How to Remove PFAS from Drinking and Wastewater: A Practical Guide for Engineers

If you are responsible for treating municipal or industrial supplies and need to know how to remove pfas from water, this guide cuts through vendor claims and gives practical, defensible options. It presents a decision framework tying PFAS chemistry and influent concentrations to proven treatment trains, including granular activated carbon, anion exchange, and membranes, and supplies specific design metrics, pilot protocols, and sampling requirements. Where destruction technologies remain experimental, the guide spells out readiness levels, concentrate and waste handling options, and the cost and regulatory trade-offs you must factor into procurement and operations.

Regulatory and site assessment to set design targets

Set design targets from the regulatory floor and the site roofline. Start by converting the applicable numeric guidance or limits into actionable targets that reference an analytical method and a reporting limit. Without that conversion you will design to an ambiguous goal and face permit or community disputes later. Use EPA PFAS resources and state primacy guidance as the baseline for what regulators expect, and use ITRC PFAS guidance to interpret likely compound profiles based on source history.

Steps to convert regulatory guidance and site data into a design target

  1. Document applicable numeric criteria. Identify federal advisories, state MCLs or guidance values, and the exact analyte list and MDLs you will use. Reference the analytical method up front – for finished water that is typically EPA Method 537.1.
  2. Do a targeted source and speciation screen. Map potential contributors – AFFF use areas, industrial dischargers, landfill leachate – and sample for the dominant PFAS species rather than only total PFAS. Species shape treatment choice.
  3. Translate into a treated water objective with a measurement limit. Specify a numeric goal tied to a method reporting limit and a monitoring frequency. For procurement and permitting, treated goal must read as treated concentration less than X ng L by Method Y with MDL Z ng L.
  4. Balance performance margin and cost. Pick a safety margin based on treatment variability and pilot results. A conservative margin reduces risk but increases capital and OPEX and may force more complex technologies.
  5. Prepare regulator engagement materials. Package site data, proposed treated target, monitoring plan, and pilot test scope for pre submission meetings with the state primacy agency or regulator to avoid surprises during permitting. Use your regulatory updates research to anticipate state level expectations.

Tradeoff to accept early. Designing to an ultralow target below routine lab MDLs imposes two hard penalties: more expensive laboratory methods or lower reporting limits through larger volumes and stricter QA QC, and treatment systems sized to remove a broader mix of short chain PFAS that drive concentrate management complexity. In practice the budget and waste handling pathway often dictate whether you aim for a conservative goal or a pragmatic one aligned with available disposal routes.

Concrete Example: A 1 MGD municipal intake downstream of an AFFF site returned PFOS at 120 ng L in source water. The state advisory was 20 ng L but not yet an MCL. The utility set a treated target of 5 ng L tied to Method 537.1 to provide regulatory margin and headroom for future limits. That decision drove a GAC pilot with extended EBCT and increased media inventory estimates, and it triggered early discussions about spent media disposal logistics with the county landfill and an incinerator vendor.

Judgment that matters in practice. Regulators will expect a defensible basis for any chosen target – raw data, method selection, and pilot evidence matter more than optimistic vendor removal claims. Do not treat regulatory engagement as a checkbox. Early alignment with the regulator on analytes, methods, and acceptable monitoring will prevent rework in design and procurement and reduce political risk during construction.

Start every design by naming the analytic method, the reporting limit, and the treated water numeric goal. Those three items determine technology suitability, pilot scale, sampling burden, and costs.

Key takeaway – A defensible design target is not a single number. It is the numeric goal tied to an analytical method, an expected compound profile from site screening, and a documented plan for monitoring and waste handling that regulators can review.

Sampling and laboratory analysis best practices

Analytical procedure quality is the bottleneck for any PFAS response: poor sampling or the wrong method produces defenseless data. Select the analytical method that matches the matrix and the decision you need to make: use EPA Method 537.1 for finished drinking water and EPA Method 533 when you are sampling source water, groundwater, or nonpotable matrices. Expect differences in the analyte lists, typical MDLs, and recommended sample volumes between those methods; design your sampling plan around the method, not the other way round.

Field controls and contamination avoidance

  • Do not use PTFE or other fluoropolymer materials for sample bottles, tubing, syringes, or faucet adaptors; these are common contamination sources.
  • Use pre-cleaned polypropylene or high-density polyethylene bottles supplied by the lab and confirm lot cleanliness before deployment.
  • Include field and trip blanks, equipment blanks, and field duplicates on every sampling event; these are not optional for defensible PFAS programs.
  • Avoid personal care products and PFAS-treated fabrics while sampling — switch to cotton coveralls and document any potential contamination events in the field log.

Tradeoff to accept: larger sample volumes and stricter chain-of-custody increase lab costs and logistics but lower reporting limits, which you will need if your treated target approaches state advisories. If you design for ultralow detection, budget for the extra sample handling, shipping, and lab QA QC.

Picking a laboratory matters more than price. Prioritize labs with documented experience running the specific EPA method you specified, with transparent MDLs, participating in interlaboratory comparison studies, and with routine reporting of lab blank results, surrogate recoveries, and matrix spike performance. Good labs will provide a recent example report for a similar matrix on request.

Concrete Example: A 25,000-person rural utility sampled three production wells near a closed landfill using Method 533, 1 L polypropylene bottles, and a field blank per well. One well returned PFHxA at 15 ng L; the lab blank showed 0.7 ng L and surrogate recoveries were within acceptance. The utility split the next sample with a second accredited lab to confirm the result, then started a monthly confirmation program while planning a GAC pilot targeted at short chain PFAS.

Practical judgment: never accept single-point detections near a lab MDL as conclusive. Require replicate or split samples, check blanks and surrogate recoveries, and treat nondetects with their reporting limits when doing mass balances or treatment performance calculations.

Key action: specify method, required MDLs, sample volumes, and QA QC checklist in the sampling work order. Make those contract items in vendor and lab SOWs so data are legally and technically defensible.

Next consideration: build split-sample and blank requirements into your pilot test RFP and budget for confirmation sampling after pilot completion. That additional step is where projects avoid costly rework and where regulatory confidence in your data is earned.

Treatment technology decision framework

Make the treatment decision around chemistry and disposal, not vendor removal percentages. Choose a technology because its removal mechanism matches the dominant PFAS in your water and because you have a defensible path for handling the resulting concentrate or spent media.

Simple decision logic

Source profile Primary constraint Preferred approach Practical tradeoff
Long chain PFAS dominant (PFOS, PFOA) at low to moderate ng/L Limited waste management options GAC with lead-lag or expanded EBCT Lower capital, predictable changeout schedule, but poor short chain performance
Mixed or short chain PFAS dominant (PFHxA, PFBS) Tight treated water goal or ultralow MDL Anion exchange resin or engineered adsorbents High selectivity, resin regeneration logistics, variable capacity with organics
Very low influent but zero tolerance policy or high spec product water Available concentrate disposal and high OPEX acceptable RO or NF membrane system Highest removal across species, generates brine requiring specialist disposal

Practical insight: media selection must follow pilot breakthrough curves under representative DOC, hardness, and PFAS speciation. Lab bench data are fine for screening; full pilot runs show real capacity loss from natural organic matter and competing anions.

  • Check disposal first: If you cannot place spent media and brine with an approved receiver, do not design for RO scale concentrate volumes.
  • Design to the compound mix: short chain compounds often force you toward IX or membranes; GAC will underperform unless you increase EBCT dramatically.
  • Include realistic fouling factors: size contactors and vessels to allow backwash, prefiltration, or pretreatment identified in pilot studies.

Concrete Example: A suburban utility with 0.75 MGD inflow measured PFHxA and PFBS at 60 ng L combined and had no local incineration option. The supplier recommended RO, but the utility instead ran a 90 day pilot using a selective anion exchange and a downstream GAC polish. The IX delivered steady removal to below the target for 10 000 bed volumes before regeneration; the utility arranged vendor regen and offsite thermal treatment for the spent regenerant, avoiding a high capital RO install and a permanent brine disposal contract.

Judgment that matters: pick the simplest train that meets your site constraints. In practice a two stage solution – selective IX followed by GAC polish – often outperforms single technology claims because it reduces media replacement frequency and narrows concentrate handling needs.

Do not accept a single percent removal figure without the water quality context and a pilot plan. Require vendors to supply pilot protocols tied to your DOC, hardness, and target analyte list.

Key action: document your disposal pathway and include it as a pass fail item in procurement. If you cannot show an offsite thermal or permitted landfill destination for wastes, scale and selection will change.

Detailed design guidance and pilot testing protocols

Start pilot design from the worst operational conditions you expect, not the average. Pilots that run under calm, low-DOC periods give false confidence. Specify the seasonal high-DOC window, peak turbidity events, and any upstream operational changes as part of the test envelope so the pilot stresses fouling, competing anions, and real PFAS speciation.

Key design parameters to set before field work

Define these parameters in the scope of work: influent PFAS speciation list and concentrations tied to an analytical method (e.g., EPA Method 537.1 for finished water), target treated concentrations, pilot flow as a percent of full scale, prefiltration and pretreatment trains, and sampling QA QC. For media systems give planned EBCT, bed depth, and expected bed volumes to breakthrough; for IX list resin type and expected regeneration chemistry; for RO state recovery and concentrate percent.

Practical tradeoff: longer EBCT or larger media volume reduces risk of early breakthrough but multiplies pilot logistics and media handling costs. Pilots should therefore demonstrate diminishing returns near your practical budget limit rather than chase theoretical removal curves.

A compact pilot protocol (work order checklist)

  1. Objective and acceptance criteria: state numeric treated goal tied to method and an acceptance rule (for example, treated PFAS less than X ng L by Method 537.1 for 95% of samples over the test window).
  2. Representative feed: run the pilot on raw source water during at least one expected high-load condition and confirm with split samples to an independent lab (EPA PFAS resources).
  3. Duration and sampling cadence: run until you observe a meaningful portion of the breakthrough curve; typical practical pilots last 6–16 weeks depending on loading — sample at higher frequency near expected breakthrough.
  4. Operational envelope tests: include at least one planned perturbation such as a DOC spike, backwash frequency change, or a resin regeneration cycle to see recovery behavior.
  5. Data and QA QC: require field blanks, lab blanks, surrogate recovery reports, and a data package with raw chromatograms and method detection limits.

How to scale pilot results to full scale. Use observed bed volumes to breakthrough but apply a safety factor for variability and aging; in practice design full scale contactors to deliver 1.2–1.6 times the pilot-derived capacity depending on source variability and regulatory conservatism. For RO, translate pilot recovery and fouling rate into projected annual downtime and cleaning chemical demand rather than just percent rejection.

Common failure modes and what to watch for. Bench spike tests overestimate capacity; high DOC and high sulfate/chloride loads reduce adsorption capacity and resin selectivity; incomplete sampling plans hide short-chain PFAS. Insist on native compound measurements and matrix spikes — not only fortified samples — during the pilot.

Concrete Example: A 0.5 MGD utility ran an 8-week skid pilot on its raw intake during spring runoff. The pilot combined a selective anion exchange ahead of a GAC polish. After a heavy rain event DOC doubled and the IX exhausted faster than vendor estimates. The utility adjusted full-scale media inventory upward and added a simple coagulation-floc pretreatment to protect the media, avoiding premature replacements in year one.

Minimum acceptance items for a pilot report: influent/effluent time series for key PFAS by method, raw chromatograms, breakthrough bed volumes, fouling and pressure data, regeneration mass balance, QA QC summary, and a recommended scale-up factor with justification.

Next consideration: build contractual language into pilot RFPs that ties vendor payment to delivering the full data package and an agreed interpretation template so the results are defensible for permitting and procurement. See our pilot testing resources for templates at pilot test guides.

Concentrate, regeneration handling, and disposal strategies

Straight talk: managing the PFAS-loaded outputs is where most projects fail; choosing a high-removal train without a proven waste pathway simply shifts the compliance problem downstream. Treatment selection must be paired with a documented, executable plan for the liquid and solid residues before you sign a purchase order.

What to quantify first

Quantify mass, not only concentration. Calculate annual PFAS mass in each waste stream (mg or g per year) using measured influent concentrations, plant throughput, and the expected percent removal from pilot tests. That mass number determines whether vendor regen, offsite thermal treatment, or a landfill is practical and what permits or contracts you need.

  • Estimate volumes: expected brine or regenerant volumes per week and storage tank sizing needs
  • Characterize chemistry: total PFAS suite, DOC, salts, and co contaminants in the concentrate or spent media
  • Logistics: number of truck shipments per year, required transporter certifications, and staging area capacity

Pathway tradeoffs. Thermal destruction at licensed facilities is currently the most defensible destruction route but comes with high unit cost, transport complexity, and acceptance criteria; secure landfilling is cheaper short term but increases long-term liability and often requires extensive leachate monitoring. Deep injection or industrial receivers may work in some jurisdictions, but acceptability varies and requires early regulator engagement.

  • Onsite regeneration versus disposable media: Onsite regen reduces waste volumes but generates concentrated regenerant that needs treatment; vendor regeneration externalizes processing but requires secure chain-of-custody and pre-approved receivers.
  • Brine minimization tactics: operate RO at lower recovery, use staged trains (IX then GAC), or recover and treat a smaller high-strength stream to reduce disposal frequency at the cost of higher freshwater use or more complex hydraulics.
  • Contract controls: include waste acceptance limits, analytical verification, certificates of destruction, and indemnity language in procurement and disposal contracts.

Practical constraint that shapes design: many utilities underestimate storage and emergency containment needs for concentrated regenerant or brine. Temporary on-site storage tanks require secondary containment sized for worst-case truck loading and spill response plans aligned with state hazardous waste rules. Plan for at least 30 days of storage while awaiting offsite pickup to avoid last-minute operational shutdowns.

Concrete Example: An industrial water plant producing 100 m3/day of RO concentrate averaged 1.2 kg PFAS/year after pilot scaling. The owner rejected continuous offsite shipments because of transport risk and instead contracted a permitted thermal processor. They installed a 45 m3 lined storage tank, staged weekly shipments, and required the processor to provide post-treatment mass balance reports and a certificate of destruction before final payment.

What people get wrong: vendors often quote media regeneration cycles without the full accounting for the secondary wastes produced during regen. That regenerant frequently contains higher PFAS mass per liter than the original feed and can be classified as hazardous; treat vendor regen claims as an operational hypothesis to be validated by pilot mass balances and a written acceptance agreement with the recycler or incinerator.

Practical rule: before final design, lock down an actual waste receiver and a written acceptance procedure. No receiver, no procurement.

Key action: require a pilot-derived waste characterization report that lists PFAS mass, co-contaminants, expected shipment frequency, and an identified TSDF or thermal processor with a signed acceptance letter. Use that package in permitting discussions and procurement files.

For more on regulatory expectations and how to document your disposal pathway, see EPA PFAS resources and include the waste handling plan in your pilot and procurement SOWs—templates are available in our pilot testing guides.

Operations, monitoring, maintenance, and lifecycle cost considerations

Operational reality: long-run costs and regulatory risk are driven by how you monitor and operate the train, not the vendor removal claim. Build operations and monitoring into design: specify routine sample points, response triggers, and who signs off on a media change or regeneration event before you buy equipment.

Monitoring and trigger logic

Monitoring cadence: start with weekly treated-water samples for the first 6 months after commissioning, move to biweekly or monthly only once results and trend analysis show stability. Always tie each sample to an analytical method (for finished water typically EPA Method 537.1) and require field blanks and surrogate recoveries on every event.

Action triggers: define objective, measurable triggers in the O&M manual — for example, if treated PFAS concentration rises above 50 percent of the treated goal for three consecutive samples, escalate to increased sampling and a media integrity check; if a single sample exceeds the treated goal, place affected production on hold and perform split-sample verification with an independent lab.

Practical limitation: online surrogates such as TOC or conductivity can warn of feed changes but do not reliably predict PFAS breakthrough. Use them for operational housekeeping and early warning, not as a substitute for PFAS lab data.

Maintenance, spare parts, and common failure modes

Component Routine task Typical frequency Critical spares
GAC contactor Backwash, monitor headloss, bed depth inspection, media replacement planning Backwash weekly–monthly; media inventory review quarterly Backwash valves, media samples, spare GAC bags for emergency top-up
Anion exchange resin Monitor effluent speciation, scheduled regeneration or resin swap, fouling checks Regeneration or swap based on bed volumes seen in pilot; resin integrity check semiannually Resin cartridges or replacement resin, regeneration chemicals, pump rebuild kits
RO/NF system CIP, pressure and leak checks, membrane integrity tests CIP frequency driven by fouling; typically monthly to quarterly Membrane elements (spare set for 1–3 months), seals, high-pressure pump parts

Common operational pitfalls: underestimating DOC impacts on media capacity, neglecting secondary containment for spent regenerant storage, and insufficient spare membrane elements that force emergency downtime. Address these in spare parts and contingency budgets rather than hoping for vendor responsiveness during a crisis.

Lifecycle cost framework and tradeoffs

Cost drivers to budget for: initial capital for vessels and control skids, recurring media or membrane replacement, analytical lab fees for operational monitoring, disposal or thermal destruction of wastes, energy and chemical use (RO), and labor for regeneration and sample handling. Which driver dominates depends on your train: adsorption systems often have high media and disposal costs; RO shifts cost to energy, membranes, and brine handling.

Inspection point: before procurement, run a 10-year cashflow with at least three OPEX scenarios (low, mid, high fouling) and include explicit line items for analytical monitoring and waste shipments. Projects that skip this step under-budget the program by 20–40 percent in our experience.

Operational example: a 3.2 MGD community plant installed a selective anion exchange ahead of a GAC polish. After commissioning they increased treated-water sampling to weekly for three months when seasonal runoff elevated DOC. That early monitoring showed accelerated resin exhaustion; the utility negotiated a vendor-regeneration schedule and added a coagulation pretreatment to extend resin life and reduce annual disposal tonnage.

Takeaway: lock monitoring frequencies, response triggers, spare parts, and the waste receiver into procurement documents. Operational discipline and an enforced escalation ladder are what keep PFAS programs affordable and defensible over their lifecycle.

Case studies and practical procurement language

Plain fact: procurement language determines whether a PFAS project delivers clean water or a liability. Vague RFPs that ask for generic percent removals or promise compliance without defining the analytical method, reporting limits, and waste handling responsibilities invite disputes and change orders. Write metrics, deliverables, and waste pathways into the contract up front.

What to require from the vendor: Specify the analytical method (for finished water typically EPA Method 537.1), the minimum detectable limits the lab must meet, and that QA QC packages accompany every performance report. Require full chromatograms, blank results, surrogate recoveries, and split-sample verification with an independent lab as a condition for milestone payments.

Copy-ready procurement items to include in an RFP or contract

  • Performance milestone: Payment tied to meeting a numeric treated-water concentration by Method 537.1 for a defined fraction of samples (for example, >= 95% of weekly samples over a 12-week acceptance window).
  • Data deliverable: Vendor must deliver raw chromatograms, MDL calculations, and a QA QC summary within 14 days of each sample event; all data uploaded to a shared project portal.
  • Waste handling: Vendor to provide a signed acceptance letter from the intended TSDF or thermal processor before final award; if vendor regenerates onsite, they must document storage, secondary containment, and a disposal manifest trail.
  • Liability and remedies: Liquidated damages or retainage for missed milestones; explicit requirements for certificates of destruction or a mutually agreed remediation plan if post-construction monitoring shows exceedances.

Practical tradeoff: Strong performance guarantees reduce your operational risk but raise vendor pricing and reduce market competition. If your waste receiver is limited, prefer a procurement structure that shifts disposal responsibility to the vendor (with proof of receiver acceptance) rather than attempting a low-cost purchase with no downstream solution.

Concrete Example: A 1.5 MGD utility made final payment contingent on three consecutive weekly effluent samples below the treated target by EPA Method 537.1, verified by a second accredited lab. The vendor missed the first round; the contract retained 10 percent until the vendor completed additional optimization and provided a certificate of destruction for spent regenerant. That retainage forced timely corrective action and avoided extended operations with substandard performance.

Judgment that matters: Do not accept vendor-provided lab data as the sole basis for acceptance. Independent verification is not optional; it is the difference between defensible regulatory filings and later re-testing demands. Also, treat waste acceptance letters as pass/fail procurement items — if a processor will not sign, the proposal should be nonresponsive.

Include an explicit acceptance path for wastes and a third-party lab verification clause in your RFP — these two items reduce project cost escalation more than any media discount.

Procurement quick checklist: 1) Define method and MDLs; 2) Milestone-based payments tied to verified samples; 3) Data package contents and turnaround; 4) Signed waste receiver acceptance; 5) Holdback/retention and remediation steps; 6) Indemnity for off-spec events.

If you want templates and pilot RFP language, include a reference to your pilot scope and QA requirements rather than reusing vendor boilerplate. See our pilot resources for templates and examples at pilot test guides and check regulatory expectations on monitoring and methods at EPA PFAS resources.