Tertiary “polishing” in oil and gas turns on a stark choice: capture trace organics with granular activated carbon or destroy them with advanced oxidation. The data show both can hit very low residuals—yet high‑end reuse typically demands a final membrane barrier.
Oil and gas effluents—from produced water to refinery streams—still carry residual dissolved organics after primary and secondary treatment. Tertiary polishing aims to drive those down to very low levels for discharge or reuse. Two main options dominate: adsorption via granular activated carbon (GAC, granular activated carbon media that captures dissolved organics) and advanced oxidation processes (AOPs, radical-based treatments such as Fenton, ozone, or UV/peroxide that oxidize organics in situ). Both can deliver very low effluent organics, but they diverge on cost, complexity, and byproducts. For facilities targeting high reuse, especially process or injection water, a membrane step—typically reverse osmosis (RO, a pressure-driven membrane that removes dissolved salts and organics)—is often required as a final barrier.
GAC adsorption: proven, passive, capacity‑limited
In tertiary service, GAC is installed after clarification and filtration to adsorb residual non‑polar and covalent compounds—hydrocarbons, phenolics, and trace organics—lowering COD (chemical oxygen demand, an aggregate measure of oxygen‑consuming organics)/toxicity and improving color and odor. It does not degrade organics; adsorbed material is periodically desorbed or regenerated. It is ineffective on inorganic salts or very soluble organics (Carbotecnia).
Performance is strong at low concentrations. One industry guide reports secondary effluent COD of 30–100 mg/L can be polished to ~2–10 mg/L with tertiary GAC (Carbotecnia). In municipal tests, GAC achieved ~90% removal of a broad suite of trace organic micropollutants (MDPI). In an oil refinery pilot, GAC used as a post‑AOP step lowered effluent TOC (total organic carbon, carbon bound in organics) to ~4–8.5 mg/L, while the AOP pre‑step alone removed only ~10–18% of TOC (ResearchGate). GAC is especially effective on persistent aromatics; one review cites >90% phenol uptake (depending on contact time and carbon type) (ResearchGate).
Capacity is the constraint. Under high load, breakthrough can occur in weeks. A refinery BAC (biological activated carbon) system ran 84 days at ~65% removal, whereas virgin GAC saturated in ~28 days (ResearchGate). At a municipal plant, GAC removal fell from 97.6% to 80.7% over 13 weeks (2,184 bed volumes) as the media saturated (MDPI). GAC beds also demand low turbidity (<5 NTU, nephelometric turbidity units) and no free oil or oxidants (chlorine) to avoid fouling.
Design and economics are straightforward: typical footprints treat 1–5 gal/min‑ft² (0.004–0.02 m³/min·m²) with EBCT (empty bed contact time) of 10–30 minutes for high removal. Reducing COD from ~50 mg/L to <5 mg/L is routine (Carbotecnia), and residual VOC/phenol levels often drop below detection. Energy is minimal (pump power). The main cost is the carbon—~$1–2/kg (with $/m³ dependent on regenerations)—and GAC typically lowers trace organics by 50–100×. Tertiary adsorbers in this role are commonly specified as activated carbon filters.
AOPs: adjustable chemistry, higher complexity
AOPs generate highly reactive radicals (mainly •OH) to oxidize organics into CO₂, water, and small acids. Common variants in oil and gas include Fenton (Fe²⁺ + H₂O₂ at pH≈3), ozone (O₃ or O₃/UV), UV/H₂O₂, persulfate, and TiO₂ photocatalysis. They tackle non‑biodegradable or toxic compounds that pass biological treatment.
Performance spans a wide range. In bench tests on oilfield water, Fenton achieved modest COD reduction when used alone—~33% COD reduction and ~60% TOC reduction in one study (ResearchGate). A refinery wastewater study found Fenton could mineralize ~53% of DOC (dissolved organic carbon) under optimized conditions (RSC). Tailored AOP can push much higher for specific micropollutants: a UV/H₂O₂ pilot hit 97.1% removal using 40 ppm H₂O₂ and 10 kJ/m² UV, while average removal at baseline conditions was ~76% (MDPI). Ozone‑based AOPs often achieve 70–80% TOC reduction in oily wastewater with sufficient contact, and photo‑Fenton or UV/TiO₂ have degraded phenolics and aromatics by ~60–90% under strong irradiation.
AOPs rarely deliver near‑zero organics alone; they’re typically paired with adsorption or filtration. One case ran Fenton (with only ~50% COD removal) and then used GAC or UF (ultrafiltration, a low‑pressure membrane for colloid/pathogen removal) to meet strict standards (PMC). In the refinery example above, an AOP step yielding ~10–18% TOC removal was followed by biological GAC to reach TOC <10 mg/L (ResearchGate); pre‑oxidation also prevented runaway GAC capacity loss by destroying refractory matter (ResearchGate). Ozone is often applied upstream of GAC to fragment complex molecules for easier adsorption.
Operations and cost reflect the chemistry. Fenton uses cheap reagents (H₂O₂ and iron) but generates iron hydroxide sludge—typically ~0.15–0.5 L wet sludge per liter treated at high removal (RSC). UV/H₂O₂ requires electricity for UV lamps and peroxide; ozone needs an ozone generator and can form bromate if bromide is present. OPEX (operating expenditure) for chemicals is on the order of $0.1–0.5 per m³, depending on dose, and processes often need neutralization or quenching. AOPs avoid saturated solid wastes (apart from sludges) and dosing can be adjusted in real time—tuning UV intensity or peroxide feed. UV‑based systems are commonly paired with UV reactors, and precise oxidant feeds are managed via a dosing pump.
Design rules are short contact times—seconds to minutes—with Fenton typically 5–30 minutes. At optimal pH, Fenton reduces DOC by ~50–60% (RSC). UV/H₂O₂ removes 70–90% of target micropollutants with ~20–40 mg/L H₂O₂ and ~5–10 kJ/m² UV (MDPI). Ozone doses of 1–20 mg/L·min deliver similar orders of magnitude for aromatic COD reduction. The tradeoff is steep: cost scales roughly linearly with removal, so pushing from 80% to 97% can nearly double chemical energy (MDPI).
Head‑to‑head data and hybrid trains
In a municipal head‑to‑head on trace organics (e.g., endocrine disruptors, pharmaceuticals), GAC outperformed UV/H₂O₂ on average: 90.0±4.6% removal via GAC vs. 76.4±6.2% via AOP (MDPI). A well‑managed GAC filter started at ≈97.6% removal and slid to ≈80.7% after ~2,184 bed volumes (13 weeks) as it saturated (MDPI). With higher oxidant dose, the same study boosted AOP removal to 97.1% using 40 ppm H₂O₂ and 10 kJ/m² UV (MDPI).
In refinery‑specific tests, coupling AOP + GAC achieved effluent TOC of ~4–9 mg/L—well below secondary effluent levels—while AOP alone delivered only a modest drop (ResearchGate). AOP transformations can leave low‑molecular‑weight organics (acids, aldehydes) that still contribute to COD and may need further polishing (e.g., biofiltration or adsorption), whereas GAC captures organics without reagent inputs. Many facilities therefore adopt hybrids: an AOP step knocks down refractory compounds, followed by a GAC column as a polishing stage; one refinery process using H₂O₂/UV + GAC reported <10 mg/L TOC (ResearchGate). Ozone is often run upstream of carbon for the same reason (PubMed).
On cost and complexity, passive GAC beds are simpler and their OPEX hinges on media replacement and regeneration frequency. The carbon price range cited is ~$1–2/kg (with $/m³ dependent on regenerations), and in comparative notes a range of $0.5–$2 per kg of carbon with ~1,000–10,000 m³ per rebuild nominal service life also appears. AOP OPEX comes from electricity and chemicals: a typical UV/H₂O₂ plant may consume ~5–20 kWh/m³ in UV power and ~20–50 mg/L H₂O₂ (roughly $0.05–0.15/m³ chemical cost) to reach ~90% removal, alongside ozone destruction or acid/base handling requirements (MDPI).
RO for high‑end water reuse
When the target is very high‑quality reuse (e.g., process cooling, boiler feed, or recharge), RO is typically added as the final barrier. RO can remove virtually all dissolved species—salts, metals, and organics—producing near‑demineralized water. In a mixed industrial wastewater case study, RO lowered aluminum to <3 mg/L (from >40 mg/L), nitrate to <20 mg/L, and sodium and chloride to near brackish levels, with permeate meeting stringent national discharge criteria (MDPI).
Modern systems typically reject >99% of TDS (total dissolved solids, a salinity proxy) and organics >100 Da. Final permeate COD often falls <5 mg/L with conductivity <50 µS/cm, making RO effectively the final polish after GAC/AOP. Facilities weighing options often evaluate membrane systems or specify service‑range units like brackish-water RO; high‑salinity feeds may push designs to seawater RO ranges.
Energy and cost scale with salinity: brackish water (2–6 g/L TDS) typically requires ~1–3 kWh/m³; seawater ~3–6 kWh/m³ with energy recovery. Produced water can be far saltier (up to 70–100 g/L), in which case multi‑stage RO or evaporative brine concentration may be needed, driving energy >10 kWh/m³. Reported OPEX for brackish RO is ~$0.4–0.7 per m³ for TDS ~2–6 g/L, and capital costs often run ~$500–3,000 per m³/d of capacity; maintenance (membrane replacement, chemical cleaning) is significant (PMC).
Pretreatment, fouling, and concentrate handling
RO feeds must be extremely clean to protect flux and membrane life. Typical pretreatment for oily or colored wastewater includes coagulation/flocculation plus media filtration, oil‑water separation, ultrafiltration, and often AOP/GAC. One pilot combined Fenton → clarification → GAC → UF and then ran RO (or UF) with minimal fouling; in that case, UF achieved <50 mg/L COD (PMC). Media beds like a sand/silica filter and upstream oil removal using free‑oil separators often set the stage for membrane steps, with a tight pore barrier added via ultrafiltration modules.
Scaling (Ca/Mg/silica) and organic fouling are primary risks; pH control and antiscalants are mandatory. Plants typically provision specialty aids—see membrane antiscalants and, for periodic clean‑in‑place, membrane cleaners. In one reuse scheme, RO was coupled with upstream ozone + activated carbon to minimize biofouling (PubMed).
Effluent routing also changes with membranes: RO concentrate contains rejected salts and organics at ~10–50% of the feed flow and needs disposal or secondary treatment. Integrating AOP/GAC before RO helps by removing organics upstream so that RO rejects mainly inorganic salts, simplifying concentrate handling.
What the data supports: deployment guidance
GAC is a proven polisher of residual organics: for example, it lowers COD by ~90% relative to secondary effluent to single‑digit mg/L (Carbotecnia) and removes ~90–98% of trace organic micropollutants (MDPI). It has modest O&M but requires frequent column changes (carbon at ~$1–2/kg; elsewhere noted $0.5–$2 per kg with ~1,000–10,000 m³ per rebuild). The sweet spot is relatively stable, moderate‑quality water (COD 20–100 mg/L) that must meet strict discharge targets or partial reuse, especially for non‑polar organics (phenols, BTEX, etc.).
AOPs offer versatile, high‑removal treatment and—with sufficient dosing—can degrade >90% of recalcitrant organics. Examples include ~97% micropollutant removal in UV/H₂O₂ pilots (MDPI) and ~53% DOC mineralization via Fenton (RSC). They handle sudden spikes by adjusting chemical feed and avoid saturated solid waste (apart from sludges), though cost per ton COD removed is higher and partial oxidation may require follow‑up. The paper’s configurations often integrate AOP upstream of adsorption, biofiltration, or membranes.
RO membranes (or UF/NF + RO) are the only route to the highest reuse levels. The cited data show RO can meet stringent reuse specs—metals down to ppm or below, organics near zero, hardness near zero (MDPI; PMC). Expect OPEX on the order of $0.5–$1/m³ for moderate‑salinity feeds (brackish) plus significant CAPEX; any facility targeting near‑zero discharge or ≥80% reuse should plan for RO with robust pretreatment. AOP/GAC ahead of RO can greatly reduce fouling and concentrate load—an integrated AOP + UF + RO train in a pilot cut COD to <50 mg/L at ~50% additional cost (PMC). Where RO is used, the data underscore that practically all regulated pollutants (metals, COD, hardness) can be driven below limits (data from [33†L80–L88]).
Internationally, many petrochemical refineries aiming for ≥80% water reuse have installed RO (often two‑stage) as the final step; Brazil’s Petrobras reported reuse of >70% via UF+RO in a refinery scheme. Regulatory trends (e.g., lowering COD limits) are also pushing more polishing (Carbotecnia).
