"Combined selected samples into a new sample" with standard drift correction

[Gseward]

John,

What happens with respect to calibration drift when analysing an unknown that has been created using the "combined selected samples into a new sample" option, specifically in the situation of combining unknowns with different elements (and potentially different conditions) into a combined sample? e.g. If unknowns x(el1, el2, el3) @t1, y(el4, el5, el6) @t2 and z(el7, el8, el9) @t3, are combined into sample xyz(el1 ..el9),  with standards collected at t0 and t5. The time stamp for the 'combined sample' is assigned t1 (I think). Will elements acquired at t1, t2 and t3 be assigned a drift corrected intensity associated with t1 (the 'combined sample' timestamp)? or does each element maintain a timestamp independent of the combined sample timestamp?

perhaps the answer is to use a different data collection strategy

Cheers,

Gareth

[John Donovan]

John,

What happens with respect to calibration drift when analysing an unknown that has been created using the "combined selected samples into a new sample" option, specifically in the situation of combining unknowns with different elements (and potentially different conditions) into a combined sample? e.g. If unknowns x(el1, el2, el3) @t1, y(el4, el5, el6) @t2 and z(el7, el8, el9) @t3, are combined into sample xyz(el1 ..el9),  with standards collected at t0 and t5. The time stamp for the 'combined sample' is assigned t1 (I think). Will elements acquired at t1, t2 and t3 be assigned a drift corrected intensity associated with t1 (the 'combined sample' timestamp)? or does each element maintain a timestamp independent of the combined sample timestamp?

perhaps the answer is to use a different data collection strategy

Cheers,

Gareth

Hi Gareth,
The short answer is that the real time stamp for the standard intensity drift correction occurs on a data "point" basis. That is all the elements for a single data line/point acquisition. I could have made the acquisition date/time stamp on an element basis but I didn't.

That said, Cameca and JEOL don't even have a standard intensity drift correction at all, so just be happy you have any standard intensity drift correction!  Seriously, assuming your various separately acquired, but subsequently combined for quant samples, don't have severe instrument drift issues, it shouldn't be a problem (the beam drift correction *is* on an element by element basis so you are OK there!).

If you are having severe instrumental drift issues affecting the standard intensities, you might, as you suggested above, acquire the samples to be combined later for quant, in groups, to avoid any really nasty standard intensity drift issues.

Can you give me an example of why you are acquiring samples separately for subsequent combining for quant?  Are there crystal flips involved?
john

[Gseward]

Crystal flips in my current issue, but sometimes just and attempt to reduce the time per analysis when different kVs and beam currents are being used.
I don't

Cheers,

Gareth

[John Donovan]

Crystal flips in my current issue, but sometimes just and attempt to reduce the time per analysis when different kVs and beam currents are being used.

Yes, balancing these considerations can be problematic.

If you are having drift issues on the instrument (room temperature related?), and one absolutely had to acquire elements using separate sample setups, I would try to combine as many points as reasonable in each digitized sample, and assign to each digitized sample, the two (or more) sample setups required.  That is using the "Multiple Setups" button in the Automate! window. Which is normally utilized for thin film multi-voltage analysis acquisition, but of course works just as well for this situation.

That way, at least most of the additional overhead from crystal flips or beam conditions changes are limited to every n points (per sample). And, the sample setups (for each subsequently combined analysis), will have been acquired relatively close together in real time. Hopefully close enough to have minimized any potential standard intensity drift.

This post describes how to use the "Combine the Selected Samples into a New Sample" button to just duplicate a single selected sample:

https://smf.probesoftware.com/index.php?topic=573.msg12816#msg12816