[kiwi8kiri]

If you are using an ICAP, and your calibration data gives u 0.9999 for its r2 value, however the RSDs are exceeding 50% some 10 - 90 % do you need to re-run your data to ensure red's below 1? or ? Any idea why this could happen also?

[Arkcon]

If I understand you, the standard R² is fine, but your sample RSD is 50%?  I suppose it depends on to what standard your analytical results are help by any of a variety of judgement boards but ... yes, no matter what the analytical method, and RSD of 50% is just awful. I know, I've gotten them myself on an HPLC run, and you just have to disappoint people with the grim reality that you don't have an answer for them, and you do have to start over.

Now, retuning to ICP, there can certainly be matrix effects with the analytical samples that are distinct from standards.  Or this could be a sample handling error.  Or a sudden drift in the instrument.  But you really haven't given us enough details to figure this out.

[kiwi8kiri]

Thanks very much.

It is an aqua regia (14%) digest of steel waste samples. Is there a way i can check the instrumental drift? (Unfortunately we don't have an expect are our university at the moment and i am trying to learn as i go along). I have tried multiple wavelength frequencies with each element (using a 28 multi element standard) and i only have 5-6 elements that have 1-2% RSDs.

[Arkcon]

Until an expert in ICP comes along (and that's not me) I'd suggest re-running some standards in between a few repetitions of samples.  You'll want these bracketing standards to remain within some level of the average of the calibrating standard -- say ± 2%.  Then we can be sure they're not drifting.

Aqua regia or nitric are typical solvents for analyzing samples by ICP.  You say these are steel waste samples, but what elements are you looking for?  Say, for example, you want to quantitate chromium, nickel and silicon.  Your standards are what, solutions of those elements in aqua regia?  What about your unknowns, are they mostly solubilized iron?  Could that be interesting with the ability of the ICP to relibily ionize the elements you want to see?  Do you standards have elements within them that aren't wanted, but are in the samples?  Things like that are what I meant by matrix.