Per- and polyfluoroalkyl substances (PFAS), colloquially known as “forever chemicals,” are a group of synthetic materials used in various industrial applications and consumer products. Characterized by their robust durability and resistance to heat, water, and oil, PFAS compounds have been used since the 1940s in everything from non-stick cookware to stain-resistant fabrics¹. But despite their widespread use, PFAS are increasingly being recognized as persistent environmental pollutants with significant risks to human health²,³. In fact, at least 45% of the nation’s tap water is estimated to have at least one type of PFAS material, risking reproductive harm, cancer, and other health complications for millions of Americans⁴.

The growing body of scientific literature and resulting public health concerns have led the U.S. Environmental Protection Agency (EPA) to mandate stringent regulations designed to remove these chemicals from drinking water. On April 10th, 2024, the EPA announced the final National Primary Drinking Water Regulation (NPDWR) targeting six PFAS compounds: PFOA, PFOS, PFNA, PFHxS, HFPO-DA (GenX), and PFBS⁵. This marks the first-ever national regulation of its kind with enforceable standards aimed at significantly reducing the allowable levels of these contaminants in public water supplies. These new rules demand rigorous monitoring and compel water systems to adopt effective treatment solutions by 2029 to ensure PFAS levels meet the new standards.

To this end, artificial intelligence is expected to play a central role in improving the management and remediation of PFAS-contaminated water. These data-driven technologies can help optimize treatment processes through predictive modeling and advanced data analytics, more rapidly developing innovative solutions and supplementing current systems with new capabilities. Their applications extend to real-time monitoring systems that can instantly detect deviations from normal operating parameters to suggest immediate corrective actions. The potential for AI to transform water treatment and rehabilitation has spurred further widespread academic and commercial attention, increasingly hedging their time and resources toward AI-driven water resource management⁶.

Turning the Tide on PFAS Contamination

The EPA’s announcement of the NPDWR for six PFAS compounds is an unprecedented step toward the future of safe and sustainable water resource management. The mandate introduced new, previously absent, legally enforceable maximum contaminant levels for the designated PFAS materials. This includes perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS), which have been given particularly stringent MCLs of 4.0 parts per trillion due to their high toxicity and near ubiquity in water supplies worldwide. In fact, the EPA’s new regulations follow earlier regulations like those in the EU, which have been established for over a decade⁷.

The regulation is likely only the first of many and was informed by an extensive and ongoing engagement between the public, policymakers, healthcare professionals, and water specialists. So far, the result of these collaborative efforts has concluded that swift and substantial actions must be taken, starting by targeting some of the most prolific and well-understood PFAS contaminants. Beginning with a narrow targeting of the most problematic contaminants will enable the most cost-effective and high-impact means of mitigating the risks posed by PFAS materials. Moreover, to accomplish the broader goal of remediating water from all types of PFAS compounds, this selective beginning will start the process of erecting foundational infrastructure, provide technical staff with invaluable experience, and facilitate operational excellence by allowing engineering challenges to be solved early.

The changes catalyzed by these regulations will soon be in effect, mandating that monitoring begin by 2027 and require full compliance to be achieved by 2029. The EPA has also allocated $1 billion from the Infrastructure Investment and Jobs Act (Bipartisan Infrastructure Law) to aid states in implementing PFAS testing and treatment technologies. For regulations targeting public services where access variability is high, this financial support is critical to ensure all communities can access safe drinking water regardless of location or socioeconomic status.

How Can PFAS Be Removed From Water?

The EPA lists three primary technologies used for PFAS remediation: granular activated carbon (GAC), ion exchange resins, and high-pressure membrane systems⁸. Each treatment option offers distinct advantages and is chosen based on several factors.

Granular activated carbon is made from high-carbon, often waste-sourced materials like coconut shells and is commonly used in various types of filtration systems. GAC is highly effective at absorbing organic compounds, including PFAS, influenced by the contact time between the water and the activated carbon and the properties of each specific PFAS compound. In a different manner, ion exchange resins are similarly adept at removing PFAS and allow a higher degree of selectivity compared to activated carbon.

A third highly effective, although costly, treatment process is with high-pressure membrane systems. This includes nanofiltration and reverse osmosis, which both fundamentally function by forcing water through ultrafine membranes to remove small molecular contaminants like PFAS. While expensive, this approach is more effective at targeting wastewater that is especially high in contaminants or when particularly stringent standards have been set.

There is no one treatment technology to rule them all, and the chosen process or processes depend on local water chemistry, PFAS compositions and concentrations, and evolving regulatory requirements. As research and development to address these challenges increase, new technologies are expected to augment current capabilities, providing ever-larger sets of tools and techniques for water treatment and rehabilitation.

AI is Revolutionizing PFAS Detection and Treatment

One of the most promising applications of AI in water rehabilitation is found in optimizing the operational parameters of the aforementioned existing treatment technologies⁶. AI models can analyze vast amounts of operational data to predict the optimal times for media replacement or regeneration, extending the life of treatment systems while ensuring maximum removal efficiency. Additionally, AI-driven systems can model different contamination scenarios to assist in designing facilities that maximize contaminant removal based on local conditions and regulatory requirements. Moreover, AI can enhance real-time monitoring and control of water treatment processes, detecting subtle changes in water quality that may indicate PFAS breakthrough or system malfunctions, prompting preemptive maintenance and avoiding costly downtime or non-compliance with regulatory standards.

Looking ahead, AI models are also being developed to design entirely new materials and methods for removing PFAS and other emerging contaminants. Promising new AI approaches, like NobleAI’s Science-Based AI, can be combined with molecular simulations made possible by Microsoft Azure Quantum Elements to design novel adsorbent materials that offer more selective and efficient PFAS capture. While the technology continues to evolve, these advances hope to deliver next-generation solutions that are both cost-effective and environmentally sustainable, ushering in a new standard for water safety and accessibility.

Seizing the Moment for Safer Water

As the world confronts increasingly precarious water resource challenges, water specialists and stakeholders across the value chain are looking to AI for solutions. The recent EPA regulations mark the first step in the right direction, addressing these chemical hazards to the benefit of millions of Americans.

Stakeholders, from municipal managers to industry leaders, must now embrace the opportunities presented by AI, leveraging these tools to uphold their responsibility to provide the most essential of public services. Collaborating with experts like NobleAI can help accelerate the adoption of these advanced solutions, ensuring that water treatment facilities are both compliant and at the forefront of innovations in the sector. For more information or to schedule a discovery call, visit our website or contact us today!

‍

References

1. An overview of the uses of per- and polyfluoroalkyl substances (PFAS)†" Environ. Sci.: Processes Impacts, 2020, 22, 2345.
2. "Grouping of PFAS for human health risk assessment: Findings from an independent panel of experts" Regulatory Toxicology and Pharmacology 134 (2022) 105226.
3. "A Never-Ending Story of Per- and Polyfluoroalkyl Substances (PFASs)?" Environ. Sci. Technol. 2017, 51, 2508−2518.
4. https://www.usgs.gov/news/national-news-release/tap-water-study-detects-pfas-forever-chemicals-across-us
5. https://www.epa.gov/newsreleases/biden-harris-administration-finalizes-critical-rule-clean-pfas-contamination-protect
6. XGBoost model as an efficient machine learning approach for PFAS removal: Effects of material characteristics and operation conditions" Environmental Research 215 (2022) 114286.
7. https://echa.europa.eu/hot-topics/perfluoroalkyl-chemicals-pfas
8. https://www.epa.gov/research-states/pfas-treatment-drinking-water-and-wastewater-state-science

‍