Tuesday, 1 March 2011

ICAM 2011 presentation on SEM-EDS mineral identification

Our team at FEI Australia Center of Excellence for Natural Resources is going to present a number of talks at a range of interesting conferences this year. In particular, we will have a strong presence at the 10th International Congress for Applied Mineralogy (ICAM 2011) in Trondheim, 1-5 August 2011. Here is a preview of our talk on a new QEMSCAN protocol for mineral identification using the new Spectral Analysis Engine.

"SEM-EDS based protocol for subsurface drilling mineral identification and petrological classification"
by David Haberlah,Michael Owen, Pieter W.S.K. Botha and Paul Gottlieb

Integrated scanning electron microscopy and energy-dispersive x-ray spectroscopy (SEM-EDS) solutions are widely employed in the mining sector. In subsurface drilling for hydrocarbons, SEM-EDS systems are less common and often only applied to the analysis of thin sections of core samples. Coring wells is more expensive and takes significantly longer than drilling wells. The downside of drilling versus coring is the produced slurry of fine rock cuttings and drilling fluid additives that prove difficult to interpret in terms of rock properties. Thus, drill cuttings are rarely analysed beyond qualitative microscopic descriptions by the mudlogger. Automated SEM-EDS solutions have great potential by providing quantitative cutting-by-cutting data. One difficulty is to develop robust mineral and petrological identification for the cuttings and the drilling mud. Here, we present a fully automated quantitative SEM-EDS based approach, measuring pixels along a predefined grid and comparing measured energy-dispersive x-ray (EDX) spectra with a library of mineral compositions. The analytical protocol is based on the new QEMSCAN® Spectral Analysis Engine (SAE) employing the elemental concentrations method. Mineral phase identification is accomplished by a multi-layered approach using iDiscover™ software v.5.0.

First of all, elemental peak positions and relative intensities from the measured EDX spectra are compared with spectra of elemental standards measured on the same system. The new SAE can perform identification and quantification of up to 72 elements. However, in subsurface drilling applications, the majority of reservoir and seal rock-forming minerals can be adequately discriminated by less than 20 elements. Subsequently, a position-dependent, multi-layered approach to mineral classification is applied. The elemental composition from the measurement point is compared to elemental ranges calculated from synthetic mineral spectra and high-count spectra measured on mineral standards. The elemental ranges are typically based on 100 iterations or more, simulating statistical variability in measured low-count spectra. Variation in the composition of mineral phases and across mineral groups is accounted for by a second layer, merging end-member definitions into single combined mineral (group) expressions (e.g. orthoclase, sanidine, and anorthoclase into alkali feldspars), and by adding common accessory and substitute elements into a “may have” category. Next, the elemental ranges are adjusted interactively on measured samples of known mineral composition. Finally, a few broad definitions trapping poorly defined components such as the mounting medium, inorganic drilling fluid additives, and organic matter can be defined. Once the mineralogical composition of the individual cuttings is fully mapped, they can be further categorised into discrete classes such as rock types. Micro-lithotype classes are based on expressions that take into account textural attributes, such as mineral associations and grain sizes. Rock characteristics of particular interest to reservoir modelling, e.g. the presence of pore-filling cements, can be identified and numerically reported.

Our results suggest that SEM-EDS based mineral identification and petrological classification of drill cuttings can significantly reduce the need for expensive coring, reduce the size of cuttings that can be analysed, and overall improve petroleum reservoir characterisation and modelling efforts. The direct quantitative cutting-by-cutting measurements can also provide an independent means for calibrating gamma ray wireline log data, by reporting the presence of low-potassium clay minerals (i.e. kaolinite and chlorite) in seal formations, and high-potassium feldspars in reservoir rocks.

No comments:

Post a Comment