exoplanet* characterization withhighresolution irspectroscopy · exoplanet* characterization...
TRANSCRIPT
Exoplanet characterization with high-‐resolution IR spectroscopy Nikolai Piskunov and Erik Aronson Uppsala University
Transit spectroscopy
Transit spectrophotometry Fi7ng FORS observa=ons of GJ1214b to selected models:
Can we do it without assuming a model?
In principle, yes: Matching observa=ons with tellurics removed against normalized stellar flux minus intensity affected by the planet plus intensity passing through planetary atmosphere: This really does not work: • We subtract two large values that are marginally different • Where telluric features are strong we will be dividing noise by zero
• We have all systema=cs (telluric, flux, intensity and instrumental profile) working directly against us
X
�,�
!�,� ·⇥O�,�/T�,� �
�F�,� � �
core
· I�,� + �atm
· I�,� ·Atr
�
�⌦ �
inst
⇤2
= minX
�,�
!�,� ·⇥O�,�/T�,� �
�F�,� � �
core
· I�,� + �atm
· I�,� ·Atr
�
�⌦ �
inst
⇤2
= min
Equations: alternative Observed flux is normalized by the theore=cal stellar flux and matched against the rela=ve contribu=on of the planet: This is beTer: • Systema=c errors with instrumental profile cancel out for the exoplanet part.
• The telluric spectrum acts as a weight: wavelengths with strong telluric absorp=on have less contribu=on to the total.
• An accurate analy=cal model of telluric spectrum allows reducing the dynamic range between observa=ons and planetary signatures.
X
�,�
!�,� ·⇢
O�,�
F�,� ⌦ �inst
� T�,� ·1� �
core
I�,�F�,� ⌦ �
inst
+ �atm
(Atr
� · I�,�)⌦ �inst
F�,� ⌦ �inst
��2
= min
X
�,�
!�,� ·⇢
O�,�
F�,� ⌦ �inst
� T�,� ·1� �
core
I�,� ⌦ �inst
F�,� ⌦ �inst
+ �atm
(Atr
� · I�,�)⌦ �inst
F�,� ⌦ �inst
��2
= min
Numerical experiments Jupiter-‐size planet in front of a solar-‐type star. Using 120 exposures during a single transit (S/N=250 per exposure)
Numerical experiments
Earth-‐like planet passing in front of an M5 dwarf. BoTom panel: simulated observa=ons with CRIRES+. Top panel: CO2 spectrum in planet atmosphere (blue) and its reconstruc=on from 10 transits (red).
Numerical experiments Resolu=on and spectral coverage: R=100000 R=10000
Test object:HD209458 CRIRES observations
−0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2Phase
−200
−100
0
100
200
Rad
ial V
eloc
ity (
m/s
)
HD 209458 bP = 3.5 d
exoplanets.org
Data reduction: λ-‐scale
Shift to star rest frame
Remove telluric and stellar features
Shift to the planet rest frame … CO
2.324 2.326 2.328 2.330 2.332 2.334Microns
0.997
0.998
0.999
1.000
1.001
We can look at many molecules
… and in other spectral regions, but we are not there yet with current the data
Well, may be there is something there …
CRIRES+ (ready 2017) • Massive increase in wavelength coverage
• BeTer throughput • BeTer sampling • BeTer detectors • Spectropolarimetry • BeTer wavelength calibra=ons
Conclusions Ground-‐based high-‐resolu6on (IR) spectroscopy has a great poten6al for true (model independent) characteriza6on of exoplanetary atmospheres. What we need is: • High (>300) S/N in a short exposure with a good duty cycle • High-‐resolu=on (R>=100000) to see between telluric lines and take advantage of the Doppler shic varia=on between exposures
• High wavelength coverage to look at different species and play with combining lines
• Reliable calibra=ons (wavelength, flat, background, blaze) • Stability on =me-‐scale of a transit • Repeatability to use day-‐=me calibra=ons and connect different transits
This can be achieved with CRIRES+ at the VLT • All of the above for tens of targets! This is a job for the E-‐ELT + HiRes