On the spectroastrometric separation of binary point-source fluxes

¹ Astrophysics Research Institute, Liverpool John Moores University, Twelve Quays House, Egerton Wharf, Birkenhead CH41 1LD, UK e-mail: jmp@astro.livjm.ac.uk
² School of Physics and Astronomy, University of Leeds, Leeds LS2 9JT, UK
³ Astrophysics, University of Oxford, Denys Wilkinson Laboratory, Keble Road, Oxford OX1 3RH, UK

Received: 14 November 2003
Accepted: 11 August 2004

Abstract

Spectroastrometry is a technique which has the potential to resolve flux distributions on scales of milliarcseconds. In this study, we examine the application of spectroastrometry to binary point sources which are spatially unresolved due to the observational point spread function convolution. The technique uses measurements with sub-pixel accuracy of the position centroid of high signal-to-noise long-slit spectrum observations. With the objects in the binary contributing fractionally more or less at different wavelengths (particularly across spectral lines), the variation of the position centroid with wavelength provides some information on the spatial distribution of the flux. We examine the width of the flux distribution in the spatial direction, and present its relation to the ratio of the fluxes of the two components of the binary. Measurement of three observables (total flux, position centroid and flux distribution width) at each wavelength allows a unique separation of the total flux into its component parts even though the angular separation of the binary is smaller than the observations' point-spread function. This is because we have three relevant observables for three unknowns (the two fluxes, and the angular separation of the binary), which therefore generates a closed problem. This is a wholly different technique than conventional deconvolution methods, which produce information on angular sizes of the sampling scale. Spectroastrometry can produce information on smaller scales than conventional deconvolution, and is successful in separating fluxes in a binary object with a separation of less than one pixel. We present an analysis of the errors involved in making binary object spectroastrometric measurements and the separation method, and highlight necessary observing methodology.