Northeastern Section–41st Annual Meeting (20–22 March 2006)

Paper No. 9
Presentation Time: 1:00 PM-4:00 PM

EVALUATING THE RECOVERY OF CRYSTAL SIZE DISTRIBUTIONS FROM COMPUTER GENERATED MICROSTRUCTURES


ROBERTS, Sarah M.1, WERTZ, Paxton D.2 and AMENTA, Roddy V.2, (1)Department of Geology & Environmental Studies, James Madison University, MSC 7703, James Madison University, Harrisonburg, VA 22807, (2)Department of Geology & Environmental Studies, James Madison University, Harrisonburg, VA 22807, amentarv@jmu.edu

The link between the kinetics of crystallization and the resulting ideal crystal size distribution (CSD) has been well established, but the microstructural and stereological problems related to the recovery of the CSDs from rocks needs to be further explored. (1) The rock microstructure may modify the real crystal sizes relative to their ideal sizes, (2) the recovered CSDs are usually derived by “corrections” to apparent crystal sizes measured from thin sections, and (3) crystal forms may bias the apparent crystal sizes. We used a computer to model crystallization of microstructures with equations that would predict ideal linear CSDs that would serve as known standards. Such experiments could only be done with computers. The ideal CSDs were then compared to the real or actual CSDs in the microstructures, and these in turn to the CSDs recovered from digital thin slices. Each microstructure is composed of crystals of a single form (1:B:5) such as tetragonal prisms or plates or orthorhombic cuboids. The real crystal sizes, as distinct from the predicted ideal ones, are calculated from their actual volumes in the microstructure. CSD charts for each microstructure show that the real population of crystals in each bin size may be greater or less than the ideal population predicted for the bin size. The real CSD calculated by linear regression is remarkably close to the predicted ideal CSD suggesting that microstructure does not significantly degrade the CSD information. The CSDs recovered from microstructures composed of prisms are linear, from plates weakly curvilinear and from cuboids moderately curvilinear. We confirm the findings of others that intersection widths correlate with smallest diameters of tetragonal prisms and intersection lengths with intermediate diameters of tetragonal plates. Results with cuboids are mixed. In the class of orthorhombic plates the intersection lengths correlate with the intermediate crystal diameters. However, in the class of orthorhombic prisms neither intersection widths nor lengths correlate with any crystal diameters.