HR: 08:15h
AN: V11A-02 INVITED     [PDF]
TI: Crystal Size Distributions in Igneous rocks: Where are we now?
AU: * Higgins, M
EM: mhiggins@uqac.ca
AF: Sciences de la Terre Universite du Quebec a Chicoutimi, 555 blvd de l'universite, Chicoutimi, Qb G7H2B1 Canada
AB: Modern Crystal Size Distributions (CSD) studies started in 1988 and have expanded since then, albeit somewhat slowly. We have now measured CSDs in a variety of different compositions and for both plutonic and volcanic rocks. However, the subject still lags far behind chemical petrology and we need many more studies. CSD methodology has advanced considerably, both for 3D and 2D methods, but it is unfortunate that some 2D studies still do not use appropriate stereological conversions or publish their raw data. The nature of the lower size limit is very important, real or measurement artefact, but is not commonly stated. All this is especially important for comparing data with earlier studies. Individual CSDs of minerals are not always very informative. A much better approach is to look at suites of related CSDs. For instance, different minerals within a single sample, ensembles of related whole rock samples, comparison of late and early textures as preserved in oikocrysts, dykes or volcanic rocks. As more data become available it will be possible to compare usefully unrelated suites of rocks. Straight or nearly straight CSDs in volcanic rocks can be produced by steady-state crystallisation. If the growth rate is known then the residence time can be determined. In some rocks there is a good agreement with other chronometric techniques, but others show no such concordance. In the latter case another model may be more appropriate, such as textural coarsening. This model has been applied in some cases in inappropriate situations, which has cast doubt on the whole subject of CSDs. For plutonic rocks exponentially increasing undercooling can also produce straight CSDs. However, many CSDs are slightly curved and other models are possible, especially if no small crystals are present. Within ensembles of straight CSDs the slope and intercept are commonly correlated. This is mostly accounted for by closure and hence this correlation is not significant, although the variation in either slope of intercept is significant and can be related to other parameters. Concave down CSDs, with no small crystals, are commonly encountered in porphyritic, oikocrystic and plutonic rocks. This texture may be produced by textural coarsening (Ostwald ripening, annealing): this occurs when the magma is maintained close to the mineral liquidus. In this situation the nucleation rate is zero, but growth rates are significant. The classic LSW model is not the only solution possible: more modern solutions, such as Communicating Neighbours may be more appropriate. Variable degrees of textural coarsening will produce CSDs that appear to rotate about a single point. This again reflects closure. Concave up CSDs with no lower size limit are very common. They do not generally have a lognormal or fractal size distribution. They can be produced by mixing of two or more magmas, or crystallisation under several different conditions of undercooling. They can also result from alternations of nucleation and growth followed by textural coarsening. Crystal accumulation and fraction should modify existing CSDs in a predictable manner. An exact solution to this problem has not yet been developed, but simplistic models suggest that CSDs should rotate upwards about the size origin for accumulation and downwards for fractionation. However, clear evidence for such effects has not yet been observed, even in well-layered rocks. There are many igneous systems still to be explored using CSDs. An exiting new domain may be the application of CSDs in experimental petrology.
UR: http://wwwdsa.uqac.uquebec.ca/~mhiggins/CSD.html
DE: 1749 Volcanology, geochemistry, and petrology
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2003 Fall Meeting