Refinery wastewater’s toughest choice: CAS vs. MBBR vs. MBR

Refinery effluent is not ordinary sewage. It is often stronger than 2,000 mg/L COD (chemical oxygen demand, a measure of oxidizable organics) and laced with phenols, heavy hydrocarbons, and sulfides. One review is blunt: petroleum refinery wastewater “containing many recalcitrant compounds… is difficult to eliminate during regular wastewater treatment processes” (pmc.ncbi.nlm.nih.gov), and biological processes alone often fail at COD > 2000 mg/L (pmc.ncbi.nlm.nih.gov).

Still, three options dominate: conventional activated sludge (CAS), moving bed biofilm reactors (MBBR), and membrane bioreactors (MBR). Bench‑scale tests show promise — a conventional reactor removed ~94–95% COD from refinery wastewater (researchgate.net) — but full‑scale plants must navigate toxic shocks, variable loads, and space constraints.

Most facilities start with oil and solids removal. Primary trains can include screens and separators; packaged options for screens, oil skimming, and grit are common in physical separation systems. Dissolved air flotation and similar oil‑water steps are often added ahead of biology; a refinery‑specific DAF unit typically targets free oil and suspended solids before biological treatment.

Conventional activated sludge: mature but space‑hungry

CAS — the baseline suspended‑growth process — is widely deployed and flexible. In lab work, removal followed a pseudo–first‑order model with k≈0.055–0.059 L·mg⁻¹·day⁻¹ (researchgate.net). Well‑operated units often top 90% COD removal under stable feeds (researchgate.net). For refineries, a CAS stage is typically arranged in multi‑zone anoxic/oxic trains with long sludge ages (SRT, solids retention time) to support nitrifiers and slowly growing degraders.

Reality’s catch: footprint and vulnerability. CAS needs big aeration basins and secondary clarifiers, with hydraulic retention time (HRT) tuned against MLSS (mixed liquor suspended solids). It also produces substantial sludge. Robustness is only moderate — suspended flocs have no protected niche, so phenol or metal spikes and hydraulic shocks can crash performance. Many refineries resort to nutrient dosing (e.g., supplemental nutrients) and add pretreatment and polishing. One full‑scale plant treating 2550 mg/L refinery influent required an ICBAF–A/O–MBR–O₃/AC train to reach ~50 mg/L COD (pmc.ncbi.nlm.nih.gov), with post‑treatment such as activated carbon used alongside ozonation.

MBBR: compact biofilms with shock tolerance

MBBR retains biomass on moving plastic carriers, creating dense biofilms that tolerate spikes and hold specialized degraders at high concentrations. Studies emphasize “less space requirement, higher surface area for mass transfer, [and] higher organic loading rates at lower HRTs” (sciencedirect.com). The footprint advantage is a key reason about 200 Norwegian plants use MBBR specifically to save space (accesswater.org).

Performance on petroleum streams is solid. A pilot MBBR with 6 h HRT and 0.6 carrier fill treated oil refinery effluent with 69–89% COD removal (effluent 40–75 mg/L) and 45–86% ammonia removal (NH₄–N 2–6 mg/L) (pubmed.ncbi.nlm.nih.gov). Those figures approached Indonesian limits of COD≤160 mg/L and NH₄–N≤8 mg/L (id.scribd.com). A refinery‑tailored MBBR stage often shows lower sludge yields than CAS and strong nitrification under lower oxygen demand.

Trade‑offs remain. MBBR effluent still carries sloughed solids, so a downstream clarifier or flotation is required. Carriers introduce moving parts and aeration mixing; oil and grease can accumulate on media and need periodic cleaning. And while biofilms buffer toxins, extreme phenolics or sulfide can inhibit MBBR too, prompting pre‑oxidation in many trains.

MBR: small footprint, reuse‑grade effluent

MBR couples activated sludge with membranes, decoupling SRT from HRT and running at high MLSS (often 8–15 g/L in practice). The result is a compact reactor with complete biomass separation and consistently clear permeate. In petrochemical tests, MBR achieved >80% COD removal and essentially full nitrification (>99.9% NH₄–N removal) (sciencedirect.com), routinely meeting strict reuse and discharge requirements (sciencedirect.com). For space‑constrained refineries, a membrane bioreactor can shrink the biological footprint, with reactors often 2–5× smaller than CAS for the same loading and no large secondary clarifier needed (sciencedirect.com).

The biological side is robust thanks to the large retained biomass and long SRTs that favor slow‑growing degraders, including those targeting complex aromatics. But membranes add vulnerabilities: oily refinery organics foul surfaces, as oil droplets coalesce into a continuous oil‑film layer distinct from a typical bio‑cake, driving cleaning frequency and energy (mdpi.com). Design tweaks (e.g., hydrophilic surfaces, air scouring) can mitigate, at the cost of higher aeration energy and maintenance.

Comparative footprint and resilience

Across effluent quality and compactness, the hierarchy is clear: MBR > MBBR > CAS (sciencedirect.com) (sciencedirect.com). All three can be robust when designed well, but MBBR’s protected biofilms and MBR’s high biomass inventories give them an edge on toxic or shock loads compared with CAS. CAS remains attractive on capex and simplicity for steady, moderate‑strength loads.

Footprint deltas are material: MBBR plants can be 30–50% smaller than equivalent CAS due to higher volumetric loading (sciencedirect.com). MBR is smaller still; its “distinctive advantage” combines minimal area with high effluent quality (sciencedirect.com). One example from municipal practice: swapping a CAS clarifier plus sand filter for submerged membranes cut the secondary clarifier footprint by 80%. In refineries, MBRs are often piloted as polishing stages or for small sidestreams where space savings justify the cost.

Handling toxins and complex organics

No single biology is a silver bullet at extreme loads; pretreatment is standard. Free oil and heavy organics are typically settled or floated ahead of biology; an anoxic or hydrolytic step is sometimes used before aerobic treatment. MBBR has built‑in nitrification capacity and can co‑metabolize aromatics under aerobic/anoxic cycling. MBR’s very long SRTs allow specialized, slow‑growing strains to establish, improving breakdown of recalcitrants. One combined system used filtration plus ozone and activated carbon to nearly fully remove phenolics and complex organics (pmc.ncbi.nlm.nih.gov).

Outcomes at a glance (units and terms)

Definitions: COD (chemical oxygen demand), BOD (biochemical oxygen demand), NH₄–N (ammonium‑nitrogen), HRT (hydraulic retention time), SRT (solids retention time), MLSS (mixed liquor suspended solids).

• COD removal: CAS ~90–95% in well‑operated units (researchgate.net); MBBR ~70–89% at 6–20 h HRT (pubmed.ncbi.nlm.nih.gov); MBR >80% (often 85–95%) (sciencedirect.com).
• Ammonia removal: CAS needs separate nitrification stage (often 90%+ under long SRT); MBBR 45–86% NH₄ removal at 6 h (pubmed.ncbi.nlm.nih.gov); MBR ~99.9% nitrification (sciencedirect.com).
• Effluent quality: CAS ~50–150 mg/L COD; MBBR ~40–75 mg/L (pubmed.ncbi.nlm.nih.gov); MBR routinely <50 mg/L. Indonesian refinery standards: PoBOD≤80, PoCOD≤160, O&G≤20 ppm (id.scribd.com).
• Footprint (relative volume for same load): CAS 1.0; MBBR ~0.5–0.7 (sciencedirect.com); MBR ~0.3–0.5 (sciencedirect.com).
• Sludge production: CAS 0.5–0.8 kg/kg COD removed; MBBR lower (0.3–0.5); MBR yields are similar per kg COD removed but wasting is less frequent.

Bottom line

CAS can work — often >90% COD removal under controlled conditions — but demands large basins and careful acclimation to survive refinery toxins (researchgate.net). MBBR adds resilience and saves area (sciencedirect.com). MBR offers the smallest footprint and the best effluent quality (sciencedirect.com) — but with membranes that foul on oil, raising energy and cleaning needs (mdpi.com). Many refineries now deploy hybrids (e.g., IFAS, CAS plus MBBR polishing, or MBR after aerobic stages) to balance robustness, footprint, and compliance goals.

Sources: peer‑reviewed journals and regulatory materials underpin the data and examples cited here, including bench/pilot/full‑scale results (researchgate.net; pubmed.ncbi.nlm.nih.gov; sciencedirect.com), reviews (pmc.ncbi.nlm.nih.gov; sciencedirect.com), and Indonesian Ministry of Environment discharge standards (id.scribd.com).