Academic Staff Member
Contact: Lab manager
East Wing - Ground floor Room N029 (Tel: 3997)
In electron probe microanalysis a finely-focussed electron beam impinges on the sample and the chemical composition is derived from the wavelength and intensity of the lines in the emitted X-ray spectrum. It is capable of quantitative analysis with accuracy of about ±1% for major elements, and detection limits ranging from 20 to 1000 ppm, depending on element, matrix phase, and time per analysis. The spatial resolution is approximately 1 µm. Element distribution line scans and 2-dimensional 'maps' can be produced as well as point analyses.
Location & Bookings
The Dept. electron microprobe is a Cameca SX100 located in room 29 (E. Wing, ground floor). Enquiries about using it should be addressed to Iris Buisman. First-time users should arrange to meet and discuss the measurement strategy appropriate to their research and techniques of sample preparation.
Peak demand for the laboratory typically occurs from October to January. At other times of the year, the SX100 is usually over-subscribed and waiting times are typically 2-5 weeks. The Electron Probe bookings is displayed on the intranet, 4 weeks in advance.
Sample PreparationAccurate analysis is possible only if the sample is well polished. To analyse rocks in thin section they must be prepared specially for the electron probe, which entails mounting a cut slice on a glass slide with epoxy resin, grinding to the right thickness and polishing the surface (instead of attaching a cover slip). Such sections can be viewed with the polarised light bench microscope in the usual manner. Slides 46-48mm long x 10-30 mm width fit the specimen holder of the electron probe. For some materials (e.g. ore minerals or separated grains) it is more appropriate to embed the sample in an epoxy resin block 26 mm in diameter.
Samples that are not electrically conducting (including most rocks and minerals) must be coated with a conducting layer to provide a path for the electron beam. For this purpose carbon is applied by vacuum evaporation. This operation is carried out the technician; sections for coating should be given to the lab manager. Please allow several days before your probe session if possible.
Finding specific features for analysis in the electron probe can be difficult and time-consuming: users are therefore strongly encouraged to equip themselves in advance with 'maps' and/or photomicrographs (note that photomicrographs should preferably be reversed in printing, to match the probe microscope image).
Operation of the Electron Probe
Having mounted the specimens in the holder, this is loaded into the instrument via an airlock; after the vacuum has recovered (in a few minutes), the beam can be turned on. The specimen stage is digitised in steps of 1 µm and can be moved either by pointing and clicking on the optical image of the sample with the mouse, or by keyboard commands.The instrument has a built-in optical microscope with zoom capability between approximately x100 and x1000, with polariser and analyser, and transmitted or reflected light. The image is displayed on a colour monitor. Scanning electron images can also be obtained: these cover a wide range of magnification and can be saved as digital files or printed out. Backscattered electron (BSE) images show compositional (mean atomic no.) contrast; secondary electron (SE) images show mainly topographic contrast. (Note that these images serve mainly to assist in locating points for analysis: for producing high-quality images for publication it is preferable to use the scanning electron microscope).On finding a feature of interest and switching the beam to spot mode, the X-ray spectrum can be observed by means of the energy-dispersive (ED) detection system. The elements present can be identified from the positions of the peaks on the energy scale [see figure], enabling mineral phases to be identified very rapidly. By accumulating X-ray counts for around 1 minute and processing the spectrum, a quantitative analysis can be obtained. This mode is limited to relatively simple applications (e.g. major elements in silicate minerals). In most cases, wavelength-dispersive (WD) spectrometers are used, but these are more time-consuming to set up and calibrate. They have higher spectral resolution, obtained by Bragg reflection from crystals of suitable interplanar spacing, and can thus separate the lines in the relatively complex spectra of 'heavy' elements. Their higher peak-to-background ratio also allows smaller concentrations to be detected than by the ED method.
Point analyses can be carried out one at a time, with the user selecting each point using the microscope or scanning images as described above. However, this involves waiting for each analysis to be completed, which typically takes ~3-6 minutes. A more efficient approach is to store all the points required and leave the instrument to run them automatically, thereby saving the user's time and making use of time out of normal working hours. This is particularly appropriate for recording lines or grids with large numbers of points. However, it is not suitable for analysing very fine scale features, where positioning of the beam is critical, since positional reproducibility on long automated runs cannot be guaranteed to better than a few microns.
Treatment of Results
The output of the probe is saved on the hard disk of the computer controlling the instrument and can be downloaded in a form suitable for reading into a spreadsheet such as Excel. Data will be mailed to users or can be exported onto CD or memory stick. The results are calculated initially in units of element weight percent, from which oxide weight percent, atomic percent, etc., are derived. For oxides a valency is assumed for each element. In the case of Fe a value of 2 is used by default and it is up to the user to carry out any recalculation required to allow for some or all of the Fe being in the trivalent state. A total oxide wt.% between 99% and 101% is acceptable, variation within this range being attributable to counting statistics. A total which is significantly low may be caused by surface topography; alternatively it may indicate one or more missing elements, or the presence of water (note that water content estimated from the deficit in the total is only approximate).
In reporting results, concentrations below the level of statistical significance should be given as zero (not detected). For major-element analyses (including the standard ED routine used for silicates) it is undesirable to express percentages with more than 2 digits after the decimal point. More digits are required, however, for lower concentrations as obtainable by WD analysis.
Users are trained and supervised by the Electron Probe Technician. Operating instructions for the equipment are clearly set out. Provided these instructions are followed, no special hazards exist. The probe may only be operated out of normal working hours by users who have received permission and completed the "Out of Hours Permission Form" for this equipment.
Any instrumental fault should be reported to Iris Buisman. The room is air conditioned, doors must be kept shut.
Under no circumstances should users attempt to rectify faults on the SX100 or associated computers. No files or programmes may be loaded onto the computer.
New probe users are encouraged to read the following book in order to gain a somewhat deeper understanding of the technique:
- Electron Microprobe Analysis and Scanning Electron Microscopy in Geology' by S.J.B. Reed, Cambridge Univ. Press, 2nd ed. 2005 (Dept. Library no. B.23.164.1).
Other useful links: