The scanning of SSTR by a sampling method

Abstract

A technique for scanning etched track recorders is described in which the track density is estimated by sampling the surface at a fixed grid of points, and counting the proportion of points which is covered by features. The technique is applicable to track densities up to at least an order of magnitude higher than can be analyzed by conventional counting. The probability distribution for the number of grid points covered by tracks and background features has been derived for arbitrary distributions of track area and background features. A method of estimating the track density, together with its uncertainty, is proposed. Predictions of the statistical model are compared with experimental results obtained using alpha tracks in CR-39. The validity of the technique has been established to an uncertainty level of less than 1.0%, for track recorders in which the fraction of the surface covered by tracks ranged from 0.01 to 0.99.