Inside FCC emissions: what actually works for SOx, NOx and dust
The fluid catalytic cracker’s regenerator drives up to ~90% of a refinery’s SOx. Additives can cut SOx by 30–>95%, low‑NOx designs trim NOx by tens of percent, and ESPs finish the job on particulate—while SCR is the pricy lever for ultra‑low NOx.
The FCC regenerator is the emissions hot spot: it accounts for up to ~90% of refinery SOₓ release, because about 10–20% of the feed sulfur ends up as coke on the catalyst and is oxidized to SO₂/SO₃ during regeneration (www.digitalrefining.com) (decarbonisationtechnology.com) (www.digitalrefining.com). Sulfur in the coke burns to roughly 90% SO₂ and 10% SO₃ (decarbonisationtechnology.com).
To control that, refineries toggle between very low‑sulfur feed, deep hydrotreating, wet gas scrubbing (WGS; a caustic/absorption “end‑of‑pipe” system), or—more flexibly—SOₓ reduction additives injected into the regenerator (decarbonisationtechnology.com). What follows is a ground‑truth read on what each pathway actually delivers for SOₓ, NOₓ, and particulate matter (PM), with the numbers and trade‑offs cited inline.
SOx reduction additives in FCC
SOₓ reduction additives are metal‑oxide catalyst particles—often Ca/Mg‑Al spinels or layered double hydroxides (hydrotalcite‑type materials)—continuously injected into the regenerator. In the hot regenerator, they react with SOₓ (especially SO₃), converting it into forms that are stripped as H₂S in the reactor. That H₂S is then quenched in the dry‑gas unit and recovered as sulfur (decarbonisationtechnology.com).
In practice, additive programs report 30–>95% SOₓ removal, depending on feed sulfur, burn mode and dosage (ptqmagazines.digitalrefining.com). One mixed‑feed, full‑combustion unit with low feed sulfur (<0.3% S) routinely achieves stack SOₓ <25 ppm using additives only (ptqmagazines.digitalrefining.com). By contrast, partial‑burn units or high‑sulfur feeds may only see 30–70% reduction even with additives (ptqmagazines.digitalrefining.com).
Performance is not linear. As targets tighten, the “pickup factor” (lb SOₓ removed per lb additive) falls and required dosage grows rapidly (ptqmagazines.digitalrefining.com). Industry guidance notes that beyond ~90% capture, consumption skyrockets and a WGS scrubber may become more economical (www.digitalrefining.com) (ptqmagazines.digitalrefining.com). Tom Ventham notes that to reach ultra‑low levels (e.g. <5 ppm SOₓ), a WGS is often chosen instead of additives (www.digitalrefining.com).
Modern additives lean on high‑MgO supports (hydrotalcite‑type). Hydrotalcite‑based additives allow ~50% higher MgO content than traditional spinels, boosting SOₓ capacity, and about 80% of current SOₓ additives are hydrotalcite‑type (decarbonisationtechnology.com). In a refinery trial, Johnson Matthey’s “SSG X2” cut SOₓ additive consumption by 11–25% for the same emission control versus the incumbent additive (www.digitalrefining.com). Johnson Matthey also reports that an optimized additive can substantially reduce the cost of meeting a given SOₓ limit (www.digitalrefining.com).
Economically, additives are generally favored: they require modest capital (an injection system) and add marginal OPEX, while WGS or deep feed desulfurization require large capex and ongoing expense. The additive approach also lets the FCC run high‑sulfur feeds (on sweet crude or heavy oil) while meeting emissions, improving margins (www.digitalrefining.com). In the supplier’s words, “SOₓ reduction additives are the preferred solution for many FCC units” (www.digitalrefining.com). In short, additives typically achieve tens to >90% reduction—often delivering feed S to stack SOₓ conversion below ~25 ppm—and are cost‑effective for most limits, with WGS reserved for the very lowest targets (ptqmagazines.digitalrefining.com) (www.digitalrefining.com) (ptqmagazines.digitalrefining.com).
NOx control: burner staging and SCR
NOₓ in FCC flue gas arises mostly from fuel nitrogen bound in coke (not thermal NOₓ from air N₂). Only a few percent of the feed nitrogen ends up as stack NOₓ, since much of the coke‑N reduces back to N₂ in the regenerator’s reducing environment (www.digitalrefining.com) (www.digitalrefining.com). Conventional FCC units typically emit on the order of 100–200 mg/Nm³ NOₓ (10s of ppm) if uncontrolled.
Low‑NOₓ regenerator designs—staged air injection or specialized burner heads—lower peak flame temperatures and enhance in‑situ NOₓ reduction. Shell’s “low‑NOₓ regenerator” uses dual dense and dilute zones to foster CO and HCN reduction back to N₂ (www.digitalrefining.com). In practice, well‑designed low‑NOₓ systems have achieved stack NOₓ as low as ~40 ppm (www.digitalrefining.com). Applying low‑NOₓ burners to the downstream CO boiler can also drastically cut residual thermal NOₓ (www.digitalrefining.com).
Field experience shows these designs routinely land in the 100–200 mg/Nm³ range, with reductions of tens of percent relative to standard full‑burn operation. With ample reducing CO and good mixing, Shell notes N₂ formation is promoted so that >95% of coke‑N is not emitted as NOₓ (www.digitalrefining.com).
Selective catalytic reduction (SCR—ammonia or urea injection over V₂O₅ or zeolite catalyst) is an end‑of‑pipe fix that converts >80–90% of NOₓ to N₂ and can push outlet concentrations to single‑digit ppm under steady operation. It carries high capital (large vessels/catalyst modules) and meaningful O&M (ammonia/urea consumption, catalyst replacement, potential SOₓ‑related fouling), plus ammonia slip and catalyst longevity concerns. Accurate injection is typically handled by a dosing pump, and urea is often supplied as DEF, such as terragard-def.
The comparison is stark. Low‑NOₓ burners are relatively simple and low‑cost retrofits with minimal consumables, but they typically cannot meet the most stringent NOₓ limits alone and rely on available CO/NH₃ from the coke (www.digitalrefining.com). By contrast, SCR can reliably achieve ultra‑low NOₓ: one U.S. refinery (Tulsa) agreed to target ~20–40 ppm NOₓ at 0% O₂ by installing a two‑stage SCR system (www.digitalrefining.com). This suggests that when permits demand NOₓ below ~50 mg/Nm³, SCR may be necessary. In summary, low‑NOₓ staging can significantly lower NOₓ (often roughly halving it), while SCR can reduce remaining NOₓ by an order of magnitude at much higher cost (www.digitalrefining.com) (www.digitalrefining.com). Many companies therefore combine advanced regenerator design to achieve perhaps 50–100 mg/Nm³, and resort to SCR/SNCR (selective non‑catalytic reduction) only if further reduction is mandated by regulation.
Particulate control: cyclones and ESPs
FCC regenerators shed catalyst fines (“cat dust”). Primary control is in‑situ multistage cyclones, which knock out most >10 μm particles. Two or more stages typically clean flue gas down to ~200–400 mg/Nm³ of particulates; higher levels often indicate cyclone erosion or upsets (www.digitalrefining.com). In mass terms, cyclones handle ~98–99% of the catalyst, but they are not sufficient to meet stringent PM standards.
Electrostatic precipitators (ESPs) apply high‑voltage fields—on the order of ~10–20 kV/cm—to charge and capture the remaining submicron particles, achieving >99% removal of the dust that escapes the cyclones and yielding stack dust in single‑digit mg/Nm³. Industry experts emphasize that “the most effective” PM capture on an FCC is via a high‑voltage ESP (www.digitalrefining.com)—with one Q&A noting operation “upto13 kV at any plant.” For context, the Indonesian standard for FCC regenerator emissions is 50 mg/Nm³ for particulate (123dok.com), and many industrialized countries target 10–30 mg/Nm³ or less.
Todd Foshee (Shell) notes typical regenerator cyclones reduce dust to ~200–400 mg/Nm³—and “if particulate levels are much higher than those levels, troubleshooting is needed”—so an ESP is generally required from that cyclone‑leaveover to meet regulatory limits (www.digitalrefining.com).
The trade‑offs are straightforward. Cyclones are rugged and low‑maintenance but drop off on fine particles (<1 μm). ESPs economically capture those fines but require significant power, periodic rapping/cleaning, and re‑tubing. The combined system—cyclones for bulk removal plus a stack ESP for final polishing—is essentially universal and routinely brings PM from a few hundred mg/Nm³ down to compliance levels (<50 mg/Nm³ per Indonesian rules and often <10 mg/Nm³ in strict markets) (www.digitalrefining.com) (www.digitalrefining.com).