Refinery wastewater’s hidden risk: strip toxics upstream, buffer the bugs with carbon

Refinery process water often arrives loaded with toxics—the kind that wipe out microbes and upend compliance. Hydrogen sulfide (H₂S, a corrosive toxic gas), ammonia (NH₃), phenols, cyanides, and heavy metals such as Ni and V frequently spike from units like crude distillation, catalytic cracking, reforming, and spent caustic generators. “Sour water” condensate from these units is typically the worst for H₂S and NH₃ (climate‑policy‑watcher.org).

The biology’s vulnerabilities are well documented: H₂S is directly poisonous; ammonia in the hundreds of mg/L collapses nitrification; and even a few mg/L of metals like Ni²⁺, Zn²⁺, Cd²⁺, and Pb²⁺ inhibit nitrifiers (mdpi.com). Untreated refinery effluent can show COD (chemical oxygen demand) in the hundreds or thousands mg/L and ammonia in the hundreds mg/L, with oil and phenol levels acutely toxic to the biological process. Upstream removal is essential so those compounds don’t reach the bioreactor.

Sour‑water stripping design and performance

A central control point is sour‑water stripping: steam‑stripping that volatilizes H₂S and NH₃ before the wastewater hits biology. In practice, refinery sour water from CDU, FCC, coker, reforming, and more is fed to a steam‑heated stripper—often two stages in series (pall.com). Strippers routinely remove 85–99% of sulfides by steam distillation (climate‑policy‑watcher.org).

Under typical tray‑column conditions (2–3 bar, ~230–270 °F), ~90% H₂S can be removed even at ~100 °F, while ~90% NH₃ removal needs ≥230 °F (climate‑policy‑watcher.org). A well‑designed system can cut H₂S to near‑zero and NH₃ to roughly 50 mg/L in the effluent (climate‑policy‑watcher.org), and some plants target ~1 mg/L NH₃ as a nutrient (climate‑policy‑watcher.org). Operators sometimes acidify or use a two‑column design to separately strip H₂S and NH₃, achieving ~90%+ removal of each (climate‑policy‑watcher.org).

Side note: the stripper must be protected from oil and fines. Oil contamination causes carryover of excess oil in the stripped water, overloading the WWTP and risking fouling and capacity impacts (pall.com). Upstream oil removal and solids control—via skimming, coalescers, and screening—are critical; plants often deploy primary equipment such as physical separation systems to keep the stripper clean.

Oil and solids removal before biology

Oil and grease (O&G) are first taken out by gravity separators (API), dissolved‑air flotation, or coalescing filters. Removing free oil by flotation is routine in refineries, and O&G carry toxics that inhibit biology if not removed; even coker wastewater skimmed of oil can retain phenolics that later poison nitrifiers. Many facilities standardize on DAF units (dissolved air flotation; “removes 95%+ suspended solids and oils, 1–3 hour detention”) for this step.

Primary separation also captures a meaningful share of bulk load. Oil‑separator units deliver COD removals of 16–45% and TSS (total suspended solids) removals of 33–68% (climate‑policy‑watcher.org). Where dispersed oil is persistent, refineries add dedicated oil removal stages to reduce carryover into downstream units.

Heavy metals and toxic organics pre‑treatment

Metals (Ni, V, Cr, etc.) from certain crudes should be precipitated or sorbed before biotreatment. Chemical dosing—sulfide or hydroxide addition—can drop Ni/V to <1 mg/L; even single‑digit mg/L metals can “completely inhibit” microbial nutrient removal (mdpi.com). Accurate metering through a dosing pump supports stable pH and reagent control in this step.

Hard‑to‑treat organics—phenol, cyanide, thiocyanate—often need specialized oxidation or segregation. A common route is wet‑air oxidation (WAO): the “Zimpro WAO” unit at Sinopec Jiujiang oxidizes sulfide and phenolic content in spent caustic prior to polishing (chemistryviews.org), keeping those highly toxic components out of the activated sludge.

Equalization and pH control

Sour‑water streams swing in pH; neutralizing to ~7.5–8 keeps nitrifiers functional. Equalization basins buffer flow and load and allow residual H₂S to off‑gas when integrated with stripping. Many refineries recirculate effluents through equalization to smooth shocks before the bioreactor. pH control is typically achieved with caustic or acid and controlled dosing via a chemical dosing pump.

PAC‑enhanced bioreactor (PACT) mechanism

Even after upstream removal, dissolved recalcitrants and trace toxics remain. Adding powdered activated carbon (PAC) to the activated‑sludge bioreactor—the PACT process—adsorbs those residuals in situ. PAC provides high surface area and sorptive sites to capture hydrophobic and non‑biodegradable compounds, buffering toxicity and stabilizing the biomass; the reactor effectively becomes an activated carbon‑sludge hybrid. Plants commonly source the carbon through activated carbon supply chains already used for organics polishing.

Lab data show 50–100 mg/L PAC significantly increases removal of chlorinated aromatics; at 100 mg/L PAC, removal of poorly degradable organics such as 1,2‑dichlorobenzene exceeded 90% (nepis.epa.gov). In one experiment, a 100 mg/L PAC unit stably removed ethylbenzene (~79% removal) with no excursions, while the control fluctuated (nepis.epa.gov); steady‑state nitrification occurred in ~21 days with PAC versus ~35 days without (nepis.epa.gov; nepis.epa.gov).

Full‑scale gains and trade‑offs

At refinery scale, net COD improvement from PAC is modest but real. One EPA multi‑scale study reported that adding PAC (80 kg/d to a full‑scale aeration basin) yielded 11–33% lower effluent COD versus control—equating to only ~2–6% higher overall COD removal efficiency (nepis.epa.gov). Effluent oil & grease was marginally lower by ~0.5–2.0 mg/L with PAC (nepis.epa.gov), while phenol and total phosphorus were essentially unchanged (nepis.epa.gov).

PAC’s biggest operational value was better settling and solids handling as carbon‑sludge flocs compacted and settled faster (nepis.epa.gov). Because PAC is fine, effluent suspended solids and NH₃ can rise slightly if not fully captured (nepis.epa.gov), making downstream capture via a clarifier important.

Carbon handling and regeneration

PAC acts by adsorption and is not biologically regenerated; there was “essentially no evidence” of biodegradation of the carbon itself in a classic EPA report (nepis.epa.gov). Spent PAC is continuously removed (clarifier or hydrocyclone) and either disposed or regenerated off‑site.

Many facilities now combine PAC with on‑site regeneration. At Sinopec Jiujiang, a wet‑air oxidation unit (WAR) will destroy organic load on the carbon and recover ~95% for reuse (chemistryviews.org), reducing disposal costs; the WAR recovers most of the carbon and oxidizes the rest with bio‑solids into clean gas. Upstream, the same site uses a Zimpro WAO system to oxidize sulfide and phenolic content in spent caustic before polishing in the PACT system (chemistryviews.org), insulating the activated sludge from highly toxic streams.

Integrated outcomes and operating envelope

With robust pretreatment plus PAC polishing, the WWTP runs steadier. After API skimming and ammonia stripping, the biological influent typically drops to a few hundred mg/L COD (instead of ~1000+) and NH₃ to tens of mg/L. Oil‑separator stages alone can remove 16–45% COD and 33–68% TSS (climate‑policy‑watcher.org). Residual toxins are then buffered by PAC adsorption—100 mg/L PAC drove refractory‑organic removal to >90% in tests (nepis.epa.gov) and gave reproducible effluent for aromatic compounds (nepis.epa.gov).

PACT has delivered near‑tertiary performance in pilots: one EPA pilot achieved ~5 mg/L COD and very low toxins (nepis.epa.gov), with demonstrations meeting stringent limits on BOD, COD, color, and fish‑toxicity better than conventional schemes (nepis.epa.gov).

Economics and decision checkpoints

Cost and risk trade‑offs are case‑specific. Sour‑water strippers cost millions but remove >90% of the problem: 85–99% H₂S removal and ~90% NH₃ under design conditions (climate‑policy‑watcher.org; climate‑policy‑watcher.org). PAC add‑on costs hinge on carbon price (typically $2–4/kg) and dose; in one full‑scale trial the 80 kg/d PAC unit delivered only ~2% more COD removal overall (nepis.epa.gov) but helped avoid process upsets.

An EPA analysis put the incremental cost at about $7 per pound of extra COD removed (nepis.epa.gov). Techno‑economic pilots are recommended to confirm that toxic‑spike risks are mitigated. Case studies such as Sinopec Jiujiang’s ongoing installation (chemistryviews.org) and the broader literature show that combining sour‑water scrubbing with PAC polishing is a viable path to strict effluent standards and continuous operation. Regulators increasingly expect such multi‑barrier approaches (climate‑policy‑watcher.org; nepis.epa.gov; nepis.epa.gov; nepis.epa.gov; chemistryviews.org; pall.com; mdpi.com).

Sources: Authoritative literature and industry case studies (EPA reports, technology reviews, and practitioner articles) report performance metrics and chemical load reductions for each unit process (climate‑policy‑watcher.org) (nepis.epa.gov) (nepis.epa.gov) (nepis.epa.gov) (chemistryviews.org) (pall.com) (mdpi.com). Each cited study provides quantitative data on removal efficiencies and operational outcomes to guide design decisions.