Precision Radial Velocities in the Infrared

Abstract

Over 250 extra-solar planets have been discovered to date using a variety of techniques. The majority have been discovered at optical wavelengths from the Doppler shift of F, G and K stars induced by orbiting planets. We have constructed models simulating likely planets around M dwarfs and demonstrated the ability to recover their radial velocity signals in the infrared. We have conducted experiments in the infrared with a brass-board instrument to explore real-world issues. We are thus confident that a stabilised radial velocity spectrometer with a single-shot 1 and 1.7 microns coverage at a resolution of around 70 k can achieve an instrumental radial velocity error of 0.5 m/s. This enables the efficient measurement of radial velocities for M, L and T spectral classes. We have modelled the radial velocity information in low-mass star spectra and checked our ability to recover this signal in the face of the telluric contamination in the infrared. Including instrumental error, telluric contamination and photon noise we predict a total radial velocity error of less than 2 m/s on a typical M6V star at 10 pc. We use these results as an input to a simulated 5-year survey of nearby M stars envisaged for Gemini Observatory. Based on a conservative scaling of optical results, such a survey has sensitivity to detect several terrestrial mass planets in the habitable zone around nearby stars. It can test theoretical planet formation models, which predict an over-abundance of terrestrial-mass planets around low-mass stars. Improvements in the efficiency and sampling of searches at optical wavelengths promise long-term precisions of 0.5 m/s and 5 MEarth detections around solartype stars. While this may be the limit for CCD-based surveys of solar type stars until larger telescopes become available it is nonetheless feasible to survey lower mass primaries to achieve a corresponding smaller mass limit. Thus the lowest mass Doppler signals have been found around M dwarfs (e.g., GJ581b 5 MEarth sin i [1]) and detections down to a few MEarth detections should be feasible around mid-type M dwarfs. This is the Precision Radial Velocity Spectrometer (PRVS) approach: to search around lower-mass primary stars since the radial velocity signal will be larger for lighter primary stars (http://www.roe.ac.uk/ukatc/projects/prvs). Based on the spectral information in real and synthetic spectra of M dwarfs, PRVS is designed to measure the peak of their energy distribution. We find that for low rotation M dwarfs a resolution of around 70 000 is optimum. This leads to us