Imaging Spectroscopy for Extrasolar Planet Detection

Abstract

Coronagraphic imaging in combination with moderate to high spectral resolution may prove more effective in both detecting extrasolar planets and characterizing them than a standard coronagraphic imaging approach. We envisage an integral-field spectrograph coupled to a coronagraph to produce a 3D datacube. For the idealised case where the spectrum of the star is well-known and unchanging across the field, we discuss the utility of cross-correlation to seek the extrasolar planet signal, and describe a mathematical approach to completely eliminate stray light from the host star (although not its Poisson noise). For the case where the PSF is dominated by diffraction and scattering effects, and comprises a multitude of speckles within an Airy pattern typical of a space-based observation, we turn the wavelength dependence of the PSF to advantage and present a general way to eliminate the contribution from the star while preserving both the flux and spectrum of the extrasolar planet. We call this method `spectral deconvolution'. We illustrate the dramatic gains by showing an idealized simulation that results in a 20-sigma detection of a Jovian planet at 2 pc with a 2-m coronagraphic space telescope, even though the planet's peak flux is only 1% that of the PSF wings of the host star. This scales to detection of a terrestrial extrasolar planet at 2 pc with an 8-m coronagraphic Terrestrial Planet Finder (TPF) in ~7 hr (or less with appropriate spatial filtering). Data on the spectral characteristics of the extrasolar planet and hence on its atmospheric constituents and possible biomarkers are obtained naturally as part of this process.