2013A&A...551A..52S

The ESPRI project relies on the astrometric capabilities offered by the PRIMA facility of the Very Large Telescope Interferometer for discovering and studying planetary systems. Our survey consists of obtaining high-precision astrometry for a large sample of stars over several years to detect their barycentric motions due to orbiting planets. We present the operation's principle, the instrument's implementation, and the results of a first series of test observations. We give a comprehensive overview of the instrument infrastructure and present the observation strategy for dual-field relative astrometry in the infrared K-band. We describe the differential delay lines, a key component of the PRIMA facility that was delivered by the ESPRI consortium, and discuss their performance within the facility. This paper serves as reference for future ESPRI publications and for the users of the PRIMA facility. Observations of bright visual binaries were used to test the observation procedures and to establish the instrument's astrometric precision and accuracy. The data reduction strategy for the astrometry and the necessary corrections to the raw data are presented. Adaptive optics observations with NACO were used as an independent verification of PRIMA astrometric observations. The PRIMA facility was used to carry out tests of astrometric observations. The astrometric performance in terms of precision is limited by the atmospheric turbulence at a level close to the theoretical expectations and a precision of 30µas was achieved. In contrast, the astrometric accuracy is insufficient for the goals of the ESPRI project and is currently limited by systematic errors that originate in the part of the interferometer beamtrain that is not monitored by the internal metrology system. Our observations led to defining corrective actions required to make the facility ready for carrying out the ESPRI search for extrasolar planets.