Development of etch pit size distributions on dissolving minerals

Abstract

We have applied the theory of crystal size distributions to the evolution of etch pit size distributions (PSD's) on dissolving minerals. Population-balance equations are derived based on nucleation, growth and annihilation of etch pits. PSD plots of log n (number of pits per size per area) vs. W (etch pit width) are used to compare etch pit populations. We have used PSD analysis to study the development of etch pits on calcite dissolved in a rotating disk experiment (far from equilibrium, 25°C, initial pH = 8.6, 0.7 M KCl). A maximum is present on the PSD plot in the early stages of dissolution and appears to result from nucleation of pits at all available sites on the initial cleaved surface over a finite time interval. As pits grow larger, become flat-bottomed and annihilate, the peak moves to larger sizes and decreases in magnitude to become insignificant after 17.5 hr. After 0.5 hr, the PSD plots become linear at small pit widths marking a population balance as new pits nucleate within flat pits, grow larger, become flat-bottomed and finally annihilate. This population of short-lived, small (< 100 μm) etch pits possibly results from nucleation at point-defect or impurity clusters. A linear PSD for long-lived, larger pits appears at 17.5 hr, possibly reflecting etch pits at dislocations. The overall PSD plot at 17.5 hr is kinked due to these two populations. Calcite dissolution rates calculated from the coefficients of the PSD are within a factor of two of the dissolution rate measured from solution chemistry. Such estimates will be more accurate if the duration of dissolution does not greatly exceed the mean lifetime of etch pits. PSD plots for hornblende and microcline grains from a loess soil that has weathered for 12,500 yr show similarities to the PSD's from our calcite experiments. The hornblende PSD shows a peak and the microcline PSD is linear or possibly kinked. Bulk dissolution rates for the grains calculated from the PSD's are slower than rates measured for these minerals in the laboratory. Dissolution rates estimated from the PSD coefficients for the hornblende and microcline soil grains are 3–4 and 1–2 orders of magnitude slower, respectively, than the corresponding laboratory rates.