Quick answer: Sizing a PFAS treatment train from vendor curves and published isotherms is engineering — the outcome is determinable in advance. A core R&D activity may exist in the narrower case where no established basis predicts whether a separation-plus-destruction train can reach a defined performance target on a specific site matrix, because co-contaminants, the short-chain fraction and the non-targeted organofluorine pool govern it, and the answer comes only from a systematic progression of work conducted to generate new knowledge. The company must assess each activity against the statutory tests: uncertainty about site performance does not by itself make treatment-train selection a core R&D activity.
24 August 2026 — this article describes the current rules. The 2026-27 Federal Budget announced R&DTI changes; the ATO states the measure is not yet law, and industry.gov.au states the changes will apply to income years starting on or after 1 July 2028.
Two things about per- and poly-fluoroalkyl substances make treatment unusual. They are not readily destroyed by the processes water treatment normally relies on, so almost every practical train separates and concentrates first and destroys second. And the analytical method most projects use measures a list of about thirty compounds, while the material actually in the water is a much larger and largely uncharacterised mixture. Those two facts together are why a train that performs beautifully on spiked laboratory water can miss its target on a site matrix, and why the question of whether the outcome was knowable in advance is a real question rather than a formality.
A boundary this article does not cross
Nothing here endorses the efficacy or safety of any technology, and nothing here is guidance on clean-up criteria, discharge limits, exposure or human-health risk. Assessment against regulatory standards and human-health risk assessment are matters for accredited environmental and health professionals working under the applicable jurisdictional framework — the PFAS National Environmental Management Plan and the relevant state regulator — not for this article. What follows is about one narrow question: whether a particular technical outcome was genuinely not knowable in advance.
The Statutory Test, Stated Once
Eligibility under the R&D Tax Incentive is assessed activity by activity, not project by project. Under s 355-25(1) of the Income Tax Assessment Act 1997, a core R&D activity is an experimental activity that satisfies the statutory requirements. These include that its outcome cannot be known or determined in advance on the basis of current knowledge, information or experience, but can only be determined by applying a systematic progression of work based on principles of established science, proceeding from hypothesis to experiment, observation and evaluation, and leading to logical conclusions. The activity must also be conducted for the purpose of generating new knowledge, including new knowledge in the form of new or improved materials, products, devices, processes or services (business.gov.au; ITAA 1997).
Two elements of that definition do the work in contamination projects, and they are worth naming rather than paraphrasing:
Grounded in established science
Interfacial adsorption thermodynamics, sorption equilibria and mass transfer, electrochemical and radical kinetics, and analytical mass balance are settled bodies of principle — but the requirement is that the particular progression rests on them, not that the field possesses them. A run-it-and-see campaign of plant trials does not satisfy it however carefully the results are logged.
Outcome genuinely indeterminable in advance
This is where most PFAS treatment work falls out: the technology is established, the vendor has design curves, and the matrix is characterised.
Why a Site Matrix Breaks Predictions Made on Clean Water
Three mechanisms do most of the damage, and each is measurable:
Sorption capacity is set by competition, not by the PFAS concentration: Granular activated carbon retains long-chain perfluoroalkyl acids largely through hydrophobic partitioning of the fluorinated tail. Dissolved organic carbon competes for the same pore space and hinders transport into it, so a Freundlich capacity fitted on reagent water spiked with a single analyte can substantially overstate field bed life. By how much is a property of the particular water rather than a constant — which is precisely why it has to be measured on that water. Single-use strong-base anion exchange behaves differently but has the same problem in a different currency: sulfate, nitrate and bicarbonate compete for the fixed charge, and in a brackish groundwater those competitors are present in orders of magnitude more mass than the target.
The short-chain fraction is a thermodynamic problem, not a contact-time problem: Perfluorobutanoic acid and perfluorobutane sulfonic acid have short perfluoroalkyl chains, high aqueous solubility and low interfacial activity. They break through carbon early and they foam poorly. Because the limitation is affinity rather than kinetics, longer empty-bed contact time and taller columns buy far less than the equivalent change would buy for a C8 species. Sites with a long history of aqueous film-forming foam use can be short-chain-rich, because in-ground transformation of precursors over decades yields terminal perfluoroalkyl acids, the short-chain ones among them.
The material is bigger than the analyte list: Targeted liquid chromatography–tandem mass spectrometry quantifies the compounds on the method's list and nothing else. The total oxidisable precursor assay oxidises precursors to terminal perfluoroalkyl acids, and the uplift it produces indicates that an oxidisable pool exists outside the targeted list — it is an indicator of that pool, not a direct measurement of its mass or composition. Combustion ion chromatography goes further and measures organofluorine as fluorine. The consequence for a destruction step is decisive: a 99% fall in the targeted sum is equally consistent with mineralisation, with chain-shortening into compounds not on the list, with sorption onto reactor internals, and with volatilisation. Only a fluorine balance — fluoride released against organofluorine fed — distinguishes them.
The national assessment framework is set out in the PFAS National Environmental Management Plan, with state guidance published by EPA South Australia; Australia's National Measurement Institute is among the accredited providers of this class of PFAS measurement (industry.gov.au).
Much PFAS treatment work sits outside all of this: a characterised low-organic-carbon feed, a long-chain-dominated profile and a vendor-supported medium is a design calculation with a determinable outcome. See what does not qualify.
A Worked Example (Hypothetical and Illustrative Only)
Invented to show where the boundary falls. It is not a real project, the numbers are illustrative, and it says nothing about whether any actual claim would be accepted or whether any technology is fit for use.
A contractor at a former fire-training area in regional South Australia is developing an on-site train for impacted groundwater. Before any trial, the feed was characterised: summed 30 targeted PFAS 41 µg/L, of which the long-chain subset is 28 µg/L, the short-chain subset 11 µg/L and the remaining 2 µg/L listed fluorotelomer precursors that fall into neither; TOP-assay sum 96 µg/L, a 2.3-fold apparent uplift over the targeted sum, indicating an oxidisable pool the targeted list does not see rather than measuring its inventory; dissolved organic carbon 26 mg/L; sulfate 310 mg/L — roughly four orders of magnitude more mass than the entire targeted PFAS sum of 0.041 mg/L; chloride 480 mg/L; total recoverable hydrocarbons C10–C16 at 1.9 mg/L.
The target, recorded before the first campaign: These are process-performance criteria set by the project, not regulatory criteria. (1) Residual load: the material leaving the treated-water line — the concentrate as presented to the destruction step, plus any spent medium — no more than 0.5% of throughput by volume, with spent solid media taken at packed bulk density; and of that, no more than 0.05% of throughput consigned off site, since the point of the train is on-site destruction. (2) At least 90% removal of the summed 30 targeted analytes from the treated stream. (3) At least 90% removal of the short-chain subset specifically. (4) On the concentrate, a defluorination ratio of at least 0.85 — moles of fluoride released, by ion chromatography, per mole of fluorine fed as organofluorine, by combustion ion chromatography — and a post-treatment TOP-assay sum no more than 10% of the targeted sum fed to the step, both expressed as mass per batch so that dilution cannot flatter the result.
Held and varied: Held: each campaign drawn from one homogenised 20 m³ batch, 18 ± 2 °C, pH unadjusted at 7.4 except where pH was the variable, the same isotope-dilution LC-MS/MS method with mass-labelled surrogates, the same laboratory and the same 30-analyte plus TOP plus organofluorine suite. Varied: superficial gas velocity 0.4–1.6 cm/s, sparger pore size 10/40/100 µm, column height 1.5/3.0 m, calcium dose 0/60 mg/L, polishing medium and empty-bed contact time 4/8/12 min, and destruction chemistry and dose.
The approach that failed, and what it ruled out: Run A put carbon first, which is what the bench data supported: an isotherm on reagent water spiked to 40 µg/L predicted about 85,000 bed volumes to 10% PFOS breakthrough. On site water the column reached 10% PFOS breakthrough at 11,300 bed volumes; perfluorobutanoic acid broke through at 900 bed volumes and perfluorobutane sulfonic acid at 2,100. Doubling contact time from 4 to 8 minutes moved short-chain breakthrough only from 900 to 1,400 bed volumes. A parallel column on feed pre-treated to 4 mg/L dissolved organic carbon restored PFOS breakthrough to about 46,000 bed volumes and left the short-chain result essentially unchanged. That pair of results ruled out two hypotheses at once: mass transfer is not the cause of short-chain breakthrough, so no contactor-geometry change would fix it; and the organic-carbon competition effect is a long-chain capacity problem separable from the short-chain affinity problem.
The sequence that followed: Run B placed foam fractionation first. At 0.8 cm/s through a 40 µm sparger the column removed 96% of PFOS and PFHxS at an enrichment factor near 1,100, with foamate at 0.09% of feed volume — the two figures are the same measurement seen twice, since 96% of the mass reporting to 1/1,111 of the volume is an enrichment of about 1,070. The operable window was narrow and the hydrocarbon co-surfactant load set it, not the PFAS: above 1.1 cm/s the foam collapsed, below 0.5 cm/s liquid carryover destroyed the enrichment. Short-chain subset removal was 24%, the C4 species sitting near 20% and perfluorohexanoic acid, the most interfacially active member of the subset, at 44%. Dosing 60 mg/L of calcium lifted perfluorohexanoic acid removal to 71% and moved perfluorobutanoic acid only from 21% to 24%, taking the subset as a whole to about 30% — still nowhere near criterion (3). Divalent-cation bridging at the air–water interface was the working hypothesis for that effect, but the campaign varied calcium dose without separating the ionic-strength contribution, so the result is consistent with bridging rather than evidence for it, and it was recorded that way. Run C added single-use anion exchange as a short-chain polisher downstream and reached 94% short-chain removal to 6,800 bed volumes; a parallel column on sulfate-reduced feed ran to 19,000 bed volumes, locating the limit in sulfate competition rather than in the resin chemistry.
The destruction step, and what "destruction" turned out to mean: The foamate carried a targeted sum near 34 mg/L: about 31 of the 41 µg/L in the feed reported into 0.09% of the volume, an overall enrichment near 840 even though the long-chain species alone enriched by about 1,100. It also carried the feed's background chloride essentially unchanged at 480 mg/L, chloride being of no interfacial activity. Electrochemical oxidation at 20 mA/cm² dropped the targeted sum by 99.2% in six hours. Fluoride release accounted for 31% of the fluorine fed as organofluorine, and the TOP assay on the treated foamate returned 7.4 mg/L — 22% of the targeted sum fed, against the 10% criterion (4) allows. On a targeted-analyte report that run looks like destruction; on the fluorine balance it is mostly chain-shortening. Raising current density to 40 mA/cm² over 24 hours lifted the defluorination ratio to 0.79, still short of 0.85, and drove perchlorate formation from that 480 mg/L of chloride, which became the binding constraint. Ultraviolet-sulfite reduction at pH 10.5 reached a defluorination ratio of 0.88 in eight hours, with a post-treatment TOP-assay sum of 0.48 mg/L against 6.9 mg/L of targeted sum fed — 7%, inside criterion (4) — but only after the foamate was diluted fivefold, because its concentrated organic load, carried up with the hydrocarbon co-surfactant, scavenged hydrated electrons at full strength.
Result, and a candidate boundary: The train meeting all four criteria was foam fractionation, then anion-exchange polishing, then ultraviolet-sulfite reduction on fivefold-diluted foamate. Criterion (1) closes on this accounting, per 1,000 m³ treated: 900 L of foamate at 0.09% of throughput, diluted fivefold to 4,500 L presented to the destruction step, plus 147 L of spent resin (1,000 m³ ÷ 6,800 bed volumes) — 4,647 L in all, or 0.46% against the 0.5% ceiling, of which only the 147 L of resin, 0.015%, leaves the site. The dilution the destruction chemistry demanded consumed most of the margin the fractionation column had won. Criterion (2) closes at about 98% on the summed 30 analytes, and criterion (3) at 94%. Two pieces of knowledge came out that the bench work did not contain: the operable gas-velocity window is set by the hydrocarbon co-surfactant load, and the defluorination ratio rather than targeted-analyte removal is what separates transformation from destruction on this matrix. A defensible candidate boundary for this hypothetical would begin when the hypothesis, the four criteria and the analytical protocol were recorded and the first campaign was set up, and end when the criteria were met on this matrix or shown unattainable on it. Items sitting outside that candidate boundary — the delineation and validation sampling programme, the regulator-facing reporting and site-audit process, procurement and installation of the full-scale plant, and routine operation afterwards — would each need their own supporting-activity analysis on its own facts, rather than being swept in or out as a group. On these facts the campaign is a candidate core activity, subject to the full facts — not an eligibility outcome, which the company self-assesses.
Where the Exclusions and the Supporting-Activity Test Bite
Section 355-25(2) lists activities that are not core R&D activities, and paragraph (f) — activities associated with complying with statutory requirements or standards, including routine testing and analysis of materials, components, products, processes, soils, atmospheres and other things — may be relevant to certain contaminated-site activities where those activities are undertaken for compliance purposes. Delineation sampling, validation sampling against site criteria, waste-classification testing and monitoring required by a regulator are generally unlikely to be core R&D activities on those facts, subject to each activity's own facts and the statutory tests. That is not because their results are predictable; a validation round can fail. It is because the activity is undertaken to demonstrate compliance.
Exclusion from core does not itself establish whether an activity is a supporting R&D activity; that requires separate application of s 355-30 to the activity's purpose and its relationship to the candidate core activity. A supporting R&D activity must be directly related to core R&D activities (s 355-30(1)), and where an activity is of a kind referred to in s 355-25(2), or produces goods or services, or is directly related to producing goods or services, s 355-30(2) allows it as a supporting R&D activity only if it is undertaken for the dominant purpose of supporting core R&D activities (business.gov.au). The third of those limbs is the one that reaches furthest into a remediation contract, because a plant treating groundwater is delivering a service. Each limb is tested against the particular activity.
Where an RSP Fits
An RSP is a scientific or technical service provider registered in specific research fields that a company may engage to provide R&D services within those registered fields. (business.gov.au). Ignition Research is registered in Pollution and Contamination (ANZSRC 4105) among other fields; our work on a train like the one above is upstream of the plant — framing which quantity is actually unknown, designing the campaign so a failed run eliminates a hypothesis rather than merely disappointing, and specifying the analytical suite to the resolution the criteria require, which for destruction means a fluorine balance and not a targeted-analyte report alone. See pollution and contamination research. There is also a point worth knowing for smaller programmes: Eligible R&D expenditure incurred to a registered research service provider may be treated differently where the usual $20,000 R&D expenditure threshold is not met, subject to the requirements in s 355-100(2) and the underlying R&D activities satisfying the eligibility tests. In mechanism, where a company's total notional R&D deductions for an income year come to less than $20,000, s 355-100(2) does not shut the offset off — it substitutes a different base, working the offset out on the kinds of expenditure that subsection lists, the first of which is expenditure incurred to a registered research service provider that is not an associate of the R&D entity, for services within a research field for which the provider is registered. The offset tiers are covered in our article on the refundable and non-refundable offset.
Talk to Ignition Research before the first treatability campaign is run, while the experimental design and the analytical protocol can still be settled against the question they are meant to answer. As a Registered Research Service Provider at Lot Fourteen in Adelaide, we design and conduct the experimental programme and produce the technical record while the work is happening. We are not a registered tax agent, and we are not an environmental auditor: your company self-assesses and remains responsible for its own claim, and assessment against regulatory criteria remains with your accredited environmental professionals. Get in touch.
Frequently Asked Questions
Q: Is PFAS treatment design an eligible R&D activity?
A: Usually not, as a core R&D activity. Sizing an established medium or unit process for a characterised feed using vendor design curves and published isotherms has an outcome a competent process engineer can determine in advance. A core activity may exist in the narrower case where no established basis predicts whether a train can reach a defined performance target on a specific site matrix — typically because co-contaminants, the short-chain fraction or the uncharacterised organofluorine pool govern the result — and that can only be determined by a systematic progression of work conducted to generate new knowledge. The company must assess each activity against the statutory tests; uncertainty about site performance does not by itself make treatment-train selection a core R&D activity.
Q: Why do short-chain PFAS break through activated carbon early?
A: Because the limitation is affinity rather than kinetics. Retention on carbon depends largely on hydrophobic partitioning of the perfluoroalkyl tail; short-chain species such as perfluorobutanoic acid and perfluorobutane sulfonic acid have shorter tails, higher aqueous solubility and much weaker partitioning, so they are displaced from the bed long before the C8 species. Increasing empty-bed contact time or column height addresses mass transfer, so it moves short-chain breakthrough far less than the same change moves long-chain breakthrough.
Q: What does defluorination ratio mean for PFAS destruction?
A: It is the moles of fluoride ion released by a destruction step divided by the moles of fluorine fed to it as organofluorine, the first measured by ion chromatography and the second by combustion ion chromatography. It matters because a targeted method quantifies only the compounds on its list, so a large fall in the targeted sum is equally consistent with mineralisation, with chain-shortening into compounds not on the list, with sorption onto reactor surfaces and with volatilisation. A fluorine balance, usually read alongside a post-treatment total oxidisable precursor assay, is what distinguishes those outcomes.
Q: Is contaminated site validation sampling a core R&D activity?
A: Generally unlikely, subject to the activity's own facts and the statutory tests. Section 355-25(2)(f) excludes certain activities associated with complying with statutory requirements or standards from being core R&D activities, including routine testing and analysis of materials, processes and soils. That does not turn on the result being predictable — a validation round can fail — but on the activity being undertaken to demonstrate compliance. Whether such an activity is instead a supporting R&D activity is a separate question under s 355-30, and where s 355-30(2) is engaged the dominant-purpose test has to be satisfied.
Sources & Further Reading
legislation.gov.au — Income Tax Assessment Act 1997 — Div 355, incl. ss 355-25, 355-30 and 355-100
EPA South Australia — Per- and poly-fluoroalkyl substances (PFAS)
industry.gov.au — National Measurement Institute: PFAS testing services
Related: pollution and contamination research · what does not qualify · what an RSP is · claiming R&D under $20,000 · refundable vs non-refundable offset · R&D Tax Incentive in Adelaide · more Insights
This article is general information from a Registered Research Service Provider about the R&D Tax Incentive. It is not tax, legal or financial advice; eligibility depends on your circumstances and you should self-assess and seek your own advice.
Thinking about a project like this?
If you're weighing up an AI, software or technical improvement project and can't tell yet whether it's implementation or research, start with a quick read on where it sits.