DuPont Kept the Science and Spun Off the Bill

In 1961, a DuPont toxicologist looked at the livers of rats exposed to a compound used to make Teflon, watched them swell at low doses, and wrote that the material should be handled 'with extreme care.'⁶ That memo did not go to a regulator. It went into a file. For roughly four decades the science stayed inside the building while the chemical kept flowing out of it - eventually more than 7,100 tons of PFOA-laced sludge dumped near one plant, seeping into the Ohio River and the water of communities downstream.⁷ DuPont knew. The people drinking it didn't.

The official story is that PFAS is a cautionary tale about chemistry we didn't understand. It isn't. The chemistry was understood early, by the people who made it. What was engineered - carefully, over years - was the distance between the knowledge and the bill.

The science went into a file, not a filing

Read the internal record and a pattern emerges that is worse than ignorance, because it is the opposite of ignorance. The 1961 liver finding was followed by a 1970 Haskell Laboratory memo calling C8 'highly toxic when inhaled and moderately toxic when ingested,' and a 1979 report in which dogs given a single dose of PFOA died within 48 hours.⁶ None of this was speculative. It was DuPont's own lab, funded by DuPont, telling DuPont what the molecule did to living things. The choice at each fork was the same: keep it internal. The cost of that choice arrived in 2004, when the EPA fined the company $16.45 million for not disclosing its PFOA findings - then the largest civil penalty ever levied under U.S. environmental statutes.⁶

A fair caveat: the 1961 documents do not show DuPont knew the chemical caused cancer in humans. They show liver enlargement in rats and a handling warning - animal and occupational data, not human disease causation, which the public record did not establish until the litigation-driven science of the 2000s. But that distinction cuts the company less slack than it appears to. A firm does not need proof of human cancer to owe the public a warning that its lab found the compound 'highly toxic.' The duty was triggered by the toxicity finding, not by the eventual death certificate.

How a lawyer turned the files against them

The internalized science stayed internalized until discovery pried it loose. Attorney Robert Bilott's case against DuPont, settled in 2001, established the 7,100 tons of dumped PFOA sludge and its path into local water; the related Leach class action, filed in 2002, swept in more than 80,000 West Virginia and Ohio plaintiffs.⁷ The documents those suits produced - spanning 1961 to 2006 - now sit in a public library at UCSF, which is how the world reads the 1961 memo at all.⁷ That is the moment the model broke. Once the science was external, the liability could no longer be contained inside the file. It had to be contained inside a balance sheet. And in 2015, DuPont built one to put it in.

The spinoff that allegedly capped the unlimited

In 2015 DuPont carved its Performance Chemicals unit into a new company, Chemours, and handed it the PFAS legacy. DuPont's stated rationale was ordinary corporate ambition - it said it was 'seeking to transform itself into a higher growth, higher value company.' Chemours' later account was less generous. In a 2019 Delaware lawsuit, Chemours alleged DuPont 'orchestrated' the spinoff 'as part of a plan to off-load its historical environmental liabilities,' and that DuPont had certified maximum aggregate liabilities of no larger than $1.42 billion - a figure Chemours called 'routinely and radically understated.'³ The complaint asked for the liability limits to be removed, or for the roughly $4 billion dividend Chemours had paid DuPont on the way out the door to be returned.³

Here is the mechanism that makes this the template case. A spinoff is a clean way to assign liabilities to a fresh entity - and if the parent estimates those liabilities low, it walks away light while the new company shoulders the gap. The numbers Chemours later pointed to are the tell. DuPont allegedly pegged the sprawling PFOA multi-district litigation - some 3,500 cases - at $128 million. That MDL settled in 2017 for $671 million. A Cape Fear River cleanup it allegedly estimated at $2.09 million grew past $200 million.⁵ If those estimates are the basis for a liability cap, the cap was set far below the cost - and the spread between the two became the business model.

Didn't the courts side with DuPont?

The honest counter is that the 'liability dump' narrative is Chemours' litigating position, not a court's verdict - and the courts did not buy the case. In March 2020, Delaware Chancery Judge Sam Glasscock III ruled that Chemours had to take its dispute to arbitration, upholding a valid arbitration clause in the separation agreement; the Delaware Supreme Court affirmed that dismissal in December 2020.⁴ Parent companies, the reasoning runs, get wide latitude to structure spinoffs, and a binding agreement is a binding agreement. So the spinoff stands as lawful. But notice what the rulings decided and what they didn't. They turned on the forum and the contract's validity - not on whether the estimates were honest or the liabilities fairly assigned. 'Lawful' and 'fair to the people poisoned' are different questions, and only one of them was on the docket.

The two sides ultimately stopped fighting each other and started splitting the bill. In January 2021, DuPont, Corteva, and Chemours signed a binding MOU establishing a cost-sharing arrangement and escrow for legacy PFAS liabilities from pre-July 2015 conduct, with expenses split 50-50.² That framework is often described as a clean resolution. It isn't a cap. DuPont's own FY2024 10-K still carried a $222 million indemnification liability under the MOU as of the end of 2024.¹ And in 2025 a separate $875 million settlement of legacy New Jersey claims was announced - DuPont's share alone an estimated $311 million, booked with a $177 million pre-tax charge - with no admission of liability.⁸ The bill is still arriving.

DuPont spent forty years keeping a memo in a file and a decade keeping a liability in a subsidiary, and the through-line is the same instinct seen twice: hold the knowledge close, push the cost away. It worked, in the narrow sense that the company is still standing and the spinoff was ruled lawful. But the chemical is in the blood of people who never read the 1961 memo, and the bill keeps reappearing on the balance sheet - $222 million here, $311 million there - precisely because the thing being managed was never going to break down. You can spin off a division. You cannot spin off a molecule that lasts forever.