X-Ray Microanalysis of Burgess Shale and Similarly Preserved Fossils

Abstract

Non-biomineralized fossils from the Burgess Shale Formation are typically preserved as near two dimensional carbonaceous remains of the ­original tissues associated with films of aluminosilicate minerals; the host lithologies have experienced a long, complicated, diagenetic and metamorphic history. Such ‘layered’ substrates exhibit lateral and vertical variation in sample chemistry over micron-scale distances. X-ray microanalysis of these presents a particular challenge to the analyst, as the results are highly sensitive to the choice of operating conditions.

In a scanning electron microscope an electron beam generates an interaction volume at and extending below the sample’s surface; various signals, including x-rays, are generated from different parts of this volume. X-ray microanalysis is therefore not an analysis of the surficial chemistry of a sample. The lower the mean atomic number of the area impacted, and the higher the accelerating voltage, the further the electron beam will penetrate, the effects of which on the generated and emitted signals can be readily visualized via computer-based simulations.

The emitted X-rays can be collected via either a wavelength or an energy dispersive spectrometer and provide compositional information. Results are presented as (a) quantitative analysis; (b) an x-y plot of x-ray energy against counts (an X-ray spectrum); (c) a visual image of abundance of an element over a user-defined area (an element map).

Specimen preparation prior to analysis is designed to eliminate sample charging (most geological samples are dielectric) and topography, both of which impact negatively on X-ray microanalysis. When sample integrity must be maintained, unprepared samples may be used. Grounding part of the sample will minimize the likelihood of specimen charging; in ‘variable pressure’ SEMs the presence of gas in the specimen chamber negates the tendency for any charge to accumulate on the surface, but reduces markedly the resolution of analyses (the ‘beam skirt’ effect). Topographic artifacts are inevitable. Surface irregularity alters absorption path length; the longer such a path and the lower the energy of the X-ray the less likely it is to be emitted. Compositional analyses of unprepared samples are best considered a qualitative assessment of a sample’s chemistry.