%Recovery of analytes spiked into sample matrix

[Meerkat]

1.how is it possible to "recover" more analyte than you spiked into a verified known blank sample matrix?
are we not violating the 1st law of thermodynamics?
measured recovery = 120% !, was expecting less than 100%. any ideas?

[Consumer Products Guy]

The official answer is "no", but there are other possibilities.

There are experiemntal errors (but 120% is real high).

If you are using internal standard, loss of internal standard causes analyte to be high in comparison.

If you lose solvent preferentially in external standard assay, analyte will concentrate.

If your reference standard in external standard deteriorates, then in comparison results will appear high.

And there are more.....

[Meerkat]

thanks for replying Mr. CSG;however, there is an internal std used in this assay--and it turns out that it's peak area count is less in the test sample, than in the reference std., but shouldn't both the query analyte AND the IS retain equivalently?

[Steve Reimer]

In the analysis of organophosphate pesticides (or chemical warefare agents) it is not unusual to have a substantial loss of analyte in the inlet of a GC. The addition of matrix components will lessen this loss so it is not unheard of to have 200 to 500% recovery as compared to clean standards.
A strong selling point for matrix matched standards.

[Peter Apps]

Plainly there cannot be a greater quantity of a substance in a sample than the quantity that is actually there, but what quantity is there and what result you get when you measure that quantity are linked only by an imperfect analysis, whose (in)accuracy in the instance that you cite yields a measurement result that is 20% higher than the true value.

The list of possible causes is as long as you want to make it. A common problem in chromatography with unselective detectors is co-elution between the larget analyte and an interfering component of the matrix.

Peter

[KM-USA]

thanks for replying Mr. CSG;however, there is an internal std used in this assay--and it turns out that it's peak area count is less in the test sample, than in the reference std., but shouldn't both the query analyte AND the IS retain equivalently?

Typically, solution volumes are engineered so that the internal standard peak in both the reference and in the sample will be the same size. And almost all use the same concentrated internal standard solution to make the Rf mixture and to add volumetrically to the samples, so its actual concentration cancels out. If your IS area differs and it shouldn't, that's an issue.

Retention should not be a factor here, but if you really meant that IS doesn't RETAIN equivalently (retention times), let us know.

[Meerkat]

Typically, solution volumes are engineered so that the internal standard peak in both the reference and in the sample will be the same size. And almost all use the same concentrated internal standard solution to make the Rf mixture and to add volumetrically to the samples, so its actual concentration cancels out. If your IS area differs and it shouldn't, that's an issue.

Retention should not be a factor here, but if you really meant that IS doesn't RETAIN equivalently (retention times), let us know.[/quote]

Yes, the key issue is what you cited: the IS area differs between sample and reference std solutions--despite the fact that into both extraction tubes 20 uL of IS was added in. I am delving into how to fix this.