Two components one %CV problem

[StephenO]

Hi

I am experiencing problems on Waters 2695 with 2489 Detector, I am intermittently seeing sample sets which are failing the system suitability or bracketing standard criteria because the %CV is to high.

We first noticed this when analysing single components, When carrying out six injections from the same vial maybe one of the injection areas would of been significantly higher than the rest of the injections causing the high %CV.

We can now exclude the injector as we are noticing the same thing in multi-component methods but only one of the components is effected and the other is fine.

Anybody had any experience of this problem before

Thanks

[tom jupille]

That can easily happen if you have intermittent flow problems (e.g., air bubbles in the check valves or the head). If the flow drops off temporarily, then residence time in the flow cell increases.

Can you go back and reprocess the statistics using peak height instead of peak areas (assuming your data system measures height directly rather than deducing it from area and width). That should be flow-independent and if it looks better than the area results, pretty much eliminates things like lamp fluctuations. The possible culprits are intermittent flow (as suggested above) or integrator settings.

If you're seeing the same kind of variation with height and area, that suggests either a lamp problem (arc wander, maybe?) or working at a wavelength far enough away from an absorbance maximum that the spectrum is changing steeply.

[carlos.teixeira]

Tom's comments are very fantastic, but are you sure there are not carryover on your injector system? All the best for your Team!

[Kwet80]

Try to record the system pressure. Add it as a channel in the instrument method parameters. If any flow drop occurs, you will see a pressure drop at the same time.

Another point to check: In isocratic elution, late eluting peaks can go out of the column in a further injection and wrongly arise an analyte peak area.