Transcript Stellar Spectroscopy during Exoplanet Transits

KVA
Stellar Spectroscopy during Exoplanet Transits
Revealing structures across stellar surfaces
Dainis Dravins1, Hans-Günter Ludwig2, Erik Dahlén1, Hiva Pazira1
1 Lund
Observatory, Sweden, 2 Landessternwarte Königstuhl, Heidelberg, Germany
www.astro.lu.se/~dainis
AN ”IDEAL” STAR ?
Very quiet solar disk
(GONG/Teide)
A REAL STAR !
Granulation near the limb (towards the top) at 488 nm; Swedish 1-m solar telescope, La Palma
MODELING STELLAR SURFACES –White dwarf vs. Red giant
Snapshots of emergent intensity during granular evolution on a 12,000 K white dwarf (left)
and a 3,800 K red giant. Horizontal areas differ by dozen orders of magnitude: 7x7 km2
for the white dwarf, and 23x23 RSun2 for the giant. (H.-G. Ludwi,g, Heidelberg)
What future stellar astrophysics ?
Stellar abundances accurate to ±0.01 dex ?
Differential chemistry in stars with or without planets?
Accurate stellar oscillations ?
Stellar sizes seem not consistent if deduced from naïve models.
Stellar differential rotation ?
Stellar gas dynamics are deduced from subtle signatures of line profiles.
Magnetic and chromospheric activity ?
How do magnetic fields affect stellar convection ?
Exoplanet signatures ?
Exoplanet properties are deduced differentially to the stellar spectrum.
Precision stellar physics requires 3-D modeling
But simulations are complex. How can one verify such models ?
How to verify or
falsify 3-D models ?
Line profiles from 3-D Hydrodynamic simulations
Model predictions insensitive to modest spatial smearing
Spatially averaged
line profiles from
20 timesteps, and
temporal averages.
 = 620 nm
 = 3 eV
5 line strengths
GIANT STAR
Teff= 5000 K
log g [cgs] = 2.5
(approx. K0 III)
Stellar disk center;
µ = cos  = 1.0
(Models by Hans-Günter Ludwig, Landessternwarte Heidelberg)
Line profiles from 3-D Hydrodynamic simulations
Model predictions insensitive to modest spatial smearing
Spatially and temporally
averaged line profiles.
 = 620 nm
 = 1, 3, 5 eV
5 line strengths
GIANT STAR
Teff= 5000 K
log g [cgs] = 2.5
(approx. K0 III)
Stellar disk center;
µ = cos  = 1.0
(Models by Hans-Günter Ludwig, Landessternwarte Heidelberg)
Line profiles from 3-D Hydrodynamic simulations
Model predictions insensitive to modest spatial smearing
SUN
Profiles from CO5BOLD solar model; Five line strengths; excitation potentials  = 1, 3, 5 eV.
Left: Solar disk center µ = cos = 1. Right: Off-center disk position µ = cos = 0.59.
(Models by Hans-Günter Ludwig, Landessternwarte Heidelberg)
Simulated
intensities
approaching
the solar limb
Mats Carlsson, Oslo;
in
Å.Nordlund, R.F.Stein,
M.Asplund:
Solar Surface Convection,
Liv.Rev.Solar Phys. 6, 2
Spectral line profiles across stellar disks
Spectral lines, spatially and temporally averaged from 3-D models, change their strengths,
widths, asymmetries and convective wavelength shifts across stellar disks, revealing details
of atmospheric structure. These line profiles from disk center (µ = cos = 1) towards the
limb are from a CO5BOLD model of a main-sequence star; solar metallicity, Teff = 6800 K.
(Models by Hans-Günter Ludwig, Landessternwarte Heidelberg)
Spatially resolving stellar surfaces

Hiva Pazira, MSc thesis, Lund Observatory (2012)

Stellar Spectroscopy during Exoplanet Transits
* Exoplanets successively hide segments of stellar disk
* Differential spectroscopy provides spectra of those
surface segments that were hidden behind the planet
* 3-D hydrodynamics studied in center-to-limb variations of
line shapes, asymmetries and wavelength shifts
* With sufficient S/N, also spectra of surface features
such as starspots may become attainable
Exoplanet transit geometry
G.Torres, J.Winn, M.J.Holman: Improved Parameters for Extrasolar Transiting Planets, ApJ 677, 1324
HD 209458
Promising star for spatially resolved spectroscopy
Spectral type: G0 V, Apparent magnitude mV = 7.65
Teff = 6100 K, log g [cgs] = 4.50, [Fe/H] = 0
Vrot  4.5 km/s; slow rotator, comparable to Sun
sin i = 1 if the star rotates in same plane as transiting planet
Sufficiently similar to Sun for same spectral identifications.
Somewhat hotter, lines somewhat weaker, less blending.
Large planet: Bloated hot Jupiter, R = 1.38 RJup.
More vigorous convection for line differences to be detectable?
Spectrum of HD209458 very similar to solar
Detailed similarities between the spectrum of the G0 V star HD209458 and the
well-studied solar spectrum enable numerous line identifications
Challenge of extremely high S/N
* Retrieving good spectra from behind exoplanet
covering 1% of star, requires S/N 10,000 !!
* Transit time short: Largest telescopes required
* High wavelength stability desired but loses light
* High spectral resolution desired but loses light
* Single spectral lines insufficient: Need averages
HD 209458 extensively observed
Several observatory archives searched
(ESO/VLT/UVES, ESO/3.6m/HARPS, Keck/HIRES, SUBARU/HDS, etc.)
Data used here originate from one ESO/VLT/UVES program:
Program ID: 077.C-0379(A), “A POWERFUL NEW METHOD
TO PROBE THE ATMOSPHERE OF EXOPLANET HD209458B”
PI/COI: SNELLEN/ COLLIER CAMERON/ HORNE
Observing date: 14.08.2006
UVES spectrometer slitwidth 0.50 arcsec
Spectral resolution  60,000
Maximum S/N = 480
Averaging many ‘similar’ photospheric lines
26 Fe I ‘similar’ photospheric lines in HD 209458
Their average ‘synthesizes’ a representative profile (ESO/VLT/UVES/REDL)
Changes in representative profile during transit
14 successive exposures during [part of] exoplanet transit
Line profile is average of 26 Fe I photospheric lines
Simulated line changes during exoplanet transit
Line profile changes during exoplanet transit. Red: Ratios of line profiles relative to the profile
outside transit. This simulation sequence from a CO5BOLD model predicts the behavior of an
Fe I line ( 620 nm,  = 3 eV) during the first half of a transit across the stellar equator by a
bloated Jupiter-size exoplanet moving in a prograde orbit, covering 2% of a main-sequence
star with solar metallicity, Teff = 6300 K, rotating with V sin i = 5 km/s.
Observed line changes during exoplanet transit
Retrieving spatially
resolved stellar
line profiles
Reconstructed line profiles: H-beta
Reconstructed line profiles: H-alpha
Stellar Spectroscopy during Exoplanet Transits
* Now: Marginally feasible with, e.g., UVES @ VLT
* Immediate future: PEPSI @ LBT
* Near future: ESPRESSO @ VLT
* Future: HIRES @ E-ELT ?
Anytime soon: More exoplanets
transiting bright stars 