We are pleased to have with us Elizabeth Denley, PFAS initiative leader and chemistry director at TRC. Elizabeth Denley leads the TRC Center of Research and Expertise PFAS team, which is a group of scientists devoted to staying informed on current PFAS issues, science, and regulations. She currently works on many different types of PFAS investigations, with a specific focus on chemistry, sampling procedures, data interpretation, forensics, QAQC, and analytical methodologies.

She has been a leader on the ITRC PFAS team, serving as co-leader on the history and use and nomenclature section, and as a trainer on PFAS chemistry, sampling, and analysis. We’re also pleased to have with us today Maureen Dooley, Vice President, Industrial Sector at Regenesis. Maureen Dooley has over 25 years’ experience in many aspects of the remediation industry, including project management, research and development, senior technical oversight, remedial design and laboratory management.

In her current role at Regenesis, she provides technical leadership for complex soil and groundwater remediation projects, including PFAS groundwater contamination treatment throughout North America, as well as remediation design, strategy, and business development. All right, that concludes our introduction. So now I will hand things over to Elizabeth Denley to get us started.

Elizabeth Denly:

Great. OK, so today we are going to take a deep dive into some PFAS analytical data. So most of us receive PFAS data from our analytical laboratories in data packages, which are typically described as a level two data package, meaning that the sample results and the quality control results are provided on summary forms without any instrument raw data.

So again, the level two packages do not contain raw data, meaning that many of us are using data that we evaluate from these level two data packages for decision-making purposes without doing any kind of review of the lab raw data. And that’s typically referred to as a Level 4 data package when we receive the raw data. Now, that can be acceptable, of course, but TRC, we’ve discovered potential data anomalies when we dig deeper into these Level 4 data packages.

And these anomalies have actually been fairly random, not necessarily laboratory-specific, And the anomalies can impact the final reported detections or the final reported concentrations of PFAS in our samples that we’re all using for decision making on our sites. So today I’m going to show you some specific examples of these anomalies. Today’s presentation is going to dive a little deeper than normal into the PFAS analytical method.

So I’m going to try and speak slowly and clearly so that everybody can understand. So I’m going to start with two different examples where false positive results were reported for PFOS, and each of these examples occurred at a different laboratory. So before I get into the specific examples and show you some data, I do want to explain the concept of confirmation ions in the PFAS analysis.

So most of you know that the instrumentation that we use for this analysis uses a liquid chromatograph and dual mass spectrometer. The actual time that these peaks elute from that liquid chromatograph is one line of evidence used as confirmation of identification of a specific PFAS chemical. So each PFAS chemical has its own unique retention time on that liquid chromatograph.

The mass spectrometers then break down the parent mass of each PFAS chemical into a unique set of these primary and conformation ions. And the way the methods are currently written, the labs are required to use this primary ion and the conformation ion to identify most PFAS of interest, and the primary ion is typically used to generate the concentration of the specific PFAS, and the conformation ion usually just needs to be present as further confirmation that that PFAS hit is real. So if there’s a detection for a specific PFAS, the raw data is going to give you the area count of each of those ions, the primary ion and the confirmation ion, and then the ratio of that primary ion area to the confirmation ion area has to be within a certain limit.

If it’s not, then the identification of that specific PFAS becomes suspect. So just kind of looking at the table here as an example, you can see the first row, have PFBS. The retention time from the liquid chromatograph here is 4.79 minutes and that’s specific to PFBS.

The parent mass of PFBS is 299 and the mass spectrometer is going to break this down into primary ion 80, conformation ion 99. So this again is specific to PFBS and the lab uses the area count of the primary ion 80 to figure out the concentration of PFBS and the conformation ion-99 has to be present for that positive identification. And then the ratio of the primary ion-80s area count to the confirmation ion-99 area count has to fall within a certain range to confirm that detection.

And then let’s just look at PFOS to kind of drive the point home. Again, the retention time for PFOS from that liquid chromatograph is 7.59 minutes.

Again, that’s unique to PFOS. The parent mass of PFOS is 499, and the mass spectrometer breaks this down into primary ion 80 and conformation ion 99, again, specific to PFOS. And the lab will use the area count of the primary ion 80 to figure out the concentration of PFOS.

This conformation ion 99 has to be present for a positive identification, and then the ratio of the primary ion 80’s area count to the conformation ion 99’s area count has to fall within this certain range to confirm the detection and the same logic follows for PFOA and other PFAS. Now there are six PFAS and EPA methods 1633 that do not have confirmation ions so I always like to call attention to these as they could be potential false positives and we should really be careful about making decisions on the basis of these PFAS until we have confirmed their presence. Now, interestingly, the examples I’m showing you today are not one of these six PFAS listed here.

And finally, on the bottom here, you can see that the two heavily used methods, so EPA 1633 for non-potable water and soil sediment and EPA 537.1 for drinking water, require the lab to use these confirmation ions for positive identification of PFAS. EPA method 533, which is an isotope dilution method for drinking water like EPA 1633, does not require the use of conformation ions for any PFAS chemicals, so keep that in mind.

Okay, so I want to first discuss some issues with potential false positive results for PFOS. And this particular example was actually a pesticide sample. So due to the odd matrix, the sample was first analyzed using a modified version of EPA method 533.

And remember, EPA method 533, it does use isotope dilution, which is like the gold standard for PFAS quantitation, but EPA method 533 does not require the lab to look at confirmation ions. Another issue with EPA method 533 is that the analysis runtime can be shorter. So the resolution between different PFAS and also potential interference may not be as good.

So with this method, 533, each PFAS is identified and it’s quantified with only one, that one primary ion. The chromatogram that we saw for this particular detection showed one peak for that primary ion of PFOS, and it quantified at almost 2 ,700 nanograms per liter of PFOS. And that was suspect to us and not expected for this pesticide sample.

So we asked the lab to reanalyze this sample using EPA method 537 modified. So again, it was using isotope dilution like EPA method 533, but also now using conformation ions, which as we discussed is again, it’s another layer in ensuring, you know, the accuracy of the identification of a specific PFAS chemical. So let’s look at the chromatogram here on the So, this chromatogram down here is a typical chromatogram of PFOS in the calibration standard.

So you can see here, there’s a larger linear isomer of PFOS here on the right side, okay? And you can see there’s a peak here for the primary ion on top and a peak for the conformation ion on the bottom. And then the smaller branched isomers are on the left here.

So the same thing, you see peaks for the branched isomers for the primary ion on the top and peaks for the branched isomers for the conformation ion on the bottom. So this here is a calibration standard, but the presence of these primary and conformation ions at defined ratios and at certain retention times provides assurance that this is a positive detection of PFOS. And we use this same process when we’re identifying PFOS or any PFAS chemical in the sample.

So now let’s look at the sample here up on top. the PFOS peaks in this sample did not produce a similar ion ratio pattern to this calibration standard. So if we look at the top here at the primary ions, we see something a little strange, right?

We see a very large peak for this branched isomer of PFOS on the left here, and a very small peak for the linear isomer on the right. Now, this is not necessarily impossible, but it’s a bit different from what we normally see with this linear isomer typically being the predominant peak. But then let’s look at the conformation ions.

So we don’t even see a peak for the conformation ion for this branched isomer on the left, and the linear isomer has a peak for the conformation ion, right, but it’s a much larger peak than the primary ion. So no conformation ion here for the branched isomer, making this ion ratio not within acceptance criteria, and the ion ratio for the primary ion is also outside criteria because this confirmation ion peak is much larger, right, than this primary ion peak. So this gets even a bit more confusing because there’s about 87 different branched isomers of PFOS, and not all of the branched isomers produce the same transition confirmation ions.

So it could be possible that this branched isomer is a real PFOS peak, but that maybe we’re not monitoring the right confirmation ion for this particular branched PFOS isomer, which is maybe why there’s no peak underneath. So maybe the confirmation ion is not this 99 here, that’s what was monitored. But this level of evaluation is really not within the scope of our analytical methods.

So when this occurs, okay, when you have a missing confirmation ion or you have an ion ratio outside the acceptance limits, how should or how does the lab report this? So when this was just EPA method 533, okay, remember it does not monitor confirmation ions, the lab is going to see the linear and branch peaks for PFOS from the primary ion on top here, they’re going to sum them, you get the total, and that’s where they get the original 2 ,700 nanograms per liter. So there’d be no reason to think that this was suspect per EPA 533, and it would be reported as is with no qualification because, again, this method does not even look at confirmation ions.

But now we use EPA 537 modified with confirmation ions, and the identification has now become questionable. So there’s no standard guidance really out there on what labs should do when, one, there’s no confirmation ion present at all, or two, the ion ratio is outside acceptance limits. Some labs may report this at the non-detect because there was no confirmation ion showing up for this major peak.

Some labs may report it as is and just qualify the result noting that the ion ratio was outside the limits. I will note that Massachusetts DEP, they recently released their Compendium of Analytical Methods document for PFAS. And in MassDEP’s protocol for PFAS, they require that the labs report this result as a non-detect if there’s no conformation ion present.

So what happened here? So after much work with the manufacturer of this pesticide, it was discovered that there was the potential for bile acids to be present in this sample, which we know from EPA 1633 can interfere with the PFOS peak. And this PQC on the top chromatogram here turned out to not be PFOS.

So these are things that are only going to be discovered if you, one, you know, review the raw data, but two, more importantly, ask the lab questions when a result does not make sense to you. Having this reanalyzed with a slightly better method helped us here. And now that we have EPA 1633, some of these types of issues may be less common, but when you’re looking at data that was generated prior to EPA 1633, you should keep this in mind.

And just to expand a little bit more on these bio-acid interferences, it’s important to understand that some of the more complex matrices can result in suppression or enhancement of PFAS signals. And we’re aware of these known bio-acid, again, interferences with PFOS. And you can see that these bio-acids, they share the same primary ion with PFOS.

So the new EPA method 1633 requires that the labs actually analyze this bile acid standard with these bile acids in it with PFOS to demonstrate that the laboratory can separate these bile acids from PFOS and demonstrate that they’re not gonna interfere with PFOS. So there shouldn’t be any false positives for PFOS because of bioassets by the newest method. But again, something to keep in mind when you are looking at PFAS data.

All right, so now we’re going to look at the second example of a false positive result for PFOS, or maybe not necessarily a false positive, but an exaggerated concentration or high bias for PFOS. And the purpose of the sampling here for this example was for compliance with a monthly monitoring requirement for PFOS and PFOA in the facility’s NPDES permit. And for this example, we collected an effluent sample.

So at the time this occurred, EPA method 1633 was not required. So we submitted the sample for analysis using EPA 537 modified, which was common at the time and close to the requirements of EPA 1633, again, using isotope dilution and confirmation ions for identification. So with the EPA 537 Modify, we’d received results back of 29 nanograms per liter for PFOA, 78 nanograms per liter for PFOS.

And in this case, this PFOS result was qualified by the lab due to the ion ratio being outside of the acceptance criteria. And we kind of thought this PFOS result was suspicious just based on our knowledge of this facility. So, we asked the lab to reanalyze the sample to confirm this PFOS result.

The lab did reanalyze the sample, this time on a different instrument, and you can see PFOA stayed relatively the same at 22 nanograms per liter compared to 29 originally, but PFOS went from 78 to 21 nanograms per liter. So why was this? So if we look at the right column here of this table, you’ll see that the initial and The continuing calibrations were within the acceptance criteria in both the original and the confirmation analysis.

The extracted internal standards, which are spiked into each sample and they’re used for the isotope dilution, were within the acceptance criteria in both the original and confirmation analysis. But again, that ion ratio for PFOS was only within the acceptance criteria in the reanalysis, not the original analysis. So what happened? So let’s look at the first analysis here on the left here.

So you can see the bottom picture, okay, and you see the linear and branched PFOS isomers in the calibration standard like we saw in the previous example. So here the primary ion peaks are on the left side for the linear and branched PFOS isomers and the conformation ion peaks are here on the right side for the linear and branched PFOS isomers. This is the calibration standard.

So we see a strong linear isomer peak and a smaller branched isomer peak for both the primary and the confirmation ions. Now let’s look at this analysis of this sample on the same instrument on the top. So you can see the left side here shows kind of what we saw in that previous example with a much larger peak for the branched isomer for the primary ion on the left here.

And the conformation ion on the right shows the exact opposite, with a larger peak for the linear isomer than the branched. So these ion ratios, the primary over the conformation, are definitely outside limits here. Remember that this PFOS result was 78 nanograms per liter.

So as I mentioned, the lab reanalyzed this sample on a different instrument. And this instrument was set up with EPA 1633 conditions. So the calibration standard is on the bottom, again, with the linear and branched PFOS isomers for the primary and conformation transition ions, similar to what we saw over here on the first instrument on the left.

Again, we see that strong linear isomer peak and the smaller branched isomer peaks for the primary conformation ions. And if we look at the analysis of this same extract now on this different instrument, we see a pattern a little more similar to the calibration standard. With the linear isomer now as the predominant peak, and we see this in both the primary and the confirmation ion transitions.

So if you compare this to the calibration standard now on the bottom, you see a much better match. And in this case, the ion ratios were within the acceptance criteria. So what happened here?

So what we saw on instrument number one here on the left with this large variation in the ion ratios versus the expected ion ratios for PFOS was an indicator of possible interferences with PFOS and this observation was confirmed when we looked on the instrument number one on the left and remember we saw the branched isomer component of PFOS in this sample which differed significantly from the reference standard below here containing branched PFOS isomers. And when we see ion ratio and chromatographic anomalies of this nature, it can mean that an interfering compound, like the bioacid TdCA that we just looked at, could be present. So as a reminder, TdCA and these other bioacids here are frequently observed in tissue matrices.

But they can also occur in any sample impacted by biological activity, which could be a sample that’s gone through a wastewater treatment plant, like the one in this example. So when the lab reanalyzed the sample on the instrument that had conditions set up to separate the bile acids from PFOS, as required in EPA 1633, this analysis successfully resolved the interference peak, as we can see, because there are now ion ratio and chromatographic pattern that was in conformance with the PFOS calibration standard in the reanalysis. Now, it’s not conclusive evidence that TDCA was present as the interfering compound on PFOS in that original analysis, but it’s highly likely that that was the case.

So remember a few things here with both of these examples I showed you. We had to ask the lab for the reanalysis. If we had not, we would have been reporting false positives or biased high results for PFOS.

And seeing interference in ion ratio anomalies with PFOS can be common. And think about this, you know, when you’re reviewing data for PFOS that are flagged as having ion ratio issues, think about this for historical data that you’ve generated prior to EPA method 1633. And if you’re using EPA method 533, which is mostly used as a drinking water method, remember that there are no requirements to look at confirmation ions or ion ratios.

So if you see results for PFOS or really any PFAS chemical that looks suspicious, you may to request analysis using a method which does look at ion ratios. Okay so now we’re going to move, we’re going to look at an example where the lab reported false negative results. So for this project we had the level two report we were using for our data evaluation and this project only required an evaluation of level two data and as part of the evaluation we were reviewing field duplicate variability.

You can see here that everything looked good in this groundwater field duplicate for comparability, with the exception of PFNA. So PFNA was 290 nanograms per liter in the original sample, and it was not detected in the field duplicate sample with a reporting limit of 1.8 nanograms per liter.

So obviously this looked weird to us. So for this particular project, although only level two validation was required, we also had in our possession the level four report. So we went into the level four report, the raw data, and checked out PFNA for these two samples.

And on the left here, you can see that the integration of the primary ion on top and the confirmation ion on the bottom for PFNA were okay in this original groundwater sample. But lo and behold, however, you can see here that the lab missed the integration of PFNA in the field duplicate sample. There were primary and confirmation ions present at the correct retention time for PFNA, but these peaks were not integrated.

So the sample was therefore incorrectly reported by the lab as a non-detect result. So we went back to the lab. We told them what we discovered.

We asked the lab to revisit all the data for this project to see if there were any other samples with missing integrations and subsequent false negative results. And this ended up impacting about three to four different data packages for this one project. And you can see here that it turned out After the lab did another review of all of the data, on the left two columns here, 10 samples for PFNA went from non-detect to detected.

And on the right two columns, 10 samples for PFHPA went from non-detect to detected. And you can see some of these detections, which were originally reported as non-detect, were now above the regulatory criteria, which for this particular project was 20 nanograms per liter. So why the heck did this happen?

So, in this instance, it turned out that all of these errors were associated with one analyst who was not properly trained. In addition, if proper secondary review had been performed in the lab, this would have been caught prior to releasing the data to the client. So, again, remember, if something doesn’t look right, ask the lab.

Something that was really concerning to me here was that if we had not seen this issue in a field duplicate pair, I’m not sure we would have caught the error from the level two review. So we wouldn’t really have had no reason to suspect anything was a myth. So that’s also a little bit concerning.

Okay. So for this one, I couldn’t really figure out what category to put it under. It’s not really false positive, false negative, or interferences.

It was just something very bizarre that happened, and hopefully you’ll agree. So as a reminder to what isotope dilution is, true isotope dilution is when we use a carbon-labeled isotope of a specific PFAS to quantify that specific PFAS. So you can see a few examples here.

We use the carbon-labeled PFBA to quantify PFBA. We use carbon-labeled PFOA to quantify PFOA. And these carbon-labeled isotopes are supposed to mimic the behavior of the associated PFAS and therefore correct for any potential matrix interferences.

So, this carbon-labeled PFOA mimics the behavior of PFOA, carbon-labeled PFOS mimics the behavior of PFOS. And there are 24 PFAS in EPA method 1633 that are quantified by true isotope dilution, meaning that these 24 PFAS compounds have their own specific carbon-labeled isotope. The remaining 16 PFAS in EPA Method 1633 are quantified using extracted internal standards, which means that these remaining 16 PFAS do not have their own specific carbon-labeled isotope, and basically that is because they are just not commercially available.

And therefore, these 16 PFAS are quantified with an existing carbon-labeled isotope that’s closest in retention time, and most importantly, closest in chemical structure. So you can see a few examples here. So perfluoropentane sulfonic acid, PFPES, this is a 5-carbon sulfonic acid, does not have its own carbon-labeled isotope, so it’s quantified using a 6-carbon-labeled sulfonic acid.

And PFNS is a 9-carbon sulfonic acid, again, does not have its own carbon-labeled isotope, quantified using an eight carbon-labeled sulfonic acid for PFOS. 3 ,3-fluorotellomer carboxylic acid. This is a six carbon carboxylic acid.

It’s quantified using a five carbon-labeled carboxylic acid, PFPA. So these PFAS are quantified with another carbon-labeled PFAS closest in chemical structure, which would be closest to mimicking the behavior of that specific PFAS. And so there are no questions.

table 10 in EPA method 1633 clearly defines which isotopically labeled analog to use for each target PFAS, so there should be no question. So if a carbon labeled isotope does not exist for a specific PFAS, an EPA method 1633, the method is very clear on what should be used, and this helps ensure there’s consistency from lab to lab. Okay, a few more highlights of isotope dilution that are important to understand.

So how it works is our samples are spiked with a known amount of these carbon labeled standards prior to the start of sample preparation. We call these extracted internal standards, EIS. And the extracted internal standards are going through the entire preparation and analytical process as the samples.

The instrument’s calibrated for both the target PFAS chemical and the extracted internal standard. And when the samples analyze, the concentration of the target PFAS and the concentration of the extracted internal standard are calculated. But if you look down here, the actual area count of that extracted internal standard is used in the calculation of the target PFAS.

So if the actual concentration of this extracted internal standard is above or below the true value, that’s going to be reflected in the area count. And then that associated target PFAS will be proportionately corrected by that amount. So this really allows the analysis to correct for any preparation or analytical errors, but really, most importantly, for any matrix interferences.

And isotope dilution is really the gold standard for quantitation. It’s required in EPA methods 533, 537 modified, 1633. Probably OK that it’s not used in EPA method 537, since this is for drinking water.

And we don’t really expect too many matrix interferences with drinking water. So OK, that was your lesson on isotope dilution. So now let’s see what happened here.

this particular example, how the lab quantified PFAS. So remember what I just told you, those extracted internal standards used to quantify the target PFAS should have both chemical and retention time similarities because it’s supposed to mimic the behavior of the target PFAS. And as I said, EPA method 1633 makes it very clear on which extracted internal standard to use for each of those 40 PFAS.

So what I’m about to show you happened in about 15 to 20 data packages for this one particular project. So here are some examples, and we’ll start here on the left. So the lab quantified perfluoroheptane sulfonic acid.

This is a 7-carbon sulfonic acid. They quantified this using a carbon-labeled isotope of PFDA, a 10-carbon carboxylic acid. And I can tell you that sulfonic acids and carboxylic acids are not chemically similar.

The lab quantified 5 ,3-fluorotelamer carboxylic acid, again, it’s a carboxylic acid. They used a carbon-labeled isotope of PFOS, a sulfonic acid. Again, sulfonic acids, carboxylic acids are not chemically similar.

They did something similar with the PFNS, PFDOS, TRDA as well. This is a carboxylic acid was quantified with a carbon-labeled isotope of a sulfonamidoethanol. again, not at all chemically similar. And these are just some examples.

They were more like this as well. And the lab said that they did this due to interferences. But even if there were interferences, this is not necessarily a justification for quantitation to be performed like this, because any interference in these extracted internal standards are also going to affect the target PFAS in the same manner. And that is the whole point of isotope dilution.

So, let’s look at it a little bit deeper, how this actually impacted the results. So, you can see the target PFAS across the top row here. The next row shows the carbon-labeled isotope or extracted internal standard that was used by the lab to quantify that PFAS.

And then this row shows the carbon-labeled isotope that was required to be used to quantify the PFAS per EPA method 1633. So, we’ll just go through a couple to make sure you’re clear what you’re looking at. So if you look at PFHPS here, this was quantified by the lab using a carbon labeled isotope of PFDA instead of the carbon labeled isotope of PFOS.

So a seven carbon sulfonic acid was supposed to be quantified by an eight carbon sulfonic acid, which makes sense, but was instead quantified by a 10 carbon carboxylic acid. If you look down here, five, three fluorotylimer carboxylic acid, this was quantified by the lab using a carbon-labeled isotope of PFOS instead of the required carbon-labeled isotope of PFHXA. So this carboxylic acid, 5 ,3-FTCA, was supposed to be quantified by a six-carbon carboxylic acid, which does make sense, but was instead quantified by an eight-carbon sulfonic acid.

And each of these here were impacted in similar manners. And for an example, I’m gonna show you the impacts of this using one sample here. So the first row here under the sample header shows the recovery of the extracted internal standard that the lab was required to use, or was supposed to use, per the EPA method.

A couple of them are outside the limits, but mostly they’re okay, signifying no real interference. The row after this shows the recovery of the extracted internal standard that the lab used, so the recovery using the wrong carbon-labeled standard. First of all, the recoveries of the carbon label standards that the lab was required to use by the method are overall actually better than the recoveries of the wrong ones that the lab used.

So the lab stating that they did this due to interferences made absolutely no sense. The percent recoveries of these extracted internal standards would not be this good if there was any significant interference. And just remember again, isotope dilution corrects for these interferences. It’s okay if there’s interferences.

That’s the whole reasoning behind isotope dilution. But finally, the misapplication of isotope dilution and the use of the wrong carbon labeled isotopes caused the final results for our samples to be biased. And in most cases, most of the final reported results were bias high, as you can see here in the bottom row, based on the use of the wrong carbon labeled isotope.

But really, the bottom line here is that there was really no need for this to happen. And to make it even worse, each data package in which this occurred was done differently. They did not even consistently use the wrong carbon labeled isotope.

Lesson learned here, this was written up in the lab’s narrative, but it didn’t make sense. So again, question the lab. You need to have the most accurate results reported for your clients so you can make appropriate decisions.

And the final example I’m going to show you today is an instance where the reporting limits reported by the lab were lower than the lab was able to demonstrate or justify, but then there was like another little twist also. So the reporting limits or limits of quantitation reported by the lab are really these are the accurate detection limits which are based on the lowest concentration standard used in the lab’s calibration curve. These reporting limits have to be equal to or above the lowest concentration standard use in the lab’s calibration curve, the lab has to demonstrate that they can see down that low and that they can accurately quantify down to this concentration.

So this slide kind of further illustrates where the reporting limits fall in relation to calibration of the instrument. So typically, the lab is going to analyze a multi-point curve. The one shown here is at five different concentrations.

And they plot the response of the analyte here on the y-axis versus the concentration on the x-axis to generate this calibration curve. And this also determines the calibration range. As I mentioned, the reporting limit or limit of quantitation is determined by the lowest concentration standard in the calibration curve.

The MDL, which is lower than these values, kind of falls into that range of uncertainty because it’s statistically determined and not based on the lab’s calibration. So when the lab runs a sample, the concentration of that specific PFAS is determined based on the response. If the response ends up above the calibration range, the lab has to dilute the sample.

But again, that lowest concentration in the curve is used to establish the reporting limit or limit of quantitation. So what did we see? So in this example, the concentration of the low standard for PFOS in the calibration curve was 0.193 nanograms per mil.

In this example, the lab extracted about 100 mils of sample down to a final volume of 5 mils. So, the calculation of that reporting limit is shown here. It’s based on the lowest concentration standard in the curve, the 0.193 nanograms per mil, the final extract volume of 5 mils, and the volume of sample extracted, the 100 mils.

And as you can see, based on that calculation, the reporting limit would have to be greater than or equal to 9.6 nanograms per liter based on the lowest concentration standard in the curve. Now, the lab reported a reporting limit of 8 nanograms per liter.

So yes, it’s not that far off from 9.6, but it’s not within the calibration range. So why was this lab’s reporting limit below the lowest standard used in the calibration?

It turned out that the lab actually did analyze two lower concentration standards, so lower than that 0.193 nanogram per mil standard, but they dropped those lowest two points and did not include them when generating the curve. So, there were standards lower than 0.193 in that curve, but the lab eliminated them, and these lower standards that they eliminated would have supported that reporting limit they used of eight nanograms per liter.

But why did they drop those two lowest standards in the curve? And the reason they dropped them, they said, is because the recoveries of several target PFAS in those two standards, including PFOS, were not within the method recommended range 70 to 130 percent. In fact, the percent recoveries of some of these PFAS, and that’s one of those standards, was greater than 200 percent. So they eliminated those points from the curve.

But then you have to ask, you know, why were the recoveries of those select target PFAS greater than 200 percent in the calibration standard? That kind of seemed odd. Well, this was another project where we luckily also had the level for raw data available.

So we kind of in and checked out that this level two standard to see why the recoveries of some of those target PFAS were greater than 200% and why they had to eliminate that standard from the curve. And lo and behold here we notice that the lab did not integrate the peak for the carbon labeled isotope for PFOS properly. So you can see here only half this peak was integrated, which now explains why the recovery of the select PFAS that are quantified with this carbon labeled isotope were greater than 200%.

Because remember, the area of this peak is in the denominator of that concentration calculation. So if it’s lower than it should be, the concentration of the target PFAS are going to be falsely higher. It’s an inverse relationship.

The lab did not even see this in their own review of the data. They went back and they revised about 20 data packages to first add those two lowest standards back into the curve, correctly report the reporting limits, And finally, we quantify all affected PFAS in the samples after adding these two standards back into the curve. So three main takeaways from today’s presentation.

This presentation is in no means meant to put down the analytical labs, because overall they are doing a really great job. But with anything, there’s potential for errors or poor judgment. And as data users, we need to be aware.

So if the data did not make sense to you, ask the lab. That’s really probably the most important message I can state today. And although reviewing Level 4 data can be a costly process, it does give you the most assurance in the accuracy of your data that you’re all likely using to make some pretty costly decisions.

So now I’m going to hand it off to Maureen Dooley. Well, thank you, Liz. That was a really interesting and important presentation.

Maureen Dooley:

I think as you say, data are so critical and we’re making decisions all the time with them and making sure it’s accurate. And I think my takeaway from that is, even from the work that we do, is if we see something that looks a little funny, make sure to ask questions and really get the data verified. So thank you for a great presentation.

I’m going to switch gears a little bit here. I want to stay on the theme of data, and I’m just going to focus a little bit more on the remediation side of things. Okay, so as we know PFAS, there’s certain challenges associated with it, and there can be many sources.

And in looking at remediation, there’s different places that we may find it where there may be a source in the vadose zone, or it eventually is going to make its way through the soil and find its way into the groundwater, and it may be concentrated or focused in the groundwater capillary fringe, but also we’ll move on and you’ll have a plume as the PFAS compounds make their way into the groundwater. What I want to focus on is talking about data and important data that are used to develop designs and remediation for a plume. So in looking at remediating PFAS in situ, for those of you that are familiar with Regenesis, what we use is a colloidal activated carbon that’s injected directly into the saturated subsurface.

And this colloidal activated carbon will create this barrier that will capture or ****** the flow or flux of PFAS compounds through the groundwater and it’s sort of to stop it and keep it from migrating offsite. So looking at this figure, focus on section D, I’m talking about designing a barrier that’s used to mitigate or stop the flow of PFAS, possibly offsite or stop at that point in the plume. And you can already think of what are some of the data that are really important?

Well, obviously concentration. We want to know how much PFAS or other contaminants that may be present and at what velocity are these traveling? So we want to understand what is the flux coming into this barrier so we know how to design it.

But even looking at this figure, it looks pretty homogeneous from top to bottom in this subsurface, but I think we all know that isn’t necessarily always the case. So again, looking at data needs when we’re trying to design a barrier, you know, specifically, some of the things that are going to be really important, you know, one is in your contamination, your movement into the target remediation zone. So it’s important to know what the concentration is, but also the velocity or the flux.

And it’s also really important to have a very detailed vertical delineation because everything isn’t always homogeneous. And these data can be really important when you’re trying to develop a design because by having a better delineation or vertical profile and the mass flux type data, that allows us to really optimize a design and ensure that we’re going to have a design that will attain the project or the specific regulatory goals. So, you know, we talk about, you know, what is mass flux?

So in essence, this is just contaminant mass moving across a unit area. And this is going to be perpendicular to groundwater flow. But as you see in this animation that, you know, the concentrations where you have darker colors, it isn’t always uniform.

quite often it may vary. And in addition, velocity across this transect may vary as well. So you may have concentrations or flux through this transect actually occurring at different rates.

And having this information is really critical if we’re designing a barrier to try to stop all of this contamination from passing through. So how does the regenesis process work when we’re putting together a design. Well, we have a proprietary model that we developed that’s used to put these designs together.

So the inputs for this are going to be your contaminants of concern. It could be PFAS, but other contaminants that may be present could be VOCs, PHCs, or even TOC. And we have compound-specific isotherms.

But another key component is obviously going to be groundwater velocity or what your contaminant mass flux is. So we’re going to take a look at what that information is to put together our design and what the outputs are really going to be, you know, how much carbon do we need? How much colloidal carbon?

Are there any vertical variations that we have to make some adjustments? What’s the thickness? And also how long is this barrier going to be effective for?

And can we achieve those regulatory goals? So an output for this model will look something like this. So on the left side you have concentration and this is an example showing PCE and TCE as well as some of the PFOS, PFOA and some of the PFOS compounds.

So we’re able to run this model and have multi-year simulations so we can evaluate what’s happening on the other side of the barrier. Is there a breakthrough or not? Or at 15 years, 30 years, are we still able to have this retention of the target compounds.

So the information that’s generally developed as part of a site investigation may be limited when you’re looking at developing a remediation design. If you know, maybe enough for understanding the plume, you know, what’s the size of the plume, but when we’re getting to the remediation stage, we really want to have some additional data and more detailed data. And not only does it help us for the design of the barrier itself, it may be helpful for injections.

We have a lot of volume that’s required to be applied in the subsurface to create this barrier and get this distribution, which is critical when you’re trying to remediate to nanogram per liter levels. So having that better understanding of variations in the type of soils and then the flux will also perhaps influence how we inject this. So there’s different types of data that can be used to gather this type of information.

I mean, we may do a pilot test or grab soil cores, or there’s high resolution sensing tools that can gather additional information. And I just want to take a quick, you know, brief moment to talk about a passive flux meter, the flux tracer that Regenesis have, and how this data is used. Now, many of you may have seen presentations on this and there was a really good technical presentation by my associate Juan Rodriguez in the last webinar.

So I’m just going to get to the surface of this. But in essence, this is a device that’s placed directly in a well and that’s going to generate data that’s gonna have more resolution and accuracy. So that’s going to allow us to have better plume characterization, data that’s used for the design.

And this may also be used to verify remediation performance. And so it’d be a two week turnaround time, something like that. As far as, when I say two weeks, I mean two weeks in the well, is really all you may need to collect the data required.

And then you just send this off to the lab. So just talking about the installation of this device, could be roughly 15 minutes per device, everything comes ready to deploy. The target contaminants that can be measured using this device.

This flux tracer device is coronated solvents, PFAS, and that’s using a method 1633, any compounds that are detected with that. Now we have BTECs, MTBE, and we have more contaminants coming. And we’re also measuring Darcy velocity.

So also to point out that these flux tracer tools are in one foot sections. So this is what gives us that additional delineation. And these can come in 10, 20 or even 30 foot devices with these separate sections.

The PFAS sections would be in a two foot section as opposed to the one foot for some of the other contaminants. The type of output that you get looks something like this where you’ll have, you know, this is the depth of the casing and this is Darcy velocity. And you see that this, you know, how this varies.

And then this is the mass flux of the contaminants or the contaminant flux. And why this is important, you can just look at this, you’re going to be able to see that there are particular zones where there’s a little more transmissivity and concentrations may be a little more elevated. So that’s certainly important for any design that we’re putting together.

And just again, very briefly, how these flux tracers work, there’s a media inside this device. And so any contaminants that flow into the device are sorbed. The media is also preloaded with a tracer, so as groundwater flows through this, the tracer moves through and moves out of the media, and so the media is directly measured for both tracer and the sorbed contaminants.

And so then using that data, we’re able to get a direct measurement of the mass flux and groundwater velocity. So again, why do design with this. This allows us to identify impacted zone, get better resolution on you know conductive zones where you have areas of higher low permeability and it also allows us to focus our designs.

I mean the cost of using these flux tracer devices are comparable to you know a slug test and we often observe that you can have 80 to 90 percent of the contaminant mass moving through just 10 to 20% of this aquifer section that we may be looking at. And so comparing this to some of the passive methods, a pump test or slug test, when you’re doing your initial site investigation, that may be just fine, but you’ll get an average across the zone. So your groundwater velocity of 135 feet per year and a concentration of 726, a flux of 20, But you look at this more resolved analysis, you can see a lot of little variation in groundwater velocity, 325, 275.

But this is also corresponding with some of the highest concentrations that were detected in this section. So this is really critical. When you’re trying to understand, I have a barrier that needs to be effective for 15 or 20 years, understanding these variations are really critical to make sure the design accommodates that.

And so we have enough reagent applied in the right materials, in the right sections. And generally, what do we observe very roughly? Probably about half of our designs are modified after we receive some of this flux tracer data.

Maybe half of that group may be a little, seepage velocities or flux velocities are a little higher than estimated and maybe another 25% may be below that. So just as an example of data that we see, this is a PFAS site that was in the Northeast. Now this was a sand.

So you think, oh, this may be pretty uniform, but you did see some variations in the Darcy velocity, but really most critical was you saw where most of the PFAS compounds were moving. So we wanted to make sure we had focus in that area with the application to ensure we can meet the project requirement of 15 years. And another example, again, we see a very targeted zone, but by adjusting our design, we were able to save $380 ,000 on this particular project because we didn’t have to apply as much material versus applying the same amount of material across the whole transect.

So just in closing, there are a lot of scenarios where this mass flux measurement can be really important. And one is just in your initial site investigation, having a good understanding of contaminant distribution and the vertical delineation can be really helpful in understanding the dynamics of that plume and really refining your conceptual site model. With remediation, there’s a lot of sensitivity, particularly with sorption or permeable barriers, to having what the loading is coming into it.

So having that understanding of what the range is, is really critical for your remediation design. Also complex hydrogeological systems where there’s some heterogeneity. Again, it’s critical to understand that for our design.

And I think finally, the use of these flux tracers or getting this type of flux information is also really important if you’re just trying to verify the performance of a remediation. and it could be an in situ application or it could even be natural attenuation. Having an understanding in measuring the flux coming through a system or off of your site and how it changes over time is going to help you assess the effectiveness of a particular remediation program or just what’s happening naturally at a particular site.

And so, what we’re saying here is, flux tracer is one way to collect that data, but in general, it’s this type of information is really quite critical when you’re putting together designs for remediation of PFAS or really any other contaminants. But with that, I’d be happy to take any questions.