Gareth Seward and I have been having a discussion about using the "auto" time constant pulse processing option in Pathfinder. I'm posting this in the Thermo board because I know that Bruker and JEOL EDS systems do not have such a feature (though I am unsure about Oxford or Edax).
Basically if you set the pulse processing time constant in Pathfinder (or NSS) to *zero*, the software will automatically adjust the EDS time constant so that the dead time is always around 40 to 50%. It's sort of nice when one is cruising around a sample where different phases can produce different dead times, or when you're using different beam currents on different samples. But we were wondering what effect it might have on quantitative analysis when utilizing the integrated WDS and EDS acquisition methods in Probe for EPMA.
Gareth and I came up with several thoughts on this. First, it should be pointed out that because a microprobe has a faraday cup, every time the faraday cup is inserted, the auto time constant mode in Pathfinder thinks "gee, the count rate is really low, I'd better increase the time constant to get better spectral resolution". But then as soon as the faraday cup comes out, the dead time shoots up to 90% or whatever, and here's the key point: it takes about *15 seconds* for Pathfinder to adjust the time constant back down so that the dead time is back to 40 to 50%.
And second, during automation, by default, as soon as the faraday cup comes out, PFE starts the EDS and WDS acquisition. But here's the thing: as soon as the EDS acquisition is started, Pathfinder *stops* adjusting the time constant, not providing enough time for the auto time constant to adjust for an optimum dead time. This is actually a good thing because Pathfinder has to know what the time constant was utilized over the whole EDS acquisition for spectrum processing (sum peak removal, etc).
But if there hasn't been enough time for the auto time constant to do its thing, and one is running a fairly "hot" beam current condition, the spectrum acquisition dead time might be "frozen" at over 90%, and then the spectrum is not really suitable for quantification.
Now, one can utilize the so-called "decontamination/incubation" time delay feature in PFE, so that after the faraday cup is removed, a user specified delay is implemented (see the Acquisition Options dialog from the PFE Acquire! window for more description). But who wants to wait for 15 seconds for the acquisition to start?
That is why we wondered if simply setting the time constant to some reasonable compromise between count rate and dead time is the best way to go for quant acquisition when using integrated EDS and WDS acquisition.
What do you all think?