QC samples for EPA method 8081 8082 dual column

[h20labs]

I read through the latest versions or EPA methods 8000 and 8081 and get a clear picture of how method blanks and spikes should be run. Assume I'm using dual column confirmation on separate instruments. Am I required to run method blanks, spikes and duplicate pairs on both columns? For instance if my method blank has no hits on the primary column do i still need to run it on my secondary column? Also, if I run a sample, sample spike, and sample spike duplicate on my primary column am I required to confirm each of these on the secondary column?

If confirmation is required how do I report the spike results?

[Steve Reimer]

You don't need to run a confirmation on a non-detect in a blank. The duplicates and the spikes should be run on both columns to demonstrate that the confirmation actually confirms known target hits. That is the purpose of running them.

[h20labs]

Thanks Steve

So when reporting a blank spike (LCS) should I create on report w/ spike recovery from primary and secondary column or do I report the spike for each column individually? I know for reporting samples you evaluate results from each column and report the lower of the two flagging results that are >40% RPD. This approach doesn't seem appropriate for LCS, MS&MSD spikes. Thanks for any help, or even if you could point me to a document or method that talks about dual column QC.

[Yama001]

Our approach has been to follow the same practice as for samples - report the lower of the two columns.

[Steve Reimer]

There are times, such as obvious signal suppression on one column, where it is appropriate to report the higher column results. I would see this when analyzing heavy oils for PCBs.

[Yama001]

Yep, that makes sense, Steve.

[MCAbraha]

For dual column confirm, typically all samples will be ran on both columns to evaluate possible co-eluting interferences and to confirm actual hits. Then you report off the better of the 2 columns in terms of CCV recovery and in the case of 8081, better pesticide breakdown.

In our case, we designate one column as primary and the other as confirm and typically will report from primary unless there is a significant issue with those results. We don't calculate RPD unless "true dual column" and we don't report lower/higher unless "true dual column".

For true dual column, the RPD comes into play and both results are reported and the RPD is calculated between the two. All samples and QC must be ran on and reported from both columns.

Just my 2 cents.