XRF Laboratory Tips – Drift Correction
XRF Laboratory Tips – Drift Correction
Drift Correction (Monitor)All XRF laboratories regularly run drift correction (monitor) and check/quality control samples. Many analysts simply look at the data in isolation. For example, has the monitor changed by more than x% since last run or is the result of check sample within specified limits ? However the data contains a lot more information that can be used proactively to control the XRF analytical process. The regular monitor and QC results need to be seen as a snap shot on an ongoing process.
QC/Check Samples
In many labs QC/check samples are run at say, the beginning of the day shift. The results are recorded, checked to see if they are “ok” and filed. Action is only taken if they are outside of the specified range. Rarely are the results compared with previous results. Even in labs with on-line control charts, there is often a tendency to “just worry about today's result”. Examining the data for patterns and trends can provide the analyst with much valuable data. For example, consider the QC/check sample results in Figure 1. Although all the QC/check sample results still lie within the control limits, there is a clear upward trend in the data. If the trend continues the QC/check sample result will go outside the control range. Obviously once this happens, corrective action must be taken and assay results can not be reported to the customer – i.e. the XRF is off line until the problem is fixed. Why wait until a crisis develops ? The QC/check sample results in Figure 1 clearly show a pending problem. If the analyst recognises this, pro-active corrective action can be taken beforethe XRF results become unusable.
QC/check sample data can also show instrument instability (for example: step changes, large fluctuations in results) and can provide important leads for the Service Engineer (eg are the problems confined to analytes analysed using a specific crystal and/or detector).
Monitor Results
Examination of the Monitor results over time can also provide the analyst with useful instrument performance data. The most obvious information is long term instrument stability. Large fluctuations in the Monitor are usually a sign of instrument instability, although a poor laboratory environment can also contribute to instrument instability. Some information about the condition of the X-ray tube can also be gained from the Monitor.
Falling Monitor counts can be caused by deposition of material on the X-ray tube window (eg powder from friable pressed powder samples falling onto the window or tungsten from the tube filament depositing on the inside of the window).
Sometimes before the X-ray tube filament fails the tube X-ray output will rise significantly and this will be reflected in a corresponding rise in Monitor counts .
XRF Spectrometer Precision Tests
All XRF laboratories should, as part of their on-going Quality Assurance program, utilise Australian Standard AS 2563-1996 (Wavelength dispersive X-ray fluorescence spectrometers – Determination of precision). This Standard details procedures to test individual components, for example the crystal changer, of a wavelength dispersive spectrometer at a selected precision level. The standard can be obtained from the SAI Global website (/http://www.saiglobal.com/).
A well maintained spectrometer housed in a well controlled laboratory environment can easily operate at a much higher precision than required for routine analysis. This allows the XRF analyst to use these tests for two purposes:
Conclusion
The data obtained from the normal day to day QC/QA testing protocols contains a wealth of information that allows the XRF analyst to take a pro-active approach to calibration and instrument problems. Do not just look at the individual “today” results. Look at the overall picture.
Ken Turner
Ken Turner Consulting
Drift Correction (Monitor)All XRF laboratories regularly run drift correction (monitor) and check/quality control samples. Many analysts simply look at the data in isolation. For example, has the monitor changed by more than x% since last run or is the result of check sample within specified limits ? However the data contains a lot more information that can be used proactively to control the XRF analytical process. The regular monitor and QC results need to be seen as a snap shot on an ongoing process.
QC/Check Samples
In many labs QC/check samples are run at say, the beginning of the day shift. The results are recorded, checked to see if they are “ok” and filed. Action is only taken if they are outside of the specified range. Rarely are the results compared with previous results. Even in labs with on-line control charts, there is often a tendency to “just worry about today's result”. Examining the data for patterns and trends can provide the analyst with much valuable data. For example, consider the QC/check sample results in Figure 1. Although all the QC/check sample results still lie within the control limits, there is a clear upward trend in the data. If the trend continues the QC/check sample result will go outside the control range. Obviously once this happens, corrective action must be taken and assay results can not be reported to the customer – i.e. the XRF is off line until the problem is fixed. Why wait until a crisis develops ? The QC/check sample results in Figure 1 clearly show a pending problem. If the analyst recognises this, pro-active corrective action can be taken beforethe XRF results become unusable.
QC/check sample data can also show instrument instability (for example: step changes, large fluctuations in results) and can provide important leads for the Service Engineer (eg are the problems confined to analytes analysed using a specific crystal and/or detector).
Monitor Results
Examination of the Monitor results over time can also provide the analyst with useful instrument performance data. The most obvious information is long term instrument stability. Large fluctuations in the Monitor are usually a sign of instrument instability, although a poor laboratory environment can also contribute to instrument instability. Some information about the condition of the X-ray tube can also be gained from the Monitor.
Falling Monitor counts can be caused by deposition of material on the X-ray tube window (eg powder from friable pressed powder samples falling onto the window or tungsten from the tube filament depositing on the inside of the window).
Sometimes before the X-ray tube filament fails the tube X-ray output will rise significantly and this will be reflected in a corresponding rise in Monitor counts .
XRF Spectrometer Precision Tests
All XRF laboratories should, as part of their on-going Quality Assurance program, utilise Australian Standard AS 2563-1996 (Wavelength dispersive X-ray fluorescence spectrometers – Determination of precision). This Standard details procedures to test individual components, for example the crystal changer, of a wavelength dispersive spectrometer at a selected precision level. The standard can be obtained from the SAI Global website (/http://www.saiglobal.com/).
A well maintained spectrometer housed in a well controlled laboratory environment can easily operate at a much higher precision than required for routine analysis. This allows the XRF analyst to use these tests for two purposes:
- Establishing that the spectrometer is functioning with sufficient precision to provide the required precision in the final analytical result. Testing should be carried out at the precision level required for routine analysis.
- A pro-active tool to identify components within the spectrometer that may require maintenance. Testing should be carried out at a higher precision level than required for routine operation. It is important to keep records of the tests and to compare present results with previous tests.
Conclusion
The data obtained from the normal day to day QC/QA testing protocols contains a wealth of information that allows the XRF analyst to take a pro-active approach to calibration and instrument problems. Do not just look at the individual “today” results. Look at the overall picture.
Ken Turner
Ken Turner Consulting