Reading list... Fischer et al. chapter in PPVI, Exoplanet Detection Techniques. Chapter 2 in Exoplanets (Fischer & Lovis) pg

Transcription

1 Reading list... Fischer et al. chapter in PPVI, Exoplanet Detection Techniques Chapter 2 in Exoplanets (Fischer & Lovis) pg (RV method)

2 Radial Velocity (RV) Detection: Early 20th century RV precision ~ 1 km/s By late 80s / early 90s ---> 10 m/s (3 m/s by 1995, detection of 51 Peg b) By 2005, ~ 1 m/s.

3 RV discoveries (up to 2014) Fischer et al. also transits

4 General orientation of orbit: = descending node orbiting body z r ν(t) pericentre reference plane ellipse focus centre of mass Ω = longitude of ascending node ω i ϒ = reference direction = ascending node apocentre orbit plane to observer reference plane tangent to celestial sphere

5 circular VCM V2 V2 V1 VCM V

6 unknowns

7 Still the circular (e = 0) case Usually express the observable (RV semi-amplitude) K as function of P (or a), Mpsin(i), M*. K is directly proportional to Mpsin(i).

8 = descending node orbiting body z r ν(t) pericentre reference plane ellipse focus centre of mass Ω = longitude of ascending node ω i ϒ = reference direction = ascending node t 3 centre of mass acceleration = GM p /r 2 t 1 acceleration = GM /r 2 M p t 2 apocentre orbit plane to observer t 2 M t 1 t 3 (rad. vel. semiamplitude) note: reference plane tangent to celestial sphere

9 systemic vel. + systematics e = 0.03, ω = 340 e = 0.53, ω = 170 e = 0.90, ω = 42 Radial velocity Orbital phase Orbital phase Orbital phase

10

11 mas / 10 pc 0.2 yr 12 yr μas / 100 pc 1 mas / 100 pc Planet mass, M p (M J ) μas / 10 pc 30 nas / 100 pc J S 1% m/s U m/s 0.01% E Semi-major axis, a (AU) detection limits obtained from earlier equations.

12 minimum mass mpsin(i) If planetary systems are randomly oriented (i between 0 and 90 o ): Average of sin(i) = 2/pi ~ 0.6 probability of i being between two values: 87% chance that mpsin(i) is within a factor of 2 of actual mp

13 start here

14 Measuring RVs (basic idea) Radial&Velocity&Technique&

15 Doppler Shift Relativistic terms are usually ignored. However, they can vary by ~ 0.1 m/s over a year (Earth s orbit). note: vel. > 0 for motion away from observer

16 must correct for Earth s motion (barycentric correction) This requires precise clocking of observations (including precise knowledge of where your telescope is located).

17 Measuring RVs (basic idea) You must have a reference spectrum to determine the wavelength calibration. use telluric lines (Griffin & Griffin, 1973) gas-cell (Walker & Campbell, 1979, HF), Iodine is the modern choice. HIRES/Keck simultaneous references spectrum, Thorium-argon lamp (lamp + star spectrum recorded simultaneously) ELODIE, SOPHIE,CORALIE, HARPS other fancy methods...

18 cross-dispersed echelle spectrograph wavelength orders star+lamp spectrum (51 Peg) Lamp Spectrum

19 Quality Factor and theoretical RV limit totally made up example! Bouchy et al. 2001

20 Bouchy et al Quality Factor and theoretical RV limit

21 Spectral Resolution: Resolving Power: R = λ / Δλ = c / Δv R =100,000 --> dlam ~ 0.05 angst, in the optical. 1 m/s shift ---> 1/3000 of a resolution element, for R ~ 100,000. Typical RV instruments disperse light such that 1 m/s doppler shift corresponds to a shift on the detector of ~ 1/1000 of a pixel.

22 Quality Factor and theoretical RV limit Q versus Spectral Resolving Power

23 Measuring tiny fractions of a pixel requires many spectral lines. Usual spectrum might contain ~ 1000 good lines. Modeling and cross-correlation techniques allow greater RV precision than is achieved on a single line. 30 m/s per line can result in 1 m/s (for high SNR spectrum with many lines).

24 cross-correlation of many orders k j low SNR échelle orders 1.0 i h g f e d c b a high SNR cross-correlation function Radial velocity, V r (km s 1 )

25

26

27 RV precision depends on: number of spectral lines / wavelength FWHM of the lines (contrast with continuum) signal-to-noise (SNR) of the data stability of the instrument and wavelength reference

28 The goals are to measure relative RV (not absolute) to high precision and have repeatability, night-to-night, for many years. The instrument must be ultra-stable or the calibration near-perfect and repeatable (preferably both).

29 What are the things you might control? changes in spectrograph variation in instrument illumination

30 What is the source of this wavelength shi9 (~ 1.3 angst.)? 30

31 What is the source of this wavelength shi9 (~ 1.3 angst.)? n(air) ~

32 Wavelength shi9s are around 200 to 300 m/s / K change in temperature (at fixed P) and ~ 100 m/s / mbar change in pressure (at fixed T). need climate- controlled vacuum chamber 32

33 A 1-meter optical bench made of aluminum expands or contracts by ~ 20 microns for every 1 K change in temperature. Thermal stability 20 microns is about the size of one CCD pixel. Such shifts are comparable to, or larger than, the RV change one wants to measure.

34 Image Stability slit wavelength A changes in the illumination across the entrance of the spectrograph can produce wavelength shifts exceeding 100 m/s. Telescope guiding never good enough.

35 Fiber-scramblers: < 10 um (perfect scramble) um (good scramble) input perfect multi-mode fiber (1 is good, 2 is better, aka double-scrambler) output (intermediate) output (to spectrograph) (Queloz 1999)

36 Very stabilized instruments: 3.6m at ESO/Chile Spectrograph on a rigid bench, which is housed in a vacuum tank.

37 Very stabilized instruments: 3.6m at ESO/Chile Spectrograph on a rigid bench, which is housed in a vacuum tank. The tank itself is housed in a climate controlled room that is never opened. Pressure controlled to 10-3 mbar OpOcal bench controlled to 1 mk

38 Very stabilized instruments: 3.6m at ESO/Chile Spectrograph on a rigid bench, which is housed in a vacuum tank. The tank itself is housed in a climate controlled room that is never opened. Light is coupled from the telescope with fiber opocs that scramble the light.

39 Very stabilized instruments: 3.6m at ESO/Chile Spectrograph on a rigid bench, which is housed in a vacuum tank. The tank itself is housed in a climate controlled room that is never opened. Light is coupled from the telescope with fiber opocs that scramble the light. A second fiber feed a simultaneous calibraoon source.

40 Very stabilized instruments: 3.6m at ESO/Chile observed spectrum Cross- CorrelaOon Technique Original reference: Tonry & Davis 1979, AJ, 84, 10 CCF template spectrum radial velocity

41 Gas cell technique

42 Radial Velocity Technique

43 Radial Velocity Technique

44 Radial Velocity Technique iodine lines

45 Gas cell technique: It is a lidle more complicated (involves forward modeling of spectrum) I 2 star (template) instrumental profile Obs & Model Butler et al. 1996, PASP, 108, 500

46 Calibrated instruments: 10m Keck telescope in Hawaii Spectrograph not parocularly stabilized.

47 Which stars to observe? 47

48 Vega Optical

49 Solar-type (The Sun) stars:

50

51 A, F, G, & K-types M dwarfs

52 quantum-efficiency

53 quantum-efficiency

54 Rotational Broadening: Normalised flux HD 1404 (A2V) 110 km/s HD (A3V) 115 km/s HD 6961 (A7V) 91 km/s HD (F0V) 63 km/s HD (F5V) 110 km/s HD (F6V) 5 km/s Mp sin i (MJ) A0V 150 km/s A0V 80 km/s A0V 40 km/s A5V 150 km/s A5V 80 km/s A5V 40 km/s Wavelength (nm) Wavelength (nm) Wavelength (nm) Period (d)

55 young stars have larger rotational velocities, and are more active.

56 Rotational Broadening (vs SpT): Affer et al. (2008)

57 spectral richness near peak of QE low Vsin(i) metal-rich old FGK-type = sweet-spot

58

59 Problem: areas on the surface with different velocioes and brightnesses + changes with Ome

60 P- modes (acousoc pressure waves) The sun oscillates with a characterisoc Omescale of five minutes. The Omescale scales as sqrt(density), so lower- mass stars have longer Omescales, and higher- mass stars have shorter Omescales. Amplitude is approximately 1 m/s.

61 GranulaOon ConvecOve cells on the surface. The bright center is the upwelling of hot gas, and the dark edges are the downward mooon of cooler gas. Relative intensity granular region intergranular region upper region lower region { { line bisector Length ~ 1000 km Timescale ~ 10 minutes Velocity ~ 1 km/s Number ~ 10 6 Wavelength Wavelength

62 Spots

63 A few words about Fourier Transforms Fourier Transform Inverse Fourier Transform

64 see Gray, 3rd ed. chapter 2

65 The professor s watch Amplitude Time

66 The professor s watch Power Frequency (Hz)

67 Sub- saturn mass planet HD3651 Fischer et al. (2003) Power Spectrum (or Lomb-Scargle periodogram, Lomb 1976)

68

69

70 MulZ- planets Example: Example from HARPS southern survey, some planets orbiong acove star BD (Hebrard et al. 2010)

71 MulZ- planets Example: How many planets do you see?

72 MulZ- planets Example: How many planets do you see?

73 MulZ- planets Example: With 238 day signal removed

74

75 Combined Fits to the data

76 What about the planet at 1 day?

77 This is an active star

78 Seven planets around a single star (HD10180) Lovis et al. 2011, A&A, 528, 112

79 Seven planets around a single star (HD10180) Lovis et al. 2011, A&A, 528, 112

80 AcOvity and RVs HD166435: spots masquerading as a planet! Line- bisector mean line- profile Queloz et al. 2001, A&A, 379, 279

81 AcOvity and RVs HD166435: spots masquerading as a planet! Queloz et al. 2001, A&A, 379, 279

82 RV planets as a funczon of date