Internal stds for references materials, but not for unknowns

ramorri

Hi iolite team,

More a curiosity question - regarding internal standards, what happens in DRS when the reference materials have an input internal standard (i.e., Si), but the unknowns accidentally get 'crunched' with the internal standard set to 0 ppm? Is iolite actually running those unknowns thinking they have Si ppm = 0 ? I am currently going back and fixing this, was more just curious as the unknown data produced actually looks pretty good, or better than I would have imagined with the internal standard set to 0 ppm for unknowns (and for a major element like Si).

Thanks!!
Rebecca

Joe

ramorri

It depends on which DRS you're using. If 3DTE and you have the internal value set to nothing or 0 then the values for that selection will remain semi-quantitative. If using one of the other trace elements DRS, an internal value of nothing results in nothing for the ppm output and 0 results in 0.

All the best,

ramorri

Ok great, thanks. Yes I am using 3DTE, and that's what I figured (semi-quantitative results), basically I'm reducing the data for those samples with no internal standard set.

Thanks!!
Rebecca

ramorri

Hi iolite team,

Sorry to bother you again. I have a few more internal standard questions.

1 ) I updated my unknowns to have an internal standard under 3DTE (first screenshot), but for some reason when I crunch the data (and export) some internal standard values are missing, which seems strange (second screenshot). Do you know why that might be?

2) Maybe another flag for me, but if I look under the results (iolite), and click on the 'inputs' for the internal standard for unknown samples, shouldn't all of those input values be = to the internal standard value? Instead I have varying input values for Si that are different from the internal standard I input.

I did not see another forum post related to this, but let me know if this is somewhere in a webinar or previously asked question. Sorry again for all the Q's!

Rebecca

ramorri

Joe

ramorri

Can you please send or share your io4 file with support@iolite-software.com? It is hard to tell from the screenshots what is happening.

Many thanks,

Joe

ramorri

Thank you for sending your session!

The issue here is that after all of the arithmetic, each time slice of the Si29_ppm channel for those selections will be scaled to the specified IS value, but due to way computer arithmetic works, you can have small fluctuations at the limit of precision for the number representation being used. We are doing 64 bit floating point math so we end up with small fluctuations around 2e-16. These tiny fluctuations are causing problems for the outlier rejection, and therefore you end up with no value for those channels/selections.

I've just made the outlier rejection a bit smarter so that it handles this case properly. In future releases, it'll report the value as expected. For now, you can be sure it is the expected value by looking in the time series view (the channel will be at the IS value but the result for the selection will be invalid), and you can fill in the IS value in the exported table.

All the best,