WDS start-up time

[Mike Matthews]

Does anyone know how long it takes for a gas flow counter WDS, from beam on, to start counting?

I've measured C Ka peak counts and absorbed current for 70s at 0.5s intervals on C coated Bi samples. The absorbed current is showing a sharp initial increase in the first 1 - 2s. If this initial increase is due to a local reservoir of C under the beam spot being cracked and deposited (lowering the BSE coefficient under the beam and thus increasing the absorbed current) it isn't being reflected in the x-ray data. Either the absorbed current isn't showing a deposition of C (in which case what the heck is it showing), or the x-ray detector is somehow missing the first couple of seconds, even though the two sets of data are apparently being recorded simultaneously (using the JEOL chart recorder). For normal analyses (i.e. beam stable samples) such a delay wouldn't be an issue and would give the beam time to settle before counting starts, but for beam sensitive samples t=0 counts actually being t=2s counts might be important (some of those volatiles can move really fast in the first few seconds!)

[Probeman]

Does anyone know how long it takes for a gas flow counter WDS, from beam on, to start counting?

I've measured C Ka peak counts and absorbed current for 70s at 0.5s intervals on C coated Bi samples. The absorbed current is showing a sharp initial increase in the first 1 - 2s. If this initial increase is due to a local reservoir of C under the beam spot being cracked and deposited (lowering the BSE coefficient under the beam and thus increasing the absorbed current) it isn't being reflected in the x-ray data. Either the absorbed current isn't showing a deposition of C (in which case what the heck is it showing), or the x-ray detector is somehow missing the first couple of seconds, even though the two sets of data are apparently being recorded simultaneously (using the JEOL chart recorder). For normal analyses (i.e. beam stable samples) such a delay wouldn't be an issue and would give the beam time to settle before counting starts, but for beam sensitive samples t=0 counts actually being t=2s counts might be important (some of those volatiles can move really fast in the first few seconds!)

Hi Mike,
I would have thought that it's almost instantaneous.

This is not that closely related to your question, and I haven't looked at absorbed current with the carbon signal, but this post here (see the attached image) here:

https://probesoftware.com/smf/index.php?topic=48.msg160#msg160

which shows a slight drop in carbon in the first few measurements using 0.1 um steps and 0.5 sec count times (which I had interpreted as burning off native hydrocarbon layer), followed by an steady increase in carbon (which I assumed was the beam starting to encroach on the deposited carbon ring from the previous measurements).

Can you show some plots of your data?
john

[Mike Matthews]

I'm measuring exactly the opposite! Static 25nA focussed beam on a C coated Bi at 3.4kV and I see (sometimes) a small initial increase then a steady decrease as the C coat is eroded by the beam. The example plot (which I think I've attached) shows no initial increase in C but does show a rapid increase in absorbed current which transforms into a linear decrease. Over the decreasing portion I can correlate the absorbed current with the C Ka counts, but this doesn't appear to hold for the increasing portion - and it's holding me up from finishing my paper!

[Probeman]

I'm measuring exactly the opposite! Static 25nA focussed beam on a C coated Bi at 3.4kV and I see (sometimes) a small initial increase then a steady decrease as the C coat is eroded by the beam. The example plot (which I think I've attached) shows no initial increase in C but does show a rapid increase in absorbed current which transforms into a linear decrease. Over the decreasing portion I can correlate the absorbed current with the C Ka counts, but this doesn't appear to hold for the increasing portion - and it's holding me up from finishing my paper!

That makes perfect sense that you're seeing the opposite of my measurement. Your sample is carbon coated, mine was not carbon coated, just clean Fe metal. The decrease in carbon signal in your sample is probably the eroding away of the carbon coat. Eventually I suspect you will see an increase in carbon signal as the carbon starts to deposit a ring around the beam impact point.  Unless you have a really clean vacuum system.

[Mike Matthews]

I’m seeing both erosion and deposition at the same time. Deposition dominates to form the classic ring around the beam point, but under the beam erosion dominates and so the measured C counts decrease. If the beam doesn’t drift the C counts should decrease until all the coat is removed. You should hopefully soon be able to read all about it in my paper (I use TDI in PfE to extrapolate back to the starting coating thickness). I’m just trying to get the last few edits done before sending it back to the reviewers. I’d like to be able to explain the initial rise in the absorbed current though.

[Probeman]

One possibility is that the JEOL pico ampmeters are getting slower with each instrument model.  The old 8900s had very fast response, but the 8200/8500 was slower and one needs to wait sometimes 2 seconds on the 8230/8530 models for a stable beam (or sample) current measurement.

I wonder if you're just seeing the slow pico ampmeter response time on your 8500?