Inside the refinery fight against furnace coke: chemicals, coatings, and faster decokes

Call it the paradox of cracking: push hydrocarbons hard enough to make value — and risk laying down solid carbon. Coke (solid carbon deposits) forms on the inner surfaces of cracking furnace coils when long‑chain hydrocarbons crack and polymerize. The rate of coke deposition depends strongly on feed quality (aromatics, sulfur content, metal contaminants) and coil metallurgy (researchgate.net) (pubs.acs.org).

In practice, fouling can limit run lengths to a few weeks or less. High coke buildup raises coil pressure drop and furnace metal temperatures, forcing shut‑downs for cleaning. Selecting corrosion‑resistant alloys or applying anti‑coking coatings can cut coke rates dramatically: one study noted >10× differences in coking rate between alloys (pubs.acs.org). In fixed‑tube steam crackers, coke completely insulates the tube, so operators shut down for decoking when pressure drop or skin temperature limits are reached. Preventing coke — and efficiently removing it when it forms — is critical to maximizing unit availability.

Chemical anti‑coking additives (definitions and dosing)

Refineries often inject specialty chemicals (“anti‑foulants”) at low concentrations (tens to hundreds of ppm) into the cracking feed to inhibit coke. These additives typically combine passivators (compounds that form a thin metal oxide layer on the tube surface, blocking catalytic sites for polymerization) with radical scavengers (molecules that react with radical intermediates in the gas phase to prevent long‑chain formation). An Arkema patent describes injecting dimethyl sulfide/disulfide (DMS/DMDS) along with an aromatic radical scavenger (α‑methylstyrene dimer, terpinolene or thiophene) at ~100–300 ppm each in the feed (patents.google.com) (patents.google.com). Typical dosing is on the order of 10–1000 ppm of each component (most fields use ~50–500 ppm total) (patents.google.com).

Field trials have shown large benefits. In a commercial ethylene furnace study, adding ~5–20 ppm of a proprietary inorganic salt blend (with water) tripled run length: the heater operated ~150 days with additive vs only ~40 days without (ogj.com). During 180 days with additive, coil pressure drop stayed flat and ethylene/propylene yields were ~2% higher (ogj.com). Analysis showed ~20–30% higher CO₂ production in the flue gas, indicating the additive was actively gasifying deposited coke as it formed (ogj.com). In another run, heavy tar (condensable side‑products) increased such that the recovered steam from transfer‑line exchangers rose ~30% (due to slower fouling) (ogj.com).

In summary, additives at tens of ppm slowed coke formation by 2–4×, extended run lengths by factors of 3–4, and even activated partial on‑stream gasification, all with negligible impact on product quality (ogj.com) (ogj.com) (ogj.com). Common commercial anti‑coking formulations include sulfurized oils or salt mixtures; examples cited are Nalco’s “COKE‑LESS” and Chevron‑Phillips’ “CCA‑500” (researchgate.net). Operators monitor coil pressure drop and add more chemical as needed to hold fouling in check; because additives often end up in bottom oils, environmental impact is described as minor — one field trial reported that added coke (and chemicals) reported entirely to the tar oil and was later burned or used for carbon black, with no effluent issues (ogj.com) (ogj.com). Dosing at these low levels makes accurate chemical metering central to operations (for example, via a dosing pump).

Coil material and coatings (preventative measures)

Selected coil materials or coatings can greatly reduce coke. Advanced alloys or ceramic coatings form stable oxide layers that inhibit coke adhesion. AlcroPlex (an alumina‑based coating) doubled furnace run lengths in one ethane‑cracking test (30→60 days) (pubs.acs.org). A spinel‑coated alloy (Kubota ANK 400) extended three commercial furnaces from ~33 days to >400 days between decokes (pubs.acs.org).

These treatments reduce catalytic coke growth and carburization. Combining good metallurgy with feed additives yields the best outcome: slurry feeds or heavy naphtha cracked in coated coils still ran far longer than uncoated tubes. Selection of alloy or coating must consider cost: high‑chrome or silicon‑carbide materials raise capex, but permit much longer cycles and can be operated at significantly higher conditions (pubs.acs.org) (pubs.acs.org).

Steam–air decoking procedure (process steps)

When coke does accumulate, it is usually removed by a controlled steam/air burn. The furnace is taken offline and each coil pass isolated. A dedicated decoking system routes medium‑pressure steam and air into the coil, and vents the tube effluent to a quench pot. Operators typically inject steam only into the cool coil while firing the burner externally, creating thermal shock: the sudden heating/contraction cracks the coke layer (1library.net). The steam flow then sweeps most of the loose coke fragments into the decoke drum (where spray water quenches them). When effluent runs clear, a small flow of air is added (usually premixed with steam or after a steam purge) and the flame is re‑ignited; the oxygen burns the remaining coke in‑place, indicated by a rise in coil metal temperature and CO/CO₂ in the exhaust (1library.net) (cokebusters.com).

The process is controlled to avoid hotspots or mechanical stress: only one pass is cleaned at a time, tube metal temperature is monitored, and steam is applied to non‑decoking passes to prevent dry‑outs (scribd.com) (scribd.com). Steam–air decoking typically takes 1–2 days per heater (depending on fouling). Textbook procedures report about 0.7 kg/cm² ΔP per 100 ft of coil when steaming (as a guide) (scribd.com). Total decoke times can range from ~6 h (light fouling) to 1–3 days for heavy coke (scribd.com).

Drawbacks include emissions — the steam burn releases CO–CO₂–H₂ into the stack (cokebusters.com) — so many plants collect the off‑gases to flare or treat them. Procedures emphasize safety (e.g., open coil to air only after thorough steam flush to avoid explosive mixtures). Crew costs are high: typically 1–2 dedicated operators must attend the decoke continuously.

Other decoking methods (offline pigging, spalling, chemical)

Mechanical “pigging” (a dense piston with metal bristles pumped through the coil) has become a preferred alternative in many refineries. It removes nearly 100% of internal fouling, recovering virtually the full heat‑transfer area (cokebusters.com). One refiner reported pigging two vacuum‑heater coils in 36 h and 24 h (two runs) versus 96+ hours needed for the equivalent steam decoke (heatflux.com). Another achieved coke removal for a 4‑pass furnace in 18 h per pass (compared to ~96 h by steam) (heatflux.com) (heatflux.com).

Heat‑transfer performance after pigging matched steam decokes (within measurement error) (heatflux.com) (heatflux.com), while emissions are negligible because pigging is a closed, pumped process; plant personnel can focus on other tasks (steam decokes typically tie up operators for days) (heatflux.com) (cokebusters.com). Downsides include equipment cost and the need to track pigs carefully — one refinery noted that lost pigs have been found lodged after the process, so inventory control is vital (heatflux.com).

In‑line spalling (on‑stream pulsed steam) uses high‑velocity steam bursts to cause rapid heating‑cooling “shocks” that can dislodge some coke without a shutdown; in practice it gives only a partial cleaning (mostly in the radiant section) and is not a full substitute for off‑line decokes (cokebusters.com). Chemical cleaning (circulating solvents or acids) is generally avoided in cracking heaters because it is slow, complex, and produces hazardous waste. For example, an acid wash might dissolve coke but then leaves the coils corroded and requires neutralization.

Performance outcomes and trends (run length and yields)

Data consistently show large improvements from advanced decoking strategies. In one industry survey, pigging cut heater downtime by 60–80% compared to steam burns (heatflux.com) (heatflux.com). Some refineries report 2–3× run‑length extensions: a crude visbreaker heater in Europe went from ~5‑day steaming cycles to ~2‑day pigging cycles (with equal heat recovery) (heatflux.com), and fuel consumption dropped ~1.8% (due to cleaner tubes) (heatflux.com).

In ethylene units, additives plus coatings have enabled furnaces to safely operate well over 100 days between decokes — the AlcroPlex case doubled runs to 60 days (pubs.acs.org), and modern working alloys like ANK 400 claim >400‑day cycles (pubs.acs.org). Key metrics: aftermarket anti‑foulant dosing (~100–300 ppm) yields measurable payback — published trials showed ethylene yield rising ≈2% and furnace duty (steam keyed‑off) rising ~2–3% under additive treatment (ogj.com) (ogj.com).

Extended run lengths directly translate to less maintenance cost and more throughput. The field trial above achieved 3–4× longer runs (ogj.com). Pigging’s ROI is similarly clear: one operator noted that avoiding a 96‑h steam burn (3–4 shift crews) in favor of a 36‑h pigging saved essentially 3 shifts of labor and avoided noise and emissions issues (heatflux.com).

Regulatory and operations context (emissions and safety)

Although major decoking emissions are CO/CO₂ (flammable but typically managed by flares or scrubbers), cracking plants must comply with local air permits. For instance, Indonesian refinery regulations require controlling any CO, NOx or particulate emissions during maintenance operations. Minimizing steam burns (by pigging or partial on‑stream decoke) helps reduce such flaring. Care is also taken to avoid flooding quench drums with petroleum contaminants.

Recommendations (layered approach)

The practical playbook is layered: minimize coking precursor levels by good feed purification (dry, low‑metal feeds) and optimized steam dilution; maintain coatings or high‑alloy tubes on heavily fouling heaters; inject anti‑foulant additives continuously or on a demand basis (10–500 ppm) to suppress coke; and plan decokes proactively — schedule pigging over steam‑air burns when possible — to cut downtime. With these measures, operators typically see significant run‑length gains (often 2–4×) and can extend furnace campaigns to several months with minimal coke accumulation.

Source notes and citations

Cause and impact of coke formation, alloy and coating effects, and run‑length differences are documented in researchgate.net and pubs.acs.org (pubs.acs.org) (pubs.acs.org). Additive chemistry, dosing ranges (10–1000 ppm each; ~50–500 ppm total typical), and specific components (DMS/DMDS with α‑methylstyrene dimer, terpinolene or thiophene at ~100–300 ppm each) are from patents (patents.google.com) (patents.google.com), while field performance (tripled run length to ~150 vs ~40 days; ~2% higher ethylene/propylene yields; ~20–30% higher CO₂ in flue; ~30% rise in recovered steam from TLE; 2–4× slower coke formation; 3–4× longer runs; environmental note that additives reported to tar oil later burned or used for carbon black; “neither acidic nor basic characteristics”) is reported by ogj.com (ogj.com) (ogj.com) (ogj.com) and researchgate.net.

Steam–air decoking steps, controls, and timeframes (1–2 days per heater; ~0.7 kg/cm² ΔP per 100 ft guide; ~6 h to 1–3 days total) and emissions (CO–CO₂–H₂) are detailed in 1library.net, scribd.com (scribd.com) (scribd.com) (scribd.com) and cokebusters.com.

Pigging performance (nearly 100% removal; 36 h and 24 h vs 96+ h; 18 h per pass vs ~96 h; matched heat‑transfer performance; negligible emissions; operator time saved; note on lost pigs) is drawn from cokebusters.com and heatflux.com (heatflux.com) (heatflux.com) (heatflux.com). In‑line spalling limitations are noted by cokebusters.com.

All figures and procedures are from these engineering publications and field evaluations.