PCBs in transformer oil vs. solvent

[MichaelVW]

Hello,

I'm doing my first PT on PCBs in Transformer oil. (8082; GC-ECD)

For years, I've done PCBs in soil and water. In the previous cases, the PT chromatograms looked perfect - like an analytical standard - so identifying the aroclor was easy.

This time, it's not a perfect match for any aroclor. I tried a 1/10 dilution in hexane, and a 1/20. The latter looked a little closer to the pattern that best matches. I had read somewhere online that hydrocarbons in mineral oil can repress the signal of the ECD.

Maybe I should get cal standards in transformer oil for this method going forward? Though I tried spiking an analytical standard with a proportional amount of mineral oil and did not see any effect.

Any suggestions?

[ajc92]

Hello,

I'm doing my first PT on PCBs in Transformer oil. (8082; GC-ECD)

For years, I've done PCBs in soil and water. In the previous cases, the PT chromatograms looked perfect - like an analytical standard - so identifying the aroclor was easy.

This time, it's not a perfect match for any aroclor. I tried a 1/10 dilution in hexane, and a 1/20. The latter looked a little closer to the pattern that best matches. I had read somewhere online that hydrocarbons in mineral oil can repress the signal of the ECD.

Maybe I should get cal standards in transformer oil for this method going forward? Though I tried spiking an analytical standard with a proportional amount of mineral oil and did not see any effect.

Any suggestions?

We used to have the techs dilute oils 1/10, put a few mL of that through sulfuric acid cleanup, then dilute it some more at the bench OR put it through silica cleanup.
Oil samples always always shifted retention time on our ECDs. You could usually confirm how much shift by running a CCV right after it.
Does the PT specify their list of aroclors they pick from to spike it? I know some labs don't include 1268 in their list and some PTs do.

[MichaelVW]

There is a slight RT shift but it's the ratios of the 1st and 2nd largest peaks that is more confusing. The pattern I'm seeing is halfway between what two aroclors typically look like. If this was an environmental sample I'd have thought it was a bit weathered. But the PT is spiked with one (and only one) of the seven standard aroclors.

I think I'll post again with the actual chromatograms after I turn in my results and get the answer back.

[MichaelVW]

Update: I identified the correct aroclor, and my results were close to the assigned value. I'd welcome comments though, because I'd like to understand these potential matrix effects better.

Here are three chromatograms: 1232, the PT sample, and 1242.

https://drive.google.com/file/d/1sBoyEf ... drive_link

1232 and 1242 look very similar on an HP-5 type column. One difference is the height/area ratios of the two peaks marked with purple arrows. 1232 has two big peaks, 1242 has one huge peak. If you just look at peak height, it appears that the PT falls somewhere in between 1232 and 1242; that is, the peaks are closer to the same size in the PT than they should be if the aroclor is 1242, but not close enough for what you'd expect with 1232.

However... when I compared actual peak areas, ratios in the PT were close to what they are in the 1232 standard. Maybe the presence of hydrocarbons caused the peak to be broader and shorter in the PT sample?

Another of 1232's distinguishing characteristics is the presence of an early peak that does not appear in other aroclors. I've noted that with purple arrows. The PT has it, but it seems too large - which made me concerned that what I was seeing may just be a coeluting containment, and I didn't want to reply on it.