Dating Basalts using the $^{40}$Ar/$^{39}$Ar Method- Improvements Based on Electron Microprobe Evaluation of Samples

AU: * Dunbar, N W
EM: nelia@nmt.edu
AF: New Mexico Bureau of Geology and Mineral Resources, 801 Leroy Place, Socorro, NM 87801 United States
AU: McIntosh, W C
EM: mcintosh@nmt.edu
AF: New Mexico Bureau of Geology and Mineral Resources, 801 Leroy Place, Socorro, NM 87801 United States
AU: Love, D
EM: dave@gis.nmt.edu
AF: New Mexico Bureau of Geology and Mineral Resources, 801 Leroy Place, Socorro, NM 87801 United States

A study of 85 samples of basaltic lava and cinders from the Los Lunas volcano, central New Mexico, demonstrates that the accuracy and precision of $^{40}$Ar/$^{39}$Ar dates for young basalts can be maximized by selecting samples that contain K-feldspar as a late-crystallized phase, and contain little or no glass or alteration phases. On the basis of optical observations of thin sections, it has long been recognized that abundant groundmass glass is detrimental to K-Ar and $^{40}$Ar/$^{39}$Ar dating of basalts. The electron microprobe offers a means of assessing and quantitatively analyzing basaltic groundmass features that are too small to be assessed optically. X-ray mapping and quantitative analyses accurately distinguish among crystal and glass phases as small as 2 $\mu$m. The crystallinity of smaller areas can be qualitatively assessed by multiple analyses, where homogeneous compositions indicate glass and heterogeneous composition indicates multiple crystal phases. Samples from Los Lunas volcano were examined by electron microprobe, with the first step in the evaluation process being collection of potassium x-ray maps of polished surfaces. Once the locations of K-bearing phases were determined, the compositions of the phases were measured using quantitative analysis. Some otherwise well-crystallized samples were found to contain small amounts of groundmass glass in interstices between crystals; in these cases, virtually all K was concentrated within the glass. In other samples the groundmass was fully crystallized. In these cases K resided in K-feldspar, either as thin ($<$10 $\mu$m) rims on plagioclase or as fine ($<$15 $\mu$m) groundmass crystals intergrown with albite. This K-feldspar would be impossible to identify optically due to the small grain size and intergrown nature of the crystals. Samples were rated on the basis of glass content versus K-feldspar content. After excluding samples containing abundant ($>$10\%) glass, 19 samples were dated by the $^{40}$Ar/$^{39}$Ar resistance furnace incremental heating method. The shape of spectra and the accuracy and precision of plateau and isochron ages correspond closely to the ranking assigned based on the microprobe evaluation. Glass-free, K-feldspar-rich samples have generally flat spectra, relatively high radiogenic yields (\%$^{40}$Ar*), and ages with precision values ($\pm$2 $\sigma$) ranging from 1-5\%. These ages agree well with stratigraphic order. Samples with increasing amounts of groundmass glass have increasingly disturbed spectra, reduced radiogenic yields, and low precision ($\pm$5-20\%) plateau or isochron ages, which in some cases violate stratigraphic order. The consequences of even small amounts of groundmass glass are attributed to various combinations of atmospheric argon acquired during hydration, $^{40}$Ar or K$_{2}$0 loss, and $^{39}$Ar recoil artifacts . $^{40}$Ar/$^{39}$Ar results from higher ranking samples indicate that Los Lunas volcano consists of a 3.81 $\pm$ 0.10 Ma southern cone overlapped by a 1.25 $\pm$ 0.02 Ma northern cone; this relationship agrees with geologic map patterns. Within the northern cone, geologic mapping has delineated a sequence of five eruptive events. The ages of high ranking samples from the northern cone all agree within analytical error, suggesting that the entire period of eruptive activity spanned less than 30,000 years. The precision and accuracy of data in this study was much enhanced by using microprobe observations of groundmass glass and K-feldspar to aid in the sample selection process. Sample selection based on only thin section observations would have yielded lower quality data.
DE: 1035 Geochronology
DE: 8494 Instruments and techniques
DE: 9350 North America
DE: 9604 Cenozoic
SC: V
MN: 2001 AGU Fall Meeting