Examples of charging effects on the spectral quality of X-ray microanalysis on a glass sample using the variable pressure scanning electron microscope.

The performance of X-ray microanalysis in the variable pressure or environmental scanning electron microscope (VP-SEM or ESEM) is limited by skirting. Under certain conditions, charging effects can occur and change the X-ray emission, which affects the X-ray microanalysis. The conventional way to evaluate charging is to calculate the Duane-Hunt limit by fitting the X-ray intensity region located below the energy cut-off. Nevertheless, this method appears to have serious limitations for instance in the case of strong insulators. A perfect example of this limitation is to study the evolution of composition of an alkali glass with time. This paper reports on the evolution of the sodium X-ray intensity with time depending on accelerating voltage, pressure and presence of a surface coating. For certain conditions, a decrease of sodium X-ray intensity with time was observed but for other conditions the reverse behavior was noticed. The increase of sodium X-ray intensity with time was obtained when the force created by the surface electrons was stronger than the force generated by electrons trapped in the interaction volume, whereas the decrease of sodium X-ray intensity occurred when the force generated by electrons trapped in the interaction volume was the stronger. The variations of sodium X-ray intensity were also compared to the variation of the Duane-Hunt limit, the determination of which is studied in detail in this article.

Skirting: a limitation for the performance of X-ray microanalysis in the variable pressure or environmental scanning electron microscope.

The variable pressure or environmental scanning electron microscope (VP-SEM; ESEM) has become the microscope of choice for many scientists and technologists. Hence, the development of robust methods for X-ray microanalysis, limited by skirting, has become critical. In this paper, two pressure variation correction methods (Doehne and Gauvin) are compared. Both of these methods appear to be effective; the results were found to be well within 10% of the values obtained at 0 Pa. The Doehne method is dependent on an empirical factor (D), therefore the accuracy of the results will depend on the accuracy of this value. Also the Doehne method is compromised by the nonlinearity of the response with pressure. The Gauvin method is more user-friendly and more precise when considering the total range of pressure.