New EPA Data: Huge Amounts of PFAS Underreported and Burned

On October 20, 2021, EPA released final Toxics Release Inventory (TRI) data for 2020, which included the first-ever reporting on many per- and polyfluoroalkyl substances (PFAS) chemicals.

Credit: 123RF

On October 20, 2021, EPA released final Toxics Release Inventory (TRI) data for 2020, which included the first-ever reporting on many per- and polyfluoroalkyl substances (PFAS) chemicals. The data has many limitations, yet still show significant PFAS pollution in the U.S.

We found:

  • 49 facilities from 37 parent companies and 8 industry sectors reported a total of 6,282,928 pounds of PFAS waste produced in 2020.
  • At least 56% of reported industrial PFAS waste is either burned (incinerated) – which has not been shown to destroy PFAS – or discharged directly into air, water and land.
  • Reported PFAS waste includes 39 different PFAS substances.
    • We found data on 6 PFAS in the TRI dataset that were not labeled by the EPA as PFAS - yet they contain at least one fully fluorinated carbon atom and therefore were incorporated into our analysis – these PFAS made up around 87% of the total reported PFAS production waste.
    • Over half of the 39 chemicals (20) are either regulated or proposed to be regulated (mostly at the state level), are listed under an international agreement for restriction or elimination (Stockholm Convention or Montreal Protocol), or are supposed to have been voluntarily phased out of use.
    • The rest with few to no health studies.
  • There is a concentration of PFAS waste production in Southern states – highest reporting states include Alabama, Kentucky, West Virginia, Texas and Louisiana. 
  • Industry sectors that reported generating PFAS waste are: Chemicals, Hazardous Waste, Machinery, Food, Nonmetallic Mineral Product, Chemical Wholesalers, Computers and Electronic Products, and Petroleum.
    • Hazardous Waste and Chemicals industries made up 97% of PFAS waste reported.

However, we note some key limitations of the data, which indicate that the problem is likely far worse than even these data show, and that these data represent only a slice of the full problem. These serious data limitations include the following:

  • a high reporting threshold, meaning that many facilities with significant emissions don’t have to report; 
  • only a limited number of PFAS must be reported— less than 200 PFAS out of thousands of PFAS must be reported, meaning that major PFAS chemicals in use, such as PFBS, PFBA, and fluorotelomer sulfonates, are not reported; and
  • to add to the problem, EPA has failed to categorize several reported PFAS as PFAS by using a narrow definition of PFAS. This artificially makes the problem seem smaller than even the current limited TRI data indicate.

These limitations have resulted in known egregious emitters not having to report PFAS waste production to TRI.

Given these findings, it is crucial for the EPA to immediately close the reporting threshold loophole, revise its definition of PFAS to align with the consensus definition of PFAS, and require much more comprehensive reporting on PFAS to ensure that hundreds of PFAS chemicals are not excluded from monitoring and management efforts and emitters are held accountable for the PFAS pollution they create.

EPA PFAS definition excludes almost 90% of reported releases

The widely accepted definition of PFAS is a chemical substance containing at least one fully fluorinated carbon atom. This definition has been used by the scientific community at large, state governments, Congress, and is in line with the definition used by the Organisation for Economic Co-operation and Development (OECD). It has recently come to light that several EPA programs are using a “working definition” of PFAS that is narrower and is out-of-step with this widely accepted definition. EPA’s definition excludes hundreds of PFAS chemicals that are contributing to widespread pollution of the environment.

In line with the widely accepted scientific definition of PFAS, our analysis of TRI data included a search for and inclusion of any chemical in TRI that had at least one fully fluorinated carbon atom. Our analysis found that of the 6,282,928 pounds of PFAS reported to TRI as production waste in 2020, only 841,443 pounds (13%) were labeled as PFAS chemicals by the EPA in the TRI dataset.

Majority of production-related PFAS chemical waste is burned

TRI mainly requires companies to report quantities of production-related waste and provide information on how much is managed through recycling, energy recovery (burned), treatment, and disposal or other releases. Quantities of total releases, including how much TRI chemical waste is released directly into the environment via air, water, or land, is also reported.

Of the 6,282,928 pounds of PFAS production waste reported, facilities stated that 86.4% of it was treated, 9.2% of it was recycled, 0.3% was burned for energy recovery, and the remaining 4.1% were released or disposed to air, water, or land. Details shown in Figure 1 below.

Figure 1 Percentages of PFAS waste management categories and releases

While it appears only a small percentage of the total PFAS waste was released or burned (those that are combusted for energy recovery), a more detailed look at the data reveals that the majority of PFAS waste categorized as “treatment” were also burned. Existing thermal waste treatment technologies were not designed or verified to break down the chemical bonds in PFAS. Rather they are likely distributing PFAS chemicals, possibly in altered forms, back into the environment. Therefore, our more detailed analysis shows that at least 56% of the industrial PFAS waste is either directly discharged to air, water and land or indirectly emitted through incineration methods. See Figure 2 below.

Figure 2 Production-Related PFAS waste management methods

10 PFAS substances contribute to 99% of the reported waste

The majority—over 99%—of production waste reported is from 10 PFAS substances: 


Almost 90% of the total PFAS production waste reported are from PFAS chemicals that are not labeled as “PFAS” in the TRI reporting system, including tetrafluoroethylene used to make fluoropolymers and ozone-depleting fluorocarbons usually used as refrigerants and being phased out under the Montreal Protocol. Tetrafluoroethylene is used to make fluoropolymers such as PTFE (Teflon) and is “reasonably anticipated to be a human carcinogen” by the National Toxicology Program. For the substances that are labeled as PFAS chemicals in TRI by the EPA, many are known to be toxic and are either 1) regulated or proposed to be regulated (at the state level); 2) are listed under the Stockholm Convention for restriction or elimination; 3) or are supposed to have been voluntarily phased out from U.S. domestic production. Details shown in Figure 3 below.

Figure 3 PFAS compounds reported in production waste in 2020 TRI data

Chemours tops list of waste producers

Companies with highest production waste reported are presented in Figure 4below. Quantities of production waste from facilities that are under the same parent company were added together. 


Figure 4 Quantities of production waste and management reported by companies in 2020 TRI data

Highest amounts of PFAS waste reported in Southern states

PFAS waste production and emission is widespread, with the highest amounts found in Louisiana, Kentucky, West Virginia, Alabama, and Texas (see Figure 5). Of note, Louisiana, Alabama, Kentucky, Georgia, and Illinois produce PFAS waste that do not fit into EPA’s narrow definition of PFAS. This is another case where if EPA does not align its definition with the consensus definition, many of the states producing the highest amount of PFAS waste would be perceived as not producing any PFAS waste at all.

Figure 5 Total production waste reported by state in 2020 TRI data. Hover over state for facility-specific emission data.

The finding that the majority of reported PFAS waste is incinerated or directly released into the environment is cause for serious concern. In a recent EPA review of PFAS disposal technologies, the agency concluded that the available data do not show that thermal technologies (incineration) can safely destroy PFAS. Instead, PFAS waste and its incineration byproducts are likely being released into communities and the environment around incineration facilities. Furthermore, a significant amount of the PFAS released are known bad-actor chemicals, linked to multiple health harms. Legacy PFAS, such as PFOA and PFOS, have a plethora of data linking them to multiple health harms and the newer PFAS technology, known as GenX, has been found by EPA to be linked to similar health harms associated with the legacy PFAS that it has replaced.

Further, only eight industry sectors reported generating PFAS waste—yet we know many more industries use PFAS. A recent paper cataloged over 200 use categories for PFAS. This raises serious questions about the quality of reporting to EPA, and we urge EPA to investigate the accuracy of this first round of TRI PFAS data and to close the reporting threshold loophole immediately. Finally, EPA should immediately revise its definition of PFAS to align with the consensus definition of PFAS and require much more comprehensive reporting on the PFAS class to ensure that hundreds of PFAS chemicals are not excluded from monitoring and management efforts.

Click here to see the technical support document for data analysis covered in this blog. 

Related Blogs