This paper explores tools for modeling and measuring the compositional heterogeneity of a rock, or other solid specimen. Intuitive ``variation per decade'' plots, simple expressions for containment probability, generalization for familiar error-in-the-mean expressions, and a useful dimensionless sample bias coefficient all emerge from the analysis. These calculations have also inspired subsequent work on log-log roughness spectroscopy (with applications to scanning probe microscope data), and on angular correlation mapping of lattice fringe images (with applications in high resolution transmission electron microscopy). It was originally published as Appendix E of a dissertation on ``Microcharacterization of interplanetary dust collected in the earth's stratosphere''.
- Pub Date:
- February 2004
- Condensed Matter - Materials Science;
- Mathematics - Probability;
- Physics - Geophysics
- 9 pages (3 figs, 14 refs) RevTeX, comments http://www.umsl.edu/~fraundor/covariances.html