V51A-0323
Challenge to Electron Microprobe Analysis by Nano-features in Geo-materials
With current high-resolution analytical scanning electron microscope and transmission electron microscope technology, nano-features (i.e., inclusions, exsolutions, pores) are being discovered in many common geo- materials routinely analyzed by EPMA. Most EPMA instruments have a thermionic electron gun with a tungsten "hairpin" type filament and operate at such high probe currents that they do not produce the high resolution imaging that field emission SEMs do. Such nano-features present difficulties to electron microprobe analysis of the host materials. It is an accuracy issue. Presented here are a few examples. Nano-inclusions like magnetite or pyroxene are common in volcanic glass. Massive rose quartz contains nano-fibers of a dumortierite-related phase that is pink, which is the cause of rose color and optical star effects. Blue quartz has nano-inclusions of ilmenite. Nano-exsolutions and nano-inclusions occur in some feldspars (moonstone and sunstone). Nano- pores in corundum and opals are observed. Thin coatings of nanocrystals on grains of hematite bring Al and P (or As) into the analytical picture. In each of these cases, electron microprobe analyses of host materials were contaminated or affected by the nano-features. For a complete analytical description of a geo-material it is necessary to first examine probe samples under an optical microscope (400× minimum) and a FE-SEM. If the sample contains nano-features, it would then be necessary to combine FIB and TEM-EDS-EELS to comprehensively analyze the host composition.
V51A-0324
Rapid Cation Depletion During Electron Microprobe Analysis of Uranium Phosphates
Alkali cation loss during electron microprobe microanalysis (EPMA) is a well-known phenomenon. Our recent work on uranyl phosphates, however, has revealed extraordinarily rapid and extensive depletion of interlayer Cu in metatorbernite [Cu(UO2)2(PO4)2·8H2O]. Minerals in the autunite and meta- autunite groups, e.g. metatorbernite, are important sinks for uranium in oxidizing environments such as contaminated soils at Hanford, WA and Oak Ridge, TN. EPMA is a valuable tool for investigating the fine-grained minerals in soils and sediments, allowing for imaging and chemical analysis of the same grains. However, the autunites and meta-autunites are hydrous and beam sensitive, making EPMA challenging. Upon exposure to vacuum and the electron beam, water is lost. In addition, exposure to the electron beam results in the loss of interlayer cations from the analysis volume. We present results from the Cu, Ca, and Ba-bearing meta-autunite minerals, metatorbernite, meta-autunite, and meta-uranocircite, respectively. Copper is lost much more rapidly than calcium, which is, in turn, lost more rapidly than barium. While the migration of alkalis during EPMA was first described more than forty years ago, to our knowledge this is the first report of the migration of a transition metal. Volatile element correction schemes that use simple exponential functions to fit dropping count rates through time are insufficient for the accurate calculation of metatorbernite compositions. This result has implications for the analysis of metatorbernite found at contaminated sites such as Hanford. More fundamentally, it has the potential to enhance our understanding of the bonding environment of Cu in metatorbernite.
V51A-0325
Electron Microprobe Measurements of Nitrogen in SiC
Methods have been developed for the measurement of low abundances of nitrogen in SiC films. These techniques were developed for measurements of synthetic thin-film samples prepared by materials scientists but the technique can also be applied to natural SiC grains in meteorites. One problem associated with measuring nitrogen at low abundance levels is the low count rates due to strong absorption of the nitrogen signal in the matrix material. In thin film samples, (SiC deposited on elemental Si) it is preferable to limit x-ray production and emission to the overlayer. This eliminates the need for data reduction using thin-film methods. Thin film data reduction is inevitably less accurate than bulk material data reduction methods. In order to limit x-ray emission to the film layer, data has been collected at 5 kV and 3.5 kV accelerating voltage (depending on film thickness estimates provided by scientists who prepared these samples). These low beam energies also promote production of x-rays in the shallow region of the samples, and this minimizes strong absorption, leading to more abundant nitrogen x-ray detection, which improves counting statistics and overall precision. The CASINO monte carlo modeling program was used to model electron penetration and x-ray production as a function of beam energy and depth in the sample in order to ensure that the excited volume is limited to the film. The beam was set to 200 nA beam current. This high beam current also improves counting statistics by providing more abundant count rates. One drawback of these beam conditions is the limited spatial resolution provided. In our Cameca probe, a 5 kV, 200 nA beam is approximately 10 microns in diameter. SiC samples and standard were not carbon coated (they are conducting). AlN was used as the nitrogen standard. These films contained 0.3 to 0.7 wt. per cent nitrogen, with analytical uncertainties in the range of 10-20 per cent relative errors. The Si:C ratios were very near 1:1 indicating that little if any Si signal originated in the substrate of the film.
V51A-0326
Keeping Track of the Selenide zoo. A Combined Optical Microscopy - EPMA Study of Complex Selenides
Hunting for new mineral phases is a fascinating scientific activity. This kind of research not only serves the replenishment of mineralogy textbooks with new mineral names but also the industry with new potential semiconductors, laser crystals and other 'high-tech phases'. Chemical analyses using the electron microprobe are an essential intermediate step in the course of the description of a new mineral. The study of a great number of different Cu-Pb-Ag-As-Hg-Tl-Sb-Bi-Cd selenides from the former uranium deposit Schlema-Alberode (Saxony) in the German part of the Erzgebirge represented a twofold challenge to us. The complex genetic and age relations of the ore minerals and gangue minerals entailed the development of very complex microstructures in a tight space. Typical features are symplectitic intergrowths, exsolutions, fine lamellae, zoned crystals and the development of pseudo- and paramorphs. Altogether we found 34 different selenide, sulfide and arsenide minerals, including 6 dimorpheous phases. Many of these minerals are indistinguishable by electron-optical methods used during different stages of the study. Dimorpheous minerals like bellidoite and berzelianite are as much indistinguishable like intergrowths of the minerals berzelianite, mgriite and lollingite using backscattered and secondary electron images. Optical microscopy is the key to overcome these problems. We show that the step-by-step combination of polarized light microscopy, phase contrast microscopy and differential interference contrast microscopy using transmitted and reflected light allowed a secure discrimination of the different minerals and unknown phases. 16 elements are incorporated into these phases either as main, minor or trace elements. A multitude of overlapping peaks, potenzial fluorescence effects caused by adjacent phases and different matrices in the minerals demanded the development of several specific methods optimized for the analysis of Pb-Se-, Cu-Tl-Se-, Cu-Fe-Zn-As-Se-, Hg-Pb-Ag-Se-, and arsenide phases. The electron microprobe studies will be followed by EBSD, TEM and crystal structure determination by X-ray diffraction using ultra-small amounts of material taken with the help of a New Wave micromill.
V51A-0327
Large Errors in Electron Microprobe Analysis of Laphamite due to Uncertainties in ZAF Correction Parameters
Laphamite, As2Se3-xSx (x~0.75), was analyzed by electron microprobe in three different laboratories, using different acceleration voltages (15kV and 20kV) and a number of different standards for As (As2S3, GaAs) and Se (Se, PtSe2, Bi2Se3). Lines used were Kα for S and Lα for As and Se. All analyses resulted in totals between 104 and 109 weight-%. Satisfactory results were obtained, however, by using As2Se3 as standard for both As and Se, giving totals between 100.1 and 100.8 weight-% Since all three laboratories used JEOL instruments with similar manufacturer-supplied software, a problem with ZAF parameters was suspected. One set of analyses was recalculated with different mass absorption coefficients (MACs), substituting the FFAST set derived from data by Chantler [1] for the Heinrich-Henke set in the instrument software. This resulted in changes of ZAF factors of -4.2% for As, -11.1% for Se, and -8.1% for S, reducing the analysis total from 107.9% to 99.3% (average of 6 analyses). The recalculated result as well as the analyses obtained with As2Se3 as standard were also close to the composition inferred from the X-ray diffraction study. To further elucidate the source of the discrepancy we are also performing calculations using the Monte Carlo program PENEPMA for comparing theoretical K-ratios to the ones obtained by electron microprobe. The theoretical K-ratios will then be processed with the same ZAF correction procedures as the microprobe data. [1] Chantler, C.T., J. Phys. Chem. Ref. Data 29, 597-1048 (2000).
V51A-0328 INVITED
Peak Shifts in Al, Mg, Si and Na Ka in Geologically Important Materials
In the course of troubleshooting slightly low analytical totals in common silicate minerals, problems with the ROM peaking routine were found, i.e. it did not always find the peak center, a central plateau, e.g. ~ 5 sin theta units wide for Al Ka. Ultimately chemical peak shifts in Al, Mg. Si and Na Ka of common minerals were found. There are published references to some Al Ka peak shifts between some minerals, but never before to shifts within the feldspar family. There is an Al Ka peak shift of 7 sin theta units (0.6 ev) between anorthite and Na/K feldspars, with most rock forming minerals falling between these end members. A shift of 7 sin theta units on a correctly centered peak could cause ~2 % error (doubled if O by stochiometry); on an incorrectly centered peak the error could be 2-3x. These experiments were done on an SX51 (#485) operating at 15 kV, on TAP crystals. Consistent results were found on two separate crystals, over a period of 2 years (4 experiments). Examination of Al Ka in garnets shows there is a 5 unit sin theta shift in the Al Ka peak of pyrope relative to almandine. Most silicate Si Ka peaks are close together, although there is a small distinct shift for K-feldspar relative to other silicates. For Mg Ka, there is a wider gap, spanning 10 sin theta units from chromite to pyrope (e.g., MgO is 7 units from pyrope). Initial work suggests a similar occurrence for Na Ka, with a 10 unit sin theta span between albite and jadeite. Here, however the peaks are much wider (18 units for albite), the count rates lower, and precision lower. The accuracy of this experimental technique was evaluated: the shift between SiO2 and Si, 8-9 sin theta units (0.5-0.6 ± 0.1 eV) is the same as has been found by other more sensitive techniques (0.62, 0.655 ev). One reason these issues may not have been recognized before is that minor errors in Z (optical) focus yield comparable errors to peak shifts.
V51A-0329
Problems in Trace Element EPMA: Modeling Secondary Fluorescence with PENEPMA
In EPMA the incident electron beam scatters, producing the "electron interaction volume". However, X-rays produced within that volume may spread tens to hundreds of microns and produce secondary fluorescence in other materials. This can be ignored in many/most cases where major element compositions are being measured. However, when minor or especially trace elements are being measured by EPMA, secondary fluorescence needs to be considered. In some cases it can be eliminated by separating the materials and mounting them by themselves, or by reducing the kV to minimize fluorescence. However, in many cases these efforts are not possible. In those cases, modeling the EPMA experience by Monte Carlo simulation is useful. PENEPMA is a Fortran program based upon the PENELOPE radiation transport model of Salvat et al (2006). It differs from other MC electron interaction programs in that it follows each electron and photon and records all interactions at each point in the "particle's path". It also permits complicated geometries with many materials. Both continuum and characteristic secondary fluorescence can be tracked, and at the end of a run, compared to x-rays produced only from primary electrons. This has important applications in geology and petrology. Llovet and Galan (2003) used it to correct calculations for olivine-cpx thermobarometry. Recent attention has been given to Ti in zircon as a geothermometer. PENEPMA has now been used to model 2 different situations: (1) rutiles present on or near zircon, (2) ilmenite present near zircon. In the first case, several geometries are modeled: a 30 micron diameter zircon (Ti-free) is surrounded by Ti-bearing silicate glass (yielding an apparent 452 ppm Ti in zircon core); with five 30 um surrounding rutiles at 15 micron distance (yielding an apparent 948 ppm Ti in zircon core); if the silicate glass were replaced by epoxy, the apparent Ti would increase to 1179 ppm. In a second case, the effect of zircon with nearby ilmenite in a rock was simulated by a simple 2-slab geometry. PENEPMA shows that at 10 um distance into the zircon there would be ~1000 ppm Ti by secondary fluorescence, gradually reducing to ~100 ppm at 40 um distance. Note that in both cases the issue is continuum fluorescence. Uncorrected EPMA measurements would yield incorrectly high temperatures using the Ti-in-zircon thermometer.
V51A-0330
Trace Element Analyses of Quartz-Two Wrongs Don't Make a Right
Microanalysis of trace elements in hydrothermal quartz can improve understanding of the chemical dynamics of ore-forming systems. Quartz is expected to vary in composition because of the wide range of formation temperatures and parent fluid compositions. Compositional variations are evident in the variations in cathodoluminescence (CL) intensity of precipitated quartz, which is a result of structural defects including those caused by trace element substitutions. Quantification of trace elements in quartz remains difficult because of their low concentrations and interference from mineral and fluid inclusions. We compared EPMA and LA-ICP-MS trace element analyses of quartz from hydrothermal ore deposits that formed at temperatures between 100 and 700 degrees C to evaluate the suitability of each analytical method. Whereas LA-ICP-MS allows the user to simultaneously acquire data on over 40 elements at less than 10 ppm detection limits, these analyses are of low spatial resolution relative to CL textures. LA-ICP-MS also included incidental ablation of fluid and mineral inclusions in the quartz. Although far fewer elements are detectable at concentrations as low as LA-ICP-MS, the spatial resolution of EMPA allows for direct comparison between CL textures and TE abundances. In particular, EPMA analysis of quartz acquired at 20 keV, 100 nA, and 20 minute count times for Ti and Al on multiple spectrometers achieves detection limits that approach that of LA-ICP-MS (7 ppm and 4 ppm, respectively) while retaining less than 5 micrometer resolution. Determination of low-level Al using EPMA is sensitive to the background correction model used. In several samples containing concentrations of Al less than 100 ppm, linear and exponential background models yield Al concentrations below detection. Analyses run using only the high background (to avoid Si tailing on the low background side of the Al peak) produce detection limits of 300 ppm. Using a polynomial fit of the background, Al concentrations measured by EPMA range from 4 to 32 ppm where LA-ICP-MS values ranged from 2 to 33 ppm. Excellent agreement between the two techniques strongly suggests that the polynomial background fit is appropriate for quantification of Al in very small quantities. Ti concentrations determined by EPMA and LA-ICP-MS also show excellent agreement. In quartz samples from two porphyry copper deposits that were mostly featureless in CL, EPMA concentrations ranged from 9-25 ppm and 26-83 ppm whereas LA-ICP-MS analyses ranged from less than 13-30 ppm and 48-69 respectively. As with Al, the complex CL texture correlates well with the Ti content variations apparent with the spatial resolution of EPMA. LA-ICP-MS analyses however show variations in Li, Na, Mg, K, Ga, Rb, Sr, Cs, and Ba, that are not detectable with EPMA, but correlate with broad variations in CL intensity. While EPMA background modeling of Ti is not an issue because of the absence of the Si tail, the analyst must also consider secondary fluorescence effects from included minerals, particularly rutile. Ti in quartz can be detected as far as 150 micrometers away from adjacent rutile (Wark 2006).
V51A-0331
Improved EPMA Trace Element Accuracy Using a Matrix Iterated Quantitative Blank Correction
At trace element levels below several hundred PPM, accuracy is more often the limiting factor for EPMA quantification rather than precision. Modern EPMA instruments equipped with low noise detectors, counting electronics and large area analyzing crystals can now routinely achieve sensitivities for most elements in the 10 to 100 PPM levels (or even lower). But due to various sample and instrumental artifacts in the x-ray continuum, absolute accuracy is often the limiting factor for ultra trace element quantification. These artifacts have various mechanisms, but are usually attributed to sample artifacts (e.g., sample matrix absorption edges)1, detector artifacts (e.g., Ar or Xe absorption edges) 2 and analyzing crystal artifacts (extended peak tails preventing accurate determination of the true background and ¡§negative peaks¡¨ or ¡§holes¡¨ in the x-ray continuum). The latter being first described3 by Self, et al. and recently documented for the Ti kƒÑ in quartz geo-thermometer. 4 Ti (ka) Ti (ka) Ti (ka) Ti (ka) Ti (ka) Si () O () Total Average: -.00146 -.00031 -.00180 .00013 .00240 46.7430 53.2563 99.9983 Std Dev: .00069 .00075 .00036 .00190 .00117 .00000 .00168 .00419 The general magnitude of these artifacts can be seen in the above analyses of Ti ka in a synthetic quartz standard. The values for each spectrometer/crystal vary systematically from ¡V18 PPM to + 24 PPM. The exact mechanism for these continuum ¡§holes¡¨ is not known but may be related to secondary lattice diffraction occurring at certain Bragg angles depending on crystal mounting orientation for non-isometric analyzing crystals5. These x-ray continuum artifacts can produce systematic errors at levels up to 100 PPM or more depending on the particular analytical situation. In order to correct for these inaccuracies, a ¡§blank¡¨ correction has been developed that applies a quantitative correction to the measured x-ray intensities during the matrix iteration, by calculating the intensity contribution from the systematic quantitative offset from a known (usually zero level) blank standard. Preliminary results from this new matrix iterated trace element blank correction demonstrate that systematic errors can be reduced to single digit PPM levels for many situations. 1B.W. Robinson, N.G. Ware and D.G.W. Smith, 1998. "Modern Electron-Microprobe Trace-Element Analysis in Mineralogy". In Cabri, L.J. and Vaughan, D.J., Eds. "Modern Approaches to Ore and Environmental Mineralogy", Short Course 27. Mineralogical Association of Canada, Ottawa 153-180 2Remond, G., Myklebust, R. Fialin, M. Nockolds, C. Phillips, M. Roques-Carmes, C. ¡§Decomposition of Wavelength Dispersive X-ray Spectra¡¨, Journal of Research of the National Institute of Standards and Technology (J. Res. Natl. Inst. Stand. Technol., v. 107, 509-529 (2002) 3Self, P.G., Norrish, K., Milnes, A.R., Graham, J. & Robinson, B.W. (1990): Holes in the Background in XRS. X-ray Spectrom. 19 (2), 59-61 4Wark, DA, and Watson, EB, 2006, TitaniQ: A Titanium-in-Quartz geothermometer: Contributions to Mineralogy and Petrology, 152:743-754, doi: 10.1007/s00410-006-0132-308
V51A-0332
Trace Element Background Modeling for Electron Microprobe Chemical Dating of Monazite
Electron probe microanalysis (EPMA) techniques have conventionally been used for major and minor element analysis in geological and material science specimens. EPMA provides both rapid and efficient quantitative measurements of elemental abundance due to the small beam-interaction volume and the ability to vary accelerating voltage and beam current over a wide range allowing for higher count rates to be obtained over small dwell time intervals. The small beam-interaction volume provides the opportunity to obtain multiple analyses within a specimen, limited only by the size of the specimen or grain, providing better statistics. Chemical dating of monazite and xenotime require analysis of the major/minor elements yttrium and thorium, along with trace element analysis for uranium and radiogenic lead. Selection of positions for off-peak background measurements and proper modeling of the continuum can affect both the accuracy and error associated with a single point analysis. An inter-lab study previously performed to correlate monazite isotopic and chemical dating techniques used published positions for major/minor and trace element background measurements, with a linear background fit. Results from that study produced chemical ages that were comparable to ages obtained using ion microprobe (IMP and SHRIMP) U-Pb and Th-Pb isotopic dating techniques, within 2-sigma error. Recent software upgrade to Probe for Windows on the Virginia Tech CAMECA SX50 has provided the ability to visually select positions for background measurements, decreasing the likelihood of inaccurate off-peak background measurements due to spectral interferences. Background modeling of the continuum for trace element measurements of uranium and radiogenic lead can be performed more precisely over a wide continuum range which may require an exponential fit or polynomial fit depending on curvature or over a narrow continuum range that would be associated only with the desired peak area. Effects resulting from wide continuum fit versus narrow continuum fit could produce a significant discrepancy between background-corrected peak intensity using a curved background continuum fit versus that calculated using a linear background continuum fit. This could yield a "reasonable" yet inaccurate geologic age. Previous investigations into the effects of background modeling for trace element analysis have contributed to the need for this detailed study of continuum modeling for EPMA chemical age determination.
V51A-0333
Analytical Aspects of EPMA for Trace Element Analysis in Complex Accessory Minerals
High-resolution microanalysis of complex REE-bearing accessory phases is becoming increasingly necessary for insight into the chronology of phase growth and tectonic histories, and in understanding the mechanisms and manifestations of growth and dissolution reactions. The in-situ analysis of very small grains, inclusions, and sub-domains is revolutionizing our understanding of the evolution of complexly deformed, multiply metamorphosed, rocks. Great progress has been made in refining analytical protocols, and improvements in instrumentation have yielded unprecedented analytical precision and spatial resolution. As signal/noise improves, complexity is revealed, illustrating the level of care that must go into obtaining meaningful results, and in adopting an appropriate approach to minimize error. Background measurement is most critical for low concentration elements. Errors on net intensity values resulting from improper background measurement alone can exceed 50% relative. Regression and modeling of the background spectrum is essential, and must be carried out independently for each spectrometer, regardless of instrument. In complex materials such as REE- bearing phosphates, high concentrations of REEs and actinides create difficult analytical challenges as numerous emission lines and absorption edges cause great spectral complexity. In addition, trace concentrations of "unexpected" emission lines such as those from sulfur, or fluoresced from nearby phases (Ti, K), cause interferences on both measured peaks and background regions which can result in very large errors on target elements (U, Pb, etc.), on the order of 10s to 100s of ppm. Characteristic X-ray emission involving electron transitions from the valence shell are subject to measureable peak shifts, in some cases significantly affecting the accuracy of results if not accounted for. Geochronology by EPMA involves careful measurement of all constituent elements, with the calculated date dependant on the resulting Pb, U, and Th concentrations. Cancellation of systematic errors, such as those arising from improper background estimation procedures, can yield "reasonable" dates despite large inaccuracies in the obtained concentrations. If not recognized, these systematic errors can result in age errors of 10s of m.y. Rigorous trace element strategies (distinct form traditional major-element protocol) must be followed in order to obtain meaningful ages from a variety of if domain compositions. Individual compositional domains in complexly zoned accessory phases should be sampled until statistical adequacy is met, and overall results for a domain should be evaluated on the basis of compositional homogeneity and the resulting propagation of counting errors. Weighted means of corresponding domains can then begin to reveal a more meaningful dataset where complete compositions can be related to calculated ages. Although reliable trace element secondary standards are not currently available, an appropriate range of age standards of differing composition can at least constrain accuracy, and reveal inappropriate procedures.
V51A-0334
Counting Data and T-test Evaluation of Detection Limits and Their Effect on Data-Group Similarities
Four arbitrary collections of detrital zircons were analyzed by EMP, using high beam currents and long counting times to produce calculated 1-sigma detection limits (DL) for U, Th, and Pb of 174, 129, and 97 ppm, respectively. The spectrometer background position was determined by qualitative scans across element peaks. 300 analyses were collected: 100 for two collections and 50 for the others, and data manipulation and Student's t-test calculations for the groups were made using a spreadsheet program. The compositional results were U: 0-3.6 wt. %, Th: 0-0.32 wt. %, and below-DL for Pb. It was assumed that below-DL net-count results should cluster around 0. Histograms of net counts below-DL had non-zero means for all three elements, suggesting that the background measurements were incorrect: the maxima were offset from 0 by +20 cts/sec/nA for U and Th, and - 80 cts/sec/nA for Pb (negative net counts were reported, but below-DL results were returned as 0 wt. % by the EMP software). Above-DL wt. % data were regressed against count rates, all the net counts were shifted to set the below-DL mean at zero counts, and the overall (+ and -) wt. % values were recalculated according to the regression. The non-zero means for U and Th were thus reduced by 40 and 9 ppm, respectively. The mean Pb was -3.2 wt. %, calculated from the net counts and Pb (ca. 64 wt. %) of the crocoite standard; this result was unreasonable, but the count-rate data (containing no compositional interpretation) for Pb was used for t-tests. The difference between t-test results from data including 0 wt. % values and from count-rate values was significant. For example, the inclusion of numerous 0 wt. % data in a group with low U concentrations biased the mean and standard deviation to suggest that the group was distinct, but the use of net count rates showed that the group was similar to another group with low but detectible U. Wt. % results for Pb were meaningless due to the calibration error, but the net-count data allowed comparison and discrimination between groups based on t- test calculations.
V51A-0335
Gallium in Feldspar Minerals - an Underutilized Source of Petrogenetic Information
The trace element gallium is routinely analyzed in all kind of rocks. Regardless of this fact it is rarely used as a petrogenetic indicator element. Arguably the best known exception is the discrimination of A-type granites according to the Ga/Al value. Interpreting these data requires our understanding of how Ga is distributed between coexisting crystal phases and liquids. Because of the great importance of the feldspar minerals for the geochemical evolution of Ga a two- step method for Ga analysis of feldspar minerals using the electron microprobe was developed. The first step includes a screening for increased Ga-values during the routine feldspar analyses. A threshold of 100 ppm Ga was used to define 'Ga-rich feldspars'. These feldspars were analyzed for Ga using a special routine with optimized conditions for trace element analysis. To minimize the well known effects of diffusion of alkaline metals and to maintain the high spatial resolution an accelerating voltage of 15 kV was used. In order to improve precision and the lower detection limit long peak counting times of 360 - 600 seconds were chosen. The background curvature and possible interferences of both Ga-Kα and Ga-Lα lines were studied using standards very poor in gallium. The influence of the beam current on the stability of different feldspar types was tested. In order to ensure the integrity of the sample and the compliance of the analytical task an optimal ratio of beam current to beam diameter has to be chosen for every analysis. Using optimal conditions with low spatial resolution a lower detection limit of 33 ppm was reached. Using this two-step method we found unexpected high Ga values in feldspars of different rock types. Characteristic examples are albite with a Ga-Ca - ratio of > 1 from albite granites in the Central Eastern Desert (Egypt) and zoned plagioclase from mafic microgranular enclaves in granites formed by magma mixing (Karkonosze Mountains in Poland) with Ga-contents greater than 1000 ppm.
V51A-0336
Electron probe microanalysis for high pressure minerals investigation
In the early 1968 in Siberian Branch of the Academy of Sciences of USSR, Novosibirsk, electron probe microanalyzer MS-46 was installed and started to operate for high pressure minerals EPMA investigation. In collaboration with Geophysical Laboratory of Carnegie Institution (Drs. F.R. Boyd, F. Schairer) a set of standards for silicates analysis was developed. Technique for quantitative analysis was developed (Lavrentiev et al., 1974, Zavodsk. Lab., v. 40, p. 657-661) and applied for the first in the USSR analyses of pyropes, associated with Siberian diamonds both as inclusions and xenoliths of diamondiferous peridotites (Sobolev et al., 1969, Dokl. Akad. Nauk SSSR, v. 188, p. 1141-1143; v. 189, p. 162-165). As a result of that research, unique Cr-rich subcalcic pyropes with high knorringite content were found in diamond-bearing kimberlites only and new mineralogical criteria for diamond exploration were developed (Sobolev 1971, Geol. Geofiz., v. 12, p. 70-80) which are still in use worldwide. Further development of electron probe instruments (JXA-5A, Camebax Micro, JXA-8100) and computers, as well as development of analysis technique led to creation of large analytical database. In another field of EPMA – determination of small concentrations of elements – for the first time importance of 0.01-0.3% Na2O admixtures in garnets (Sobolev, Lavrentiev, 1971, Contrib. Min. Petr., v. 31, p. 1-12) and K2O in clinopyroxenes (Sobolev et al., 1970, Dokl. Akad. Nauk SSSR, v. 192, p. 1349-1352) were demonstrated. Since then, determination of sodium content in EPMA of garnets and potassium in pyroxenes became a routine technique. Last generation analyzer (JXA-8100) provided record results down to 6 ppm in detection limit of Ni in pyropes (Lavrentiev et al., Rus. Geol. Geophys., 2006, v. 47, p. 1090-1093). As a result, application of EPMA for single mineral geothermometry (currently based mainly on PIXE method) becomes possible.
V51A-0337
The Combination of EMP Compositional Maps With Other Image Data to Investigate Microscale Contaminant Geochemistry
EMP compositional imaging is limited by collection times to relatively high detection levels (100+ ppm). The mobility of environmental contaminants is often controlled by sorption or ion exchange processes so that they are present at levels below the detection limit. The phases in spatial association with the contaminants can be determined by combining EMP images (including elemental abundance and backscattered electron images) with compositional images produced by more sensitive methods at lower optical resolution. The mineral-specific processes responsible for contaminant immobilization may then be inferred from phase and contaminant distributions to provide a basis for the interpretation of macroscale data from leaching or cation exchange experiments. The association of sediment phases with 90Sr and 137Cs in separate field samples was investigated using EMP images matched to autoradiographs collected with a resolution of 25 μm and sensitive to sub-Bq levels of radiation. 90Sr was associated with smectites within weathered basalt clasts, and controlled by cation exchange in sediments that were coarse-grained and contained little or no gravitationally measurable clay fraction. 137Cs was confined to micaceous minerals. Subsequent experimentation with micas and combined X-ray microprobe (XMP), TEM, and EMP analysis showed that the Cs was bound to frayed edge sites on and within the micas. XMP methods allow valence determinations with a detection limit of 1 ppm at a resolution of 10 μm, and were used along with EMP analysis to show that U contamination in some contaminated sediments was unreduced and precipitated as a uranyl silicate in microfractures in granitic sediment clasts. To investigate the indirect effects of microbiota on 99Tc, pertechnetate (99TcO4-) was exposed to bioreduced sediments, which were subsequently examined using EMP and XMP. The pertechnetate was reduced to 99TcO2 by sediment-bound Fe(II), and was preferentially bound to iron-rich phyllosilicate clasts. The combination of sensitive and high-resolution images yielded valuable data for the management of trace contaminants and extended the utility of EMP analyses.
V51A-0338
Multiple stages of growth in snowball garnets revealed by EBSD analyses
Electron backscattered diffraction (EBSD) analysis, chemical mapping, and computed tomography (CT) imaging have been performed on snowball garnet prophyroblasts from the Lukmanier pass area (central Swiss Alps). The geometry of most garnets is characterized by a core region, around which two spiral arms are wound over 360 degrees. The EBSD maps reveal that the snowball garnets are composed of several grains with distinct crystallographic orientations. The core region and the base of the spiral arms are generally formed by one single, large grain, whereas the arm terminations are made of smaller and often box-shaped crystals separated by straight grain boundaries. The central sector domain exhibits approximately 270 degrees of rotation. Crenulated internal foliations observed near the rim of the garnets indicate a change in the stress field during garnet growth. The distribution of equal Mn concentration domains, considered as time markers, reveals that the change in the growth regime coincides with the occurrence of the first subgrains at the end of the spiral. A similar approach was applied on a second population of snowball garnets collected from a neighboring outcrop. These latter garnets display an apparent rotation of 270 degrees and are formed by one single crystallographic domain for the whole spiral. No evidence of modification in the stress field during garnet growth is observed in these snowball garnets suggesting that the whole spiral grew under a rotational regime. Quantitative microprobe analyses reveal that the end of the spiral for those garnet contain the same amount of Mn as the end of the central sector domain in garnet displaying 360 degrees of apparent rotation. The similarity in geometry between the central sector domains and the geometry acquired by the snowball garnets under the rotational regime strongly suggests that as long as the growth is accompanied by rotation, the primary core orientation is preserved, but once the rotation stops the crystallographic orientation may change. EBSD data also indicate that the central domain displays a crystallographic orientation characterized by a [001] pole oriented sub-parallel to the rotation axis of the snowball garnet. Moreover, in most crystallographic sectors, one of the two other [100] poles is (sub)parallel to the orientation of the internal foliation. This feature suggests that the crystallographic orientation across the garnet spiral is not random and that a relation between rotation axis, internal foliation and crystallographic orientation does exist. In this view, EBSD data can potentially be used to distinguish between the rotational and non-rotational models.