Flattening at apex of peak

[Larry]

Hello everyone,

I am trying to diagnos a problem I am having. When injecting an unknown sample I am seeing flattening at the apex of the peak. Is this due to column overload or is there a problem with the UV detector.

Many thanks,

Larry

[khw]

I think it is detector overload. You are very likely in the no-longer proportional absorption range. I assume your absorption in the signal to be greater than about 2.0 Absorption units. This means 99% of the incoming light is blocked by the compound in your sample, only 1% gets to the detector. The detector has difficulties to discriminate between let's say 0.9% and 1.1%. Stray light effects are another reason why you never will see absorptions greater than 3.0 A.U.
Depending on what's the goal of your investigation you can either inject a lower volume/a more dilute sample or go to a different wavelength for detection.

[DR]

Prepare & inject 1/10 dilutions until you get a decent peak shape. If you aren't sure what wavelength to use, run a UV scan on either the last dilution or one up from there and make sure that the wavelength you're using corresponds to (prefereably) a local maximum. If that is not possible, at least pick a wavelength where the slope (in the scan) isn't too severe.

[khw]

DR,

I understand the reason why to choose a maximum. You want to make sure that small changes in the wavelength do not have a big impact. My question is, whether this is still an issue with modern detectors. I have done several injections of caffeine and used maximum, minimum and the slope as detection wavelengths without big differences in variation of peak area. I have to admit, however, that I was using a DAD. Would a VWD be different?

Thank you very much for enlightening me!

[DR]

VWDs are still more sensitive than DADs but not by nearly as wide a margin as they used to be. The abs/wave length issue is pretty "analyte specific" (with some compounds, you can overload an A/D by having the detector trying to put out too much voltage, with others, you get flat topped peaks welll below the maximum analog output of the detector). Given that we're talking about having too much stuff in the flow cell, I doubt that there would be any significant difference between VWD & DAD.

[HW Mueller]

Also khw,
you were obviously at a concentration which allowed very robust detection. Now dilute your caffeine so that you can barely see a peak at the max and then run this at a min..... see what I mean?

[Peter Apps]

Life as an analyst is never simple.

The perfect detection wavelength is the maximum of the spectrum because that gives maximum detectability.

BUT if the maximum is at the top of a very sharp peak in absorbance a small change detector wavelength will give a large change in signal.

For robustness to wavelength change you need a part of the spectrum that is fairly flat.

The ideal compromise is the top of a gentle hump. The worst of all possible worlds is at the bottom of a narrow valley.

Peter