RE: [Agilent Atomic Spectroscopy Forum] Can someone explain to me what a 'full-fit' calibration for ICP-AES is? And how this is different from a point to point calibration?

Question about ICP-AES calibration curves

Hello JRossi,

Thank you for sending over the screenshot.

I believe this is a nomenclature adopted by the manufacturer. For instance, in Agilent's ICP OES software we call the regression "linear", "quadratic" or "rational", each one having a specific equation to make the regression. We can also add "weighted fit" to make a weighted least squares regression, where the variance of each point is taken under consideration when the regression is made.

Full-fit in here, as long as I could understand, refers to a linear regression using the weighted least square method, because the report says that the weight is 1/Conc. I suppose it would be the same as point-to-point calibration as you mentioned, aside if this software has this option and uses another regression method.

If possible, knowing from which hardware these results come from would help to get in touch with the manufacturer and understand what kind of regression they call "full-fit". Regarding quality results, I believe this analytical curve seems very good and there will be no problems with quantification.

Hope this helps!

Rodolfo

Hello JRossi,

Thank you for sending over the screenshot.

I believe this is a nomenclature adopted by the manufacturer. For instance, in Agilent's ICP OES software we call the regression "linear", "quadratic" or "rational", each one having a specific equation to make the regression. We can also add "weighted fit" to make a weighted least squares regression, where the variance of each point is taken under consideration when the regression is made.

Full-fit in here, as long as I could understand, refers to a linear regression using the weighted least square method, because the report says that the weight is 1/Conc. I suppose it would be the same as point-to-point calibration as you mentioned, aside if this software has this option and uses another regression method.

If possible, knowing from which hardware these results come from would help to get in touch with the manufacturer and understand what kind of regression they call "full-fit". Regarding quality results, I believe this analytical curve seems very good and there will be no problems with quantification.

Hope this helps!

Rodolfo