Electron Microprobe Techniques for use in Tephrochronological AnalysesPowerPoint PresentationSlide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26Slide 27Slide 28Slide 29Slide 30Electron Microprobe Techniques for use in Tephrochronological AnalysesJ. Fournelle, K. Severin, K.Wallace, J.Beget, J. LarsenMaterial in Poster V33B-0663 presented at AGU December 2006 in “Tephra Dispersal and Sedimentation: Field Studies, Modeling, and Hazard Assessment III” sessionTephrochronology generally assumes that a layer of volcanic ash represents a snapshot of eruption/deposition and of a region within the subvolcanic magma chamber. Correlation of tephra deposits over long distances helps establish age control for other deposits (volcanic and nonvolcanic). Reliable correlations depend on establishing similarity among tephra deposits. Although multi-parameter characterization of a tephra enhances long-distance correlations, identification and correlation of unknown tephras is often done using only geochemical analyses. Techniques vary but generally deal with chemically characterizing all (bulk) or portions (glass, crystals) of the tephra layer, with various geochemical techniques at various spatial scales. Electron probe microanalysis (EPMA) is the most commonly used analytical tool for geochemical analysis and imaging of micron-size volumes of glass and crystals. However, a standard method for collecting, reducing, and reporting tephra data among and within laboratories is not common practice, making comparison of data sets problematic. There is also a range of proficiencies, from advanced to introductory: some EMP and SEM labs have mature established procedures, whereas other labs may have little experience with this somewhat specialized work (hey, someone comes and asks you if your probe will do glass analysis...of course it will).Froggatt (1992), reporting on the 1990 INQUA's Committee on Tephro- chronology, listed recommendations related to quality control of EPMA data:(3) All EMP data should be accompanied by a statement of probe make and model, whether WDS or EDS, and accelerating voltage, beam current (Faraday or absorbed on some standard), and beam diameter. If referenced to earlier work, that should be readily available.(4) All tephra analyses should be based on common working standards, similar to the unknowns. At present the best are KN18 and VG99. Also suitable minerals are feldspar OR1A, Kakanui Augite, and Engels amphibole.(5) All published EMP analyses should be accompanied by typical analyses, or the mean, and other statistics on one or more standards, especially for glass data, or reference to where these standard data are available(6) All EMP analyses of volcanic glass should use an electron beam defocussed to at least 10 um diameter to reduce potential loss of Na and other elements.(continued on next slide)INQUA/Froggatt (1992) Recommendations(7) Give single shard analysis, rather than averaging out heterogeneity.(8) All glass EPMA should be of freshly exposed (polished) internal surfaces.(9) EMP data on glass should be normalized to 100% to facilitate proper comparison. The deficiency from 100% can be recorded as 'water by difference' or the original analysis totals listed.(10) A minimum of 10 analyses on different shards should be collected.(11) All data should be examined for homogeneity, and unusual standard deviations explained. (12) Minimally publish: mean and std dev of each element, after normalization; 'water by difference' or original analysis totals; number of separate shard analyses in the mean. Also (3) and (5) should be noted.INQUA/Froggatt (1992) Recommendations-continuedTo better understand the extent of EPMA use in tephrochronology, and how EPMA operation was documented, a literature survey was done recently by searching bibliographic databases with the terms tephra + chronology + glass. Around 200 articles were then checked to find those with published glass analyses by EPMA. 52 articles were found with EPMA data, 39 published since 2000, and 46 published since 1996 (these figures are weighed toward those journals with online pdf files).There are about 2 dozen laboratories represented, of which 13 are recognized by name as established EPMA facilities with dedicated WDS electron probes (Cameca camebax, sx50/51, sx100; Jeol 733, 8600, 8900; Microscan V), the balance SEM facilities plus 6 "mystery" labs (no details given in the publications about their equipment).Survey of Recent Tephrochronology Publications Reporting EPMA Glass Chemical DataJournals used for this surveyReporting basic conditions…. some well done, others not so well done…Reporting conditions related to element mobility (e.g. Na) in glasses…. some well done, others not so well done…Reporting standards, important for quality controlNormalization of tephra glass is somewhat controversial; a good argument has been made that even if it is done, the original values should be able to be recovered.Suggestions by Hunt and HillJ.B. Hunt and P.G. Hill in a series of papers (1993, 1996, 1998) followed up Froggatt's challenge to the tephrochrology and EPMA communities.In “Tephra geochemistry: a discussion of some persistent analytical problems” (1993, The Holocene 3, p. 271) they agreed with many of Froggatt's suggestions, but questioned others, e.g., the acceptance of "low totals" via data normalization. They also pointed out that it is not just Na (and K) that are "lost" in EPMA, but other elements (Al, Si) increase under beam irradiation. They argued that some North Atlantic tephra EPMA data are of poor quality and "normalization to 100%" permits erroneous conclusions as errors in analysis are allowed.They suggested not using any analysis with a total <95%, that >2 wt% H2O implies alteration of a glass, inappropriate operating conditions, poor sample selection, or poor sample preparation. (They did not address the question of the situation where the only glass available is truly hydrated).They gave a list of 8 EPMA parameters that should be included in any publication. They repeat many of those listed by Froggatt (1992), but add counting time (peak and background), total analysis time (x-ray counting), and exposure time (beam on, on the sample).From Hunt and Hill, 1998In the 1996 paper "An inter-laboratory comparison of the EPMA of glass geochemistry"