Techniques Involving the Electron Microprobe

Experimental petrological studies of natural rocks and synthetic analog materials are carried out over the range of pressure and temperature conditions that extend from surface and crustal pressures (0.1 to 800 MPa) to conditions through the upper mantle (1 to 25 GPa). These studies are integrated with geological, geophysical and geochemical constraints to provide a framework for understanding the chemical fractionation and time scales of processes that create and maintain the chemical segregation present in the crust, mantle and cores of the earth, moon and meteorite parent bodies. Current research involves the determination of the influence of water on crystallization differentiation and mantle melting in subduction zone settings (Sisson and Grove 1993a,b). Water is an essential element in the generation of subduction zone magmatism. Quantitative measurement of oxygen in glasses provides a reliable estimate of H2O content (Sisson and Layne , 1993) and is essential in the characterization of these experiments. Research efforts in planetary science include experimental studies to elucidate the processes that control the distribution of siderophile elements in planetary materials. Information from the distribution of siderophile elements among silicate minerals, metal, silicate melt and sulfide melt can provide important constraints on intensive parameters in planetary interiors (T, P, fO2, fS2, etc.). Experimental studies are performed to determine equilibria among silicate minerals, solid metal, silicate melt and sulfide melt. As part of this work we must analyze an immiscible sulfide melt phase. These phases typically form multiphase quench growths. Gaetani and Grove (1997) reassembled the compositions and proportions of quench sulfide phases manually with the image-analysis capabilities of our Electron Microprobe Facility.

Current experimental practice is to evaluate the composition of the phases in an experimental run product by performing spot analyses of each phase to obtain an average for glass and minerals. This exercise is the first critical step in evaluating the success of an experiment. These analyses are used to perform a mass balance by multiple linear regression (e.g. Kinzler and Grove, 1992a,b; Sisson and Grove, 1993a,b; Baker et al., 1994). We determine whether or not the sample has interacted with its surroundings and gained or lost any elements. The multiple spot method is only applicable when the experiment can be quenched and the glass and minerals preserved. In our current research we are examining melting at high-H2O contents and mineral/melt equilibria at high temperatures and pressures where it is impossible to quench the experimental products. To evaluate the experiment, we estimate the proportions of all phases in the experimental product, using the large scale mapping capabilities of the electron microprobe. Mineralogically and chemically complex quench products in the glass are analyzed along with quench overgrowths on minerals. Using materials balance the abundance of the quench-modified phases are determined and the glass composition and sample bulk composition is estimated. The presence or absence of a mineral phase is also of key importance in the interpretation of experimental products. Chemical mapping of experimental products on the electron microprobe eliminates uncertainty in mineral identification. In many instances it is difficult to distinguish different minerals using conventional optical or electron imagining techniques. Chemical maps provide systematic means for determining the presence or absence of a phase in an experimental product, as well as chemical analyses of the phase and information on its textural setting in the experiment.

Electron microprobe data are also crucial to the interpretation of detailed field, chemical, and petrologic studies of magmatic processes. Recent work has focused on two volcanoes in the southern Cascades: Medicine Lake and Mount Shasta (e.g. Grove et al., 1988; Baker et al., 1994), and in the Barberton Mountainland, South Africa. Medicine Lake volcano has lava flows and lava fields that provide constraints on crustal level processes: fractionation, crustal melting, magma recharge. During the latest phase of the volcano's eruptive history (ca 10,000 to 800 ybp) a suite of lavas that range from basalt to rhyolite reveal the depth range of fractionation (Glass Mountain) and changes in mantle inputs to crust (Callahan and Paint Pot flows) in recent history of volcano (<1000 ybp). At Mount Shasta new petrologic/experimental evidence indicates that the andesites are experiencing fractional crystallization at pressure equivalent to those of the lower crust (25-30 km). These lavas contain a suite of lower crustal xenoliths. They provide a direct record of the way in which subduction zone magmas interact with the lower crust and provide a mechanism for exploring these interactions. In the Barberton Mountainland, komatiites contain fresh igneous minerals that are being used to estimate emplacement conditions of olivine spinifex units and to infer the conditions of melt generation for these important Archean magmas (Grove et al., 1996) The primary evidence for inferring these conditions comes from the analysis of mineral assemblages in lavas and inclusions by electron microprobe.