Missing peak !!!!! BIG PROBLEM, URGENT

[fantasmagorie]

Hey everyone,

Could you please help us with our problem:

1)Method for impurities, 6 known impurities + active substance
2)Column inertsil ODS 3 250x4.6x5um,
3)Mobile phase: phosphate buffer (pH6. with ACN and MeOH, gradient program finished with strong addition of ACN
4)The same equipmet: Agilent 1100 with DAD

The problem is with the mixture of all impurities with substance.
At the stage of method developing all substances had a good resolution (the lowest resolution was 2.00). After about 6 months of routine analysis 4 of our 6 columns doesn't show one of the impurities in the mixture solution. Additionaly, the peak areas of other substances don't grow and have good purity factors, which means that peaks do not co-elute. This impurity analyzed alone on the same column at the same conditionsis detected at expected retention time.

we will be thankful for any ideas

[zokitano]

Hi,

Do you sure that you put enough "non-eluting" impurity in your standard solution? The concentration of the "non-eluting" impurity in standard solution may be critical in this case.

If the final solution's concentration for "non-eluting" impurity is low, below the detection limit, you may not see it in the chromatogram of the standard solution. But when you inject the "non-eluting" impurity's stock solution, you can obtain its peak at the expected retention time. So consider about this. Just a thought.

Best regards

[fantasmagorie]

The concentration of all impurities in mixture solution is the same (0.01 mg/mL) and it is much bigger than quantitation limit equal to about 0.0005 mg/mL.

Best regards

[zokitano]

Sorry I overlooked the fact:

After about 6 months of routine analysis 4 of our 6 columns doesn't show one of the impurities in the mixture solution.

Did you use 6 different columns with the same stationary phase, length, ID, particle size...
Or you use different columns (different manufacturers), lengths, ID and so on?
You said that 4 columns don't show the impurity, but other 2 do. Are there any difference between columns ("bad" and "good") regarding age, number of applications, manufacturer, storage etc

[zokitano]

I suppose that you have problems with the columns that didn't show the impurity. Maybe they are wasted already.

[fantasmagorie]

All the columns used in the analysis come form the same manufacturer, all the parameters are the same. The only difference is the serial number of each column and age of columns, but this parameter does not have influence on our problem - one of our good column is quite old (about 50 analysis) and the other one is new.

[JM]

Have you used the same impurity mix stock solution on all 6 columns?.
if no , please consider the stability of this impurity in solution with time and storage.


JM

[tom jupille]

Let me make sure I'm reading your post correctly. When you inject the "missing" impurity alone, you see a peak. When you inject the same impurity, at the same concentration, on the same column, mixed with other impurities and major components, that peak is missing.

If that is correct, then the missing impurity is either degrading to something else that you are not seeing, or is binding to the active, or is adsorbing somewhere.

If this only occurs on some columns, but not on others (run on the same instrument), then I would suspect adsorption on the stationary phase (silanols?).

If you have been running the method for a while, can you go back and look at the history, with particular attention to things like tailing on the now-missing peak?

[Bruce Hamilton]

I suspect that one of your components has changed and is binding the compound - because the std mix solution is different, but the single sample ( at the same concentration ) isn't. If the columns are dying, you may also get inconsistent behaviour, but if only 50 samples, that's surprsing.

I'd be looking at the buffer concentration and solubility, or possible precipitation of a component that affects the peak.

Are your standards and samples dissolved in initial mobile phase?
Have the column pressures changed?,
Do you have guard columns - what happens if you change them?.
Assuming you get the same area response in the mix and single std, if you double/triple/etc the concentration of the missing impurity, does it increase proportionally on the columns you see it, and appear on the others?
Have you tried flushing the columns with highly aqueous and acetonitrile mobile phases, and check column resolution and performance with the test mix on the original CoA?
Have you changed any of the components of your buffers or mobile phases, and is the peak aree response for all peaks similar to what it was initially?.
If you spike samples with the mix, can you still see the missing peaks on the columns where it's visible?.

I assume your 1100 lamp is OK, and you're not using a low energy part of the spectrum, and std solutions are fresh..

Only you can identify what has canged, we can only suggest possible places to look.

Please keep having fun,

Edit - Tom just posted while I was writing, similar thoughts about columns.

Bruce Hamilton