The conversation around PFAS exposure and risk assessment often focuses on the prevalence of these compounds in our environment and their potential toxicity in the human body. However, the bridge that connects exposure to potential health impacts is bioaccumulation: understanding how PFAS move through the body, bind to proteins, and transform into new forms.
“Traditional risk models assume that chemicals accumulate in fatty tissues – in the lipids – and that their distribution is governed by hydrophobicity and affinity for organic carbon. PFAS don’t behave that way,” says Carrie McDonough, Associate Professor of Chemistry at Carnegie Mellon University, USA, whose work focuses on exploring the hidden biology of PFAS. “They interact with transporter proteins, bind to phospholipids, and engage in more polar interactions, so the classic POPs (persistent organic pollutants) models don’t really apply.”
In this third installment of our PFAS: New Frontiers, Emerging Solutions series, McDonough argues that we need new models that capture these unique behaviors – especially for assessing long-term, low-level chronic exposures like those from drinking water. “Otherwise, we’re not accurately representing what the body is actually experiencing over time,” she says.
The Story So Far
Chapter One: PFAS: New Frontiers, Emerging Solutions – with Chris Higgins
Chapter Two: PFAS Enters its Big Data Era – with Jennifer Field
How did you first find your way into PFAS research?
I come from a fairly multidisciplinary background – which I think is becoming increasingly common among people in environmental research. I started out in chemistry for my bachelor’s degree, then worked in industry for a while. That’s where I began doing mass spectrometry – eventually running an analytical lab at an environmental consulting firm. I absolutely loved it – especially the mass spectrometry – but I didn’t enjoy doing the same analysis over and over again. I wanted to go back to school and keep learning.
When I started looking at graduate programs, I realized that if you wanted to do advanced mass spectrometry for water analysis or environmental contaminants, those programs were rarely in chemistry departments. Chemistry programs tended to focus on atmospheric, basic, or pharmaceutical chemistry – not the kind of work I wanted to do. That’s still a challenge for undergrads today: figuring out where to go next if they want to pursue this kind of research.
So I ended up getting a PhD in oceanography – which, by the way, is where a lot of really good analytical scientists come from. I think that’s because in oceanography you study extremely low trace-level chemicals that change by tiny amounts, so you need strong analytical skills.
I worked with Rainer Lohmann at the University of Rhode Island, doing emerging organic contaminant research in an oceanography school. He wasn’t doing PFAS work at the time, so I studied a wide range of other contaminants – synthetic musks, flame retardants, organophosphate flame retardants, and so on. It was a great way to learn environmental chemistry because I had to understand how structural differences affect analysis, fate, and transport.
When I finished my PhD in 2017, nearly every postdoc position I looked at focused on PFAS. It was becoming a major concern, with lots of federal funding available, so all the opportunities were in that area. I was also really interested in high-resolution mass spectrometry. Up to that point, I had only used single quad and triple quad systems – never QTOF or other high-resolution instruments – and I wanted that experience. So I joined Chris Higgins’ lab, where most of the work was PFAS-focused. As a postdoc, I led much of the biological analysis work – including human serum analysis (a first for the lab) and mouse tissue studies from dosing experiments in collaboration with Jamie DeWitt, who’s now at Oregon State. That experience got me deeply into both high-resolution mass spectrometry and toxicokinetics and chemical biology – which remain big parts of my research today.
It wasn’t that I deliberately chose PFAS because they fascinated me – though they are fascinating – it was more that the field moved that way and I moved with it. Nevertheless, PFAS are fascinating. Due to the fact that there are thousands of compounds with different properties, you get a broad view of fate and transport – as long as you don’t focus solely on perfluoroalkyl acids. You get to use a wide range of instrumentation and consider many different properties, which I really enjoy.
Still, every time I write a proposal or plan a new project, it’s always PFAS – and then I try to include other contaminants as well. Partly that’s because I train students, and I want them to think beyond PFAS: to learn how to do PFAS analysis, but also to ask, “What’s next?” What will they be studying in 20 years if they start their own lab? That broader mindset is important, and I try to instill it in them.
What makes PFAS such a difficult – and interesting – problem to tackle from an analytical perspective?
One of the first things that really intrigued me was the unidentified organofluorine question. You can measure total organic fluorine in a sample, and then measure what you can detect with LC-MS/MS – and those numbers often disagree by a lot. There’s usually a big fraction we’re not identifying, across all kinds of environmental samples.
What I find especially fascinating is how this plays out in biological samples. When I was setting up my lab at Carnegie Mellon, we reviewed every paper we could find – from the early 2000s onward – that measured organofluorine alongside targeted LC-MS/MS, and then calculated the unidentified fraction. The range was remarkable: in human samples, some studies found 0 percent unidentified, others 100 percent. And over time, the unidentified organofluorine fraction seems to be increasing. It also appears to be greater in women than in men, and it often has an inverse relationship with total organofluorine. If you live in a highly contaminated area – say near AFFF contamination – you’re more likely to explain a larger portion of your PFAS burden. But if you’re part of the general population, exposed mainly to diffuse sources and consumer products, you’re less likely to explain much of it.
That’s what fascinates me – why our current methods still don’t close that gap. Many of the techniques I learned from Chris Higgins, and still use, haven’t fully resolved it. Moving from LC-MS/MS to high-resolution LC-MS adds thousands of compounds to your analyte list, but it doesn’t necessarily bring us closer to explaining the difference between what we can see and that unidentified organofluorine number. Some of that gap may exist because we can’t properly quantify the unknown compounds – we can do semi-quantitation, but not enough to explain the missing mass. Fluorinated pharmaceuticals are often cited as contributors, and they probably are, but they can’t explain everything.
In that same review, we saw large unidentified organofluorine fractions in aquatic organisms and invertebrates. Most pharmaceuticals break down fairly quickly in the environment – they either mineralize or degrade into trifluoroacetic acid (TFA). If they’re ending up as TFA, that’s significant, because it represents another contamination pathway.
So this whole question of unidentified organofluorine has become a compelling challenge for me: how do we measure organic fluorine and targeted PFAS together, and actually close that gap?
During my postdoc, I co-wrote a short critical review on this with Chris Higgins and Jen Guelfo. We looked at all the different methods for measuring the PFAS “universe,” and showed how the more specific your method gets, the smaller your slice of that universe becomes. If you want high certainty, you measure a small fraction of the total. If you aim to capture total organofluorine, you gain breadth but lose specificity – you don’t know what the structures are.
I still rely mainly on high-resolution approaches, but we’re pushing further – trying new sample-introduction techniques. For example, we’re starting to use GC-APCI to look for highly soluble, ultra-short-chain compounds and assess how much they contribute to the unidentified fraction. That kind of question – trying to solve the mass balance and understand what we’re missing – draws many analytically minded researchers into PFAS work.
From the high-resolution perspective, there are other exciting directions too, like using clever filtering methods: by mass defect, by homologous series, and now, with ion mobility spectrometry, by size versus mass. PFAS are smaller, but heavier than hydrocarbons of the same mass, so they cluster in a distinct “PFAS region” in ion-mobility space.
We can isolate that PFAS region from complex samples, which increases confidence in our identifications. Those kinds of smart filtering approaches make the problem feel solvable – even if it remains incredibly complex.
Since your “trade-off between selectivity and inclusivity” paper, how much progress have we made – and do we need new tools or just smarter use of existing ones?
I do think we’ve made a lot of progress since then. Almost every lab I talk to now is either already measuring ultra-short- and short-chain PFAS, or actively gearing up to do so – and that’s helping close the gap quite a bit. Many samples contain significant amounts of trifluoroacetic acid (TFA), which used to be largely overlooked. In some extractable organofluorine (EOF) measurements, TFA could even be lost, so it wasn’t always captured consistently. Now that researchers are focusing specifically on TFA and other ultra-short compounds, that’s definitely reducing the unidentified fraction.
The adoption of GC-high-resolution methods is also progressing – more slowly, but steadily – and that will help too. My sense is that a lot of the unidentified fraction in commercial products, and potentially in human tissues, could be compounds that simply aren’t amenable to LC-ESI, so we just haven’t been seeing them. As labs expand into GC-EI/CI and GC-APCI, and develop methods for highly soluble compounds using different chromatographic approaches, we’re starting to detect more of them.
^19F NMR has also come a long way. It used to be seen mostly as a way to confirm that organofluorine was present, without providing much structural detail. Now, you can get some structural information, and using NMR complementarily alongside other methods is becoming increasingly common.
Overall, I think we’ve advanced a great deal – not necessarily by inventing brand-new tools, but by expanding how and where we apply the ones we already have.
How do factors like bioaccumulation fit into the bigger picture of PFAS risk assessment?
Our work tends to focus on everything that happens before toxicity – essentially, everything leading up to the point where a compound reaches its target concentration and activates a receptor. Beyond that, I usually collaborate with toxicologists, because things get much more complex. But all the steps leading up to that point are equally important – and often overlooked.
For instance, when people run in vitro toxicity assays, they might try to mimic real-world mixtures – which is good, because we’re never exposed to just a single PFAS compound. We’re always exposed to mixtures. But what those experiments often miss is that what cells actually “see” isn’t the same as what’s in the water or exposure medium. Cells are exposed to the compounds that persist long enough to bioaccumulate – those with long biological half-lives. So, over time, the molar ratios your cells experience are very different from those in the original mixture.
A lot of our research focuses on that question: if you’re exposed to a certain mixture, what actually ends up in your cells, and how is it distributed throughout the body? That’s the toxicokinetics side of things.
Bioaccumulation has always been a core part of risk assessment, but with PFAS it’s far more complicated. People who specialize in PFAS understand this, but it’s still gaining traction more broadly in the risk assessment community. Traditional risk models assume that chemicals accumulate in fatty tissues – in the lipids – and that their distribution is governed by hydrophobicity and affinity for organic carbon. PFAS don’t behave that way.
They interact with transporter proteins, bind to phospholipids, and engage in more polar interactions, so the classic POPs (persistent organic pollutants) models don’t really apply. Because of that, we need new models that capture these unique behaviors – especially for assessing long-term, low-level chronic exposures like those from drinking water. Otherwise, we’re not accurately representing what the body is actually experiencing over time.
How close are we to understanding what all these PFAS are doing to us?
I’d say we have a good foundation – but there’s still a long way to go.
We do have strong epidemiological studies showing clear effects from certain PFAS, particularly the well-studied compounds like PFOS and PFOA. These are also the ones most frequently detected and often found at the highest levels in human blood and other tissues. In contaminated communities, where most of the measured organofluorine can be explained by these known compounds, a lot of the observed health effects and risk appear to be driven by them.
So, in that sense, we do have solid information – certainly enough to justify stronger regulatory action than we’re currently seeing. But many questions remain, especially around mechanisms and how these compounds are distributed in the body.
One major focus of my work is precursors – compounds that can transform into perfluoroalkyl acids but may have very different potencies or behaviors. If you’re exposed to neutral, volatile PFAS in cosmetics or consumer products, they can degrade into terminal perfluoroalkyl acids – but where does that transformation happen? On your skin? In your skin? In the liver? In the gut, before they enter your bloodstream? Understanding where and how those conversions occur is crucial – as is understanding the difference between being exposed to a poorly characterized precursor versus its well-studied breakdown product.
Another major area of uncertainty involves the short-chain and ultra-short-chain compounds. Much of my work takes a bioaccumulation-directed prioritization approach – the idea that compounds persisting in the body will reach higher concentrations in cells than those that are rapidly excreted. But even if short, highly soluble PFAS don’t bioaccumulate, continuous exposure still matters. If you drink the same contaminated water every day, you’re being constantly re-exposed to them.
The challenge is that most toxicology assays – particularly short-term exposure models – aren’t designed to address chronic, low-level exposure to soluble compounds. We need much more research targeting those long-term, everyday exposures. That means rethinking experimental design and moving beyond some traditional approaches. It will take time, but it’s essential if we want to truly understand the risks posed by the full range of PFAS.
From an analytical standpoint, how effective are current remediation approaches – and what are their limits?
Yes – I’ve done some work evaluating remediation technologies, and I expect to do more in the future. When I was working with Chris Higgins, and later during my three years at Stony Brook – before joining Carnegie Mellon – I was in an engineering department where a lot of PFAS treatment research was taking place.
Early on, when I began collaborating with engineers on PFAS treatment projects, I often found myself saying: “Just because PFOS is going down doesn’t mean this is a good treatment technique.” That mindset seems to have caught on now. These days, many researchers are using untargeted methods, measuring total organic fluorine, or monitoring fluoride release as part of their treatment evaluations – which is really encouraging to see.
Much of the early remediation work I saw (and sometimes contributed to) simply transformed long-chain PFAS into short-chain ones. That can make it look as though treatment is working – because the regulated compounds disappear – but in reality, you’re often just creating shorter chains that are harder to remove and may eventually be regulated themselves. We did a lot to highlight that issue, and I’d like to continue exploring it, especially from the non-targeted analytical side.
I often talk with collaborators about how to confirm mineralization – which is now seen as the “holy grail” of remediation: actually breaking PFAS down completely, releasing fluoride rather than generating TFA or other persistent by-products.
Overall, remediation is important – and we should be using the most advanced techniques available, especially for drinking water treatment. That said, I think the real solution is avoiding PFAS altogether. For the newer exposure pathways I’m studying – where PFAS exposure comes directly from products – if you’re still using PFAS, you’re still being exposed. There’s no way to “treat” that exposure.
And for contamination that’s already out there – in oceans, sediments, even stored in the Arctic – there’s no realistic way to clean it up at scale. We can treat drinking water before people consume it, but once PFAS are in fish, for example, it’s too late. They live in contaminated water. You can’t treat all the water, and once PFAS accumulate in their tissues, there’s no way to remove them. We can test and measure, but if we want to eat fish, there’s simply no method to make them PFAS-free.
So remediation has limits – and it’s important to keep that in mind when communicating with the media. I worry when I see headlines claiming, “PFAS are no longer forever – we can break them down.” Usually, that means a few PFCAs can be degraded under very specific lab conditions, using highly specialized processes. You might be able to apply those techniques in a pump-and-treat system or a drinking water plant – but they’re still early-stage and don’t solve the global problem.
I worry people will see those headlines and think, “Oh, the problem is solved – we can keep using them.” And some industries would be very happy for that perception to spread. But that mindset doesn’t solve anything.
We need to stop using organofluorines in products that don’t absolutely need them. Yes, there are essential pharmaceuticals and certain safety-critical materials where PFAS are probably still necessary – but for many other uses, I have faith in the ingenuity of synthetic chemists. They’re already finding ways to achieve those useful properties without adding a fluorinated methyl group or a perfluoroalkyl chain to everything.
Where can the next generation of analytical chemists make the biggest impact on the PFAS problem?
Definitely on non-targeted analysis. Get really solid training in high-resolution mass spectrometry – which is still something of an art – and, ideally, in machine learning and big data analysis as well. We need more people who can bridge those two worlds, because there still aren’t many researchers fluent in both. That combination could help us prioritize what to investigate, rather than getting lost in the sea of possibilities.
Right now, even in my own group, we sometimes get stuck in “analysis paralysis” – there are so many tools and workflows for non-target analysis that it can be dizzying. You start down one path, then pivot to another, and it’s easy to lose sight of the core question. Having more researchers who can design better prioritization and filtering strategies would be incredibly valuable.
For those working on PFAS analysis, I’d also say: don’t limit yourself to one sample introduction or ionization technique. The field often defaults to ESI and C18 LC columns, because those were ideal for PFOS and PFOA when we first realized they were a problem. But they’re not necessarily ideal for all PFAS. Gain experience with a range of tools – GC-APCI, GC-EI, and alternative chromatographies – especially for detecting ultra-soluble compounds that conventional methods might miss. That kind of versatility will be essential.
Another point I emphasize with students is that there’s a real need for more master’s-level training in analytical chemistry. Many of the undergraduates in my lab want to work in industry doing PFAS analysis. They often find that a bachelor’s degree isn’t enough, but a PhD feels like overkill. In reality, a master’s can open many doors – especially since a lot of instrument companies hire PhDs for roles where a master’s would be perfectly suited. For students aiming for analytical or industry-focused careers, that’s a great path to consider.
And finally, in all my classes and training, I tell students: make sure your results make sense. It sounds simple, but it’s easy to lose sight of in non-target PFAS work. If you think you’ve found an ultra-short-chain compound in dolphin blubber, stop and ask – does that make sense? Could it be something else? Bring what we already know about compound properties into your interpretation. That kind of critical reasoning is something students often aren’t explicitly trained in – but it’s absolutely essential for doing good non-target analysis.
What key message would you like analytical scientists – especially those outside the PFAS field – to take away?
The main message I’d want to share with analytical scientists – especially those who aren’t deeply involved in PFAS research and are simply looking for clear answers about contamination – is this: if you’re not performing any kind of total organofluorine analysis, you’re not seeing the full picture.
Correlation isn’t causation, and PFOS and PFOA are not the whole story. We need to stay aware that without using a variety of ionization techniques, sample introduction methods, and broader analytical approaches, we’re only observing a narrow slice of the PFAS universe.
And something people often forget: not all PFAS are immutable. Most PFAS are precursors – compounds that can transform into perfluoroalkyl acids. While the end products may be persistent, the majority of the PFAS universe consists of transformable compounds. Transformation processes really matter for PFAS – they’re often overlooked, but they’re critical to understanding the problem in its entirety.
The Story Continues
Chapter Four: Confronting the Messy Reality of PFAS Regulation – with David Megson
Chapter Five: Portable Sensors: The Next Generation of PFAS Detection – Silvana Andreescu
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
