Electron Microprobe Facility - Projects

Basaltic volcanism has shaped the Hawaiian Islands, and studying the compositions to improve our understanding of how they form has been an important aspect of our research and teaching portfolio. In addition to Hawaiian volcanism, we've been using the EPMA to measure compositions of minerals and glasses from volcanic rocks from many places on Earth, the Moon and Mars.

First a few words about data quality:

Currently, several electron microprobe studies focus on the incorporation of minor elements in olivine. In order to resolve subtle variations in the concentration of Ni, Mn and Ca, (and sometimes Na, Al P, Ti, Cr) we have optimized a high-current -- long-counting-time method that is both precise and accurate. The high current (200-500 nA) and the relatively long counting times (300 s on peak, 300 s off-peak) obviously improve the precision compared to routine conditions (20 nA, 30 s), but 2 underappreciated factors affect the accuracy and trade-offs of high-current analyses: (1) deadtime and (2) background subtraction.

(1) very high count rates at high currents require deadtime corrections. We found that at currents >250 nA, these corrections lead to inaccurate Ni concentrations (due to the high Ni in the commonly used NiO standard acquired under the same conditions). Using San Carlos olivine, or preferably Ni2SiO4 [synthetic Liebenbergite endmember], as Ni standard yields internally consistent results.

(2) all quantitative EPMA analyses require subtraction of the background countrate from the peak intensity. For routine analyses, the background is determined by linearly interpolating the background intensities on both sides of the peak. For trace elements, we use exponential fits to the background, a now-routine feature of Probe-for-EPMA software and we determine this by using long-count wavescans on standards and unknowns. This significantly improves the accuracy of trace element abundances and we employ this approach whenever we're in the <500-1000 ppm range (element dependent). For Ti and P abundances in olivine, we've also been using the "blank correction" in PfE.

X-Y diagram showing Forsterite vs NiO in San Carlos olivine standard

X-Y diagram showing Forsterite vs MnO in San Carlos olivine standard

Reproducibility of San Carlos olivine standard measured as an external standard by 4 different students between 2011 and 2013. Error bars are standard errors.

 

 


 

The ongoing eruption of Kilauea volcano on the Big Island of Hawaii has been subject of many studies. To better understand the plumbing system of this and other Hawaiian eruptions, we've been conducting high-precision olivine analyses that allow us to assess mixing and assimilation phenomena.

Prior to high-precision quantitative analyses, we commonly obtain element distribution maps, focussing on the distribution of the element phosphorus (P) in olivine. Most of the olivine contains an extremely low (<50-200 ppm) P abundance, but mapping reveals preferential P enrichments along crystallographically controlled lines and planes. We interpret that these enrichments result from rapid initial dendritic growth, following by slower low-P infilling. More about this topic can be found in Welsch et al. (2014) and Shea et al. (2015).

Olivine crystal with dendritic growth history, marked by phosphorus enrichments

Olivine crystal with dendritic growth history, marked by phosphorus enrichments (highlighted in yellow in right image). From Welsch et al. (2014)

Phosphorus X-ray intensity distribution in olivine (on the left image), and forsterite zoning map on the right image

Sketch showing two stages of olivine growth, followed by partial reequilibration

Two stages of olivine growth, followed by partial reequilibration. Mg-Fe diffusive exchange is fast and produces core-rim zoning, unlike slow-diffusing phosphorus, which preserves growth and overgrowth architecture. From Shea, Lynn & Garcia (2015).