An HST/ACS Survey of the Kuiper Belt

Gary Bernstein (University of Pennsylvania)

#ith co-I’sLynne A&en (UBC), Mike Brown (Cal Tech), Mat'

Holman (CfA), Renu Malhotra (LPL), & David Tri&ing (Penn➙Steward)

and for Astrometry: N. Zacharias (USNO) & J. A. Smith (LANL)

Background Image: section of summedACS TNO search-field data.☀Funded by GO-9433 grants

The Kuiper Belt as an accretion and dynamics laboratory

Design and Execution of the ACS Survey

TNO Discoveries & Followup

Implications for Solar System Evolution

Light Curves and Astrometry

Planet Formation Scenario:

Nebular/dust phases areobserved around other stars

Giant planets have beendetected around otherstars

Planetesimal phaseis observable onlyin our outer "fossilized"Solar System - theKuiper Belt.

Collapse of gas/dust cloudto form protostar and nebulardisk.

Coagulation of grains, runaway gravitational growthof planetesimals

Oligarchic growth of planetesimals.

Gas accretion onto largest bodies,dissipation of gas, ejection of smallbodies

Mature planetary system

Theoretical Expectation: Observational Evidence:

Known Outer Solar-System Bodies (as of 2003 April 19)

Key

Plutino

Scattered

Other TNO

Comet

Asteroid

Giant Planet

Centaur

Pluto

Neptune

Uranus

Open symbols denote orbits uncertain due to short arcs.

20 30 40 50 60 70 80 90 1001000

0.2

0.4

0.6

Semimajor Axis (AU)

Outer Solar System Objects (as of 4/03)

Neptune-crossing

Eccen

tric

ity

Uranus

Neptune

Pluto

Resonant

3:2

2:1

5:2

1:1(Trojan)

KBO Resonances and Planetary Migratio(

Pluto and many KBOs (”Plutinos”) are locked in 3:2 resonance with Neptune, crossing Neptune’s orbit yet avoiding ejection.

Malhotra (1993) explains “accident” of Pluto’s resonance by positing an outward migration by ~8 AU of Neptune during the clearing of the planetesimal disk. Resonances “sweep” through the population, acting like phase-space traps. Prediction is that many KBOs should be found in 3:2 and 2:1 resonances

KBOs have been found in following resonances (Chiang et al ‘03):1:1 (”Trojan”), 5:4, 4:3, 3:2 (”Plutino”), 5:3, 7:4, 2:1, 7:3, 5:2Resonance sweeping can only explain some of these, obviously not 1:1.

3:2 Resonances inNeptune frame(Malhotra 1995)

Orbit of 2001 QR322 in Neptune frame(Chiang et al 2003)

20 30 40 50 60 70 80 90 1001000

0.2

0.4

0.6

Semimajor Axis (AU)

Outer Solar System Objects (as of 4/03)

Neptune-crossing

Eccen

tric

ity

Uranus

Neptune

Pluto

Resonant

3:2

2:1

5:2

1:1(Trojan)

Centaurs

"Scattered

Disk"

"Classical" Belt

Cold Primordial Disk?

CAUTION:Strong

Selection Effects

Recall that distant TNOs have

Traditional response to fainter target is longer integration, but S/N gain for moving objects stops when object trails across sky background. TNOs move few arcsec per hour.

Non-sidereal tracking of telescope only works for one preselected motion vector, not appropriate for a survey

Ground-based exposures limited to ~10 minutes before trailing, or

Digital Tracking: Take short exposures, shift and add digitially to follow any candidate orbit, then run detection algorithm.

Used successfully from ground in 1992 by GMB, with Cochran et al WFPC2 data, many other times since.

Finding Faint TNOs

f ∝ r−4

mR<∼ 24

a) Original Single Exposure (c) Single Exposure, Fixed Objects Subtracted/Masked

(b) Combined ImageTrack at Sidereal Rate

(d) Combined ImageTrack at KBO rate

R=23.5

R=24.0

R=23.0

R=24.5

R=25.0

R=25.5

Co

rrec

tly tra

ck

ed

Tra

ck

ed

at

wro

ng

rate

6.5

7

8

9

10

11

Below detection th

reshold

(

R>26)

10 20 30 40 50 60 70

Heliocentric Distance (AU)

Dia

me

ter

of

Ob

ject

(km

)

300

200

100

0

Earth

Jupiter

Saturn

Uranus

Neptune

Pluto

H

Ground-Based Survey of 1.5 Square DegreesA&en, Bernstein, & Malhotra (2001)

Size Distribution of KBOs

20 22 24 26 28 30

R Magnitude

100 10Diameter at 35 AU (km, albedo=0.04)

0.01

0.1

1

10

100

1000

104

105

log

N(<

m)

pe

r squ

are

de

gre

e

Number vs Mag (~size) for

Trans-Neptunian Objects

ABM01

GL98/01

LJ98

JLT98

TJL01

CB99

Fa01

La01

CLSD/WFPC2

???

Keck data

For fixed distance, mag distribution is the size distribution.

Faintest two points on this plot known to be contaminated by false positives!

Data seem consistent with a single power law,

N(< m) = 100.63(m−23)

⇒ dN

dr∝ r−q, q = 4.0± 0.5

Theoretical Expectations for Size Distributio(

Single power law must break down to avoid divergent mass, surface brightness.

Modelling of collisional accretion and erosion requires detailed numerical study, assumptions about fracturing, dynamics, etc.

Dohnanyi shows that q=3.5 is a point of steady-state collisional cascade under scale-free fracturing laws. Expect this below some disruption scale.

t = 0 Myr

t = 5 Myr t = 25 Myr

t = 50 Myr t = 100 Myr

t = 130 Myr t = 265 Myr

t = 1000 Myr

Does this size break exist? Kuiper Belt is perhaps our only opportunity to test an accretion model!

Model by Davis et al

Outstanding Kuiper Belt Questions

What is the size (magnitude) distribution of objects?

Are the dynamical classes physically distinct?

What event(s) or process(es) heated the planetesimal disk?

Is there truly a truncation? What caused it?

Are any of the dynamical classes viable source reservoir for the observed 1-10 km Jupiter family comets?

What can we learn about the history of the solar system, and planetary systems in general?

Plan for the ACS SurveySix ACS fields, 0.02 square degrees (12x HDF)

Observe each field 22 ksec over 5 days

Sum flux over 30 exposures in candidate orbits

Wait 5 days; most TNOs pass stationary points and reverse

Observe another 15 ksec to confirm discoveries

Detection limits: m<29.2 in F606W

Expect 85 detections under extrapolation.

≈ 1014

20 22 24 26 28 30

R Magnitude

100 10Diameter at 35 AU (km, albedo=0.04)

0.01

0.1

1

10

100

1000

104

105

log N

(<m

) per

square

degre

e

Number vs Mag (~size) for

Trans-Neptunian Objects

ABM01

GL98/01

LJ98

JLT98

TJL01

CB99

Fa01

La01

Expected

ACS Result

break t

o Dohnanyi

slope?

A& it takes is

116 orbit Cycle 11 allocation (executed in 125 orbits, thanks Tony Roman!)

30,000 lines of C++ code

New Fourier-domain PSF characterization methodPSF interpolation across ACSImage registration & distortion (Anderson)Image combining with reduced PSF distortionImage subtractionDigital tracking addition & detectionOrbit calculation and trailed-PSF fitting directly to images.

Several CPU-years on P4 processors

Schedul)

Feb 2002: Phase II request

Aug 2002: Preparations begin in earnest

Jan 26 2003: First TNO observation

Feb 09 2003: Last observation

April 14 2003: All candidates identified

April 29 2003: Keck retrieval run (Magellan clouded out)

August 2003: Deep search complete

I&ustration of ACS Data Quality

140

150

160

170

180

-60 -40 -20 0 20-3

-2

-1

0

1

2

X Position of 2000 FV53

in part of ACS Data

...linear trend removed

Positio

n (

arc

sec)

-70 -65 -60 -55 -50

Time (Hours)

0.4

0.2

0.0

-0.2

-0.4predicted parallax from HST orbit at 39 AU

TNO Discoveries in the ACS Survey2000 FV53

m=23.4

2003 BF91

m=26.95

2003 BG91

m=28.15

2003 BH91

m=28.38

Single Exposure

400 seconds

Summed Orbit

2000 seconds

Summed Visit

6000 seconds

All Data

38,000 seconds

S/N=73 S/N=26 S/N=23

S/N=90

1.5 arcsec

40.26 AU2.5 deg

0.0744 km

Distance:Inclination:

Eccentricity:Diameter:

42.14 AU1.5 deg

??28 km

42.55 AU2.0 deg

??25 km

Faintest solar system objects ever detected!1 photon per 5 seconds at HST

Analysis of a& survey data

All TNOCKBExcited

0.001

0.01

0.1

10

100

18 20 22 24 26 28

R Magnitude

0.01

0.1

10

1

1

0

1

100

1000

10

Σ(m

) (m

ag

-1 d

eg

-2)

Se

arc

h A

rea

(d

eg

2)

De

tectio

ns

LarsenTB

ABM

TJL

Gladman

ACSCB

38AU ≤ d ≤ 55AU

i < 5◦

Combine data from all large, well-characterized surveys within 3 degrees of the invariable plane.

Deviation from single power law at high confidence.

Split the TNO sample into “Classical” KB and “Excited” populations. CKBs defined as:

LFs of dynamical classes distinct at 96% confidence.

Double-power-law fits to the Magnitude Distributio(

18 20 22 24 26 28

0.1

1

R Magnitude

Σ(m

) x 1

0-0

.6(m

-23

) (m

ag

-1 d

eg

-2)

Excited

CKB

(Mass p

er

mag)

Bright-end slope α1

Fa

int-

en

d s

lop

e α

2

CKBO

Excited

0.4

0.6

0.2

0.

-0.2

-0.4

0.6 0.8 1.0 1.2 1.4

What does it mean?

t = 0 Myr

t = 5 Myr t = 25 Myr

t = 50 Myr t = 100 Myr

t = 130 Myr t = 265 Myr

t = 1000 Myr

Stern (1995) gives sizes below ~200 km as subject to disruption in current collisional environment:

Crudely, this is size of potential break in size distribution.

Resembles >1 Gyr phases in Davis simulation - Excited older than CKB

Implications

14 16 18 20 220

0.2

0.4

0.6

0.8

1

R magnitude

Pro

ba

bili

ty o

f A

ny T

NO

s <

RQuaoar

2000 KX14

Pluto

Total mass of the CKB: (for p=0.04, <d>=42 AU, ρ=1000 kg/m3):

Extrapolation of Excited bright end predicts largest object in the Pluto-Quaoar range. Pluto is uniquely but not anomalously large.

Note largest CKBO is predicted (and is, so far) 60x lower mass.

0.010 M⊕ ± 15%

Are KBOs a viable source reservoir for Jupiter-family comets?

One astronomer’s noise is another’s signal

Cycle 13 Observations

Additional 6 orbits granted and executed in May 2004 (GO-10268, PI Trilling)

Retrieve the 3 new objects to determine eccentricity - really CKB objects?

Additional astrometry will give sufficient orbital info to locate these objects with JWST in 2010.

Recovery of 2003 BG91

Phased photometry of ACS TNOs

0.4

0.6

0.8

1

1.2

flu

x (

elec

tro

n/s

ec)

0.2

0.4

0.6

flu

x (

elec

tro

n/s

ec)

0 0.2 0.4 0.6 0.8 1phase

0

0.2

0.4

flu

x (

elec

tro

n/s

ec)

0.2 0.4 0.6 0.8 1phase

18

19

20

21

22

23

flu

x (

elec

tro

n/s

ec)

P=4.2 h P=9.1 h

P=2.8 h P=3.79 h

2003 BG91

2003 BF91

2003 BH91

2000 FV53

>1 mag variation!

0.07 mag variation!

Trilling & Bernstein (submitted to AJ)

A Model for 2003 BF91

Is 2000 FV53 a strengthless body?

0 0.04 0.08 0.12 0.16 0.2 0.24 0.28τ ( = K / W )

0

0.5

1

1.5

2

2.5

3

3.5

4ρ

( g/c

m3 )

(Varuna)

Pluto

Mathilde

Eros

P = photom. period

P = twice photom. period

P(θ)

JacobianMaclaurin

b ≈ cb > ca ≈ b ≈ ca = b a >> b

Exclud

ed:

Too Asym

metric

Oblate:

no light c

urve!

Too dense?

Too light?

Spots?

Rubble-pile solutions

0 0.2 0.4 0.6 0.8 1b / a

0

15

30

45

60

75

90φ (

deg

rees

)

P(θ

) =

1.0

P(θ

) =

0.1

density = 0.5

density = 1.0

density = 2.0Too

asym

metric

Ang

le o

f Rep

ose

Acceptable!

Astrometric Residuals for 2000 FV53

Bernstein, Zacharias, & Smith (in prep)

0

0

N-S

Err

or

(mill

iarc

sec)

E-W

Err

or

(mill

iarc

sec)

-4

-2

2

4

-4

-2

2

4

0. 0.5 1. 11. 11.513.3. 3.5 13.5

Time (days)

Is mi&iarcsecond KBO astrometry just a neat trick?

N. Zacharias (USNO): 5 mas tie to ICRS very practical, better over 1-degree patches?

1 mas at 40 AU = 30 km = 5 seconds of HST motion

Detectable photocenter motion on 300-km spotted body

Displacement from undiscovered solar system masses is

∆x = 1 km(

M

M⊕

) (T

1 yr

)2 (D

100 AU

)−2

Future DevelopmentsSky to R=21 is mostly covered by Palomar/Quest

NOAO Deep Ecliptic Survey, CFH Legacy survey to expand sky coverage for R<24

Full sky to R=24 will be done by Pan-STARRS, LSST - 10,000’s of objects with full orbital dynamics

Could reach R=26 over limited regions of sky

Some further coverage from ACS archive (PI: Tamblyn) or another Large program.

SNAP could repeat ACS survey to get 10,000’s of R=28 objects, detail dynamics/size correlations.

Occultation surveys (TAOS) explore <10km regime

New Horizons probe flies by 2 (?) KBOs c. 2017

Summary

ACS works flawlessly for TNO search. Background-limited detection for 29th mag moving objects.

Clear detection of a feature in the size distribution, with dependence upon dynamical state of subpopulations.

No sign of cold population beyond 50 AU, for D>40 km

Scattered disk a feasible JFC source, but still a stretch?

Suprising light curves, consistent with rubble pile in 200 km body 2000 FV53

Additional info forthcoming from followup & astrometry.

0.1 10 100 10000.1

1

10

100

1000

Distance at Discovery (AU)

1

104

105

Bo

dy R

adiu

s (

km

)

Solar System Discovery Space

Naked Eye (<1610)JupiterSaturn

Venus

Mercury

Mars

Sun

Moon

0.1 10 100 10000.1

1

10

100

1000

Distance at Discovery (AU)

1

104

105

Bo

dy R

adiu

s (

km

)

Solar System Discovery Space

Naked Eye (<1610)

Telescopic (1610-1900)

Ceres (1801)Pallas(1802)

Juno (1804)

Vesta (1807)

GalileanMoons(1610)

PhobosDeimos(1877)

Uranus(1781)

Neptune(1846)

Titan (1655)

0.1 10 100 10000.1

1

10

100

1000

Distance at Discovery (AU)

1

104

105

Bo

dy R

adiu

s (

km

)

Solar System Discovery Space

Naked Eye (<1610)

Telescopic (1610-1900)

Photo (1900-1990)

Pluto (1930)

Charon (1978)

Voyager Moons(1978-1989)Eros

NEA's

0.1 10 100 10000.1

1

10

100

1000

Distance at Discovery (AU)

1

104

105

Bo

dy R

adiu

s (

km

)

Solar System Discovery Space

Naked Eye (<1610)

Telescopic (1610-1900)

Photo (1900-1990)

1992 QB1

Quaoar (2002)

Kuiper Belt

CCD (1990-2002)

0.1 10 100 10000.1

1

10

100

1000

Distance at Discovery (AU)

1

104

105

Bo

dy R

adiu

s (

km

)

Solar System Discovery Space

Naked Eye (<1610)

Telescopic (1610-1900)

Photo (1900-1990)

New (2003+)

Sedna

2003 Bx91

CCD (1990-2002)

0.1 10 100 10000.1

1

10

100

1000

Distance at Discovery (AU)

1

104

105

Bo

dy R

adiu

s (

km

)

Solar System Discovery Space

Pan-STARRSLSST

Space Occultation Survey

Wid

e-Fi

eld

Spac

e Im

agin

g

Naked Eye (<1610)

Telescopic (1610-1900)

Photo (1900-1990)

New (2003+)

Sedna

2003 Bx91

CCD (1990-2002)