Engineering analysis
Destroying PFAS in Saline RO Reject: Why Whole-Brine SCWO and BDD Fail and Concentrate-Then-HALT Wins
None of whole-brine SCWO, raw-brine boron-doped diamond (BDD), or foam fractionation alone hits PFAS destruction targets on saline RO reject — each is disqualified by the salt matrix. The viable architecture is concentrate-then-destroy: use foam plus PFAS-selective ion exchange (with a hardness-exclusion front-end) to pull ~7.6 kg/day of PFAS out of ~53 t/day of salt, then feed the small concentrate to a subcritical Hydrothermal Alkaline Treatment (HALT) reactor that tolerates dissolved NaCl and fixes fluorine as soluble NaF. Before spending capital, retire the unknowns with ~$150K of testing over one quarter: confirm the bulk-brine disposal path, get a merchant off-site destruction quote for the ~2.8 t/yr of PFAS, and run a HALT treatability test to check that Ca/sulfate hardness does not scale the reactor.
Key takeaways
- Whole-brine SCWO fails physically: crossing the critical point drops NaCl solubility ~1000×, leaving a 3.5 wt% feed 100–350× supersaturated, and liberated fluoride forms HF that corrodes the vessel at ~600 °C.
- Raw-brine BDD fails chemically: 15,000 mg/L chloride becomes a chlorate (~6.3 mol/m²/day) and perchlorate factory against 2–6 µg/L limits, and it cannot defluorinate short-chain/GenX PFAS.
- The real fork is salt management: tolerate the NaCl (subcritical HALT at ~350 °C, ~16.5 MPa keeps dielectric ~15–20 so NaCl stays dissolved >300 g/L), but exclude the inverse-solubility Ca/sulfate hardness that can still scale the reactor.
- The governing figure of merit is $/kg PFAS destroyed, not $/1,000 gal — a whole-brine reactor spends ≥99.9% of its energy and oxidant on salt and ~0.02% on the ~7.6 kg/day of PFAS.
- Concentration is the master lever: it moves the matrix:target ratio from ~30,000:1 molar (whole brine) toward <100:1, a 3+ order-of-magnitude gain no reactor optimization can match.
- Total de-risking spend is ~$150K over one quarter to protect a $10–30M capital decision, with a 4-month gate: proceed to HALT only if concentration factor ≥50× and hardness does not scale the reactor.
PFAS destruction architectures for saline RO reject: performance is nearly tied, salt tolerance is the divider
| Criterion | A1: Separate then SCWO | A2: Separate then BDD | B: Concentrate then HALT |
|---|---|---|---|
| Long + short-chain >99% destruction | Yes (non-selective) | Long yes; short-chain risk | Yes (non-selective) |
| Perchlorate/oxychlorine risk | None | Only if chloride fully removed first | None |
| Salt-management burden | High — needs low-salt and low-volume feed | Highest — must also strip chloride | Low — tolerates residual NaCl |
| Closest real data point | Peterson SFB SCWO (low-salt) | Foam + electro-oxidation pilot | Aquagga/3M HALT on IX regen brine |
| Reactor materials | SiC/TaC/YF3 (exotic) | Electrodes | Ni-alloy if caustic-SCC survives |
| Modeled CAPEX | ~$15–30M | Higher than A1 and B | ~$8–15M |
| Frontier position | Viable hedge | Dominated | Primary candidate |
Frequently asked
Why can't supercritical water oxidation (SCWO) treat whole saline brine directly?
Above water's critical point the dielectric constant collapses (~80 to 2–6), dropping NaCl solubility ~1000× so a 3.5 wt% feed becomes 100–350× supersaturated and plugs the reactor. Liberated fluoride also forms corrosive HF at ~600 °C, and over 99.9% of the energy is spent destroying salt rather than PFAS.
What is HALT and why does it tolerate salt?
Hydrothermal Alkaline Treatment (HALT) is a subcritical process running at roughly 350 °C and 16.5 MPa with excess NaOH. Because it stays below the critical point, water's dielectric constant remains ~15–20, so NaCl stays fully dissolved above 300 g/L and fluorine exits as soluble NaF instead of forming corrosive HF.
Is chloride or hardness the bigger problem for a subcritical PFAS reactor?
Once you drop the supercritical premise, dissolved NaCl stops mattering to the reactor. The species to fear is the Ca/sulfate hardness, the inverse-solubility component that can still scale even a subcritical reactor — which is why a hardness-exclusion front-end and a treatability test are the decisive pivot.
Should PFAS be destroyed on-site or shipped off-site?
The PFAS mass is only ~2.8 t/yr, so a merchant off-site destruction or regeneration quote in $/kg-PFAS may beat any on-site build. Get that quote before committing capital; a one-week phone call can obsolete the entire on-site train.
Does the reactor need to meet the sub-4 ppt PFAS limit?
No. The <4 ppt target is a train-level product-water spec met by RO plus a GAC/IX guard bed, not a reactor spec. The reactor's job is >99% mass destruction of the concentrate; chasing the last two 9s inside the reactor over-sizes it by orders of magnitude.
Why does raw-brine BDD create a new regulatory problem?
The 15,000 mg/L chloride in RO reject is oxidized into chlorate (~6.3 mol/m²/day) and accumulating perchlorate, both regulated against 2–6 µg/L limits, which BDD cannot self-remediate. It also cannot reliably defluorinate short-chain and ether PFAS such as PFBA, PFBS, and GenX.
This answer is distilled from a full Sparlo analysis — the root-cause trade-off, cross-domain paths, and vendor-ready options.
Read the full analysis →