Boil, bugs, or bleach: how coke plants drive ammonia down to single‑digit mg/L

Coke‑oven effluent is loaded with nitrogen, mostly as ammonia liberated during desulfurization. One real‑world snapshot: raw coke‑oven condensate at SSAB Oxelösund (Sweden) contained about 240 mg/L total nitrogen (TN), mostly as ammonia with some thiocyanate (SCN) (file.scirp.org). Regulations typically demand single‑digit mg/L ammonia in treated effluent; Indonesian industrial standards often cite around 8 mg/L NH3–N (pt.scribd.com), similar in spirit to the SSAB consent of 30 mg/L TN.

The wastewater is often complex—high biochemical oxygen demand (BOD), phenols, cyanide/SCN, and salts. The technology choice hinges on the NH3–N concentration, pH (which sets NH3/NH4+ speciation), biodegradable carbon, and inhibitors such as metals or cyanides. High‑strength, low‑oxygen water typically favors steam stripping or chemical oxidation; moderate ammonia with good BOD favors biology; extreme polishing can need combined methods (file.scirp.org).

Steam stripping (thermal ammonia distillation)

Steam stripping raises pH (around 11) to convert ammonium (NH4+) to free ammonia (NH3), then volatilizes NH3 into steam for condensation and recovery. It is essentially a heated distillation, requiring energy and caustic dosing, but producing a concentrated ammonia solution and minimal off‑gas: the ammonia‑laden vapor is condensed, so no separate off‑gas scrubber is needed (mdpi.com). Plants typically feed alkali via precise chemical dosing; in practice, facilities rely on equipment such as a dosing pump for pH control upstream of the stripper.

High‑gravity gas–liquid contactors, or rotating packed beds (RPBs), have recently outperformed conventional towers. Yuan et al. (2024) used an RPB at pH 11 on 5,000–20,000 mg/L NH4–N streams, achieving ammonia removal efficiencies (ARE) up to about 98% (mdpi.com) and recovering an ammonia liquor of roughly 22.9 wt% NH3 (mdpi.com). The RPB mass‑transfer coefficients KLa reached ~12.3–18.4 h⁻¹ versus 0.42–1.2 h⁻¹ for packed towers (mdpi.com).

In one pilot, a 1,000 mg/L NH3 feed stripped in an RPB left less than 1% NH3 in the treated water, with gaseous NH3 absorbed in sulfuric acid to yield ammonium sulfate, (NH4)2SO4 (mdpi.com). Under optimal conditions—steam:liquid near 0.175 kg/kg at 20,000 mg/L feed—the recovered liquor reached 22.88 wt% NH3 with ARE around 98% (mdpi.com). Increasing steam beyond the optimum returned little extra benefit; ARE plateaued near 98% (mdpi.com).

The trade‑off is energy and chemicals. An EPA report estimated that stripping roughly 757 L/min of process condensate at 1,000 ppm NH3 would consume about 5,440 kg/h of steam (≈$288,000/year at typical steam costs) (nepis.epa.gov). Heat integration using excess plant steam can partly offset this. The system footprint is compact, but packing or RPB costs are higher. Seasonal effects are minimal compared with biological processes, and steam stripping is robust to toxic organics or metals because it relies on volatility; the wastewater must still be heated and caustically adjusted, and residual chloride (if any) must be managed (e.g., acid after stripping) (mdpi.com).

Biological nitrification/denitrification performance

Conventional biotech converts ammonium to nitrate (nitrification) and then to nitrogen gas (denitrification). With proper design—long solids retention time (SRT, the average time biomass stays in the system) and enough carbon—coke‑oven effluent can be treated reliably. Systems like extended‑aeration activated sludge or sequencing batch reactors are used (file.scirp.org); many plants implement this with standard configurations such as an activated‑sludge train or a packaged sequencing batch reactor.

At SSAB Oxelösund, Morling et al. (2012) reported more than 90% total‑N removal: feeding 130–250 mg/L NH4–N, the team brought effluent TN below 20 mg/L, and at times down to 5 mg/L (file.scirp.org). Observed nitrification rates were about 1.4–1.8 g N/(kg‑VSS·h) (VSS, volatile suspended solids) (file.scirp.org), and SRT had to be high—around 40–50 days—to acclimate nitrifiers to high NH3/SCN toxicity (file.scirp.org). An equalization basin helped buffer peaks (file.scirp.org).

Notably, the coke wastewater contained enough organic carbon that little or no external carbon was needed for denitrification (file.scirp.org). Under stable operation—complete nitrification plus limited denitrification with added methanol—the plant consistently met TN discharge below 20 mg/L (file.scirp.org). Biological systems also degrade co‑pollutants: cyanide and SCN were removed (via fermentation/oxidation and precipitated metals) (file.scirp.org).

Limits do apply. At very high ammonia loads the required reactor volume becomes excessive, and performance drops sharply at low temperature (nepis.epa.gov). Biological treatment requires careful pH control (around neutral) and dissolved oxygen management, and it produces waste sludge.

Breakpoint chlorination chemistry and doses

Breakpoint chlorination chemically oxidizes ammonium using chlorine (as gas or sodium hypochlorite). As Cl2 is added, monochloramine (NH2Cl), dichloramine (NHCl2), and nitrogen trichloride (NCl3) form; with enough dose, chloramines collapse and ammonia is driven nearly completely to nitrogen gas (N2), with some nitrate (NO3−) by‑product (nepis.epa.gov, nepis.epa.gov). In practice this demands a high chlorine‑to‑ammonia mass ratio—on the order of ~8–10:1—and controlled pH around 7–8 (nepis.epa.gov, nepis.epa.gov). Accurate chemical feed is essential, so operators commonly specify metering devices such as a dosing pump.

Data from EPA pilot work at Blue Plains showed 95–99% of ammonia‑N converted to N2, with very low residual NH3 (often ≪1 mg/L and around 0.1 mg/L in tests) (nepis.epa.gov, nepis.epa.gov). Not all nitrogen ends as N2—small fractions become nitrate and nitrogen trichloride—and at pH 6–8, nitrate residuals are minimized and N2 is maximized (nepis.epa.gov). The stoichiometric requirement is heavy: roughly 7.6 kg Cl2 per kg NH3 to convert purely to N2 (nepis.epa.gov).

Drawbacks are significant. High chlorine demand, formation of disinfection by‑products (DBPs), and chlorination of organics (phenols, PAHs) can generate toxic compounds; nitrogen trichloride and chlorinated nitrogen species also form and typically require neutralization. Post‑treatment dechlorination—often using a dedicated dechlorination agent—is usually needed to prevent toxic discharge (nepis.epa.gov). Because of these issues, breakpoint is rarely used on raw industrial wastewaters today; it appears as a final polishing step after biological nitrification when very low discharge (≪10 mg/L) is required (nepis.epa.gov).

Comparative performance and costs

All three methods can meet typical discharge limits under the right conditions. Steam stripping commonly delivers more than 95% ammonia removal—often ~98% in optimized RPB systems (mdpi.com). Biological systems under steady‑state have similarly brought TN from ~240 mg/L down below 20 mg/L (>90% removal, as low as 5 mg/L) in coke‑oven service (file.scirp.org). Breakpoint chlorination can exceed 95% removal and leave roughly 0.1–1 mg/L NH3 in the effluent (nepis.epa.gov).

Influent drives selection. When NH3 is very large (≫500 mg/L), biological reactors struggle because required SRT/volume grows unwieldy (nepis.epa.gov) and breakpoint demands enormous chlorine. RPB‑enabled steam stripping targets “ammonia‑rich” streams (e.g., in the 5,000–20,000 mg/L NH4–N tests) (mdpi.com). If wastewater has strong BOD/COD and moderate NH3 (<200–300 mg/L), biology is attractive and can leverage in‑situ carbon for denitrification. Where inhibitors are present (e.g., cyanide), biology can adapt over long SRT while steam stripping is inherently robust; breakpoint can unpredictably chlorinate organics (file.scirp.org).

Operating economics diverge. Biological plants trade capital for air and mixing power and produce benign N2; capex is moderate (tanks and blowers) and O&M centers on aeration. Steam strippers are compact but energy‑intensive; one EPA analysis put steam expenses at roughly $0.6/kg NH3 removed (≈$288k/year for 14 kg/s NH3) (nepis.epa.gov). Breakpoint plants are compact yet chemical‑heavy: on the order of 8–10 kg Cl2 per kg NH3 removed, with corrosion‑resistant materials and downstream quenching (nepis.epa.gov). None of the three produce significant solids except biology, which yields about 0.5 kg sludge per kg N removed.

By‑products and compliance matter. Steam stripping yields ammonia liquor for potential fertilizer salt recovery and introduces no chlorine. Biology yields N2, plus nitrate if denitrification is incomplete. Breakpoint yields N2 but also chlorinated nitrogen and potential halogenated organics; under Indonesian standards that focus on NH3–N and often require dechlorination of effluent, chlorine residuals from breakpoint could violate water quality unless quenched (pt.scribd.com).

Decision framework for coke‑oven effluent

In practice, many sites converge on hybrids: biological nitrification/denitrification does the heavy lifting, followed by steam stripping or breakpoint for the last few milligrams per liter. A practical framework starts with characterizing influent NH3 (mg/L) and the effluent target, then checking heat, power, and chemicals on hand.

• If influent NH3 ≫ 1000 mg/L: lean toward steam stripping for bulk removal, including high‑gravity RPBs; some flows consider combined nanofiltration (NF) plus steam stripping for pre‑concentration (mdpi.com). Where membranes are part of the plan, facilities typically specify nano‑filtration upstream to condition the stream.
• If influent NH3 is moderate (<300 mg/L) and BOD is adequate: use biological treatment; target SRT around 40+ days and consider equalization. Expect >90% TN removal, with effluent TN down to <20 mg/L and occasionally to 5 mg/L, as in the SSAB data (file.scirp.org). For configuration, operators commonly adopt an activated‑sludge or a sequencing batch reactor scheme within a broader biological digestion train.
• If effluent must be <1–5 mg/L: consider breakpoint chlorination as a polishing step after biology. Plan for high chlorine doses (~8–10:1 by weight of Cl2:NH3) and downstream quenching (nepis.epa.gov, nepis.epa.gov).
• If toxic inhibitors are present: steam stripping is favored because microbes struggle with cyanide/phenol at normal operating conditions, while RPB stripping relies on volatility. Biology can adapt slowly at long SRT, but breakpoint may chlorinate organics unpredictably (file.scirp.org).
• Cost and footprint: where land is constrained, steam or breakpoint’s compact equipment can beat large basins; where energy is limited, biology is preferred. Steam stripping’s energy can be mitigated by heat integration or leveraging plant steam (nepis.epa.gov).

Sources and citations

Authoritative studies and reviews were used for performance data and criteria. For steam stripping: recent RPB experiments demonstrated ~98% NH3 removal and ~23% NH3 recovery, with design analyses providing packing/KLa and cost figures (mdpi.com, mdpi.com, mdpi.com, nepis.epa.gov). For biological treatment: pilot data on coke wastewater show nitrification/denitrification reducing TN from ~240 to <20 mg/L under 40–50 day SRT (file.scirp.org, file.scirp.org). For breakpoint chlorination: EPA reports show >95% NH3→N2 conversion and required ~8:1 Cl2:NH3 mass ratios (nepis.epa.gov, nepis.epa.gov). Indonesian MOEF “Baku Mutu Air Limbah” tables reflect ~8 mg/L NH3–N limits for general industry (pt.scribd.com).

Specific citations include: Morling et al., “Biological Removal of Nitrogen Compounds at a Coke‑Oven Effluent Stream,” J. Water Resour. Prot., Vol.4, No.6 (2012) (file.scirp.org, file.scirp.org); Yuan et al., “Steam Stripping for Recovery of Ammonia from Wastewater Using a High‑Gravity Rotating Packed Bed,” Environments 11, 206 (2024) (mdpi.com, mdpi.com); EPA, Treatment of Ammonia Plant Process Condensate Effluent, EPA‑RTP‑600/2‑77‑200 (1977) (nepis.epa.gov, nepis.epa.gov); Teichgraber & Stein, “Nitrogen elimination from sludge treatment reject water – comparison of steam‑stripping and denitrification,” Water Sci. Technol., 30(1) (1994): 41–51 (mdpi.com); EPA Breakpoint Chlorination Report, EPA‑670/2‑73‑058 (1973) (nepis.epa.gov, nepis.epa.gov); Kinidi et al., “Recent Development in Ammonia Stripping Process for Industrial WW,” Int. J. Chem. Eng., 2018, ID3181087 (onlinelibrary.wiley.com, mdpi.com); and Indonesian MOEF “Baku Mutu Air Limbah” tables (pt.scribd.com).