GEOSCAN, résultats de la recherche


TitreQuantitative determination and mapping of trace element concentrations in geochemistry
AuteurJackson, S E
Source2012 winter conference on plasma spectrochemistry, abtracts; 2012 p. 122-123
Séries alt.Secteur des sciences de la Terre, Contribution externe 20110234
Réunion2012 winter conference on plasma spectrochemistry; Tucson; US; janvier 9-14, 2012
Sujetsanalyses des éléments en trace; géochimie des éléments en trace; éléments en trace; géochimie
ProgrammeDeveloppements methodologie, Initiative géoscientifique ciblée (IGC-4)
Résumé(disponible en anglais seulement)
With the recent advent of sophisticated data processing software (e.g., Iolite [1]), 2-D element mapping by LA-ICP-MS has become routinely possible [1, 2]. In the geosciences, there has been particularly strong interest in mapping element distributions in rock samples because it elegantly presents information on the behaviour of trace elements during mineral crystallisation and precipitation processes, especially for trace elements that typically do not form mineral phases.
Most published LA-ICP-MS element maps to date have been produced from time-resolved signal intensity data acquired during multiple, parallel, continuous line-ablations across selected areas of petrographic sections. The data were then deconvoluted and digitally combined to generate elemental signal intensity or, more rarely, concentration maps. There are, however, three shortcomings of the continuous line scan ablation approach: (1) it results in degraded spatial resolution (> ablation spot size) because new ablation aerosol is continuously mixing with previously ablated material; (2) it results in different spatial resolution in the X and Y directions, which must be accounted for during data processing; (3) it does not allow appropriate handling of surface contaminants, which can be the source of significant error, especially when samples containing multiple sulfide phases are mapped. Unlike common rock-forming silicate minerals, which are composed largely of 10 or so elements, sulfide ore minerals commonly contain 10¿s of percent of elements that are trace constituents (ppm) in most other minerals. Surface contamination occurs both by ¿smearing¿ of the softer sulfides during sample polishing and also by condensation of ablation products. If not dealt with appropriately, this contamination can result in erroneous signal enhancements of several orders of magnitude.
An additional limitation of most published element maps of multi-mineral samples is that they present signal intensity data only, presumably because of the extreme complexity of calibrating data using the conventional internal standardisation approach, which would require that a major element concentration be applied to correct the data for every point in the sample map.
In this study, we present new techniques for element mapping of geological samples of mixed mineralogy that eliminate the problems of surface contamination and aerosol mixing, resulting in true spot-size-limited spatial resolution. Furthermore, the techniques produce quantitative element concentration maps with no requirement for prior knowledge of the concentration of elements in the mineral grains. We then apply this technique to determining the trace element paragenesis of samples of gold ore from the Colosseum Mine, California.
Sampling was performed using a Photon Machines Analyte.193 laser ablation sampler coupled to an Agilent 7700x quadrupole ICP-MS. A new sampling protocol is presented which involved ablating lines of individual (14 or 17 µm) square spots ('line of spots' sampling - LOS), aligned edge to edge (Figure 1), on 14 or 17 µm line spacing. Each LOS analysis was a separate acquisition that consisted of 30 seconds of background data acquisition (laser off), followed by analysis of up to >150 separate spots. Each spot analysis (75 or 90 laser pulses at 30 Hz laser repetition rate) was preceded by 3 pulses, followed by 1.5 s of flushing, to rid the sample surface of smeared and ablation contaminants, and allow this material to be flushed out of the sample cell.
Signal intensity vs. time data files for each LOS were processed in three stages using in-house developed software: (1) peak-recognition software detected the signal peaks related to each spot ablation, rejected data associated with the cleaning pulses, and integrated the remaining mass sweeps for each spot into a single reading for each element; (2) background correction was applied and the net signal intensity data were converted to concentrations on a spot by spot basis. Standardization was achieved using calibration against an external basaltic reference standard (USGS GSD-1G or GSE 1G [3]) and internal normalization to 100% total element abundance [4]. This method is uniquely suited to analysis of multi-mineral samples since it requires no prior knowledge of the concentration of elements in the mineral grains and it effectively deals with instances where two or more mineral phases are ablated simultaneously (grain boundaries, inclusions). It does, however, require that all elements present in the sample at significant concentrations (percent level) are measured or can be calculated (e.g., C, O); (3) the concentration data for each LOS were converted to concentration maps by digitally combining the data and assigning colors using a variety of user-selectable color scales (e.g., linear, logarithmic, percentile, etc.).
The new methodologies have been tested on several samples of gold ore from the Colosseum gold deposit, Clark Mountains, CA. This deposit is hosted in a mineralized breccia pipe within a 102 Ma rhyolite plug. The pipe is composed of rock fragments in a matrix of hydrothermally deposited minerals. These include early, coarse grained, fractured, euhedral pyrite and later, fine-grained pyrite and associated base metal sulfides (largely chalcopyrite, sphalerite and minor galena, arsenopyrite and tetrahedrite-tennentite) and gold in a matrix of siderite (Fe-Mg carbonate) ± sericite. Gold, when visible, is paragenetically late and is associated with late fine pyrite and in fracture fillings ± base metal sulfides in coarse early pyrite.
The element concentrations ranges shown in the Colosseum element maps (Fig. 2) agree with independent LA-ICP-MS analyses performed using sulfide standards and conventional internal standardization techniques. The maps confirm exactly the paragenesis determined by petrographic observation and provide quantitative information on element behaviour, most importantly those elements not represented in the sample mineralogy (e.g., Co, Ni, Ga, Ag, Te, In, W, Tl, Bi).
Figure 2. Contrasting element behaviour in Colosseum gold ore. Te is enriched in early pyrite, and shows single pixel width (14 µm) growth zoning, while Au is enriched (ca. 1 ppm) in late pyrite and in fractures in early pyrite, confirming its late paragenesis. Concentration scales are in ppm.