ESPRI science

ESPRI science goals

To date, more than 200 stars are known to have planetary-mass companions (see the Extrasolar Planet Encyclopedia). Almost all of them are solar-type main-sequence stars, i.e., their spectral types, stellar masses, and ages are comparable to that of the Sun, which is the result of a biased sample selection for radial-velocity surveys caused by the applicability of this method. In order to explore the diversity of planet-bearing stars and to overcome the limitations of the radial-velocity method, which measures only one dimension of the stellar reflex motion and provides only a lower limit to the planet masses, existing search programmes have to be complemented and extended with other detection techniques.

Precise astrometry is a powerful tool for detecting planetary companions around stars of different mass (spectral type) and age and for measuring their orbits and masses. The VLTI instrument PRIMA (Phase Referenced Imaging and Microarcsecond Astrometry), when equipped with Differential Delay Lines, will perform high-precision astrometry at the 10 micro arcsecond level, and will be capable of detecting Saturn-mass planets around nearby main sequence stars of any spectral type (down to Uranus-mass planets around nearby M dwarfs) as well as around pre-main sequence stars, which are not suitable for the radial-velocity method. At a distance of 10 parsec, a Saturn-mass planet at 1 AU from the parent star of M = 0.5 M(sun) produces an astrometric signal of ~50 micro arcsec.

The MPIA, Geneva Observatory and Landessternwarte, ZAH (Zentrum für Astronomie) have agreed with ESO to build and deliver the Differential Delay Lines, data reduction software and observing strategies for PRIMA, and then to carry out a planet search program, called Exoplanet Search with PRIMA (ESPRI).

Interferometric phase-referenced (i.e., differential) astrometry relies on a secondary star very close to the target star. The actual target of an astrometric measurement is therefore always a pair (or triple), consisting of the usually bright exoplanet target star (the primary star, which will be used for fringe tracking) and one or more fainter astrometric reference stars (secondary stars).

Both primary and secondary star have to fulfill stringent requirements on separation, limiting magnitude, distance, astrometric stability, physical stellar parameters, and K-band color index difference (for OPD corrections), as described on the Method webpage.

Therefore, extensive preparatory studies have to be carried out to select and characterize well-suited target/reference star pairs, before astrometric measurements at the unprecedented accuracy of 10 micro arcsec can be performed. In particular, no exisiting catalogs or data compilations can provide the information required to ensure astrometric stability of the target/reference star pairs at the 10 micro arcsec level.

Figure 1 — an example target candidate for PRIMA. This target has three possible astrometric reference stars. In a PRIMA observation, the motion of the target star will be compared to that of the nearby reference stars over a period of several years. Exoplanets around the target star can then be detected from the reflex motion of the target star.

Besides a general (but precise) characterization of the planet search target stars, it is important to collect information on, e.g., stellar/substellar companions, stellar activity (starspots, pulsation, etc), of both target and reference stars. In addition to investigating all terrestrial and instrumental effects and requirements, the PRIMA-DDL astrometric planet search science team is therefore carrying out an extensive preparatory program which consists of five principle steps:

1.	Pre-selection of potential target/reference star pairs from existing catalogs according to a well-defined set of pre-selection criteria. In most cases, potential reference stars (for K-band astrometry) can only be pre-selected from surveys at visible wavelengths.
2.	Modeling all sources of systematic and random astrometric effects and explore the parameter space for suitable target/reference star pairs and necessary preparatory observations (thereby using the experience from the SIM preparatory program).
3.	K-band photometric imaging to finally select and characterize target and reference star pairs according to actual brightness and separation requirements for K-band astrometry.
4.	Detailed (mainly spectroscopic) study of all target and reference stars pre-selected until step 3 and fully characterize all relevant stellar parameters and the expected astrometric noise.
5.	Build up a data base for the astrometric planet search program.

We have categorized our target stars for ESPRI into the following groups:

a.	Stars with planets from the radial velocity surveys. The precise astrometric measurement of the host star will allow the characterization of the planetary system. In particular, it allows to obtain the real mass of the planet. Moreover, the astrometric method is complementary to the radial velocity survey, since it is more suitable for the detection of the longer period planets.
b.	Young stars. The majority of young stars possess high-level stellar activity, which render them to be inappropriate for the radial velocity planet search programs. For the active young stars precise astrometry is indeed more effective than the radial velocity method. The astrometric noise due to the stellar activity (spots, pulsations) can be of few micro to milli arcseconds (Launhardt et al., in preparation). Precise astrometry will allow planet search around young stars to address early evolution and formation of the planetary systems.
c.	Nearby stars. Astrometric planet search around these stars is complementary to the radial velocity surveys, especially on the sensitivity and spectral types, especially those which are not (or only very limited) covered by the radial velocity surveys, such as A- , early F-type and low-mass stars.

Following the strategy for the preparation of the astrometric planet search program we have started an extensive spectroscopic study of the target star candidates with FEROS at the 2.2m MPI/ESO within the MPI guaranteed time (step 4). From the spectroscopic measurements we determine the basic stellar parameters of the targets. Many stars in the input catalogs, especially young stars do not have accurate determination of these parameters or there are no sufficient information at all. Radial velocity measurements have been carried out to sort out the spectroscopic binaries from our sample. These objects are not suitable for the astrometric planet search, because they can produce large astrometric noise.

For the last step in the preparatory program we are currently building a target-reference star database (step 5), called ASTRID. Further details are given on the software webpage.