Chromatography Forum

Today we had a curious problem in the lab and i guess someone might have seen something similar and may help me out with a quick answer...

We had a standard solution from which we couldn't get a good peak area RSD after several repeat injections.
The system was stabilised for quite some time and a performance qualification was done less than a month ago.
The method stated that the solution should be transferred to a polypropylene vial (for no apparent reason as other test methods for this compound used glass with no problems!) so after we couldn't get a good area RSD from the plastic vials we tried to run it again from glass vials but nothing improved.

Finally just before we left i had a solution of anthracene injected a couple of times (using the column and mobile phase that were already set up), and to my amazement we got an RSD of 0.12% from about 5 runs.

What solution charateristic could possibly cause repeat injections to have variable peak areas.
I looked at the solution and no undissolved particles were observed, it is well mixed and the vial has a septum so it is not solvent evaporating off the vial causing concentration changes...

any wild theory welcome....Thanks!

When this type of problem is encountered, the first step is to check your daily horoscope and the phase of the moon. Seriously, my first question is, has this compound given good results in the past using exactly the same conditions or is this the first test of this compound?

Maybe it's a problem with your compound not being as soluble in the mobile phase as the sample solvent. If your sample concentration is right at the saturation point in your mobile phase you might get a small amount of sample coming out of solution. This may result in a small shoulder on the compound peak if there is a significant amount coming out of solution.

I've also seen weird variability problems with samples that weren't mixed well. The samples that gave us problems were samples that had been frozen and were stratifying in the tubes. However, if you've been transferring from one vial to the next, this is less likely.

It sounds like a very frustrating problem. Good luck!

Karen

Oliver,

Two things come to mind with the problem you describe. The first one is that there could be some incompatibility between mobile phase, sample solvent, and wash solvent (autosampler). This is what I would check first. great differences in polarityu, PH , organic content, etc, could lead to that problem.

The second thought is that it is rare but it happens, that some compounds are sensitive to the injection valve rotor material. Is most cases this is Vespel, but for some compounds Tefzel can be better.

I will keep thinking, maybe there is something I am overlooking.

Good Luck,

josebenjamin

Thanks for the feedback every one!

The idea of diluting solvent vs mobile phase and solubility incompatability seems the most promissing, it didn't come to my mind at the time, i was actually thinking why the standard is dissolved in 30%ACN while the samples were in 0.1N HCL (FYI this is a dissolution), but the analyst told me that in assays there were no problems. As soon as i get back to work I'll check that out, by comparing the diluents and mobile phases of all tests on this material. I don't recall a soulder though.
Before we left yesterday we mixed the solution again in the volumetric, resonicated the standard solution and filled new plastic vials and set some runs overnight. I'll let you know on the results....

as regards previous work, we did work on it previously although rarely and it has a bad reputation, afaik we always have problems with this method.

I'll check the material of the LPV, i know its black and Vespel is a familiar word i might have seen it in the manual, but i hope its not that reason!

hello!!!
would you let us know what wevelength you are working on? and detail about your mobile phase? also do let us know whether your peak area is decreasing or increasing during repeated injection.

I would suggest to look carefully, the Peak shape, tailing and integration. There might be a change in integration due to baseline problems.

JM