Chromatography Forum

On our system we analyse for riboflavin using a flourescence detector and the problem is that after changing the sample handling the peaks on the flourescence detector are 2 to 3 times too big.

Before we froze the sample in eppendorf tubes and stored the samples a couple of days before transfering the sample to a vail and running the analysis. And here we get the expected sized peaks.

Now we are freezing the sample direcly in the vails instead of the eppendorf tubes and now when we run the samples the peaks are much larger then before. And the weird thing is that if we run the samples again from the same vails the peaks have the expected size.

We have tried vortex mixing the vails before running them but this didn't help. The first sequence still had peaks 2 -3 times larger then expected... and the second sequence with the same vails had fine peaks.

Can anyone explain why our peaks are larger in the first sequence when the sample has been frozen in the vails?

And the weird thing is that if we run the samples again from the same vails the peaks have the expected size.

What did you do with the samples between the two sequences?

Regards Bert

Riboflavin is very poorly soluble in water, and worse in everything else. I think you are precipitating the riboflavin onto the walls of the eppendorf vial and not redissolving it thoroughly. If this is the case, you should expect large r.s.d. for replicate preparations of the same sample.

A second possibility is that the sample vials don't seal well in the freezer and you are losing solvent. Again you should expect a large r.s.d. for replicate samples.

The only explanation I can think of is that, with the new procedure, your samples are still cold when you do the analysis. I believe this could give you a larger signal with flourescence: do to less non-radiative decay (or something like that).

This would also explain why "normal" results are obtained upon re-analysis at a later time. Because, by then, the samples have had time to warm up.

If the standards and the samples are at the same temperature you should, theoretically, get accurate results.

Adam.

adam, the samples keep their temp through the column??
Another possibility: How stable is riboflavin? Maybe you decomposed some by the methods which give you low results? A internal standard, this time with considerably different chemistry (solubility, stability) than the riboflavins, might have helped to identify the problem.

A second possibility is that the sample vials don't seal well in the freezer and you are losing solvent. Again you should expect a large r.s.d. for replicate samples.

This was also one of my first thoughts.. but with the peaks 2 to 3 times larger I would expect a visible amount of solvent gone from the vails... and it dosen't fit together with the fact that on the second run on the same vails the peaks are back down to "normal" levels.

Bert: Nothing is done with the samples.. I just press start and run the sequence again.

HW Mueller: Can you suggest an ISTD? My solvent is a 20 mM acetate buffer.
Also i don't think riboflavin decomposition is the problem since it is the low peaks that are the correct (theoretically) results. But an ISTD could at least answer som questions about the injektion amount and such like.

Thank you all for your responses... Working with the problem alone was/is driving me mad!

This looks like a real puzzle I suppose that the vials in the freezer are capped? If so, a difference between the first and second injection could be the pressure in the vial. Can this disturb your injection?? I once weighted vials before and after injection, to evaluate the injection volume.

Supposing you are using glass vials, did you ever try to store your samples in polypropylene vials?

regards Bert

The vails are capped...

I have also thought about internal pressure in the vail but I didn't consider it relevant. I didn't think it would be able to influence the results so much... but maybe... Again an ISTD would help answer this question.. I must try this!

I am using glass vails.. What difference could there be in a PP vail? Do you think the built up pressure would be less? Or is there another reason for you to suggest PP vails?

Kind regards
Leadazide

In cases like yours, after having found a solution, it is much easier to explain what went wrong .


I would start to look what is different now, and that is ao the storage conditions. It may be helpful to know, if this is influencing your results (without a priori knowledge of a possible cause......) Also, in the original situation, you handled your samples prior to injection (mixing, pipetting etc.) and now they come directly from the freezer into the sample tray (after defrosting of course). That is why I thought of the pressure in the vial.

regards Bert

Any stable compound with the wanted rt around the lab may do for this.
That your low values are the correct ones is highly unlikely, you really would have some strange going ons if it were so, can imagine only that inadvertendly you inject more at the situatons where you get the high peaks. If one sticks with the more likely proposition that the larger peaks are ~ the correct ones (note that everybody assumed this) then there is also the possibility that your actions selectively precipitate the riboflavins.

It makes no sense that the internal standard is the problem. She said the peaks were bigger - not just that the results were higher.

Internal standard degradation (or loss by any mechanism) would give higher results, but not larger peaks.

It seems to me -as Adam suggested- that the problem may be caused by the differences in the temperature between the standard and the samples. If the samples were cold they will be more dense, and, therefore, more amount of the analytes will be injected to the column and the detector.

What do you think about this?

When did leadazide use an INST? The suggestion was that he may use one to locate his problem.
A factor 2 - 3 in density??

That looks complicated. Some peaks 2-3 times bigger than injections from the same vial just a few hours later.
Run a sequence with subsequent injections from the same vial, that could give some information. Injections with different injection volumes would be interesting, just in case the systems "eats" the sample.
Look at the solvent. With some samples we have obtained large rsds when using fresh solvent. That was due to oxygen dissolved in the solvent (yes, we used online vacuum degassing). After leaving the solvent half a day in the bottle results in the fluorescence became stable.
I would also try different vials. I have obseved differences in recoveries from glass vials from different vendors, especially at very low concentrations. What are your concentrations?

would like to know if only one peak increases or some other peaks as well?? like solvent or matrix peak?? the ans to this question will determine , whether you have injection volumn problem or not?

JM