Does dead time vary as a function of emission energy, or gain or bias?

[Anette von der Handt]

Here are also some graphs of the change in bias settings needed to keep the PHA peak centered at 6V with increasing beam currents. Plotting it against count rates would be better but this might be still interesting to others as well.

All data was acquired at 15kV on Si metal again

TAP, GF


TAPL, GF


PETL, GF

[Probeman]

Here are also some graphs of the change in bias settings needed to keep the PHA peak centered at 6V with increasing beam currents. Plotting it against count rates would be better but this might be still interesting to others as well.

This is excellent data, very interesting. Note the new plot item "Raw On Counts" in the Output | Output Standard and Unknown XY Plots menu which now allows plotting the x-axis as intensities uncorrected for dead time and beam current (cps).

It is very interesting to see that pulses above 10v are also still counted on JEOL instruments, just as they are on Cameca instruments.

Therefore, is it really necessary to adjust the PHA peaks as a function of beam current to keep them "centered"? Since pulses above the PHA "maximum" voltage are still counted when in INTEGRAL mode, isn't it just easier to set up our PHAs *once* so that for a given maximum beam current and highest expected concentration (for a particular run) we merely make sure that our PHA peaks are above the baseline? Again, here is Si Ka on a LTAP at 200 nA on SiO2:



Again: all pulses above 5v are still measured in INTEGRAL mode. In other words, let's say in a particular run I don't expect to measure Si Ka at greater than 200 nA, nor do I expect concentrations of Si higher than what is found in SiO2. So I simply adjust my PHA settings at the highest beam current expected and on the highest concentration expected, and merely ensure that the PHA peak is above the baseline setting.  What could be easier?

Therefore when we do measure at lower beam currents or lower concentrations, the PHA peak is only going to shift slightly towards the right and all pulses will still be measured as shown here:



However, because the JEOL noise peak needs to stay below the baseline at all beam currents, maybe I am missing something... can you do some PHA scans for us on your instrument with Si Ka peaked up on all spectrometers using your default gain setting for each spectrometer on Si metal at say 30 nA?

Then set your beam current to 200 nA and adjust your bias (as one usually does on a JEOL instrument) on each spectrometer to ensure that the PHA peak is completely above the baseline and that the baseline is set properly.

Then without changing any of your PHA settings, do two more PHA scans, one at 100 nA and one at 10 nA and let's see what happens to the noise peak. 

I expect the PHA peak to shift somewhat to the right as it does on the Cameca, but what happens to the noise peak at lower count rates?

[Probeman]

Anyway, here's another gain test (proxy for count rate) where I acquired a number of samples over a range of gain settings in integral mode, and one can see the intensities are essentially flat, except for spc 3 (again), but it is much flatter than before:

Silly me, I just thought of a simple way to perform gain and bias testing without having to specify sample setups with different gains or biases!

And that is, from the PHA dialog in PFE just adjust your PHA peak so that is is entirely above the baseline level (including the escape peak if present) at whatever count rate you are at, and then just specify a range of gain values for your gain scan (if a Cameca instrument) starting at your current gain setting and continuing upwards, and for JEOL instruments, just specify a range of bias values starting at your current bias value and continuing upwards.  Of course these tests must be performed in INTEGRAL mode!



The only downside is that one can only test one spectrometer at a time...

And you'll also probably want to set a reasonably long count time for these gain/bias scans to get decent counting statistics.  I thought of this while testing the PHA scan methods in simulation mode and accidentally had the TDI correction turned on and saw this output for a bias scan:



I mean it was modeling a beam sensitive sample, right?  So obviously perform these gain/bias scans on a sample that is not beam sensitive!