Imaging faint companions with
interferometric closure phases
Olivier Absil
Chargé de Recherches FNRS
AEOS group
3rd ARC meeting – March 4th, 2011
3rd ARC meeting
04/03/2011
Why interferometry?
 Enables physical characterisation


Photometry / spectrum  temperature, composition
3D orbit  mass (when combined with RV)
 Complementary to single pupil imaging


Search region ~ 200 mas in radius
Milli-arcsec resolution  access to hot Jupiters
 Could replace RV for inappropriate stars
VLTI
0.1
3rd ARC meeting
VLT (+ adaptive optics)
1
10
Distance [AU], for star at 40 pc
100
04/03/2011
The interferometric view of binaries
 Sum of 2 offset fringe packets
 Global size increased  visibility affected
 Photocentre shifted  phase affected
 “Resolved” when Δθ > λ/2B
λ/2B
B
Δθ
3rd ARC meeting
04/03/2011
Effect on the (squared) visibility
 Modulation vs baseline

Period = λ/Δθ
 V2 deficit w.r.t. single star

1% contrast at 10 mas
Up to 4× flux ratio
 Achievable dynamic range


Up to ~1:250 (at 3σ)
Assumes 0.5% accuracy on
cumulated V2
 Contrast/position
ambiguity

Many V2 measurements
3rd ARC meeting
04/03/2011
Effect on the (differential) phase
 Absolute phase lost due
to turbulence
 Wavelength-differential
phase can be measured


Non-zero if star and
companion have different
spectra
Position/spectrum still
ambiguous
λ
photocenter
photocenter
M0V at 10 mas of A0V
(100 m baselines)
 Affected by dispersion

Contrast limited to ~1:100
3rd ARC meeting
04/03/2011
A better solution: closure phase
 Ψ123 = φ12+ε1 + φ23 + φ31−ε1


External perturbation
removed
≠ 0 only when object departs
from point-symmetry
turbulence
ε1
1
 Case of a binary object

Ψ = ρ (sin α12+sin α23+sin α31)


CP proportional to flux ratio ρ


where αij=2πBij∙θ/λ
contrast of 1:100  CP ~ 1°
Contrast/position ambiguity
mitigated by spectral
dispersion
3rd ARC meeting
φ12+ε1
φ31−ε1
2
3
φ23
04/03/2011
VLTI/AMBER observations of β Pic
 24 – 28 January 2010
 10+ hours of observations
 VLTI
 ATs: A0 – G1 – K0
 Good seeing (0.8")
 AMBER
 MedRes (R = 1500)
 K1 band (1.93 – 2.27 µm)
 FOV ~ 420 mas FWHM

128m
Radius ~ 4 AU for β Pic
 FINITO
 Fringe tracking
3rd ARC meeting
96m
96m
04/03/2011
Closure phase stability
 Calibration star

HD 39640 (G8III, K=3)
At 1.2° on sky
 Unresolved (V2~0.9)

 Data reduction


Amdlib 3.0
No special options/scripts
 CP stability

~0.3° in 2.00 – 2.27 µm
3rd ARC meeting
04/03/2011
Fitting a (high-contrast) binary model
Best fit: 1.8×10-3 ± 1.1×10-3 at 14 mas (χr2=0.87)
3rd ARC meeting
04/03/2011
VLTI/AMBER sensitivity profile
 AMBER performance



Optimal search zone:
2 – 100 mas
Median error bar = 1.2 × 10-3
3σ limit at 50% completeness


1:300 (= 29 MJup)
3σ limit at 90% completeness

1:200 (= 47 MJup)
 Reduced performance
beyond 100 mas


Time smearing
FOV limitation
Galland et al. 2006
3rd ARC meeting
Boccaletti et al. 2009
04/03/2011
VLTI/PIONIER observations of Fomalhaut
 New visitor instrument at
VLTI


Developped in Grenoble
First light: Oct 2010
 Combines 4 telescopes
 4 CPs at a time!
 CP stability ~ 0.2°
 First test on Fomalhaut
 2×2h of observation
 Dynamic range up to 1:500
 Corresponds to ΔK=6.7
 Confirmed by double-blind
test
3rd ARC meeting
04/03/2011
Confirmation with double blind test
< 3σ
3rd ARC meeting
> 3σ
04/03/2011
But we do have detections, too!
 Snapshot on δ Aqr



A3V at 50 pc
G5V companion detected
2% contrast  easy
3rd ARC meeting
04/03/2011
Perspectives with PIONIER
 Survey of young main sequence stars
 1:500 around AU Mic (M0V, 10pc, 10Myr)  9 MJup planet
 Fill the AO/corono “blind” hole (< 10 AU)
 Explore the shores of the brown dwarf desert
 Towards hot Jupiters?
3rd ARC meeting
04/03/2011