Chromatography Forum

I've noticed that in assessing Suitability, wherein a criterion is 5 consecutive injects of a STD, that on injects #4,5 the peak areas are evincing a systematic small decline--in comparision to injects #1,2,3. A persusal of this Chrom Forum disclosed several similar comments/observations. Why this occurs is open to question and has inspired postings of great variety; however, the issue I wish to inquire upon is the following:
I've noticed that if I pick up the ALS vial after an injection and give it brief manual swril/mix that on the subsequent injects the peak areas are more consistent than if I'd have left them sitting on the tray un-disturbed.
Is it kosher to perform this inter-injection mixing seeing that it has improved the replicate peak area determinations?

Does your procedure state that you mix the vials immediately before injecting them?

This is part of the timing of sample preparations I have mentioned before in the context of HS analysis, but if can refer to any analysis.

If you don't have the mixing in your procedure then future analysts may not be able to reproduce your results. If you are aware of a weakness in your validation, fix it immediately. Revise the existing procedure noting the sample preparations must not stand for longer than xxx minutes before analysis, or add a mixing step.

best wishes,

Rodney George

I would be wondering why thereis such an effect.

Peter

After a sample has sat for 30 to 40 minutes prior to analysis (i.e., being injected into the GC) there is the possibility that a mild level of non-homogeniety could result which would be expressed as a noticeable change in peak area response. This could be due to either analyte settling/adhering onto the suspension particles of an extract or other kinetic variables that occur to the analyte in solution. I certainly believe that this is not a new and novel observation on my part; therefore, I questioned whether it was acceptable practice to invert/mix the sample several times then set it back in it's tray slot prior to injection. I have observed that the cited practice improved the Suitability %RSD.
As Chrom 1 suggested, if during the method validation process this practice was observed to improve repeatability, then it ought to be written into the validated method so that the next Analyst is not ''spinning their wheels" and fretting over why there was a continual decline in peak area response upon subsequent injections of the STD from the same ALS vial. I submit that the least likely explanation for this phenomenon is degradation of the analyte and that the most plausible (and parsimonious) explanation is that the analyte has become non-uniformly distributed in the liquid matrix. And, that this latter factor can be very simply corrected for by a brief mixing/inversion prior to injection.

A more elegant solution might be to set some extra pumps on the syringe fill cycle, then the sample will get mixed immediatley before injection with no manual intervention - who is going to shake the samples during overnight runs ?

Peter

I like Peters first answer the best. Something needs a correction here.

Why do you even remotely consider deomposition, do you have an idea how shaking can reverse decomposition?

I am not conjecturing that "decomposition" was a factor at all--this is refuted by the observation that the peak areas increased upon each subsequent injection. Likely was not carryover either because solvent blanks injected at the end of the 5 consecutive injects showed no peaks. Still a mystery?
Will try to add more sample pumps prior to inject--currently it is set at two.

No mystery, but utterly confusing, for instance quoting:
". . . fretting over why there was a continual decline in peak area response upon subsequent injections of the STD from the same ALS vial. I submit that the least likely explanation for this phenomenon is degradation of the analyte and that the most plausible (and parsimonious) explanation is that the analyte has become non-uniformly distributed in the liquid matrix."

Decrease or increase of peaks? Least likely explanation or not a possibility??