AbstractThe aim of this study was to investigate low voltage microanalysis of radioactive (RA) materials. Field emission electron sources provide access to higher beam currents at smaller beam diameters down to lower accelerating voltages than tungsten sources. Surface layers rapidly contribute errors with decreasing accelerating voltage. Correcting for these can produce acceptable analyses down to ~7 kV. The high oxygen affinity of RA materials results in a surface oxide which also needs correcting for. Uranium has a wide range of possible oxide stoichiometries but substrate quantification is relatively insensitive to the oxide state. Modelling oxides as UO2 is sufficient.
Accurate measurement of layer thicknesses is the dominant factor at low voltages. The method of measurement has an impact on quantification. The tooling factor for in-coater film thickness monitors (FTM) is a source of error: Sensors co-located with the sample provide more accurate measurements than side-located sensors. Thicknesses calculated from measured k-ratios give more reliable results than the FTM.
The coating material also has an impact. Carbon can be significantly eroded by the electron beam but this can be mitigated by defocussing the beam to 10 m where the sample geometry permits. Carbon also requires correction for interference with the uranium N6-O4 line. High conductivity metals such as copper and silver can provide thermal protection for beam sensitive samples, but do incur higher absorption corrections, raising the lower limit for acceptable to analyses to ~10 kV. Aluminium offers low absorption of soft energy x-rays but the coating density was found to be lower than bulk metal and produced higher quantification errors compared to copper coatings.
Bi has been proved to be an effective non-RA surrogate for uranium.
|Date of Award||26 Nov 2020|
|Sponsors||Atomic Weapons Establishment|
|Supervisor||Benjamin Buse (Supervisor) & Stuart L Kearns (Supervisor)|