PFAS Contamination - The NeverEnding Story

One of the most challenging aspects of PFAS analysis is the general ubiquity of PFAS compounds, which can make them very common contaminants. Solvents are often filtered with Teflon, or have Teflon lined caps on the bottles. DI water system resin beds often contain fluoropolymers. Plastic containers are prevented from sticking to their molds with fluoropolymer-based mold release compounds. Lab coats are often stain or chemical resistant, which is accomplished through the use of PFAS compounds. The list is seemingly endless. With all these potential sources of contamination, it’s no surprise that many labs struggle with PFAS contamination, especially when performing low level drinking water analyses. The methods themselves don’t make it any easier on labs. EPA methods 537.1 and 533 require that the PFAS background in blanks be less than 1/3 the minimum reporting limit (MRL) for all analytes. My colleagues Gary Oden and Mike Chang recently covered some potential internal standard interferences from high level samples, and we have literature providing advice for PFAS free workflows for method 537.1, but given the ubiquity of PFAS contamination is something labs still struggle with.

I’ve been relatively fortunate in that in my work I’ve only consistently seen background from one PFAS compound, perfluorobutanoic acid (PFBA). This compound has an example lowest concentration minimum reporting limit (LCMRL) in method 533 of 13 ng/mL. A lab that has similar MRLs would be allowed up to 4.3 ng/mL in their blanks. Assuming a 250 mL sample volume and a 1 mL final extract volume, blanks on instrument should be 1.08 µg/mL. and they were well below this level, with lot 1 showing 0.36 µg/mL background and lot 2 showing 0.54 µg/mL. However, I later tested lot 1 again This time it had PFBA at an order of magnitude higher, 4.8 µg/mL, which was concerning. All of my instrument blanks were showing no PFBA, so I started to dig in to other possible sources of contamination to see if the cartridges varied lot to lot, or if something else was up.

The conditioning, loading, and elution steps all use 10 mL of phosphate buffer, ammonium acetate solution, methanol, and a 2% ammonium hydroxide in methanol solution. I took 10 mL of each of them and blew them all down to 1 mL. The phosphate buffer and ammonium acetate solutions tested at 0.2 to 0.5 µg/mL. The methanol I was using tested at 0.08 µg/mL, and several other brands of methanol in the lab were between 0.12 and 0.92 µg/mL. At first, I was glad I was using the best methanol available in the lab, but I decided to retest the first brand to be sure. The retest came at 0.64 µg/mL. Just like the SPE cartridges, I was seeing an order of magnitude difference in the same lot of material when testing on different days. A test of the ammonium hydroxide in methanol came in at 0.61 µg/mL.

Next, I decided to check the containers I was using for extract collection and blowdown. I was using 15 mL polypropylene centrifuge tubes, so I added 10 mL of methanol and put them on a shaker table for an hour. I did the same thing with some 50 mL  polypropylene tubes to have a second comparison. The 15 mL centrifuge tubes showed 0.83 µg/mL PFBA, and the 50 mL tubes showed 0.56 µg/mL PFBA. Both were similar enough to the methanol results of 0.64 µg/mL that I would say they don’t significantly contribute to the PFBA background.

Next, I tested various other things that would come into contact with the sample or solvents during sample or reagent prep. 1 and 10 mL pipette tips were placed into the centrifuge tubes and shook out with methanol, and they showed 0.29 and 0.72 µg/mL PFAS, respectively. I did the same for one of the disposable SPE liners, which is an obvious potential source of PFAS contamination as it is made of Teflon, but it only showed 0.23 µg/mL. Finally, I cut the fingers off a large glove that I was using, as well as a different brand of small gloves that an intern who was assisting me was wearing. The large gloves showed 3.04 µg/mL PFBA while the small gloves showed 0.75 µg/mL. Unfortunately, the box of large gloves was almost empty, and I couldn’t find any to retest after I saw the high results.

There were several other things that were tested and re-tested, and I could go on about that. However, I think this makes the point of how difficult it can be to manage low level PFAS contamination. My blanks have been consistently below 1 µg/mL recently, so Regardless of the source, this shows that consistent monitoring of your blanks and thorough testing of all lab supplies is critical for PFAS testing. The additive nature of contamination is important to consider as well. Just in my lab workflow I have to be concerned about SPE tubes, SPE transfer lines, sample bottles, SPE liners, pipette tips, collection tubes, lab coats, gloves, DI water, methanol, phosphate buffer, ammonium acetate, ammonium hydroxide, autosampler vials, and probably several other things I’m overlooking. There are so many supplies and reagents that go into a WAX SPE method that even contamination well below the MRL can cascade into a problem if it is present in enough supplies.

It's also important to consider if your MRL is realistic. Method 533 states in section 9.1.4 “one should consider that establishing the MRL concentration too low may cause repeated failure of ongoing QC requirements.” Your labs best case MRL may not be practical if you can’t consistently keep your background less than 1/3 of it.