Hi there,

I have ran out of possible causes and solutions to the following problem, any help would be most appreciated.

We currently analyse for the US EPA 16 PAHs using HPLC with diode array and fluorescence detection.

Normally the analysis runs well, however recently we have been experiencing problems with only 1 of the 16 compounds, benzo (a) pyrene. As the run progresses the response for the QC standards (both area and height) increases on both channels. The repsonse for all other compounds remains constant. In some cases the response has increased by 150% by the end of the run. Peak shape, and resolution are fine throughout.

My first thought was carryover, however blanks are clear from any interference on both channels at the specific retention time. Also there were samples in the run for which BAP was absent.

An increase in lamp intesity for both detectors has been ruled out as well as all other compounds are unaffected.

If it was injector related then i assumed that all compounds would be effected equally which is not the case.

If it was column related then i would expect poor chromatography, however again this is not the case.

I have also thought it may be standard related, either the solutions themselves or the way they are prepared for analysis. I have ruled out the solutions as the first QC in the run is always within expected limits. I have also excluded the preparation. The calibration and QC stds are prepared fresh for each run using a Hamilton diluter, any volume errros would effect all componenets not just one.

As i said above any help on this would be much appreciated.

Kind regards

David

If only one compound is effected, I think that problem is column-related.
Does the problem disappear when you use another column?
The only explanation I can think of is that BAP was partially absorbed by the column and right now the column is saturated with BAP.
This explanation indicates that when reanalysing older samples you have to find the same result for BAP, regardless of changes in response.

When PAHs can be flushed from the column with THF purge the column overnight with a mixture of THF/ACN (1:1 or 2:1 - ACN for lowering pressure).
Next day repeat the last sequence and the response should be as normal as before.

Tricky problem.

Although it still sometimes is seen that a certain number of injections are required to saturate a column, in order to get stable peak response, it seems to be a less likely explanation since only one compound is affected.

There are 16 PAH's injected and they bear quite similar properties and would all contibute to a "deactivation" of non specific sites on the column support, not only B(a)P.

What about the retention time does that change as well? What type of vials are used for the samples that I suppose are in a autoinjector? Are the samples exposed to ambient light?

Andreas/ Einar,

Thanks for your replys.

Andreas, i will certainly try another column however if it were the fact that the column was saturated with BAP then why is the BAP not leaching in the blanks or the samples? The QC stds are bracketing approx 4 samples, some of which show no BAP. Thanks for the flush solvents.

Einar, the RTs remain unchanged. We use amber, crimp top vials. The system is a Waters 2695, the sample compartment is not temperature controlled however the lab is.

Anymore suggestions welcomed.

Cheers

David

I have occasionally seen strange and selective adsorption of compounds on injector seals. This should show up in blank runs, but it requires that the blank injection uses the same solvent as the standard injection. Did you do this?

Hi Uwe,

The samples are in 100% ACN, as are the blanks.

Hi David,

saturation don't have to give a peak on a blank run. With basic compounds this happens, caused by reversible interaction with acidic silanol groups, but with unpolar compounds I don't know the mechanism behind this phenomen.
The rare occasions I have seen this were independend of the hplc system.
How old is your column?
I have seen this happening with columns which were used for some time and with the same application.