From flue‑gas scrubbers that target ≈90–95% sulfur dioxide to vapor recovery units capturing 95–99% of hydrocarbons, the refinery emissions toolkit is mature, data‑heavy, and full of trade‑offs. The bigger question is cost, complexity, and how far regulations push plants toward premium controls like SCR.
Refinery stacks still face three stubborn letters: SO₂, NOₓ, and VOCs. The hardware to tame them is well proven—think wet limestone/lime scrubbers, low‑NOₓ burners, selective catalytic reduction, and vapor recovery units—but performance depends on design choices, reagent strategies, and how tight the limits are.
The headline numbers are striking. Wet flue gas desulfurization (FGD) is routinely designed around ≈90% SO₂ removal (EPA), with field tests in the 88–96% band—and even >99% under ideal conditions (EPA field study).
For NOₓ, combustion tweaks via low‑NOₓ burners typically deliver reductions in the tens of percent—roughly 40–70% in many deployments (Modern Power Systems)—while selective catalytic reduction (SCR) is the high‑horsepower option, with about 80–90%+ removal and “90 percent” in EPA techno‑economic assessments (EPA).
On the vapor front, recovery systems on tanks commonly capture ∼95% of hydrocarbons (Chemical Engineering; IPIECA), while loading‑rack units can hit “99%-plus” recovery (Zeeco).
SO₂ control: wet versus dry flue‑gas scrubbing
Refinery furnaces and boilers firing high‑sulfur fuel typically install flue gas desulfurization (FGD) units. In a wet limestone/lime scrubber, SO₂ is absorbed in an alkaline slurry; systems are engineered for very high removal—design targets are often ≈90% SO₂ reduction (EPA). Pilot testing confirms this: one field study showed SO₂ removal in the range 88–96% (EPA field study), and even >99% under ideal conditions (EPA field study). Modern wet FGDs on coal/oil boilers routinely achieve the mid‑90s percent removal; for example, a high‑alkalinity FGD at Paducah, KY ran >99% particulate capture and 88–96% SO₂ reduction (EPA field study), and dual‑alkali designs can exceed 98–99% efficiency at high pH (EPA dual‑alkali data).
In contrast, dry or semi‑dry FGD schemes (spray dryers, circulating fluids) typically hit only ~70–90% removal due to lower gas–liquid contact. Wet FGDs are often specified around ~90–95% removal (EPA design target; EPA field study), and advanced absorption towers with high L/G ratios (liquid‑to‑gas rates) or dual‑stage scrubbing can push toward 98%+.
Byproducts matter: limestone scrubbing produces gypsum (CaSO₄), which can be sold (drywall, cement) or landfilled, offsetting some costs. To maintain the alkaline environment (high pH) cited in dual‑alkali performance, operators meter reagent precisely—often via a dosing pump—to keep scrubbing chemistry on spec.
Costs and trends: FGDs have high capital and O&M cost but yield large SO₂ cuts. In markets with strict SO₂ caps—Indonesia’s thermal units often face ~300 mg/Nm³ (milligrams per normal cubic meter) SO₂ limits (Cybertig)—adding FGD is the standard solution. New Indonesian power plants (e.g., PLTU Sumsel‑8) have adopted FGD to comply with emissions rules. In refineries, sulfur tends to be removed via process units and a sulfur plant rather than vented, but any fuel‑fired boilers will follow similar FGD practice.
NOₓ control choices: LNB versus SCR
Refinery heaters, boilers, and furnaces generate NOₓ primarily via thermal mechanisms. Two main approaches dominate: low‑NOₓ burners (LNB) that modify combustion, and post‑combustion selective catalytic reduction (SCR). Both aim to meet stringent standards (often ~50–100 mg/Nm³ for new units), but effectiveness and cost diverge.
Low‑NOₓ burners stage the fuel/air mix to limit flame temperature and oxygen, reducing thermal NOₓ. Retrofits typically reduce NOₓ by a few tens of percent, with published examples roughly 40–70% under many conditions (Modern Power Systems). With clean refinery fuels (natural gas/light oil), LNB might achieve the lower end of that range. Combustion staging like over‑fire air can add another ~15–25% reduction (Modern Power Systems), but very low NOₓ limits (<50 mg/Nm³) often cannot be met by burners alone. The upside: modest cost and complexity.
SCR injects ammonia (or urea) into the flue gas upstream of a catalyst. In the right temperature window (typically 300–400°C), NO and NO₂ convert to N₂ and H₂O. Modern systems remove the vast majority—typically 80–90% or more—with an EPA study stating ammonia‑SCR “can achieve 90 percent reduction of NOₓ emissions” (EPA), and industry reviews noting “large NOₓ reductions (>80%)” (Modern Power Systems). Reagent strategy includes ammonia or urea solution; many plants handle urea in forms similar to AdBlue (diesel exhaust fluid), and industrial‑grade DEF options such as terragard-def align with the urea pathway referenced here. Trade‑offs are clear: catalysts and NH₃/urea supply raise capital and O&M, and ammonia slip management is an added task (venting small slip into downstream burners or a “slip catalyst” is often required; Modern Power Systems).
Bottom line comparison: LNB provides a moderate (tens of percent) NOₓ cut at modest cost, whereas SCR provides a very large (≈80–90%+) cut but at significantly higher capital/O&M. Combining sources: Mitsui Babcock reports LNB yield ~40–70% NOₓ reduction on coal burners (Modern Power Systems), whereas a US EPA report cites 90% for ammonia‑SCR (EPA). In practice, many refineries use LNB as the first line of defense and add SCR only when needed for strict limits. In OECD power plants and gas turbines, wide SCR adoption has led to >99% of large coal units in some regions using SCR, but in refineries SCR is typically on only the largest furnaces or boilers subject to stringent regs.
VOC capture: storage tanks and loading racks
VOCs at refineries come mainly from storage tanks and product loading operations. Vapor recovery units (VRUs) or vapor balancing capture vapors displaced from tanks or loading connections and either condense or adsorb them, preventing release to atmosphere. Typical performance is very high: recovery efficiencies usually exceed 90–95%. For tanks, simulation and field data indicate VRUs capture ∼95% of tank vapors (Chemical Engineering), and IPIECA notes onshore VRUs “recover more than 95% of the hydrocarbon vapours from tanks” (IPIECA).
Real‑world outcomes back this up: one detailed case reported that after installing a VRU on aromatic storage tanks, benzene/toluene/xylenes emissions fell ~73–86%, with ambient BTX concentrations down ~63–88% (PMC study). Floating roofs likewise prevent most breathing losses, but VRUs can further capture vapors during filling and product movement. Overall, a well‑designed tank VRU will remove generally >90% of evaporative losses.
Loading racks (truck/rail/ship) also lean on VRUs. Vendors report recovery efficiencies up to ~99% (Zeeco), often corresponding to recovering ~1–2 L of product per 1,000 L loaded (Zeeco). IPIECA notes that VOC recovery on tankers or FPSOs can cut >90% of emissions (IPIECA).
The economics are straightforward: installing VRs typically eliminates the vast majority of what otherwise would vent—and turns it back into sellable product. One vendor notes payback in a few years from recovered product alone (Zeeco). National programs (e.g., US Stage II vapor recovery rules) have long required rack VRUs, and increasingly stringent VOC laws worldwide are pushing similar deployments at terminals. In Indonesia (and globally), standards on VOC (often ~30–100 mg/L loaded, milligrams per liter of product loaded) effectively mandate these systems.
Key performance figures (recap)
FGD: wet scrubbing is designed for ≈90% SO₂ capture and commonly achieves ~90–95% removal (EPA design; EPA field), with advanced designs pushing toward 98%+; dry/semi‑dry typically ~70–90%. LNB: roughly 40–70% NOₓ reduction in many cases; over‑fire air can add ~15–25% more (Modern Power Systems). SCR: typically 80–90%+, with “90 percent” cited by EPA (EPA). VRUs: tanks >95% recovery (IPIECA; Chemical Engineering), case‑study tank VOC cuts of ~70–85% with ambient BTX down ~63–88% (PMC); loading racks “up to 99%-plus,” about 1–2 L recovered per 1,000 L loaded (Zeeco).
Data example note: a US EPA assessment highlighted ammonia‑SCR—“can achieve 90% reduction of NOₓ”—as a reference point for post‑combustion control efficacy (EPA). By analogy, high‑end FGD systems are routinely sized for ≈90% SO₂ removal, with higher capacities (e.g., >95%) required if fuel sulfur or flow rates are above normal (EPA design; EPA field).
Performance and cost data are drawn from industry and regulatory studies (EPA; Modern Power Systems; EPA FGD studies; PMC case study; IPIECA; Chemical Engineering; Zeeco), and recent regulatory/market reports. These figures are typical of current technology in use; actual results may vary by fuel, operation, and equipment design.
