Linearity range is usually specified for UV detectors by their manufacturers, e.g. 0-1.5 AU. I wonder how it is determined. What affects detector linearity range for UV and fluorescence detector? Are they analyte-and system-specific?

Many thanks in advance.

Although we think of them as measuring absorbance, UV detectors actually measure the ratio of the amount of light *transmitted* through the sample and reference cells (absorbance it the negative log of that ratio). At low absorbance, the detector is looking at a small difference in two large amounts of light, so the noise characteristics of the photodetectors and electronics predominate. At high absorbance, the limitation is typically stray light. Since absorbance is a logarithmic function, 1% stray light means that you will never see an absorbance greater than 2.0 (and non-linearity will set in well below that).

Fluorometers are more sensitive at least in part because they only have to measure the absolute amount of light emitted (no dealing with ratios and logs). At the low end, they are probably limited by photodetector "shot noise" (essentially thermally generated rather than photo-generated electrons). I suspect stray light may come into play here as well as in UV. I'm not sure whether the limitation at the high end comes from the detector itself or from sample chemistry (quenching of fluorescence if concentration gets too high).

-- Tom Jupille / LC Resources Inc.

Many thanks to Tom!
I still wonder if the linearity range of a UV detector is related to the wavelength at which the detector is working. Let's say, the detector is linear for compound A in the concentration range of 1 - 500 ng/mL at 260 nm. Would it be linear also at 205 nm for the above concentration range?
IS the linearity range related to molecular extinction coefficient?

Yes, linearity is related to the extinction coefficient, and no, for compound A which is linear in the concentration range 1-500 ng/ml at 260 nm the linearity of response range won't necessarily be the same at 205 nm.
Beer's law is
A=ecl
In a UV detector, l (length of the light path) is fixed (normally 10 mm). Since e (molar extinction coefficient) varies with the wavelength, for a given concentration c the absorptivity A varies also with the wavelength.
In LC-UV, normally one works at the wavelength of the UV maximum, where e is the largest