1
$\begingroup$

Can analytical precision be finer than a lower detection limit?

For example, if the lower detection limit of an analytical method is reported as 0.010 ppm. Is it possible to have an analytical precision of 0.001 ppm, and therefore a measurement like 0.116 ppm?

Obviously this depends significantly on the method, but in theory is this possible?

Improve this question
$\endgroup$
8
  • 2
    $\begingroup$ There's no inherent contradiction, I think. Detection limit doesn't necessarily stem from lack of precision. $\endgroup$
    – Mithoron
    Commented Jun 11, 2020 at 1:08
  • 2
    $\begingroup$ @JoshuaWright, The question is how did you determine the detection limit? The problem is multiple methods will yield different detection limits for the same analyte or the same analytical method. There is no true "definition" of detection limit nor there is any ideal way of determining it. $\endgroup$
    – ACR
    Commented Jun 11, 2020 at 2:23
  • 2
    $\begingroup$ Yes this is obviously possible. For many methods of interest, the method will be precisely tailored to a particular measurement range. For example, a technique might have a precision of 0.001 ppm when analyzing solutions that are between 0.114 and 0.118 ppm in concentration, and have rapidly declining precision outside of this range. Isotope ratio mass spectrometry comes to mind as a related example. $\endgroup$
    – Curt F.
    Commented Jun 11, 2020 at 15:44
  • $\begingroup$ @CurtF. Thanks for your comment, I suspected this was the case. Any chance you know where I could find a isotope ratio mass spectrometry data set that shows this? $\endgroup$
    – Joshua Wright
    Commented Jun 11, 2020 at 16:16
  • 1
    $\begingroup$ @CurtF. Wouldn't that allow you to spike a sample to increase the concentration of an analyte and then, by extrapolating from higher concentration, to determine an analyte concentration in the original sample that is below the detection limit? For instance, if I can determine the concentration of original sample + spike as $x+dx$ and that of spike solution as $x$, then by simple subtraction I know $dx$. If what you write is true $dx$ could be below the detection limit $\endgroup$
    – Buck Thorn ♦
    Commented Jun 11, 2020 at 17:45

0

Reset to default