Shipping’s scrubber math: Wet beats dry on SOx, SCR owns NOx — but watch the wastewater

Since the IMO Annex VI global cap of 0.5% sulfur in marine fuel took effect in January 2020, operators have faced a binary: burn pricier low-sulfur fuels or install exhaust gas cleaning systems. The market largely chose scrubbers that strip sulfur oxides (SOx) from stack emissions and selective catalytic reduction (SCR) for nitrogen oxides (NOx). The upshot: wet scrubbers dominate SOx control at roughly 95% removal efficiency, while SCR delivers about 80–90% NOx cuts for Tier III compliance — if vessels budget for reagent and handle water and waste streams to spec (nepis.epa.gov) (wartsila.com) (file.scirp.org).

These choices are changing bunker economics: as of 2024 about 5% of the global fleet — roughly 5,000 ships — sails with scrubbers, largely the oil-thirsty tankers, bulkers, and boxships that account for around 25% of high-sulfur fuel use (sciencesources.eurekalert.org). Analysts estimate scrubber investments typically pay back in 3–5 years on fuel spreads; one study found over 95% of scrubber‑equipped ships recouped their costs within five years, producing a €4.7 billion net benefit by 2022 (dieselnet.com).

SOx scrubbing technologies compared

Wet scrubbers — the predominant marine choice — spray water into the exhaust to dissolve SO₂ into sulfuric/sulfate form. In open‑loop seawater systems, the ocean’s natural alkalinity neutralizes the acid; in closed‑loop freshwater systems, the washwater recirculates and is neutralized with caustic soda or lime; hybrid units switch modes at sea or near port (wartsila.com) (wartsila.com). For closed‑loop units dosing alkali precisely, metering via a dosing pump is standard practice.

On performance, wet systems are proven: U.S. EPA testing reports about 95% SOx removal, with Wärtsilä citing roughly 97.1% efficiency when converting 3.5% fuel sulfur to 0.1% (nepis.epa.gov) (wartsila.com). They also remove most direct sulfate particulate matter from the exhaust stream (nepis.epa.gov).

Dry scrubbers work differently. They inject powdered or granular alkaline reagents (sodium or magnesium compounds) into the exhaust; the resulting dry solid sulfates are collected in a baghouse or cyclone. They avoid liquid effluent but produce solid waste that must be handled ashore. In practice, dry capture generally lands in the 80–90% SO₂ range versus roughly 95–99% for wet scrubbers, and dry units require bulky fixed beds and particulate filters that add pressure drop and OPEX; guides note dry scrubbing “is more costly” and cannot capture all contaminants (nepis.epa.gov) (marinediesel.co.in). Unsurprisingly, virtually all marine SOx scrubbers sold to date are wet; no major manufacturer is marketing pure dry systems for ships.

Energy, water, and solids: operating realities

Open‑loop wet scrubbers move water at scale. A 50 MW installation typically circulates about 4,500 m³/h of seawater, driven by ~500 kW of pumping power through roughly 30‑inch piping; back‑pressure climbs about 1–3% on the engine (file.scirp.org) (file.scirp.org). Closed‑loop systems consume on the order of 0.1–0.2 m³ of freshwater per MWh and bleed 0.1–0.3 m³/MWh; they typically recirculate washwater and polish it, discharging only a small stream (wartsila.com) (file.scirp.org). Dry systems skip water entirely but consume reagents — roughly 10–30 kg/MWh — and require disposal of the spent solids.

Capital outlays for SOx scrubbers broadly align: retrofits typically run about $1–5 million per vessel, with OPEX driven by pumping power and occasional packing replacement for open‑loop units; closed‑loop units add caustic consumption and bleed handling, and dry systems bear continuous sorbent supply and solid waste management. Where polishing filters are specified, vessels often standardize around robust elements such as a cartridge filter to capture fine particulates.

Adoption and fuel economics

The scrubber cohort — about 5% of the fleet, or ~5,000 ships — accounts for an outsized 25% of high‑sulfur fuel consumption, skewed toward large tankers, bulk carriers, and container ships (sciencesources.eurekalert.org). Economics have favored the move: with typical fuel spreads, multiple analyses state scrubber CAPEX is often recovered in 3–5 years; indeed, over 95% of installations paid back within five years, delivering an estimated €4.7 billion net benefit by 2022 (dieselnet.com).

NOx control via SCR systems

To meet IMO Tier III NOx limits — an 85% cut from Tier I for engines placed on or after 2016 in NOx Emission Control Areas (NECAs) — vessels have largely standardized on SCR. SCR injects a urea solution (often AUS32, also known as diesel exhaust fluid) into hot exhaust; urea thermolyzes to ammonia, which reacts on a vanadium or zeolite catalyst to convert NO and NO₂ into N₂ and H₂O. In service, SCR systems remove about 80–90% of NOx, the level required for Tier III; industry and academic sources call it the “best technology to achieve Tier III compliance” (file.scirp.org) (chiefengineerlog.com).

SCR effectiveness hinges on exhaust temperature (typically ~300–400 °C) and space velocity. High‑pressure SCR (upstream of the turbocharger) and low‑pressure SCR (downstream) are both used; the latter can require fuel scrubbing or particulate traps to protect catalysts. Ammonia slip is minimized with precise injection control and catalyst design. Urea consumption runs on the order of ~5–10% of diesel fuel by mass; one IMO‑aligned analysis estimated a 10 MW engine using SCR for ~1,500 hours annually would require about A$950k of urea over 25 years (iaccsea.com). For assured reagent quality, operators frequently procure marine‑grade DEF such as AUS32/DEF.

Costs scale with hours in NECAs. For a 10 MW engine on a 20,000 DWT ship spending ~1,500 h/yr in an NOx ECA, modeled CAPEX is about $500k (hardware and install), urea roughly $950k over 25 years, catalyst recharges about $450k, and maintenance ~$150k — totaling approximately $1.8–2.0 million across the life, or around $75k per year (iaccsea.com) (iaccsea.com). A year‑round NECA profile (≈8,000 h/yr) drives urea costs toward ~$5 million over 25 years (iaccsea.com).

There are operational upsides: SCR lets engines run at optimal combustion settings, with the slight back‑pressure penalty offset by timing, and little impact on CO₂. Unlike exhaust gas recirculation, SCR does not limit engine load or increase particulates. Wärtsilä notes SCR can be retrofitted on existing engines, and over the past decade it has become standard for Tier III ships; in practice, NECAs drive NOx “burn rates” to ~10–20% of pre‑2016 levels, achievable via SCR or a fuel switch to gas/LNG (file.scirp.org).

Washwater composition and onboard treatment

Wet scrubbers trade emissions to water. The resulting washwater is blended to near‑neutral pH yet carries dissolved sulfates and salts, heavy metals such as iron, copper, nickel, and vanadium (from fuel and corrosion), polycyclic aromatic hydrocarbons (PAHs), and particulate solids; the systems also remove most direct sulfate particulate matter, which then shifts to the liquid phase (nepis.epa.gov) (nepis.epa.gov). Tests report PAHs often in the hundreds of parts per billion (nepis.epa.gov).

Onboard treatment lines typically start with multicyclones or hydrocyclones to remove oil and soot, followed by settling tanks for solids; the remaining water is neutralized by blending a reaction water bleed or by alkali dosing before any overboard discharge (nepis.epa.gov). Where free oil separation is part of the train, shipyards slot in compact units designed for oil removal. Closed‑loop scrubbers recirculate and polish, bleeding only a small stream for treatment; open‑loop models discharge continuously after treatment. All systems continuously monitor effluent quality — MARPOL requires real‑time pH, oil/PAH content, and turbidity measurements (turbidity is light scatter from particles, reported as FNU/NTU) (puc.overheid.nl).

Some designs now incorporate advanced oxidation or membrane units to polish washwater onboard, though with added cost and energy. Where membranes are selected, engineers frequently specify ultrafiltration for fine solids and emulsions management or a broader membrane system architecture as a compact, high‑rejection polishing stage.

Discharge limits and regional restrictions

IMO guidance MEPC.259(68) sets numeric limits. Turbidity must be ≤25 FNU on a 15‑minute average, measured downstream of the onboard treatment; PAHs are capped via a flow‑dependent formula (often expressed as phenanthrene equivalence) (nepis.epa.gov). In practice, restrictions are tightening: as of 2023, around 45 jurisdictions had moved to prohibit or limit washwater discharges — especially from open‑loop units — via 93 distinct measures documented across ports and EEZs; countries including Denmark, Germany, France, Portugal, Turkey, and China restrict open‑loop discharge outside deep sea (theicct.org) (sciencesources.eurekalert.org).

Indonesia’s current position allows scrubbers to meet the 0.5% sulfur rule for Indonesian‑flag ships, while requiring compliance with MARPOL discharge rules in its waters (manifoldtimes.com).

Sludge handling and port reception

Scrubber sludge — the solids captured from washwater — is hazardous, for example categorized under EU waste code 10 01 18* due to toxic metals. MARPOL Reg. 17 obliges ports to provide reception facilities; shipowners typically pump liquid sludge ashore or dewater to reduce volume, balancing the cost of liquid versus dried transport (alfalaval.com) (alfalaval.com).

Typical flows, monitoring, and polishing

Operationally, scrubber washwater outputs land around 0.1–0.5 m³ per MWh of engine output, depending on design. For a 50 MW engine running 20 hours, an open‑loop scrubber might process about 90,000 m³ of seawater (most discharged), while a closed‑loop system might bleed only ~2–5 m³ of concentrated waste for later offload; sulfuric acid is neutralized on the fly, but metals and PAHs require filtration and capture (file.scirp.org) (wartsila.com). Continuous monitoring of pH, PAH, and turbidity is mandatory per MARPOL (puc.overheid.nl), and well‑treated discharge targets neutral pH (~6–9), suspended solids below the 25 FNU/NTU turbidity threshold, and PAHs below the MEPC formula limits (nepis.epa.gov).

Where operators want additional polishing beyond cyclones and settling, compact elements and housings designed for marine service are often specified; for example, stainless housings paired with fine media are common, as are systems oriented to primary separation. In closed‑loop trains, precise reagent dosing and polishing filter maintenance are central tasks; many yards specify DEF logistics alongside SCR hardware, including marine supply chains for urea such as high‑quality DEF.

What the data implies for retrofits

The evidence is consistent. Wet scrubbers deliver ≥95% SO₂ reduction and dominate marine installations, but they generate wastewater that must comply with tight standards (e.g., turbidity ≤25 FNU) and with local restrictions that now span dozens of jurisdictions (nepis.epa.gov) (nepis.epa.gov) (theicct.org) (sciencesources.eurekalert.org). Dry scrubbers avoid liquid discharge but yield solid residues and lower removal rates, making them rare at sea. SCR removes ~90% of NOx, enabling Tier III, at a lifetime cost modeled around $1.8–2.0 million for a 10 MW engine running ~1,500 hours per year in a NECA, or about $75k annually; reagent costs escalate with hours (≈$5M over 25 years at ~8,000 h/yr) (file.scirp.org) (iaccsea.com) (iaccsea.com) (iaccsea.com).

For planning, that translates into engineering scrubber trains with continuous monitoring, onboard treatment, and reception arrangements for hazardous sludges — and, where necessary, onboard polishing using membranes or oxidation. In these trains, specifying maintainable media and dosing hardware matters; it is why compact components like a polishing filter and accurate chemical dosing recur in shipyard bills of materials.

Sources: IMO/EPA guidance, peer‑reviewed and industry reports, and manufacturer manuals support all figures cited, including scrubber efficiencies and energy/water metrics (nepis.epa.gov) (wartsila.com) (sciencesources.eurekalert.org) (file.scirp.org) (puc.overheid.nl) (alfalaval.com). All URLs are as cited.