Does Varying ELSD Light Intensity affect sensitivity/resolut

[mikejungle]

Hi,

So we've got a normal phase method which we run with a gradient (mp1: Hexane, mp2: Hexane, IPA, EtOH), which we use to detect free fatty acid content in various oils, and I was wondering if there are any issues with non-proportional decrease in sensitivity with decreased LED intensity.

We've basically bottomed out with PMT Gain set at 1, and an injection volume of 1 uL, so the next step, i assumed, would be the light intensity, which was at 100%.

So yes, I want to know if reducing it (anywhere across the whole range, from 1%-99%), would result in a proportional decrease, or if it would be disproportionate. Mind you, all the FFA's come out in one peak...if that matters.

I don't think it would, based on spectrophotometry formulas and whatnot (I/Io, absorbance and transmittance), but a colleague has said that it would decrease unpredictably.

Thanks a lot.

[tom jupille]

Assuming you were within the working range of the detector (which is inherently non-linear, by the way), then a decrease in lamp intensity should give an approximately proportional decrease in the signal. That said, if you are trying to decrease the signal, that suggests that you're overloading the detector. In that situation, you might have so much dust/aerosol that it's actually blocking the light; in which case reducing the light intensity won't help. And, if you're overloading by that much, the presence of your analyte may affect the dynamics of aerosol formation. If it were my problem, I'd look first at diluting the sample.

[Consumer Products Guy]

we've got a normal phase method which we run with a gradient (mp1: Hexane, mp2: Hexane, IPA, EtOH), which we use to detect free fatty acid content in various oils...

Mike - why do you use HPLC for this instead of a typical AOCS nonaqueous titration in hot neutral alcohol to phenolphthalein endpoint with NaOH titrant? Is it important for you to also know the composition of the FFA as well as the total amount?

[mikejungle]

nice, someone's who familiar with AOCS methods

Consumer guy, we do perform the titration method to determine FFA, but the operator variation is so...variable, that we use it more as an industry standard type measurement and don't accept it as a reliable indicator, because one person may perceive one pink color as the endpoint, and another may perceive another pink as the end point.
And with our current method, the FFA peaks don't resolve, so we don't find out much about composition, just about the more subtle changes that might occur.

Tom, we are overloading the detector...it is on the top end of our calibration curve, and even at 30% LED intensity, it looks as if it's at the limit of detection.

If it were a sample and not the calibration curve, i'd perform a dilution and be done with it, but since it's at the top end of our calibration curve, I'm not quite sure how to fit it in.

I thought of maybe diluting the whole curve by 1/2 and correcting via dilution factor, but is that a viable way to get the range I want or am I over-thinking this? Can you guys outline how I would do this best? I'm not sure what the best way is.

Our samples rarely read as high as our high point, but sometimes they do wander close, and I think it's analytically sound to want my samples to read in the middle of the curve.

Thanks a lot for the input so far.

[tom jupille]

I thought of maybe diluting the whole curve by 1/2 and correcting via dilution factor, but is that a viable way to get the range I want or am I over-thinking this?

That would be my choice.

[HippyLabRat]

Consumer guy, we do perform the titration method to determine FFA, but the operator variation is so...variable, that we use it more as an industry standard type measurement and don't accept it as a reliable indicator, because one person may perceive one pink color as the endpoint, and another may perceive another pink as the end point.

Spectrophotometer with flow cell loop/and a parastalsic pump should cure this...

titration beaker -> photometer flow cell -> pump -> titration beaker

That way you can get a hard number as an endpoint.

All you need to do is make sure your photometer can read in realtime.

[Petrus Hemstrom]

If you do not have issues with detection limmit with the lowest concentration in your calib curve I would reduce the injection volume or use a larger column to dilute the sample. If you are overloading an ELSD at gain 1 you are most likely also overloading your column.