yrs) short compared to the collision
time (~107 yrs). Thus, the zodiacal particles are
not whittled down to smaller sizes by collisions
with other particles.
At HD 23514 and BD+20 307, the strong
silicate emission features indicate the dust
particles are of micron size (due to a
As a result, at these stars, PR lifetimes
from <~1 AU, are very short. How can
there be so much dust in such tight orbits
around the two stars?
Very recent collision of two
• To account for the estimated dust mass at BD+20
307, one must pulverize a 300 km diameter object
(e.g., Davida, the 5th largest asteroid) into micron-size
• Perhaps something analogous to the collision
postulated to explain Earth’s moon has occurred
recently within a planetary system at BD+20 307.
When World’s collide?
Initially, BD+20 307 might have been
regarded as a “miracle”
But discovery of a Pleiades star with a
comparably large tau, has changed
BD+20 307 from a miracle into a statistic.
We can now estimate how common
collisions of young planets may be.
By comparing the number of adolescent-
age (70-700 Myr old), solar-type, stars at
which IRAS could have detected the “very-
dusty” phenomenon, with the number of
such stars that were actually found by IRAS
to be very dusty, i.e., two, we estimate that
about one star in 1000 is very dusty and,
thus, the lifetime of the very-dusty
phenomenon at a typical solar-like star is a
few 100,000 years.
But the lifetime of the observed, micron-size
particles is orders of magnitude shorter.
To interpret our observations, we
considered a model of colliding “planetary
Dust particles => planetesimals =>
planetary embros => planets
C. Agnor and E. Asphaug (UCSC)
considered collisions of large bodies in the
late stages of planetary formation in the
terrestrial planet zone.
To make the terrestrial planets, requires a
minimum of many hundreds of planetary
embryos of dimensions >~1000 km.
Embryo collisions can result in coalescence
or, alternatively, fragmentation of the smaller
embryo into smaller objects along with the
ejection of copious debris.
A typical large fragment size is ~100 m.
Collisions of embryos continues for as long
as a few 100 Myr,
i.e., to the age of HD 23514
N(a)da N O a da
4 3 3.5
a a da ~ a1/ 2
(a) ~ N(a)a da ~ a a da ~ a
2 2 3.5 1/ 2
tc P / ~ a 1/ 2
Thus, the smallest particles (~1 um) collide
the fastest, for HD 23514 ~50 years at 1 AU.
And the lifetime of the large, 100 m,
fragments will be ~500,000 yrs.
Thus, catastrophic disruption of large
planetary embryos (mass estimate to follow
next) can supply material for a sufficient time
to account for the event lifetime indicated by
the observations of HD 23514
A collisional cascade will whittle
particles down until, when “a” ~
0.5 m, radiation pressure blows
them away from HD 23514.
Are there any loss mechanisms that can
compete with collisions followed by
The PR effect that dominates
the orbits of the zodiacal dust
in our solar system is too slow
(~1000 yrs) to compete with
collisions at HD 23514
But stellar wind drag might be sufficiently fast
to compete -- if so, then a “sweeper planet”
might be necessary to account for the
absence of hot dust at HD 23514.
To determine how rapidly mass is lost
due to either collisions and radiative
blowout or stellar wind drag one can
estimate the minimum dust mass (Mmin)
needed to intercept 2% of the light
emitted by HD 23514.
M min 16aR / 3
Mmin ~ few x 1022 g,
with mass loss rate ~ 1013 g/s
In a few 100,000 years, the total mass lost will
be ~1026 g -- the mass of Earth’s moon.
Or for an average density of a few g/cm3, an
object with radius ~2000 km.
Not possible to distinguish from
our data and this model if most
solar-like stars undergo very dusty
episodes, or if only a (significant)
minority do so, but perhaps with
catastrophic conversion of even
more mass to dust particles than
indicated on the previous slide.
How old is BD+20 307?
~300 Myr (Song et al 2005)
Based on lithium content, Galactic
space motion (UVW), and upper limit
to ROSAT X-ray flux
Michael Muno & I received
Chandra X-ray time to measure
the X-ray flux.
Gregory Henry joined us to measure the
rotation period of BD+20 307 from
photometric observations with an
automated telescope at Fairborn
Observatory (in southeast Arizona)
Both the Chandra-measured
X-ray flux (just below the
ROSAT upper limit) and the
rotation period (~3.5 days)
indicated a youthful star, of
age a few 100 Myr.
All looked consistent and
In late October 2007, Alycia
Weinberger made the surprising
discovery, from observations on 3
successive days, that BD+20 307
is ~3.5 day period double-line
This was surprising because the 2004
epoch, Song et al echelle spectrum showed
no suggestion of binarity
So, Frank Fekel and Mike
Williamson joined our team
and conducted a 45 day
earlier this year (at Fairborn
Observatory) to measure the
Here are their results:
Where do things stand now?
Because of the short orbital period (in
conjunction with the [weak] 6708 A lithium
absorption line strength), we can be confident
that the orbital and rotation periods are
synchronized and that neither the rotation
period, nor any activity indicators that depend on
it, (e.g. X-ray flux), tell us anything about the age
of BD+20 307
Instead, for an age estimate,
we must use other indicators,
such as lithium line strength,
Galactic space motion (UVW),
metallicity, and placement on
the HR diagram
In summary, BD+20 307 is
probably a few Gyr old, and
possibly considerably older.
Can we be witnessing the aftermath of a
collision of two terrestrial planets orbiting
such an old star!?
Is there an unseen faint star or
substellar object whose
gravitational field has
destabilized the orbits of two
Some suggestion in the radial
velocity data of the presence of a
3rd object of considerable mass
Whatever its age, if the
massive amounts of dust do
point toward the presence of
terrestrial planets, then BD+20
307 is the first known example
of planets of any mass in orbit
around a close binary star.
Gemini/Michelle spectrum of HD
To finish, now a brief
discussion of a
isolated, stellar IRAS
A little bit of history…
First mention of BP Psc:
object #202 in Stephenson’s
(1986) list of high-latitude
H-alpha emission-line stars
15-20 years ago, various astronomers
were investigating high Galactic latitude
IRAS sources to determine if any might
be nearby T Tauri stars (rather than,
say, galaxies). For example, this is the
way the first members of the TW Hya
Association were identified.
Thus, was BP Psc “rediscovered”
In 1996, my then student Richard Webb
(whose Ph. D. thesis first firmly
established the TW Hya Association as a
substantial group of very nearby T Tauri
stars), my former Ph. D. student Joel
Kastner, Thierry Forveille (Grenoble),
and I obtained optical and radio line data
on BP Psc (Galactic latitude -57 deg).
Before 2006, BP Psc never was a
subject of serious investigation
When included as just one star in large surveys it
has been generally classified as a pre-main
sequence star but with a wide range of spectral
types. But, in the absence of a measured proper
motion or distance, even classification as a distant,
luminous, post-AGB star seemed conceivable to us
Thus, BP Psc was relegated to a back burner…
Then, in 2006 we recognized that the Tycho proper
motion, ~50 mas/yr, ruled out such a distant star.
Thus began an extensive
Carl Melis, Joseph Rhee, Eric Becklin
(UCLA), Inseok Song (Gemini), David
Meier (NRAO), Alycia Weinberger
(Carnegie), Bruce Macintosh, Christian
Marois (LLNL), James Graham, Geoff
Marcy (Berkeley), David Wilner (Harvard),
Thierry Forveille (Grenoble), Joel Kastner
(RIT), Travis Barman (Lowell), Patrick
Palmer (Chicago), Mike Bessell (ANU)
Weinberger at al 2002
Already christened “HH 999”
by Bo Reipurth (& these data
are not yet even published)!
Will this designation have to be
rescinded if BP Psc is actually a post-
main sequence star!?
Bo Reipurth to BZ:
How old is BP Psc?
How far away?
Where was it born?
Why so much warm dust?
What is the mass accretion rate?
Does it have any “fellow travelers”?
Comparison of gravity-sensitive
lines in our HIRES spectra of BP
Psc with lines from late-G and
early-K type dwarfs, subgiants,
and first-ascent giants, imply
that, most likely, BP Psc has the
gravity of a first-ascent giant star.
Is BP Psc the first known example of a first
ascent giant star with a substantial orbiting
gaseous disk, rapid accretion of gas, and
associated jets and HH objects? If so, then
what strange phenomenon has occasioned
its massive rapidly accreting disk of dust and
gas? A shredded companion star or
substellar object? Are planets likely now
forming in the dusty CO disk? (a gazillion
years after BP Psc itself formed)
W UMa binaries and “hot
Jupiters” are two known
classes of close binary
systems where the secondary
could be engulfed by the
evolving primary as it leaves
the main sequence.
We believe that the IRAS
catalog reveals a suite of such
common envelope objects,
which, apparently are often
(always?) accompanied by
dusty excretion disks.
One such dusty giant HD
233517, was modeled previously
by Mike Jura (UCLA) as a cold
excretion disk. However, we are
finding multiple examples in the
IRAS catalog of giant stars with
much warmer dust; presumably
such stars are closer in time to
the actual merger event.
TYC 4144 329 2
mid-F type -- dusty secondary of
visual binary with ordinary K-type
primary (~60 pc from Earth)
Trig Parallax of BP Psc
If pre-main sequence, then distance will be
<~100 pc. If post main sequence, then
distance will be a few 100 pc.
Galactic Space Motions
Youthful Group U V W age distance
(km/s) (Myr) (pc)
• TW Hydrae -11 -18 -5 8 ~60
• Tucana/Hor -11 -21 0 30 ~45
• Pictoris -11 -16 -9 12 ~40
• AB Doradus -8 -27 -14 70? ~30
• Cha -12 -19 -10 8 97
• Carina-Near -26 -18 -2 200 ~30
• BP Psc -12 -23 +5 20 81
• BP Psc -14 -27 +2 10 104
• BP Psc -19 -33 -2 5 139
• BP Psc -21 -36 -4 3 157
Kinematic mass vs
M(kin) = R(Vkep)2/G = R(VP)2/G(sini)2
R = radius of CO disk
Vkep = Keplerian velocity at R
VP = projected line of sight velocity
i = disk inclination angle (i = 0 is face-on)
Kinematic mass = 0.61 M(sun)
• Evolutionary mass (from Baraffe et al):
• 1.26 M(sun) (for an age of 10 Myr)
• The ratio of the kinematic & evolutionary
masses hardly varies for ages between 3 &
20 Myr (both are ~proportional to distance)
As some of you were born
*after* IRAS flew,
IRAS swept almost the entire sky at 4
12, 25, 60 and 100 microns
Gas & Dust masses
• Dust mass (Md) from submillimeter flux:
Md = SD2/kB(Td)
D = distance to star
S = flux
k = opacity (cm2/g)
Dust mass = 0.7 Earth
masses (for cold dust at 36 K)
The mass of H2 depends on the uncertain
optical depth of the 12CO lines. We have an
upper limit on the brightness of the 13CO, J =
2-1 line which implies that the 12CO optical
depth is <14. Thus, in the outer regions of
the CO disk,
6 < M(gas)/M(dust) < 100
In the past two years, five more stars of
age >~100 Myr with warm dust in the
terrestrial region have been identified
With Spitzer, Beichman et al 2005 found an ~1 Gyr old K-type star (HD
69830) with LIR/LBol ~10-4 and silicate emission features seen in the
wavelength range accessible to Spitzer’s IRS. Three Neptunes
present! (Note: excess emission at 25 micron was marginally
detected by IRAS!)
From old IRAS data, we identified two solar-mass, adolescent stars -- a
Pleiad and a field star BD+20 307. Follow-up of the latter at Keck and
at Gemini revealed a huge LIR/LBol (4%) and evidence for micron-size
crystalline and amorphous silicate particles.
Solar system time scales and ages
of young nearby stars
• Formation of Jupiter < 10 Myr
• Formation of Earth’s core ~ 30 Myr
• Era of heavy bombardment
in inner solar system ~ 600 Myr
• Cha cluster 8 Myr
• TW Hydrae Assoc. 8 Myr
• Pictoris moving group 12 Myr
• Tucana/Horologium Assoc. 30 Myr
• AB Dor moving group 70 Myr
• Carina-Near moving group 200 Myr
So why devote any time to
• In the early 90’s, a few researchers became
interested in “isolated,” high-b T Tauri stars
– Various such TTS identified in the 80’s
• First & still most famous: TW Hya (Rucinski & Krautter 1982)
• de la Reza et al., Stephenson identified more
• Existence of such stars raised fundamental
– Where were they born?
– How old/distant are they?
• Do they still retain dusty/gaseous disks?
• Are they closer than Tau/Oph/Cha TTS (i.e., d < 140 pc)?
– If so, opportunity to study star & planet formation “at close range”
Is BP Psc one of the nearest, oldest, classical,
T Tauri stars and the closest known example of
bipolar jets and HH objects, but with
uniquely(?) low lithium abundance and a
unique(?) IRS spectrum, along with a variety of
other lesser problems (not mentioned in this
A decade of wondering…
Is BP Psc an evolved star?
• Optical emission line spectrum reminiscent of
– Dusty disks/envelopes common
• Rotating, gaseous disk in such a system would be a first!
• Infrared SED reminiscent of post-asymptotic giant
– Many examples known at high latitude
– Some post-AGB (binary) systems known to harbor dusty
• Again, a rotating molecular disk would be quite a find!
• Where is the associated young group?