[ajax0604]

I am having trouble getting my head around why the hollow cathode lamp is pulsed. I do have an idea but I was hoping someone could confirm for me. This is what I understand so far.
When the hollow cathode lamp is "on", the detector will pick up the unabsorbed light coming through the sample. There will also be emitted light coming from the excited electrons as they return to their original energy states but I have read that since the light is emitted in all directions, the chances of it entering the detector is very low and won't be picked up.
So when the hollow cathode lamp is "off", what is the detector picking up then? Is it detecting any stray light ("noise") coming from the sample atoms that are in the flame? Or is it detecting the light coming from the flame itself? Or is it both? I am assuming that for both cases, the light is considered as "noise" and needs to be measured so that it can be eliminated from the calculations.
Thank you.

[Corribus]

I don't have a whole lot of experience with AAS, but in the case of other spectrometers you will always subtract a background spectrum (or multiple background spectra) that include noise from a variety of sources including both stray light (a number of origins) as well as dark current from the detector itself. By the latter I mean that most detectors work in some way by creating a measurable current in response to collisions with photons. However there is always a baseline current, called the dark current (for obvious reasons), even in the absence of photons. This must be subtracted in order to get an accurate count of photons hitting the detector.

https://en.wikipedia.org/wiki/Dark_current_(chemistry)

Note that measuring the absorption with and without light is important in addition to measuring blank samples, as the former subtracts noise due to the instrument itself, whereas the latter only measures noise due to the sample.

[Enthalpy]

According to:
"Hollow Cathode Lamps – Yesterday, Today and Tomorrow" by Agilent Technologies
discharge lamps are often "boosted", where a first pulse sputters the emitting element and a second pulse excites this element, with the added advantage that much of the element vapour is ionized hence absorbs less the useful light.

Well, as these lamps rely on sputtering, I'd say that a strong peak power is necessary to get enough metal vaporized and have the emission by that element rather than by the noble gas filling, and that at this peak power, continuous operation would destroy the lamp.

With the added benefit of being sensitive more shortly to the detector noise, sure.