Advantages and challenges of automated apatite fission track counting

Abstract

Fission track thermochronometry data are often a core element of modern tectonic and denudation studies. Soon after the development of the fission track method, interest emerged for the development of automated counting procedures to replace the time consuming labor of counting fission tracks under the microscope. Automated track counting became feasible in recent years with increasing improvements in computer software and hardware, and the development of microscopes and stage systems. One such example examined in this study is the commercial automated fission track counting procedure from Autoscan Systems Pty. and the University of Melbourne thermochronology group. We conducted experiments that are designed to reliably and consistently test the ability of this fully automated counting system to recognize fission tracks in apatite and a muscovite external detector. Fission tracks were analyzed in a series of samples with a step-wise increase in sample complexity. In addition to the automated counting procedure, the same samples were also analyzed using conventional counting procedures. We found for all samples that the fully automated fission track counting procedure using the Autoscan System yields a larger scatter in the measured fission-track densities compared to conventional (manual) track counting. This scatter typically resulted from the false identification of tracks, regardless of the image filtering procedure used or sample coating. Large differences between track densities analyzed with the automated counting persisted between different grains analyzed in one sample as well as between different samples. As a result of these differences a manual correction of the fully automated fission track counts is necessary for each individual surface area and grain counted. This manual correction procedure significantly increases the time required to analyze a sample with the automated counting procedure compared to the conventional approach.