PCB question (picture included)

[chemwipe]

Hello people-

This is an overlay of an Aroclor 1016/1260 standard at 100 ppm (in green) and a pesticide waste sample at a 4X dilution (in blue).



Is this good enough to determine whether or not PCBs are present? I didn't overlay the same standard which was run after the sample, but there was no retention time shift.

Also, some of the major 1242/1248/1254 peaks don't seem to be present (but of course, some are possibly masked by the huge peaks that are saturating the detector).

I'm aware of the possibility of weathering when it comes to PCBs - but there was a recent sample I ran which looked EXACTLY like a 1254 standard. The sample above, not so obvious. (I will try to upload it tomorrow if I can).


Any comments, suggestions? I ran the sample through a sulfuric acid wash, as well as a silica cartridge.

CW

[Don_Hilton]

I assume that this is FID? And what qustion are you trying to answer: The presence of PCB's at all or the presence of this specific mixture (unaltered) in the sample.?

With some of the very large peaks, I would be cautious that retention times for PCB's could shift. Where the blue trace goes well above the green trace, you can say nothing about the presence of the specific PCB's.

The PCB mixture that you injected is not detectable a part of the blue trace - at least from what we can see, when the PCB mixture is taken as a whole. But you could have individual compounds present at significant levels.

[chemwipe]

I assume that this is FID?

ECD.

And what qustion are you trying to answer: The presence of PCB's at all or the presence of this specific mixture (unaltered) in the sample.?

The presence of PCBs at all - This is a 1016/1260 mixture, but I went through and checked for the characteristic peaks of Aroclors 1221/1232/1242/1248 and 1254. Like I said, there are some peaks that are present (maybe a bit shifted) and some that may be hiding in the large peaks.

With some of the very large peaks, I would be cautious that retention times for PCB's could shift. Where the blue trace goes well above the green trace, you can say nothing about the presence of the specific PCB's.

Would it help if I showed overlays of the individual Aroclors?

If so, I'll post them tomorrow if I can.

Thanks!

[Don_Hilton]

An overlay will only take us so far. For the peaks that could be PCB's - or not, I'd have to spike the sample to see if the apex matched. And I'd have to try it at a least a couple of levels of spiking to be sure that I can locate them clearly (big peaks) and small - where apexes be partially obscured (or appear to be shifted) by coelutions.

Ultimately all you can do is come up with a statement that PCB's are not detected below some level where you can distuninguish a peak. And you will probably have to do several injections to scope out the limit you can reach in that mix.

I'd love to get a GCxGC column set in front of that ECD - and you could get some good peak resolution. But, one does not find parts for that in the junk drawer...

[chemwipe]

An overlay will only take us so far.

What if I overlaid the sample with higher concentration standards of individual Aroclors? Would that be more reasonable to compare?

For the peaks that could be PCB's - or not, I'd have to spkike the sample to see if the apex matched. And I'd have to try it at a least a couple of levels of spiking to be sure that I can locate them clearly (big peaks) and small - where apexes be partially obscured (or appear to be shifted) by coelutions.

OK, I will try this, the chromatogram becomes much clearer at a 40X dilution...

Ultimately all you can do is come up with a statement that PCB's are not detected below some level where you can distuninguish a peak. And you will probably have to do several injections to scope out the limit you can reach in that mix.

Unfortunately, I'm looking for a reporting limit of <40 ppm, and that seems impossible so far.

I'd love to get a GCxGC column set in front of that ECD - and you could get some good peak resolution. But, one does not find parts for that in the junk drawer...

Not familiar with GCxGC, but I'll look it up.

Thanks, Don!

-John

[Don_Hilton]

For where you are trying to get and matrix - I would take a look at trying to clean up the sample more, if possible. Can you get the sulfuric acid wash aggressive enough to hydrolize anything that can be hydrolized (like heating)? PCB's should be able to withstand heat and acid just fine - that's the proble with them... they don't go away . Silica jel - perhaps take it off in stages less polar to get rid of matrix until the PCB's finally move - or go with carbon. See if you can get pi-pi interactions to hang onto the PCB's for you while everyting else goes by.

If you have time and buget for it, GCxGC could help you out.

GCxGC is comprehensive two dimensional chromatography. It involves a clold trap at the end of your column which captures everything coming off for a few seconds and then releases it into a second column with a different stationary phase.The second column is short and narrow to allow for everthing to separate while the next slice of the first chromatogram elutes. Fancy software make it look like you did your GC in a plane - like two dimensional paper chromatograpy. It's not a build your own kind of thing. But there are companies out there that sell this: LECO, ZOEX and Thermo. (Last I saw, Thermo only had FID detection, but that may have changed.) There are some others that use a valve based system (Agilent and Schimadzu are two) but the valve based system does not provide focusing going into the second column - so you bands start out broadened. With your chromatogram I'd stick with the cryofocused systems. And believe it or not, with this direction you may even be able to skip some of the preparation you already do. If you have access to J Chrom. A, there is a nice paper by Jack Cochran on pesticides in a tobacco extract - shake and shoot. (He does use mass spec - but for that matrix and analyte list, you need all the help you can get.)

[Ron]

There is no way to confirm the presence of PCBs in this sample with an ECD without more sample preparation. Most of what you are seeing in the chromatogram is hydrocarbon.

An ECD is a selective detector, not a specific detector. It will respond to anything that can capture electrons, and that includes hydrocarbons at a high enough concentration. The selectivity of an ECD for halogenated compounds compared to hydrocarbons is typically approximately 1,000,000:1. Since you can see parts per trillion of halogenated compounds on an ECD you can see ppm of hydrocarbons on this detector.

[JLCookael]

Clean up your sample, Sulfuric acid and or permanganate is where I would go first, then silica gel if you need more separation

[EnvChemist]

Your peaks have also topped off making it hard to determine what PCBs are in your sample so you would need a further dilution, because the relative sizes of the peaks in the patterns are important. I would also agree with those who suggest further sample clean up.

[chemwipe]

. I would also agree with those who suggest further sample clean up.

Ok, here is a similar sample:



The blue chromatogram is the sample at 4X after just a sulfuric wash -

The green chromatogram is the sample digested overnight with sulfuric, washed 3 more times with sulfuric, treated with Florisil, run through a silica gel cartidge, and then diluted.

Not sure what else I should do at this point. Any suggestions?

[gpronger]

Review of your first chromatogram would seem that at best you have a possibility of a PCB with a bit of a retention time shift to later RTs for the sample. (The pattern around 22.5 minutes in your original chromatograms caught my eye.) The rest of the chromatogram has enough interferences its tough to make a call.

Have you tried the permanganate cleanup (SW-846, 3665A)? Often this will address interferences that sulfuric acid does not.

Greg

[chemwipe]

Have you tried the permanganate cleanup (SW-846, 3665A)? Often this will address interferences that sulfuric acid does not.

Greg

Using KMnO4 actually makes the hexane layer darker than it originally was, and I also see a precipitate form on the bottom layer.

The extract does not become less colored on subsequent KMnO4 washes. I'm assuming the method calls for the extract to be colorless? It's not happening.

[gpronger]

What your describing is a bit unusual and without being there all I can suggest is double-checking the cleanup method, if you don't have a copy it can be found at; http://www.epa.gov/epawaste/hazard/test ... /3665a.pdf

My experience is that the solvent should end up cleaner than it started; sorry for not having more insight though.

Greg

[chemwipe]

My experience is that the solvent should end up cleaner than it started; sorry for not having more insight though.

Greg

Left side is a sample washed with only H2SO4, right side is a sample washed with H2SO4 and KMnO4 solution. The top layer (on the right tube) is darker after the KMnO4 wash, and a precipitate forms as well.

Has this happened to anyone in here before?

John

[gpronger]

I double-checked with the prep folk here, and it does sometimes happen with really mucky samples. They would say keep at it (3 - 4 washes) and it should knock it all out.

Greg