What COD Means and Why 2026 Discharge Limits Drive Your Method Choice
Chemical oxygen demand (COD) is the mass of oxygen consumed when organics in water are oxidized by a strong chemical oxidant, typically potassium dichromate in a sulfuric-acid boil, reported in mg O₂/L. Unlike BOD₅, which captures only the biodegradable fraction over five days, COD includes both biodegradable and refractory organics, so a COD value is always equal to or higher than the corresponding BOD₅ on the same sample. Refractory COD — pesticides, azo dyes, solvents, pharmaceutical residues, and high-MW humic substances — does not biodegrade in standard activated sludge and is the main reason a "biological-only" plant fails compliance.
The 2026 compliance numbers you are designing against are concrete and jurisdiction-specific. China GB 8978-1996 (still in force 2026) caps COD at 500 mg/L for second-class discharge, with 1,000 mg/L for third-class. EU Urban Waste Water Directive 91/271/EEC holds effluent COD to 125 mg/L (as a 95th-percentile) for plants serving >10,000 PE. EPA categorical pretreatment standards under 40 CFR Parts 405–471 set industry-specific limits that commonly fall in the 100–400 mg/L window, with some subcategories as tight as 50 mg/L for direct discharge. Any decision on method selection starts with two numbers: the influent COD you actually have, and the limit you must hit on the far side of the train.
Removing COD is fundamentally a matching exercise: pair an influent concentration and matrix with a mechanism that targets that fraction. Suspended and colloidal COD calls for coagulation or flotation. Dissolved biodegradable COD calls for activated sludge or an MBR. Dissolved refractory COD calls for chemical or advanced oxidation. Most industrial plants need two or more stages in series to land inside the regulatory band.
The Six COD Removal Methods Compared Side by Side
The master table below is the anchor for the rest of this article. Every method listed is in active industrial use in 2026; CAPEX/OPEX bands are 2026 equipment pricing for a 100 m³/day reference flow and exclude civil works.
| Method | Typical Influent COD (mg/L) | Removal Efficiency (%) | CAPEX ($/m³/day) | OPEX ($/m³ treated) | Best Suited For | Main Limitation |
|---|---|---|---|---|---|---|
| Biological A/O (anoxic-aerobic) | 200–2,000 | 80–90 | 50–120 | 0.15–0.35 | Municipal + biodegradable industrial streams | Sensitive to toxicity and load shocks |
| MBR (membrane bioreactor) | 200–3,000 | 90–97 | 180–350 | 0.25–0.55 | Strict effluent or footprint-constrained sites | Membrane fouling and CIP chemicals |
| Coagulation + DAF | 100–1,500 (suspended/colloidal) | 50–80 | 60–150 | 0.10–0.20 | Food, dairy, slaughterhouse, textile desizing | Sludge handling and polymer cost |
| Fenton oxidation | 500–10,000 (refractory) | 70–90 | 80–200 | 0.30–0.70 | Dye, pesticide, pharma, landfill leachate | Iron sludge and pH swing to ~3 |
| Ozonation | 200–5,000 | 50–80 | 120–280 | 0.25–0.50 | Color removal, micropollutants, polishing | Energy at 12–18 kWh/kg O₃ |
| RO / NF membrane | <500 (polishing) | 95–99 | 200–500 | 0.40–0.90 | Water reuse, ZLD polishing | High-pressure pumping and concentrate disposal |
Two patterns fall out of this table immediately. First, no single method covers the 100–10,000 mg/L industrial range alone — biology dominates the low-to-mid band, while Fenton and ozonation handle the upper refractory range. Second, OPEX gaps between methods are narrower than CAPEX gaps, so a Fenton-plus-biology hybrid often has lower total-cost-of-ownership than an MBR over a 10-year horizon, despite higher chemical demand.
Biological COD Removal: A/O, SBR, and MBR for Industrial Loads
The anoxic-aerobic (A/O) process is the workhorse of industrial biological COD reduction, with an F/M ratio of 0.05–0.15 kg BOD/kg MLSS·day being typical for industrial activated sludge running at 2,000–4,000 mg/L MLSS. COD removal of 80–90% is routine on streams with BOD/COD above 0.4 — meaning the organics are genuinely biodegradable — and total nitrogen drops simultaneously through coupled nitrification-denitrification. Conventional A/O is the cheapest option on a $/m³ basis but is sensitive to influent toxicity, temperature swings below 10 °C, and hydraulic shock.
An MBR substitutes a submerged ultrafiltration module for the secondary clarifier and decouples solids retention time from hydraulic retention time. Operating reality in 2026: PVDF flat-sheet or hollow-fiber membranes with 0.1 μm nominal pore size paired with MLSS controlled at 8,000–12,000 mg/L routinely deliver effluent COD below 50 mg/L on industrial streams — well inside both the EU 125 mg/L limit and most reuse targets. The trade-off is membrane fouling from extracellular polymeric substances (EPS), which demands a defined cleaning-in-place schedule and, in retrofits, a 60% smaller footprint than a clarifier-based plant of equivalent capacity. For a 100 m³/day industrial MBR treating 1,500 mg/L COD, expect effluent at 30–60 mg/L COD, CAPEX in the $200K–$350K band, and OPEX dominated by aeration at 0.4–0.6 kWh/m³.
Failure modes worth specifying into your operating SOP: filamentous bulking when SVI climbs above 150 mL/g (diagnose with microscopy and capillary suction time, then adjust F/M or RAS), and irreversible membrane fouling from chronic EPS overload (control with relaxed-cycle backwash and quarterly alkaline CIP). If you are selecting a packaged unit for a retrofit or containerized plant, an MBR membrane bioreactor system with built-in CIP is the most defensible single-vendor choice. For greenfield or larger-scale industrial flows, the WSZ underground integrated sewage treatment plant covers the biological block when land surface area is the binding constraint.
Physicochemical COD Removal: Coagulation, Flocculation, and DAF
Coagulation with Al₂(SO₄)₃, polyaluminum chloride (PAC), or FeCl₃ neutralizes the surface charge on colloids; flocculation with anionic or cationic polyacrylamide (PAM) aggregates those destabilized particles into settleable or floatable floc. Jar testing with 1–2 L samples sets the dose empirically — there is no substitute — and the optimum typically lands in the 50–300 mg/L coagulant range with 1–5 mg/L polymer. Skipping this step is one of the most common reasons downstream biological or membrane stages fail: soluble COD that should have been physically removed instead loads the biology with particulates, and the membranes foul on colloids they were never sized for.
Dissolved air flotation (DAF) is the standard pretreatment for food processing, dairy, slaughterhouse, edible oil, and textile desizing streams because it strips 50–80% of suspended and emulsified COD at surface loading of 5–20 m/h, using 30–80 μm micro-bubbles generated by a saturator loop at 5–7 bar. Industrial DAF skids cover 4–300 m³/h across roughly 13 standard models, with recirculation ratios of 20–40% of throughput. A key cost lever: pairing the DAF with sludge recirculation reduces fresh chemical consumption by 30–50% versus once-through settling tanks, and the float layer can be dewatered to 18–25% dry solids on a belt press or a plate-and-frame filter. For sites where FOG runs below 100 mg/L and the goal is TSS rather than oil removal, a lamella clarifier is the lower-CAPEX alternative.
Specifying a complete train means bundling a DAF pretreatment system with an automatic chemical dosing system so coagulant and polymer feed track flow-proportionally rather than by hand. A reader weighing the trade-off between DAF and primary settling will find the full decision in the DAF advantages and disadvantages engineering guide.
Chemical Oxidation: Fenton, Ozone, and Chlorine for Refractory COD
Fenton oxidation is the default attack on refractory COD because it generates hydroxyl radicals (·OH), a non-selective 2.8 V oxidant, from the Fenton reaction Fe²⁺ + H₂O₂ → Fe³⁺ + ·OH + OH⁻ at pH 2.8–3.5. Within an optimum H₂O₂:Fe²⁺ molar ratio window of 5:1 to 20:1, increasing the ratio raises COD removal efficiency (Alipour 2020, ScienceDirect, 88 citations on industrial spent-caustic COD). A worked example from that study: a 2,000 mg/L COD effluent drops to below 200 mg/L using Fe²⁺ at 200 mg/L, H₂O₂ at 1,500 mg/L, pH 3.0, and a 60-minute reaction time at 25–35 °C. Outside the optimum ratio — too little H₂O₂ leaves organics unreacted, too much scavenges the ·OH back to water — efficiency collapses. Fenton's CAPEX line items are the pH adjustment tanks (acid and caustic) and an iron-sludge dewatering step, usually a plate-and-frame filter press sized at 2–4 kg dry solids/m²·h.
Ozonation attacks organics through direct O₃ reactions and through ·OH generated in the side chain, converting dyes, phenolics, and many micropollutants to aldehydes, organic acids, and eventually CO₂. Industrial ozone generators run at 12–18 kWh/kg O₃ at the point of production, which is the dominant OPEX line; COD removal lands at 50–80% for a 200–5,000 mg/L feed, and color removal is consistently above 90%. Chlorine and chlorine dioxide are the lowest-CAPEX option for polishing, but they form trihalomethanes (THMs) and bromate on bromide-containing streams. On-site ClO₂ generation at 50 g/h to 20,000 g/h is preferred over direct chlorine gas for hospital, food, and pharmaceutical effluent because of the safety profile and the absence of chlorinated organic byproducts at controlled dose. If you are weighing Fenton against electrocoagulation for a new refractory stream, the Fenton vs electrocoagulation buyer's guide gives the side-by-side cost and operating data. A packaged chlorine dioxide generator covers the polishing step where biology and Fenton have already done the bulk reduction.
Advanced Oxidation and Adsorption: Closing the Last 5–20% of Refractory COD
Advanced oxidation processes (AOPs) — O₃/H₂O₂ (peroxone), O₃/UV, Fe²⁺/H₂O₂/UV (photo-Fenton), and photocatalytic TiO₂ — are deployed when inlet COD is already below ~500 mg/L but the effluent still fails a 50 mg/L reuse or ZLD target. The combinations beat Fenton alone on specific UV-absorbing organics because photon input drives additional ·OH generation and shifts oxidation pathways. Peroxone is the most common in full-scale industrial plants because it slots into an existing ozone contactor and adds only the H₂O₂ dosing skid.
Adsorption handles the residual fraction. Granular activated carbon (GAC) with a surface area of 800–1,200 m²/g, dosed at 5–50 g AC per m³, reduces residual COD by 60–90% in polishing and requires periodic thermal regeneration (typically every 6–18 months at 800–900 °C in a rotary kiln) to restore capacity. Synthetic resins (Purolite, Lewatit) are the right tool when the target is a specific compound class — phenols, aniline, or PFAS precursors — rather than bulk COD. A real 2025 reference case: a textile dyehouse polishing MBR effluent at ~120 mg/L COD with ozone followed by GAC landed below 30 mg/L for loop reuse, cutting fresh water intake by roughly 40% at a dyeing water cost above $1.80/m³. The full GAC specification and operating envelope for dye streams is in the activated carbon polishing for dye wastewater engineering guide.
How to Choose the Right COD Removal Train for Your Influent
Method selection collapses to four rules. Rule 1: if influent COD is above 5,000 mg/L or the BOD/COD is below 0.2, the stream contains a high refractory fraction and you must start with Fenton, wet air oxidation, or incineration — biology will not close the gap on its own. Rule 2: if influent COD is 500–3,000 mg/L and the BOD/COD is above 0.4, biological A/O or MBR is the cheapest baseline; add DAF upstream whenever oil and grease exceeds 100 mg/L or TSS exceeds 300 mg/L, otherwise the biology will foam and the membranes will foul. Rule 3: if influent COD is already below 500 mg/L but the effluent still fails discharge, add advanced oxidation or membrane polishing rather than expanding biology — the marginal cost of pushing biology lower than ~60 mg/L COD is steep. Rule 4: if water reuse or ZLD is the target, plan for RO/NF as the last stage and size upstream biology to feed RO at less than 200 mg/L COD to prevent irreversible fouling.
| Influent COD (mg/L) | Recommended Primary | Recommended Polishing | Expected Effluent COD (mg/L) |
|---|---|---|---|
| >5,000 (refractory) | Fenton or incineration | Biological A/O | 150–300 |
| 500–3,000 (biodegradable) | DAF + A/O or MBR | None or sand filter | 40–100 |
| 200–500 (mixed) | MBR or A/O | Ozonation or AOP | 30–60 |
| <500 (failing discharge) | AOP or ozone | GAC or NF | <30 (reuse-grade) |
Worked example for a 200 m³/day food processing wastewater at 3,000 mg/L COD and 200 mg/L oil: DAF drops COD to about 1,200 mg/L, A/O brings it to 100 mg/L, an MBR polish lands below 50 mg/L, and the total CAPEX band sits at $280K–$500K with OPEX at $0.35–$0.55/m³. A greenfield version of this train typically starts with a DF series PVDF flat sheet membrane module as the MBR polishing element.
2026 CAPEX and OPEX Benchmarks for a COD Removal Train
The table below gives 2026 pricing for a representative 100 m³/day industrial flow across four representative process trains, sized by influent COD and target effluent. Numbers are equipment + installation in USD and exclude major civil works or permitting.
| Train | Configuration | CAPEX (USD) | OPEX ($/m³ treated) | Best For |
|---|---|---|---|---|
| 1 | DAF + A/O | $180,000 | 0.25 | Biodegradable industrial, FOG <200 mg/L |
| 2 | DAF + A/O + MBR | $320,000 | 0.45 | Strict discharge or footprint retrofit |
| 3 | Fenton + A/O | $280,000 | 0.55 | Refractory influent >3,000 mg/L |
| 4 | Fenton + biological + RO | $650,000 | 1.10 | ZLD or water reuse |
OPEX for Train 4 (Fenton + biological + RO) breaks down approximately as: energy 45%, chemicals 25%, membrane replacement 15%, and labor 15%. Payback framing for reuse-grade trains: at industrial water tariffs above $1.50/m³, Train 4 typically recovers its CAPEX premium over Train 2 in 2.5–4 years through avoided fresh-water purchase and reduced discharge fees. For a full RO pricing model including element replacement schedules, see the RO membrane system cost guide for 2025 industrial pricing. To keep any of these trains available rather than reactive, the predictive maintenance for wastewater plants guide explains how to wire operating data into failure forecasting.
Frequently Asked Questions
What is the most effective method to remove COD from wastewater? For biodegradable streams at 200–3,000 mg/L influent, an MBR delivers the highest single-stage removal at 90–97% and effluent below 50 mg/L. For refractory streams above 3,000 mg/L, Fenton oxidation at 70–90% removal is the most effective first stage, with biological polishing closing the remainder.
Can COD be removed by biological treatment alone? Yes, but only when the BOD/COD ratio is above approximately 0.4 and the influent is below ~3,000 mg/L. Below that ratio, refractory COD passes through the biology essentially untouched, and you will need chemical or advanced oxidation downstream.
Which method is best for high COD industrial wastewater above 5,000 mg/L? A Fenton-plus-biological hybrid is the standard 2026 answer. Fenton drops the refractory fraction by 70–90%, after which a conventional A/O stage finishes the job to compliance. Incineration is the alternative at very high concentrations or where the waste has heating value.
How much does a COD removal system cost in 2026? For a 100 m³/day industrial flow, CAPEX ranges from about $180,000 for a DAF + A/O train to $650,000 for a Fenton + biological + RO reuse train. OPEX runs $0.25 to $1.10 per m³ treated, dominated by energy in RO trains and by chemicals in Fenton trains.
Is MBR worth the higher cost over conventional activated sludge? Yes, when you need effluent COD reliably below 50 mg/L, when the site is footprint-constrained, or when the downstream process (RO, water reuse) cannot tolerate the TSS variability of a clarifier. MBR's 60% smaller footprint and stable effluent typically justify the 2–3× CAPEX premium within 3–5 years on space-constrained retrofits.
Related Equipment
- WSZ underground integrated sewage treatment plant — specifications, capacity range, and technical data
