programming fluorescence detector (WATERS)

[purnachandar]

Hi,
I am a novice in HPLC method development. I am using WATERS fluorescence detector (2475 series). I have encountered a problem while working with my assay. It's been difficult to integrate peaks of my analyte as there is toomuch fluorescence from unavoidable sample contaminant.

Is it possible to set up gain events so that i can have higher gain for my anlyte while suppressing the contaminant. To explain in detail: If my sample run time is 10 min. The highly fluorescing stuff come out at 2-4 mins and sample peak at 7 min. So is it possible to tell the detector to have a lower gain for 5 min then higher gain later on. That way it simplifies integration process. I am just curious about this.

It would be great if some one could help me.

Thanks

[Uwe Neue]

It appears as if you have set up the peak recognition criteria too high. If this is the problem, simply lower the value to get good integration of your second peak.

I doubt that there is peak overlap, but if this is the case, we need to think of another solution.

[Rob Burgess]

This is a common problem that instrument manufacturers have yet to cotton onto yet. We had a situation recently where it would have been advantageous for us if we were able to time program a change in UV attenuation at some point during the run. I did explain to Agilent what we required and they said something similar has been requested by users before.

Hopefully the more us bench scientists can hassle the HPLC manaufacturers, the more they will take heed of what we actually require from our HPLC instrumentation!

[Tony]

We use a Thermo FL3000 fluorescence detector to overcome the same problem; at 4 min into the HPLC run, highly fluorescent contaminants elute that overlaod the PMT tube and shut down the detector - the peaks we want to detect elute after 10 min. To solve this problem, we programmed a change in the emission wavelength 10 min into the run. From 0-10 min, emission was set at a long wavelength (600 nm), where there was little fluorescence from the contaminant. At 10 min, the FL3000 was programmed to switch to a lower emission wavelength (used for peak quantification) and rezero the baseline at the same time. I don't know if such programming is possible with the Qaters detector.

Tony

[HW Mueller]

Maybe I am misunderstanding something here, but generally:
If peaks do not overlap than there can not be any problem no matter how big the unwanted (I usually call it dirt) peak is.
If there is interference due to overlap, than increasing a low gain at the start of the analyte peak increases the tail of the dirt peak to exactly the same level as if the high gain was used throughout, no improvement of anything.
One possible way to get around an interference due to overlapping peaks might be to choose a different exciting wavelength, so that the dirt peak does not absorb, the analyte still absorbs, but probably not at its max wavelength. Or, choose a emission wavelength where only the analyte emits. (Sometimes a compromise among these will do.)

[unmgvar]

purnachandar

If you FL detector has a PhotoMultiplierTube (PMT) like the latest Hitachi model then you can play at attenuating or strenghtening your emission gain over the run.
But on the other hand if your dirty peak do elute before your peak of interest and with baseline separation, why is it so important to you?
can't you just decide to "turn off" the integration for that segment of time? there is no problem even, specifying that in a method which is validated correctly.

[HW Mueller]

Sorry, I must repeat, the gain will not get you anywhere, if you reduce the gain by, let´s say a factor of 10, then you reduce both dirt tail and analyte 10x, where is the improvement? (This is the same as expansion or contraction which can be done with any integration software.)

[unmgvar]

HW Mueller,
I will defenetly agree with you, but in that specific detector you can adjust the gain throu the run with the help of the PMT device. you can decide to reduce the gain over one stretch of the run and also to emphisize it on another stretch. this is currently, hitachie's answer to what Rob Burgess said in is post.
anyway i agree with you totally that if the peaks do not overlap then to simply play with the integretion parameters.

[HW Mueller]

That was understood, but it doesn´t do any good for peaks which overlap such that one can´t intgrate.