COMPOSITIONAL VARIATION AND MIXING OF IMPACT MELTS ON MICROSCOPIC SCALES

We investigated the compositional characteristics of schlieren-rich, h olohyaline impact glasses from Ries, Wabar, and Meteor Crater using a Cameca SX 100 scanning electron microprobe. This instrument is capable of producing detailed maps of major elements at spatial resolutions o f <10 mu m. The objective was to characterize the composition of an un usually large number of individual schlieren and to evaluate details o f the process that causes melts of lithologically diverse target rocks to mix on scales of micrometers. The Ries and Meteor Crater impacts i nvolved lithologically heterogeneous targets; whereas, Wabar Crater fo rmed in relatively uniform dune sand. Texturally heterogeneous, schlie ren-rich glasses from the Ries Crater illustrate that schlieren of hig hly variable color can be surprisingly similar in composition, as firs t detailed by Stahle (1972). Consistent with these earlier findings, m ost schlieren represent mixtures of diverse rock melts; their composit ions deviate only subtly from the average melt and do not resemble mon omineralic melts nor binary mixtures of major rock-forming minerals. A specific population of schlieren is enriched in mafic elements (Mg, F e: and Ca), which suggests incomplete homogenization of an amphibolite progenitor. In the case of Wabar Crater, a compositionally simple mel t of dune sand mixed with projectile (IIIA iron meteorite) materials, and specific schlieren are variable mixtures of these two progenitors. The optically homogeneous glass from Meteor Crater is compositionally homogeneous as well, which suggests ideal mixing of such diverse lith ologies as platform carbonates, sandstone, and a class IIIA iron meteo rite. The mixing of projectile and target melts at Wabar and Meteor Cr ater unambiguously demonstrates that melts initially produced in disti nctly different stratigraphic/structural locations will undergo wholes ale mixing, if not homogenization. Also, the projectile melts unquesti onably formed relatively early in the cratering process, and their dis semination throughout the prospective melt volume, albeit at variable concentration levels, suggests that the entire mixing process may be a n early cratering feature. This also follows from the fact that we inv estigated ballistic melt ejecta, which thereby eliminates all of those mixing processes that may additionally operate during the pooling and generation of massive melt-ponds following gravitational collapse of large, structurally complex craters. Substantial turbulence ranging fr om field dimensions to microscopic scales seems inescapable to accompl ish the observed degree of mixing, yet this is not readily inferred fr om current models of macroscopic material motions during hypervelocity impact.