1. High Resolution Spectroscopy of Stars with Planets Won-Seok Kang Seoul National University 2010. 10. 6. Sang-Gak Lee, Seoul National University Kang-Min Kim, Korea Astronomy and Space science Institute CHEMICAL ABUNDANCE OF PLANET-HOST STAR

2. INTRODUCTION Why we study chemical abundances of host stars –Conserve primordial abundances of planetary systems Related with planet formation process –Find the relation between abundances and planets by observations Describe planet formation process in more detail Select proper candidates with interesting planets –Super-Earths and habitable planets What we can to with GMT high-resolution spectroscopy –Perform abundance analysis for more faint star Transiting planet-host star, M dwarf –Obtain abundances and stellar parameters of more late-type stars Avoid strong molecular bands and pressure-broadened atomic lines 2GMT Workshop 2010 at SNU

3. PLANET AND METALLICITY Fischer and Valenti (2005) I –Spectroscopic analysis of ~1000 stars –For selecting planet-host stars Stars with planets were selected with period 30 m/s (gas giant planets) Stars without planets have been verified by observations of over 10 times for 4 or more years –Calculate the planet-host ratio for each [Fe/H] bin Planet-host ratios are exponentially increasing with increasing metallicity 3GMT Workshop 2010 at SNU Fischer and Valenti 2005

4. PLANET AND METALLICITY Fischer and Valenti (2005) II –Suggest the relation between maximum of total planet mass and metallicity Total planet mass is related with protoplanetary disk mass ⇒ upper limit of total planet mass is increasing with increasing [Fe/H] Planet mass from radial velocity measurement is M J sin i, which means that this planet mass is lower limit of exact value So, need to know exact planet mass 4GMT Workshop 2010 at SNU Fischer and Valenti 2005

5. METHOD OF ABUNDANCE ANALYSIS Observations (166 FGK-type stars) –BOES at BOAO 1.8m telescope –R ~ 30,000 or 45,000 / SNR ~ 150 at 5500Å –Planet-host stars : 93 (74 dwarfs) –Comparisons : 73 (70 dwarfs) ← stars without known planets Abundance analysis –Kurucz ATLAS9 model grids and MOOG code –Measure EWs of Fe lines (TAME developed by IDL) –Determine model parameters by fine analysis (MOOGFE) Iteratively run MOOG code and ATLAS9 –Estimate abundances of 13 elements (Na, Mg, Al, Si, Ca, Sc, Ti, V, Cr, Mn, Co, Ni, S) Measuring EWs of elemental lines (TAME) Comparing observational spectrum with synthetic spectrum – 5GMT Workshop 2010 at SNU

6. METHOD OF ABUNDANCE ANALYSIS TAME and MOOGFE 6GMT Workshop 2010 at SNU Tools for Automatic Measurement of Equivalent-widths Result of MOOGFE for the Sun Excitation potential Equivalent-width Fe IFe II Model parameters log eps(Fe) = 7.53 dex T eff = 5765 K log g = 4.46 dex ξ t = 0.82 km/s For accurate estimation, we selected only weak lines of Fe I Automatically find model parameters by iterations By estimating the trend of iron abundance for excitation potential or equivalent width, and the abundance difference between Fe I and Fe II

7. METALLICITY HISTOGRAM Metallicity distribution –Mean value of PHS is 0.13 dex higher than that of comparison –Planet-Host Stars are more concentrated at higher [Fe/H] Comparisons are more widely distributed overall In low-metallicity, comparisons are more than PHSs In high-metallicity, PHSs are much more than comparisons 7GMT Workshop 2010 at SNU -0.06 +0.07 Only dwarfs 74 PHSs 70 Comparions

8. METALLICITY AND PLANET PROPERTIES [Fe/H] and Planet mass, M J sin i –Increase with increasing [Fe/H] Similar result to Fischer and Valenti (2005) –HD 114762 Known as spectroscopic binary Companion is M6 dwarf at the distance of 130 AU Exceptional case or new evidence? –For verifying, more samples in the range of low-metallicity will be required 8GMT Workshop 2010 at SNU In the case of multiple-planetary system, total planet mass is indicated These planetary masses represent M J sini, which is less than M J HD 114762 b Only dwarfs [Fe/H] vs. Planet Mass Only 4 samples

9. METALLICITY AND PLANET PROPERTIES 9GMT Workshop 2010 at SNU Only dwarfs Metallicities and Planet properties –Hot jupiters are concentrated in the region of [Fe/H] > 0 It can support the relation between migration and metallicity (Livio & Pringle, 2003) A Few stars in low-metallicity region –In the region of low-metallicity, about half of host stars have relatively low- mass multiple planets. 2 of 5 planet-host stars have low- mass multiple planets X : semi-major axis Y : [Fe/H] Size of circle : planet mass HD 114762 b

10. ABUNDANCE RESULTS [X/Fe] vs. [Fe/H] –Averaged for each [Fe/H] bin –For most elements, statistical difference between two groups ~ 0.03 dex [Mn/Fe] ratio –Difference between two groups ~ 0.10 dex –Hyperfine structure It is necessary to confirm this difference by synthetic spectra and high S/N observational spectra 10GMT Workshop 2010 at SNU Chemical Abundance Trend ; [Fe/H] vs. [X/Fe] Only dwarfs Red : Planet-Host Stars Blue : Comparisons Bin size : 0.2 dex Center of each [Fe/H] bin : -0.5, -0.3, -0.1, +0.1, +0.3, +0.5

11. ABUNDANCE RESULTS 11GMT Workshop 2010 at SNU HD 114762 b [Mn/H] vs. planet mass [Mn/H] and Planet mass –Maximum of planet mass are increasing in low [Mn/H] range and decreasing in high [Mn/H] range –Turn-off point of trend is located at solar Mn abundance –It seems that the high [Mn/H] ratio has suppressed the massive planet formation

12. DIFFICULTIES 12GMT Workshop 2010 at SNU Most of planets were detected by radial velocity method –Don’t know exact mass of planet –Samples are limited to almost nearby stars –Solution ; transiting planet Transit observation gives us more accurate mass of planet Transit observation is available for faint and distant stars Lack of low-metallicity star –More low-metallicity stars are required to verify the relation between planet properties and abundances –It seems to be easier to find Neptune-mass planets in low-metallicity stars (Sousa et al. 2009) They have only three Neptune-mass samples Expect that more low-metallicity stars will be detected, in the near future

13. PRELIMINARY TEST 13GMT Workshop 2010 at SNU Transiting planets Radial velocity method Homogeneous studies of 30 transiting extrasolar planets (Southworth, 2010) –Provide the properties of planets and host stars Test the relation for only transiting planets –Maximum of planet mass is decreasing with increasing [Fe/H] –Inverse trend for the previous result of samples detected by radial velocity method –Problems No stars of metallicity less than -0.2 dex It seems that there are two groups of planet mass Metallicities were adopted from several references –Solutions More low-metallicity stars with transiting planets Perform abundance analysis in uniform method and with the same instrument

14. WHAT WE CAN DO WITH GMT Detailed abundances of host stars with transiting planets –Potential to detect new transiting planets in the near future HATNet, Kepler, CoRoT, SuperWASP, SWEEPS –There are already 37 planets detected by transit in this year –Host stars are relatively faint, V ~ 10-15 –Magnitude limit of transit observations will be fainter ⇒ GMT 14GMT Workshop 2010 at SNU http://exoplanet.eu Number of planets by year of discovery 2010 (37) 2009 (10) 2008 (17) 2007 (19)

15. WHAT WE CAN DO WITH GMT Abundances of M dwarfs using GMTNIRS –Advantages 1.Easy to detect new exoplanets or extraterrestrial lives –Host star is less massive (radial velocity method) »Less massive exoplanets (super-earths) –Habitable zone is closer to host star (extraterrestrial life) »Short period and probability of tra nsits 2.Life time in the stage of main-sequence –Enough time for life evolution 3.The large number of M dwarfs in the Galaxy –Disadvantages 1.Faint at visible wavelength ⇒ large telescope, GMTNIRS 2.Strongly pressure-broadened atomic lines, and strong molecular bands in visual wavelength range ⇒ GMTNIRS 15GMT Workshop 2010 at SNU

16. NIR spectroscopy –Spectrum of planetary atmosphere by transiting event SNR ∝ (2πR P ·h) / πR S 2 h : thickness of atmosphere Gl 581 (M3V, R S ~ 0.4R ⊙ ) V=10.6, J=6.7, H=6.1, K=5.8 –Earth( R P ~ 0.01R ⊙, h ~ 20 km ) ⇒ Factor ~ 3 × 10 -6 –Jupiter( R P ~ 0.1R ⊙, h ~ 2000 km ) ⇒ Factor ~ 3 × 10 -3 GMTNIRS can make it possible Band  m) RS/N=100 (1 hr) S/N=10 (1 hr) J1.2350,00015.518.8 H1.6350,00015.518.8 K2.2250,00015.118.5 L3.47100,00011.414.6 M4.89100,0009.111.9 GMTNIRS S/N (AO case) WHAT WE CAN DO WITH GMT