Steve Chapman Protrain

The advent of the ultra thin widow (UTW) and super ultra thin window (S-UTW) energy dispersive x-ray detectors has allowed the investigation of elements down as far as element 6 carbon.  The quantification of elements above element 11 has been vastly simplified by the EDS manufacturers through their “semi quantitative” analysis programs.  It is unfortunate that the ease at which it has seemed possible to quantify down to element six actually over simplifies the procedure creating errors in the analysis.

With very light elements (<Na) the condition of the detector itself plays a large part in the level of x-ray signal received.

The detector, window whilst being nominally very thin, is subject to contamination from the vacuum environment.  Any cold surface within a vacuum acts as a vacuum trap adsorbing material onto its surface, the window of the EDS detector being a typical example.  The build up of contamination, often found to be rotary pump fluid, may considerably reduce the energy of incoming x-rays having a more major effect the lower the x-ray energy.

The detecting crystal, itself at a lower temperature than the window, may also become contaminated through ice formation on its surface.  As the detector window ages it becomes perforated and on letting air into the microscope system moisture passes through these holes condensing as ice on the crystal.  This build up of material also reduces the energy of the incoming x-rays having a more major effect upon the light element contribution.

Evidence of contamination on energy dispersive x-ray detectors may be found through an investigation of the count rates from standard samples that contribute elements to the area below sodium in the periodic table.  A periodic test relating the copper K alpha peak (8.04keV) height to that of the copper L alpha (0.93keV) peak may act as a first indication of light element energy loss.  As the detector becomes contaminated x-rays will increasingly be absorbed by the dirt on the window, and/or the ice contamination on the detector, reducing the height of the L alpha peak.  The test requires a copper sample which is investigated at a fixed accelerating voltage, at a fixed processing time, at a count rate of 1,000cps and a live time of 100secs.

A second test, developed by Oxford Instruments, is particularly sensitive to ice build up on the detecting crystal.  This test requires a chromium specimen which is investigated at 5kV, with a long processing time, at 1000cps for 100 seconds live time.  At around 0.57keV two peaks are present representing the chromium L L line and the L alpha line.  If the x-ray detector is free from ice the two peaks are similar in height or the right hand peak is larger.  If there is ice build up on the detector there is a higher level of oxygen which produces an oxygen absorption edge over the L alpha line, positioned to the right, reducing its height.

Organisations interested in silicon to oxygen ratios should be aware that the problems outlined above have a considerable effect upon the oxygen contribution.

Investigating on a microscope and x-ray detector under one year old the O2 to Si ratio at 4kV, with a slow processing time, was 3 to 1.  Using faster processing times the ratio decreased as the processing time was reduced, but did not fall below an O2 to Si ratio of 2.4 to 1.  At 25kV the emphasis is on the higher energy peak resulting in a Si to O2 ratio of 6.2 to 1.

With a much older microscope and x-ray detector (15 years old) the result of an investigation in O2 to Si ratio at 4kV, with a slow processing time, was a ratio of 1.7 to 1.  At 25kV the Si to O2 ratio was 5.5 to 1

The solutions to these problems are two fold, the window needs cleaning and/or the detector needs to be warmed up to remove the ice.  However the actions require considerable thought.

Firstly it would be dangerous to try to wash the window (1) without consulting the manufacturer (2) without conducting the chromium test and finding no indication of ice.  If ice is indicated this must be due to holes in the window which would cause additional problems if the window was washed.  Almost certainly the washing fluid would penetrate the window and this would cause even greater ice build up problems than those that already exist.

Secondly if the detector has ice contamination this would indicate holes in the window.  Allowing the detector to warm under the correct conditions may well dissipate the ice but, when re cooling the system, it may not be possible for the detector vacuum to recover due to too high a leakage rate.  It is quite possible for a detector, even with holes in the window, to retain a workable vacuum.  However expecting such a condition to allow the vacuum to recover without re pumping mechanically is very doubtful.

X-ray detectors may be recovered through mechanical re pumping after being allowed to warm to remove ice contamination.  Specific procedures are required to carry out this warm up; it should not be attempted without advice from a qualified source.  In the case of a contaminated window the chromium test should indicate that there is no ice contamination before moving further.  In this case discussions with the detector manufacturer are strongly recommended in order that information on a suitable media and washing technique may be obtained.