PFAS behave unlike any other regulated determinand: biological treatment does not remove them, it manufactures them. Precursors biotransform into the very perfluoroalkyl acids the consent limits, so a works can discharge more persistent PFAS than it receives.

The classes that matter, and why

The engineering distinction is not between individual compounds but between terminal acids, which cannot be degraded biologically, and precursors, which can — into those acids.

ClassExampleBehaviour in secondary treatment
Perfluoroalkyl carboxylic acids (PFCAs)PFOATerminal — not biodegradable
Perfluoroalkyl sulfonic acids (PFSAs)PFOSTerminal — not biodegradable
Fluorotelomer alcohols (FTOHs)6:2 FTOHBiotransforms to shorter PFCAs
Perfluoroalkyl sulfonamides (FASAs)N-EtFOSEBiotransforms — a route to PFOS
Fluorotelomer sulfonates (FTSs)6:2 FTSPartially biotransforms

Partitioning: chain length decides where it goes

Distribution between liquor and solids is described by the solid–water partition coefficient:

Kd = Cs / CwCs = concentration on solids (µg/kg); Cw = dissolved concentration (µg/L)

The sorbed fraction then follows from the mixed liquor solids concentration:

fsorbed = KdX / (1 + KdX)X = suspended solids concentration (kg/L); at MLSS 3 g/L, X = 0.003

CompoundChain lengthKd (L/kg)Sorbed at MLSS 3 g/L
PFOSC8 sulfonate≈1,20078%
PFOAC8 carboxylate≈45057%
PFHxSC6 sulfonate≈32049%
PFBAC4 carboxylate≈185%
The split governs the treatment strategy. Long-chain compounds concentrate into sludge, so sludge disposal becomes the controlling route. Short-chain compounds stay dissolved and pass through to the effluent, where granular carbon performs poorly. Neither problem is solved by the other’s technology.

Precursor biotransformation increases the persistent load

Polyfluorinated precursors oxidise stepwise in activated sludge, terminating in a perfluoroalkyl acid that will not degrade further:

6:2 FTOH → 6:2 FTCA → 6:2 FTUCA → PFHxA → PFPeA

PrecursorHalf-life in activated sludgePrincipal terminal product
6:2 FTOH≈2 days6:2 FTCA, then PFHxA
8:2 FTOH≈2 daysPFOA
N-EtFOSE≈3 daysPFOS
6:2 FTCA≈4–5 daysPFHxA

Half-lives indicative, at roughly 15-day sludge age and 20 °C.

At sludge ages typical of nitrifying works, precursors are comfortably within their transformation half-life. A works can therefore show a net increase in perfluoroalkyl acid mass between inlet and outlet — entirely legitimately, and entirely invisibly if only terminal acids are monitored.

The practical consequence for monitoring: an influent sample analysed only for the regulated terminal acids understates the true PFAS load, because the precursor reservoir is not counted. Total oxidisable precursor assay gives the fuller picture where consent risk is material. The wider regulatory framing is in PFAS removal from water.

Separation: carbon, resin and membranes

Granular activated carbon

Adsorption follows a Freundlich isotherm, and capacity falls sharply with chain length because the driving force is hydrophobic:

qe = KF Ce1/nqe = capacity at equilibrium; KF, n = Freundlich constants

At 10–20 minutes empty bed contact time, PFOA typically runs to the order of 10,000 bed volumes before breakthrough; PFOS considerably fewer once natural organic matter competes for sites. Short-chain acids break through almost immediately — a few hundred to a couple of thousand bed volumes — which is why carbon alone cannot meet a low total-PFAS consent.

Anion exchange

PFAS carry an anionic head group, so a strong-base resin binds them by ion exchange rather than by hydrophobic partitioning:

R4N+Cl− + RfSO3− ⇌ R4N+RfSO3− + Cl−

That mechanism is far less sensitive to chain length, so resin substantially outperforms carbon on the short-chain compounds — by roughly an order of magnitude in bed volumes to breakthrough for PFBA and PFPeA. The comparison is developed in GAC vs ion exchange for PFAS.

Reverse osmosis

Rejection exceeds 99.5 per cent for PFOA on a tight polyamide composite, by size exclusion and charge repulsion together. The difficulty is not rejection but the concentrate: at 75 per cent recovery, a 10,000 m³/d plant produces 2,500 m³/d of concentrate at roughly four times the feed concentration, which must then go to thermal or electrochemical destruction. RO relocates the problem; it does not close it. See reverse osmosis system design.

Destruction, and why it is only viable on concentrate

Electrochemical oxidation

Boron-doped diamond anodes mineralise perfluoroalkyl acids completely:

CnF2n+1COO− + (2n+1) H2O → nCO2 + (2n+1) F− + (4n+1) H+ + (4n+2) e−

The specific energy follows from the charge required, the cell voltage and the current efficiency:

Esp = ne F Ecell / (M · η)F = 96,485 C/mol; Ecell = cell voltage; M = molar mass; η = current efficiency

Worked example — PFOA mineralisation energy. Taking 16 electrons per molecule at a 5.2 V cell voltage, 35 per cent current efficiency and M = 414 g/mol:
  • Charge energy: 16 × 96,485 × 5.2 ≈ 8.03 × 106 J per mol.
  • Per gram, at 35% efficiency: 8.03 × 106 / (414 × 0.35) ≈ 55 kJ/g.
  • That is about 55 MJ/kg, or 15.4 kWh per kg of PFOA destroyed.

At municipal concentrations of 100–500 ng/L that energy is spread over an enormous water volume, which is what makes direct treatment of full flow untenable — not the energy per kilogram, which is modest.

Sonolysis

Acoustic cavitation produces transient local temperatures of several thousand kelvin, sufficient to break the carbon–fluorine bond. Defluorination follows pseudo-first-order kinetics; at representative laboratory conditions, 99 per cent defluorination of PFOA needs of the order of five hours, corresponding to several hundred kWh per cubic metre. That is defensible for a small volume of concentrated firefighting-foam waste and indefensible for a works effluent.

The general rule for PFAS: separate and concentrate first, destroy second. Every destruction technology is priced per unit volume treated and paid for per unit mass destroyed, so the economics improve in direct proportion to how far the stream has been concentrated beforehand.

Treatment trains

Municipal works facing a low effluent consent

Primary and secondary treatment, then tertiary adsorption — carbon for long-chain, resin where short-chain compounds dominate — then, if the consent demands it, an advanced oxidation stage to attack the remaining precursors. Sludge goes to incineration; land application returns the long-chain fraction to the environment and should not be assumed available.

Industrial effluent with high loading

Balance the flow, then foam fractionation to exploit the surfactant behaviour of PFAS and pull the mass into a small volume, then membrane concentration, then electrochemical destruction of the concentrate. Each stage exists to reduce the volume presented to the next.

Metals and other consent determinands on the same works are covered in heavy metals removal; the sludge route in sludge incineration and energy recovery.

Design rules

  • Do not credit secondary treatment with PFAS removal. It converts precursors into regulated terminal acids and can increase the persistent load across the works.
  • Follow the chain length. Long-chain compounds go to sludge and become a solids-disposal problem; short-chain compounds stay dissolved and become an effluent problem.
  • Specify resin, not carbon, where short-chain acids dominate. Carbon capacity collapses below about C6.
  • Plan the concentrate route before selecting RO. Excellent rejection is worthless without a destination for the reject stream.
  • Concentrate before destroying. Destruction energy per kilogram is manageable; the cost is in pushing dilute water through the reactor.
  • Assume sludge incineration. Long-chain PFAS partition to solids, and land application is unlikely to remain a compliant route.

Frequently asked questions

Can a works discharge more PFAS than it receives?

In terms of regulated perfluoroalkyl acids, yes. Polyfluorinated precursors that are not themselves on the consent biotransform in activated sludge into terminal acids that are. Monitoring only the terminal compounds at the inlet understates the load arriving.

Why does granular carbon fail on short-chain PFAS?

Because adsorption onto carbon is driven by hydrophobic interaction, which scales with the fluorinated chain length. Below about six carbons the driving force is too weak to compete with natural organic matter for sites, and breakthrough occurs within a few hundred to a couple of thousand bed volumes. Anion exchange binds the charged head group instead and is far less chain-length sensitive.

Is electrochemical destruction practical for a municipal works?

Not on full flow. The energy per kilogram of PFOA destroyed is modest at roughly 15 kWh/kg, but municipal concentrations are in the hundreds of nanograms per litre, so that mass is spread across an enormous volume. Destruction becomes viable only after separation has concentrated the PFAS into a small stream.

Can PFAS-bearing sludge still go to land?

Long-chain PFAS partition strongly to solids — around 78 per cent of PFOS is sludge-associated at typical MLSS — so land application returns them directly to the soil and potentially to groundwater. Incineration at high temperature is the route that should be assumed for planning purposes.

Sources & further reading