So, I have been recently playing with the element mapping function both on the JEOL system (JEOL JXA-8900) and with Probe Image and noticed that I get spurious intensity peaks for Si and Al (on the TAP's) that I do not understand. Maybe someone has seen this before and can point me in a direction. Those make for very ugly maps and also cause totals up to 170wt% when I quantify the maps with Probe Image.
The "blibs" only occur for Si and Al maps, returning uniformly 511 counts (compared to a max of around 300-350 cts for the upper limit of the intensity distribution). However, if I run the chart recorder alongside the mapping there are no spikes that would coincide with the blibs (or at all).
These high intensity blibs occur at random positions in the maps but their percentage is reproducible (around 2%, both for Al and Si). It does not matter if I do a beam or stage map. Maps for neither Na and Mg on TAP nor for Si (or Cr) on PET show this behavior. Looking at older maps for Si and Al shows that the blibs were sometimes present but were much more rare.
I double checked the PHA settings and I ran the JEOL maps with 12nA, definitely not excessive. Running the maps on oxide standards or at a lower current (3nA) on the metal standards improves the maps. Nevertheless, the count rates at 12nA are not near the upper limit (around 23 000 for Al)
I attached a few images obtained on (metal) standards (for all my three TAP's, run with the JEOL software) and a larger map of a K-feldspar (run at 80nA with Probe image) to better illustrate what is happening.
Anyone has an idea what is going on here?
Thanks!
Hi Anette,
Were the samples last processed with colloidal silica, SiC, and/or alumina polishing compounds?
Try diamond or very meticulous ultrasonication. I find hot and cold spots in most sonicbaths-which you can feel via vibrations in the beaker. You could test the abundance of bright pixels against number of rinses and successive ultrasonication.
Hopefully that, and not something more serious with electronics,
Jared
Quote from: Anette von der Handt on October 30, 2014, 07:00:48 PM
I attached a few images obtained on (metal) standards (for all my three TAP's, run with the JEOL software) and a larger map of a K-feldspar (run at 80nA with Probe image) to better illustrate what is happening. Anyone has an idea what is going on here?
I'm not sure, but I suspect something you might not like to hear...
On your JEOL 8900 have the detectors ever been replaced? I ask because on the JEOL instrument there are generally some sealed Xe detectors and these sealed detectors will eventually get contaminated and pumped out and will start to get noisy after some number of years. Colin MacRae suggests replacing them every 5 years or less.
Even flow detectors will eventually get contaminated and need replacement after 15 or 20 years or so. I've replaced a flow detector on my ancient SX50 after it got noisy for low energy photon detection.
Hi Jared and John,
thanks for your answers.
The peaks are definitely not from sample prep because their absolute locations are different in each map and not reproducible. The samples were produced in different labs, partly by me and and the density of the blibs has increased with time (without re-polishing or such).
I am afraid that the correct answer is that the detectors are aging. These are all flow detectors but at least Spectrometer 3 gives me also a very strange PHA scan for carbon. I will run another test with C and O with maps to see what the low energy range does and then call in the service, I guess. :-\
Hi Owen,
the sealed counters seem to be fine (gain/bias in good values). Different samples show it, so it does not seem to be sample related.
There are is a test to see if the detector window is deteriorated or charging up. I did not get a good read on charging but otherwise only get a very broad distribution rather than a defined peak for carbon (other elements are fine), which makes me suspect that the window/detector has a problem.
How to check your detector window (from the JEOL manual) in case you are interested:
Using the CKa line, carry out the pulse-height analysis to check the window.
The measurement conditions are as follows:
Analyzing crystal: STE (I use LDE's obviously)
Specimen: Carbon graphite
Accelerating voltage: 10kV
- Perform peak search and set the spectrometer to the peak position
- Open the Full Scale window from the Ratemeter window and the Full Scale to 10 kcps. With the PROBE CURRENT knob on the 1st control panel set the ratemeter bar (counting rate) to read 7,000 to 8,000 cps and leave the instrument for 10 minutes.
- Measure the pulse height distribution using the SCA Scan window
- Using the HV Scan window, a subwindow of the SCA scan window, perform HV scan with Base set to 3.0 V and Window at 0.2 V and measure the voltage that makes the counting rate maximum.
- Open the SCA control window and set High V to this voltage. With this operation the pulse height of CKa is set to around 3 V.
- Open the Base L Scan window, a subwindow of the SCA scan window, and perform Base L Scan measurement with Base set at 0.2V, Window at 0.2V, Step at 0.2V and S time at 1000 msec to check the pulse height distribution of the CKa signal pulse
Inspect the window for deterioration or damage in accordance with the pulse height distribution obtained in the preceding step (see attached Figures).
As the conductivity of the window surface deteriorates (window charging), the X-ray intensity will gradually decrease even if the probe current is kept at a fixed value. Check the window charging as follows:
From the Ratemeter window, open the spectrometer window that shows the settings of the specified spectrometer whose proportional counter window is to be checked for charging. Using the SCA Control window, a subwindow of the Spectrometer window, increase the voltage applied to the proportional counter by 50 to 100 V, then decrease it to the original value. If this operation lowers the X-ray intensity below the value before raising the voltage, proportional counter window is charged.