Fitting exoplanet orbits to epoch astrometry Daniel Michalik
Fitting exoplanet orbits to epoch astrometry Daniel Michalik (Research Fellow @ estec) T. Brandt, M. Brandt, T. Dupuy, Y. Li, M. Chen, J. Faherty
Introduction to astrometry
Gaia: mapping the Milky Way
Gaia space astrometry Milky Way: 100 -300 billion stars Gaia’s goal: map 1 billion in 5 -d (6 -d) Uncertainties of tens of micro-arcseconds No pre-selection – include all point sources 3 < G < 21 Simultaneous astrometry, photometry, and spectroscopy
Astrometric fit Observations are most accurate in scan direction Repetitively scan the sky => avg. 80 transits per star Observations are most accurate in scan direction Fit five astrometric parameters Assumes no companions, a. k. a. single star behaviour
Same principle as Hipparcos ESA mission 1989 – 1993 118, o 00 stars Full epoch astrometry published (1997, and rereduction 2007) Legacy archive in
Unseen companions
Mode A: scan angles and times. Ignore observation coordinates 1)Simulate orbits incl. companions with varying parameters 2)Project covariances at observation epochs onto the sky 3)Compute the resulting proper motions 4)Compare with Hipparcos, Gaia, tgas, etc
Simulated orbit, astrometric fit h (Blue ellipses are much more stretched in reality. )
Different companion period → different proper motion h (Blue ellipses are much more stretched in reality. )
Constrains orbit & mass of Gl 229 B c. f. Brandt, … Michalik, et al. (2019; in review)
Mode B: use individual observations However: must ensure that they are independent data points Also: Reference frame calibration between two different missions is very difficult at the level of individual observations And: epoch astrometry is not yet available for Gaia Thus: limited to within one mission (code is ready for Gaia-dr 4 ; ))
Beta Pic b and c Lagrange et al. (2019) have suggested a second planet around beta Pic Orbital parameters based on RV data
Astrometry with and without second planet Small change in proper motion β Pic c compatible with epoch astrometry (after correcting T ) periastron Can qualify likelihoods of orbital elements
Astrometric fit for varying orbital elements
(1/2) Summary We implemented astrometric fits to simulated orbits plus the Gaia/Hipprcos scan geometry We also fit individual observations (Hipparcos, Gaia-dr 4) A random orbit of beta Pic c has a 20% probability to observationally agree with Hipparcos and Gaia We are using the orbit fit code for a number of scenarios (e. g. see Brandt, Michalik, et al. 2019 on ar. Xiv for the dynamical mass of Gliese 229 B)
(2/2) Outlook: Tycho-2: 2 million stars seen by Hipparcos starmappers 20 x more targets Epoch astrometry not readily available Cross-calibration is tricky
Outlook: Tycho-2: 2 million stars seen by Hipparcos starmappers 20 x more targets Epoch astrometry not readily available Cross-calibration is tricky
- Slides: 24