The X-ray Fluorescence Microprobe at the Australian Synchrotron

The microprobe facility has two microprobes, differing mainly in their achievable spatial resolution. In brief:

• KB microprobe: spot size range from 2 x 1 microns (H x V) to 10 x 10 microns
• ZP microprobe: spot size is around 100 nm (0.1 microns).

In order to determine which microprobe would be best suited to your investigation, you will need to know the size of the smallest feature that you need to resolve.

We also have two fluorescence detectors. The main differences between these are (1) speed and (2) energy range (which determines which elements you can see). In brief:

• Vortex detector
  Dwell time will usually be around 0.2-2 sec per pixel.
  Sensitive to photons with energy above 1.1 keV
  K-, L-, or M- lines of all elements heavier than Mg (Z = 12)
• Maia detector
  Dwell times of 0.5 msec - 50 msec have been demonstrated
  Sensitive to photons with energy above 3.3 keV
  - K-lines of elements heavier than Ca (Z = 20)
  - L-lines of elements heavier than Sb (Z = 51)

While the Maia detector is extremely reliable, we note that it is a research detector and so we do not carry a complete spare.  As such, it is available on a best-effort basis, and every experiment should anticipate using the Vortex detector as a back up if needed.

Note that, at present, the ZP microprobe can ONLY be used with the Vortex detector.

If you are using the KB microprobe you will need to decide which fluorescence detector would be best suited to your investigation. This decision is a trade-off between elemental sensitivity and scan speed (which will translate directly to increased measurement areas). Therefore you will need to know which elements you would like to map. If you do not need to see elements with fluorescence energies between 1.1 keV and 3.6 keV, and are using the KB microprobe, then we recommend using the Maia detector.

Estimating the scan duration

The scan duration is determined almost entirely by the product of (1) the total number of pixels in the scan and (2) the dwell at each pixel.

One parameter that we have not yet discussed is the scan area. Of course, this depends on your specimen, but is often dictated by time constraints.

Example

Let's say that you have two cell types with three treatments, and wish to have n=3 for each cell (giving a total of 18 cells).

Your study aims to investigate correlations between S and Se, so you will need to use the Vortex detector.

The cells are about 5 microns in diameter, and so we will image them using the ZP microprobe (5 x 5 pixels is not sufficient for this study!)

The scan time would be calculated as:

18 [cells] x 50 [pixels wide] x 50 [pixels high] x 1.5 [seconds per pixel] = 18.75 hours.

How much beamtime should you ask for? It is difficult to say, as other factors such as specimen mounting and instrument access will come into play. You will need to become more familiar with the instruments in order to estimate these overheads, which could easily range from 10 minutes / specimen to 1 hour per specimen. For a first visit, with a full day of measurement required, I would suggest a request for three days of beamtime. This request could be further justified with arguments such as:

1. Need to over-scan the specimen; a 6-micron scan is appropriate to measure a 5-micron cell. This will increase the required time by (60x60) / (50x50) or about 50%.
2. If trace metal concentrations are being mapped, it may be appropriate to increase the dwell per pixel, so as to collect more of the fluorescence signal, and obtain better statistics on this signal.

Selecting a good substrate

Two factors that influence the choice of substrate are:

• The substrate should not produce signal in excess of that expected from the specimen itself.
  For low-signal specimens (usually biological, but anything that is thin) this means it should be free of the elements of interest and should not produce significant x-ray scatter (this means that it should be as thin as possible, and made of reasonably light elements). For bulk specimens these issues are not so important, as the specimen itself will produce scatter. It is for this reason that thin specimens are generally preferred.
• The substrate should be relatively small so that we can:
  (a) bring the fluorescence detector very close to the specimen (useful for optimising fluorescence collection), and;
  (b) pack many close together on the specimen stage (useful for reducing change-over time).

Accordingly, we recommend;

• Silicon Nitride windows (see Silson at http://www.silson.com/index.html). These are particularly good as they come with field of view up to 5-10-20 mm (at a price).

Please contact us if you need a few windows to play with; we may soon be able to point you to a local manufacturer.

• EM grids (please be aware that metal constituents in the EM support frame may affect background, especially if the specimen is close to the gridbar or support)
• Formvar films - but how to hold them?
• LUXFilm (TM) - not yet tested, but these look promising (see Luxel at http://luxel.com/tem_supports.html). We could contact them to arrange metal-free supports.

Please let us know of any other mounts and suppliers that you think are good and we will add them to this list (and feel free to give us a sample of your favourite substrate).