The experience of BEST Heike Rauer and the

The experience of BEST Heike Rauer and the BEST Team Institut für Planetenforschung Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR) and Zentrum für Astronomie und Astrophysik Technische Universität Berlin +

The experience of BEST Berlin Exoplanet Search Telescope System Goals of BEST: - support for Co. Ro. T - detect large planets - variable stars, additional science

Berlin Exoplanet Search Telescope Specifications: Telescope Aperture Focal ratio Instrument Size Pixel scale Field of view Schmidt-Cassegrain 20 cm f/2. 7 AP-10 CCD 2048 x 2048 pixels 14 µm 5. 5 arcsec/pixel 3. 1° x 3. 1° 2001 - 2004 Thüringer Landessternwarte Tautenburg (TLS), Germany Since end 2004 Observatoire de Haute Provence (OHP), France

BEST I TEST at TLS BEST II

Observatorio Cerro Armazones, Chile Instituto de Astronomía Universidad Católica del Norte (UCN) in Antofagasta, Chile Astronomisches Institut- Ruhr. Universität Bochum (RUB), Germany.

BEST II Specifications: Telescope Aperture Focal ratio Instrument Size Pixel scale Field of view Precision < 1% BRC - 250 25 cm f/5. 0 FLI IMG-1680 CCD 4096 x 4096 pixels 9 µm 1. 5 arcsec/pixel 1. 7° x 1. 7° V=15 -16 smaller Fo. V for BEST II is compensated by less stars influenced by crowding

Modes of operation • BEST I at TLS: - obervations by observer at TLS - data reduction at DLR • BEST I at OHP: - observations via remote control from Berlin - data reduction at DLR • BEST II at OCA - „robotic“ observations (regular remote monitoring, manual interaction in case of alarm) - basic calibration at OCA, full data reduction at DLR

Performance The two critical factors for a transit search system are: 1. High duty cycle: full coverage of planetary orbits by observations. 2. 2. Large number of high quality lightcurves (e. g. rms < 1%).

Duty cycle Need: High duty cycle, full coverage of planetary orbits by observations of sufficient quality. BEST experience: duty cycle is the major limiting factor from central Europe (not a surprise ). Next: place BEST II at OCA, Chile start building a network (NEST) participate to ASTEP

Performance The two critical factors for a transit search system are: 1. High duty cycle, full coverage of planetary orbits by observations 2. 2. Large number of high quality lightcurves (e. g. rms < 1%) 3. - correction for detector effects (dark, bias, flats, hot/cold/defect pixels, …) 4. - correction of atmosphere (extinction, seeing, scientillation, …) 5. - accurate photometry (aperture/image subtraction/PSF fitting,

Detector effects For example: a star moves across a hot pixel during the night due to imperfect guiding of the telescope…. Causes transit-like signal which has to be evaluated by comparison with the original data. adds work-load on transit candidate evaluation

Detector effects Correction for detector effects (dark, bias, flats, hot/cold/defect pixels, …) - Low-quality CCD: need to check transit events for detector effects, check position of the star on CCD - varying bias, dark, etc. , adds to systematic noise residuals Recommendation: buy good h/w adapt the observing sequence to calibration needs

Atmosphere effects Correction of atmosphere (extinction, seeing, scintillation, …) - Airmass correction is critical (no filter, large FOV) restriction in airmass, depending on site and target field adapt reduction method, e. g. work on sub-fields implement filter if possible - Effect of seeing variations on crowding

Photometry Accurate photometry (aperture/image subtraction/PSF fitting, crowding) - Crowding can be a major problem improve photometric method (image subtraction ok) match the pixel scale of h/w

B. E. S. T. Candidate 3 BEST POSS-I BEST Magnitude of host star Depth [%] Duration [h] Orbital period [d] Number of detections 12. 1 2. 5 3. 0 423. 10/n 2

BEST candidate 5 depth [%] duration [h] orbital period [d] semi mayor axis[AU] number of detections target field No. host star magnitude(BEST) V magnitude reference star BEST 18. 3‘ x 18. 3‘ 1. 0 4. 5 > 10 ? ? 1 8 M dwarf 12. 56 14. 73

Crowded target fields lead to further reduction of the photometric accuracy a) diluted signals b) c) - neighboring stars are resolved - but a neighbor contributes flux within the PSF or photometric aperture d) b) unresolved stars - neighboring stars are not resolved c) a combination of a) and b) - due to varying seeing the resolution of stars changes over the night The photometric data reduction algorithm needs to be adopted. a transit signal is weakend noise is added if the neighbor is variable

Comparison of Photometric methods * Source-Extractor, SExtractor (Bertin & Arnouts 1996) Performs different kinds of aperture photometry * Multi Object Multi Frame photometry, MOMF (Kjeldsen & Frandsen 1992) Combination of aperture and PSF photometry * Image Subtraction, ISIS (Alard 2000) Subtraction of a convolved reference frame from all frames. Karoff et al. 2005 implementation of parallel approach in data pipeline (SExtractor, ISIS)

SExtractor used in less crowded fields SExtractor TLS

Comparison of both methods for the COROT field Data from 4 nights in spring 2005 obtained at OHP (COROT winter field) SExtractor ISIS Karoff et al. 2005 a reduction routine able to deal with very crowded target fields is important

Performance of BEST I at OHP after evaluation of its first regular observing season at OHP in summer 2005.

BEST at OHP: Variable stars in the Co. Ro. T center field 37 nights/142 hours observations from OHP in summer 2005. Search for variable stars and eclipsing binaries: • 83 periodic variable stars identified: 76 new discoveries • 11 variable stars with period < 120 days known in GCVS, 8 confirmed variables, for 3 stars no variability found • 37 of the variables are eclipsing binaries Karoff et al. 2006

Location of the variables in the center field mag # Amin- Amax [mag] 10 - <11 24 0. 01 -0. 29 12 - <13 27 0. 02 -0. 20 13 - <14 - There are many more stars with high rms: real but non-periodic variables and distorted lightcurves 0. 02 -0. 12 11 - <12 - Periodic variable stars are detected over the whole magnitude range. 7 16 0. 05 -0. 18 14 - <15 3 0. 10 -0. 19 See Karoff et al. 2006

How complete is our search for variables from OHP?

Number of detected eclipsing binaries as a function of period Eclipsing binaries in Hipparcos data Söderhjelm & Dischler 2005 Eclipsing binaries in BEST observations Karoff et al. 2006 The BEST survey is complete only up to 10 – 20% for periods > 1 day. But: we have indications that the detection algorithm for periodic variables is not perfect…

BEST – basic lessons from TLS and OHP Strong points: Ø Simple robust system with good capability for long-term photometric surveys of a substantial number of stars Ø Well suited to catalog bright stars in the COROT fields Ø Demonstrated ability to reach the accuracy limit needed to discover Jupiter-sized planets Ø Cost efficient way to characterize stellar variability over long time periods Weak point: Ø Transit detections limited by crowding and duty cycle (i. e. to few nights and/or to few stars)