First, I'll describe what I'm doing, so you have more information than you need:)
I measure FAD hydrolysis by HPLC. I mix McIlvaine buffer (phosphate+citrate), FAD, magnesium, DTT and my enzyme. After certain time I stop it with 100% Methanol (mobile phase B) and mix with 15mM ammonium formate (phase A) to make 15% MeOH (starting conditions).
Right now, my HPLC Nexera needs to replace plunger seals on pump A, but the serviceman said, that as long as there are no shifts in retention times, it's OK, so I haven't done it yet.
And nobody else had any problems with the HPLC so far (no changes in retention times or anything else).
The file sample.jpg shows how my usual chromatogram looks like, two main peaks, one is FAD, one is FMN. Sometimes there appeares a third one in between, but it's closer to the FMN.
Yesterday when I've run my samples, there appeared a new peak in between. However, when you compare the retention times:
FAD in standards - 2,46
FAD in sample - 2,27
new peak in sample - 2,5
FAD in OLD sample - 2,2
it rather appears as if FAD was the smaller peak and in front of it was something new, although it has similar retention time as FAD in the old samples.
FMN's retention time is unaffected, in all cases around 3,05.
My first thought was, that the column deteriorated and it's peak splitting, however
1) the column is not so old (but they run dirty samples on it, so whatever)
2) it would happen to all peaks, right?
For that reason I've run it also on longer gradient to see whether it will do the same and at first was scared because it eluted about five minutes earlier than it should. However, I've realized, that I used 150mm column for that gradient previously, while this is 50mm column :)
Anyway, the FAD peak in sample went faster than in standards again, but this time it didn't really split into two peaks, but was rather tailing (file long_grad.jpg).
So, what would you suggest? What are the possibilities and options to solve them?
Mostly, can I trust the results now? The area of the first peak seem to be approx. OK. Or should I sum both of the peaks?
EDIT
so I've checked the results and it seems that it areas should be summed to make the usual amount. So peak splitting afterall?
Hi Tomas, have you run a trial without injecting the sample. Run the same method and inject a blank or inject a solvent alone (of your molecule) and see if you notice the peak. This should rule out injector issues that could be affecting the run. Sometimes despite the peak shift, it could be that the injector has some physical issues.
I do that always. I haven't checked today, but I do not recall anything suspicious yesterday.
Hi Tomäs, what will happens if you run 1. standard 2.old samples 3. new samples and thereafter 1+2 and 1+2+3 ? I hope this test will give you answer. In some case slight differences in the solvent/buffer may led to such shifting of retention time.
I'm afraid I have no old samples (I have compared only the chromatograms. But I will try to run the standard with the sample together. Thank you for the suggestion!
Thanks! At least you can run standard and standard+ sample. I have observed quite similar problem, but finally we have found that someone prepare slightly wrong running buffer. This led to the problem of shifting in retention time.
the biggest discrepancy for me is the difference between sample and standard. I'll try to run it mixed and let you know in the afternoon/tomorrow.
OK, now I'm totaly confused.
So I've run the following samples:
- original - the ones prepared last week
- old - prepared yesterday from the reactions made last week and frozen at -20°C (I always keep them for if something happened...
- new - newly set reactions and freshly prepared etc.
I've run two samples for both and both with and without FAD added.
One of the original samples is the same as last week, while the other has more of the second peak (coresponding to FAD standard, but later than usual). Both old samples have only the later peak, while both new samples have both of them (in similar ratio as the samples last week).
All samples with FAD added have only the second peak. Although it seems wider than usual, it obviously isn't double peak, because it starts when the first peak is already at 3/4 of its height.
Now I'm really confused. How can the addition of standard shift the peak?
When I look at your first chromatogram above (normal), the peak identified as FAD has a noticeable tail and has the appearance of a co-eluting component. In the chromatogram with the "new peak", the first peak now has better shape. I suspect that in this second chromatogram you are finally resolving a component that was hidden under the first peak, and the reason could be due to an unidentified change in elution or sample solvent conditions (you already mentioned discovering a change in buffer).
Also, in your last post you show that FAD is the second peak based on the RT of the standard. I wonder if the first peak has been mis-identified from the beginning, because your FAD co-eluted with a large interference here in the original experiments and you are just now seeing them as separate components.
Correction: sorry, I see that it was Taras that mentioned the buffer problem, not you. Doesn't change my suggestion, you could still have some changes in solvent (or column, but it would be unlikely a column would change to give better resolution with time)
Thank you Robert for the suggestion.
However, I'm not sure, what to do about that. The only option is probably to run it on MS to see whether there are two peaks, right?
However, how would that explain, why the peaks coeluted after addition of the standard?
Moreover, it seems that the standard has worse tailing factor (1.55) than the sample (1.3)
I have prepared fresh mobile phase, that improved it little, but only after I prepared also fresh samples, the second peak disappeared. Or rather it merged in between.
It could be differently charged forms, right? If I'm close to pKa, I'd get two peaks right? Based on this page ( http://www.drugbank.ca/drugs/DB03147 ) it has pKa of 4,99, so that must be the problem.
You may have identified your problem. The pH of your mobile phase should be at least 2 units different from the pKa of your analyte. Try adding 0.1% formic acid to the mobile phase.
Yes, this looks like a pH vs pKa(s) issue to me, a quite frequent phenomenon.
Did you ever think on traces of impurities in your mobile phase which can causes ghost peaks in gradient analysis. This traces could come from solvents, salts or your hardware (pump, gradient mixing unit) etc. Sometimes substances like dibutyl phthalate could cause such peaks.
In such cases we use the Ghost-Guard-LC (online absorber column) to remove such impurities!
Hi everybody,
I tried to play with the pH, but got only more problems as I described here
https://www.researchgate.net/post/What_are_the_possible_causes_of_peak_shift_and_broadening_in_HPLC
so I switched to phosphate buffer pH 7.0 and now are the peaks nice and all, tailing factors are
Tomas, I think you are running into the same pH problem but in a different way. There isn´t very much buffering capacity for phosphate around pH 4. }
Phosphoric acid has three hydrogens that can be lost, with pKas of 2, 7, and 12. That means that phosphate can be an effective buffer around these pHs. Normally, it is used for pH ranges of 1-3 and 6-8. I would use a different buffer, like formate or acetate to make a well-buffered solution at pH 4.
- Badges
- Science method