The Science · Dark PFAS
The PFAS the Test Can't See
There is a "forever chemical" that makes up most of the PFAS in semiconductor wastewater — and the test everyone is relying on doesn't even look for it.
Lenny Siegel has spent decades holding polluters accountable — he runs the Center for Public Environmental Oversight and analyzed the PFAS discharge data from the GlobalFoundries fab in Vermont. He pointed our team to a compound most people have never heard of: TFA — trifluoroacetic acid.
TFA is an ultrashort-chain PFAS. It is the smallest member of the family, and that is exactly what makes it dangerous to regulate around: it slips through the standard test, and it slips through the filters. When researchers found a huge amount of "dark PFAS" in semiconductor wastewater — PFAS that was clearly there but wasn't showing up on the normal analysis — TFA turned out to be most of it.
Lenny Siegel · Center for Public Environmental Oversight
TFA was responsible for most of the "dark PFAS" measured in fab wastewater — an ultrashort PFAS not measured by EPA's Method 1633.
>60% of total PFAS
TFA's share of the total PFAS in fab effluents in the Cornell study — the single biggest compound by mass.
96,413 ng/L
TFA measured in real semiconductor fab wastewater in the Colorado School of Mines study.
Invisible to Method 1633
TFA is not a target analyte in EPA Method 1633 or 1633A — the very methods the permit relies on.
Paper 1 — the fingerprint
Cornell University · ACS ES&T Water, 2023
The dominant PFAS in fab wastewater is one the standard test ignores
Jacob & Helbling, "Rapid and Simultaneous Quantification of Short- and Ultrashort-Chain PFAS." DOI: 10.1021/acsestwater.2c00446
- In three semiconductor fab effluents, carboxylic-acid PFAS made up more than 90% of total PFAS, and TFA alone was over 60% — at levels up to 15,280 ng/L. This is the "dark PFAS."
- TFA and the other ultrashort compounds are not measured by EPA Method 1633; they require a separate, specialized method.
- One fab's sample was the effluent from its own onsite reverse-osmosis system — and it still had the highest PFAS levels of any fab tested. Bolting a membrane onto a fab did not make the discharge clean.
- Funded in part by the Semiconductor Research Corporation — the industry's own research arm documents this.
Paper 2 — filtration is not the answer
Colorado School of Mines · Water Research, 2024
Even the best membranes barely remove TFA — and never destroy it
Griffin, Bellona & Strathmann, "Rejection of PFAS… by nanofiltration membranes." DOI: 10.1016/j.watres.2024.122111
- Real fab wastewater tested at 96,413 ng/L of TFA — dwarfing every other PFAS present.
- A tight nanofiltration membrane rejected most PFAS above 99% — but TFA was the worst, at about 92%, and got worse as the system recovered more water. The treated water still carried ~2,022 ng/L of TFA.
- Critically, filtration only separates and concentrates PFAS into a reject stream — it does not destroy it. That concentrate still has to be destroyed. The paper says so explicitly.
- Carbon and ion-exchange "just filter it" answers don't work for these ultrashort compounds — fab co-contaminants inhibit them.
What it means for Oak Orchard
Put the two papers together and the conclusion is unavoidable. The largest single PFAS in semiconductor wastewater is invisible to the test in the permit, and is the hardest of all to remove — and even when a membrane removes most of it, the PFAS isn't gone. It's concentrated into a stream that still has to be destroyed, or it's left in the discharge.
That is the whole case for our two technical demands, now backed by peer-reviewed, industry-funded science:
What this proves
- Testing must go beyond Method 1633. Require EPA Method 1633A plus the TOP Assay plus a dedicated ultrashort-chain method — or the dominant compound (TFA) never gets counted.
- Destruction, not filtration. Membranes concentrate PFAS; they don't eliminate it. The permit must require a PFAS destruction technology for the concentrate — filtration alone just moves the problem.
Count all of it. Then destroy it.
Straight with the science: TFA is not currently regulated at the 4 ppt federal drinking-water limit — that limit is for PFOA and PFOS. The point about TFA is different and just as serious: it is the dominant PFAS by mass in fab wastewater, it is unregulated, it is barely removed by filtration, and it is invisible to the standard test. The concentrations above are documented in semiconductor fab wastewater in these studies (unnamed fabs) — they are not measurements of Micron's Clay or Boise discharge.
Sources: A. K. Jacob & D. E. Helbling, ACS ES&T Water 3 (2023) 118–128, DOI 10.1021/acsestwater.2c00446 · E. Griffin, C. Bellona & T. J. Strathmann, Water Research 262 (2024) 122111, DOI 10.1016/j.watres.2024.122111 · Analysis and referral: Lenny Siegel, Center for Public Environmental Oversight (cpeo.org). Both studies funded in part by the Semiconductor Research Corporation.