I've been measuring relative fluorescence quantum yield (hereafter FQY) for a photosensitive pigment. Most analytical articles and standard operating procedures (SOPs) I've read suggest taking the average of multiple FQY measurements (at least 3) before calculating the standard deviation, which seems good; however, wouldn't comparing residual standard deviations (RSDs) of the linear regression of fluorescence intensity versus the absorbance of the dye standard (e.g. rhodamine B) and dye sample (e.g. hypericin) provide thorougher validation of the calibration fits and thus FQY? In other words, since the relative FQY is calculated from the slopes of these plots, can I glean more information by comparing the RSD of the plots and SD of the FQY?
$\begingroup$
$\endgroup$
2
-
2$\begingroup$ The experts at NIST are worth knowing about: nvlpubs.nist.gov/nistpubs/jres/113/1/v113.n01.a03.pdf. $\endgroup$– Ed VCommented Jul 8, 2023 at 1:48
-
$\begingroup$ Even better would calculation of SD estimation of the slope and intercept and the confidence band of the linear regression. It then leads to ability to determine the y confidence interval from the given x value and vice versa. $\endgroup$– PoutnikCommented Jul 20, 2023 at 10:47
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
You do need to account for the absorbance of each sample otherwise the yield will be hugely inaccurate, then take the average etc. Normally you would try to make each solution of the same concentration but you could do what you suggest provided you are sure that no aggregation etc. of the molecule is present.
To get relative yield (other than just a number proportional to a voltage from your detector) you also need to measure the fluorescence of a known standard molecule, and at the same wavelength because detection sensitivity will vary with wavelength.
-
$\begingroup$ Concern for aggregation is a good point, and you give good general advice for FQY. However, the majority of the post doesn't really answer my question. Namely, is RSD a good statistic to include in a paper for validating results? Is calculating the standard deviation of FQY from RSD and error propagation is sound? You seem very intelligent. $\endgroup$– YHoshiCommented Jul 7, 2023 at 15:56
-
1$\begingroup$ Why not simply quote std deviation of slope of relative yield vs concentration, is this what you mean by RSD ? If you look at residuals to the fit then they should be randomly distributed, but you probably don't have enough data points to check this fully, but a visual check should be ok. At this point I think you need to think about what you actually want to achieve, you may not need very much statistical analysis, your result will only be as good as the standard's yield is, I'm not sure rhodamine b is that good, but you should check on this re photostability, aggregates etc. $\endgroup$ Commented Jul 7, 2023 at 16:24