Spectra saturation or detector saturation

alpom's picture

Often, acquiring spectra (UV-Vis or NIR) in the transmission mode we come across regions with very high absorbance of analyte. For example many water solutions have high absorbance round 1400nm. In this case instead the absorbance peaks we see flat areas on the top of the absorbance peaks.
Often in literature, books and research papers, such a case is called ‘detector saturation’. I wonder whether this expression is correct.
As I understand transmittance signal that comes to the detector is very low in the regions of the high absorbance of analyte. Thus, there are no reasons for detector saturation and the results mainly depend on detector sensitivity.
As to terminology, this can be called spectra saturation. Am I right?
Oxana Rodionova,
ICP RAS, Moscow

gabiruth's picture

Dear Oxana,

You are absolutely right - this isn't detector saturation - it is more of a detector "starvation" - it means that the transmission is reaching a very low value, and as a result the actual signal generated by the detector becomes practically constant. if you go to the 1930 nm peak of water it will be even worse.

Detector saturation isn't something that you will encounter too frequently - in the real word of instruments the flux of photons reaching the detector isn't enough to bring it to the "saturation". There is though a different type of saturation that appears as detector saturation - this results from the way we convert analog signal (all detectors generate an analog signal) into digital signal. Depending on the conversion - when a certain analog signal is such that the conversion reaches the maximum "digital" value possible - you will see that the digital signal flattens out at a fixed value. The analog signal can increase, but there is no digital way of presenting it - so the digital presentation of the analog becomes flat - or saturated.

The maximum digital representation of analog signal is related to what they call bits of the converter - but this is already into electronics and it is out of my expertise. Electronic engineers can explain that better.

Gabi Levin


alpom's picture

Dear Gabi,
Thank you very mutch for this comment and two additional question in connection
with this topic.
1.  Am I right that in case of, as you called, detector "starvation",
when the signal on the detector is very low, its digital presentation
is some constant value defined by the instrument engineers?

2. A strange think. When I use FT-NIR instrument this peak flattering
is clearly seen in acquired spectra. But using spectrophotometer, very
often it is difficult to distinguish between peak saturation and some
regular measurement. Spectral peak may look a little bit wider in
comparison with ordinary one. And this detector "starvation" may be
caught only by comparison of spectra with different concentration.
So, the detector response to "starvation" is not some constant value,
but some function. And this is very strange, isn't it?
Regards, Oxana