SCR and SNCR Compared—Plus other NOx-Reduction alternatives

Two mature routes, one shared chemistry

Selective catalytic reduction and selective non-catalytic reduction have both been in industrial service for decades, both are well understood, and in many plants one of them is the right choice.

Chemically they do the same thing. You add ammonia, or urea that breaks down into ammonia, to the flue gas. The ammonia reacts with nitric oxide and nitrogen dioxide, and what comes out is ordinary nitrogen and water vapour. "Selective" means the reagent attacks the NOx rather than burning in the oxygen that is also present. The difference is how you make that reaction happen: SCR uses a catalyst at moderate temperature, SNCR skips the catalyst and uses heat, and that one choice drives cost, removal rate and behaviour at part load.

How SCR works

In SCR the flue gas passes through a reactor packed with catalyst, shaped as a honeycomb or as plates, with ammonia injected upstream and mixed in before the first layer. The usual catalyst is vanadium oxide on a titanium dioxide carrier, with tungsten for stability; zeolite catalysts, often with iron or copper, suit hotter gas.

The vanadium type works between roughly 250 and 450 °C. Below that window the reaction is too slow; above it the ammonia starts to burn and makes new NOx. Iron- and copper-zeolite catalysts stay active higher, to roughly 550 to 600 °C. Inside the window SCR is very effective: 80 to 95 percent removal is normal, more with extra catalyst layers, and reagent use is close to theory at about one mole of ammonia per mole of NOx. That is why it became the standard for power plants, big engines and waste incinerators.

What SCR asks of the plant

The catalyst is the strength and the weakness at once. Sodium, potassium, calcium, arsenic and phosphorus attack the active sites. This is poisoning, and it is hard to reverse. Washing and chemical regeneration of spent catalyst recover much of the lost activity, but not all of it. Cement kilns, biomass boilers and waste plants carry these elements, so catalyst life there is far shorter than on clean gas.

Dust is a separate problem: fine particles settle in the channels and cover the surface, so gas cannot reach the active sites. That is blinding, and it is why high-dust reactors need soot blowers. Sulphur adds a third: sulphur trioxide and ammonia combine into ammonium bisulfate, a sticky salt that plugs the pores and fouls the air preheater. On sulphur-bearing gas the bottom of the window is often raised to about 320 °C.

Ammonia slip, the unreacted ammonia leaving with the clean gas, is held to a few ppm by permit. It grows as the catalyst ages and contaminates fly ash, which can make the ash unsellable to cement makers. The reagent must also be delivered and stored. Anhydrous ammonia is cheapest but toxic and pressurised; store more than 50 tonnes and the site also falls under European major-accident law, though many plants keep less. Ammonia water is safer but four times the trucked mass; urea is safest and dearest.

How SNCR works, and where it gives ground

SNCR has no catalyst. The reagent is sprayed straight into the furnace through wall lances, where the gas sits between about 850 and 1050 °C. At that temperature the reaction runs on its own. The hardware is little more than a tank, pumps, lances and controls, so it is cheap and fits into a normal maintenance stop.

The price is that temperature window. Below about 850 °C the reaction is too slow and most of the reagent leaves as ammonia slip. Above about 1050 °C the ammonia burns to NOx and makes things worse. The band is only about 200 degrees wide and it moves as load changes, so plants fit several injection levels. Part load stays the hard case.

Removal is lower and less steady than with SCR. Thirty to fifty percent is the normal expectation; cement precalciners with long residence time and a steady profile do better. Reagent use is much higher, one and a half to two and a half times theory, because much never meets NOx at the right temperature. With urea, part of it forms nitrous oxide, a greenhouse gas about 270 times stronger than CO2.

Comparing the two on cost and flexibility

On capital cost the gap is wide. An SNCR system commonly costs a fifth of an SCR system or less: no reactor, no catalyst, no mixing grid, no extra fan for the pressure drop. For a plant with a moderate NOx target and a tight budget, that gap decides it.

On running cost the order can reverse. SCR uses far less reagent per tonne removed, but adds catalyst washing or replacement every few years, disposal of spent catalyst as hazardous waste, higher fan power, and reheating in tail-end layouts where the gas has already cooled. SNCR carries none of that catalyst cost and its equipment is simple to maintain, but it consumes reagent steadily and that bill never stops.

On real gas each has its own weak spot. Dust and poisons hurt SCR and barely matter to SNCR, whose reagent goes into the furnace where dust is part of the process. Load swings hurt SNCR and are tolerable for SCR, as long as the gas stays in the catalyst window.

What happens to the nitrogen

Both methods destroy the NOx: the nitrogen leaves as N2, the hydrogen ends up in water vapour, and the stack gas is cleaner. That is a real result, and it is why both are written into permits.

The ammonia or urea you inject was made by the Haber-Bosch process. A tonne of ammonia takes roughly 30 to 40 gigajoules to make, and most of that goes into producing the hydrogen from natural gas rather than into the synthesis loop itself. Energy is spent to bind nitrogen, the product is shipped and stored under safety rules, and then used to turn another nitrogen compound back into plain air. Nothing is recovered.

Cutting NOx in the flame

Before any cleanup device there is the combustion itself. A furnace makes NOx in three ways. Thermal NOx forms when nitrogen and oxygen from the combustion air combine at high flame temperature, and the rate climbs steeply above about 1300 °C; it dominates on gas and oil. Fuel NOx comes from nitrogen bound in the fuel, and on coal, waste and biomass it usually dominates instead, often 70 to 80 percent of the total. Prompt NOx, from hydrocarbon radicals attacking N2 in the flame front, is the smallest share.

Low-NOx burners shape the flame so the hottest zone is short of oxygen and the oxygen-rich zone is cooler, cutting NOx by 30 to 50 percent. Air staging, usually overfire air, holds part of the combustion air back so the burner zone runs fuel-rich and forms less NOx, then adds it higher up to finish the burnout. Fuel staging, or reburning, works on NOx that already exists: extra fuel injected downstream makes a reducing zone that turns it back into N2. Flue-gas recirculation works differently again: cooled flue gas goes back into the combustion air, diluting the oxygen and soaking up heat, cutting thermal NOx by 20 to 50 percent on gas and oil.

These in-furnace measures are the cheapest per tonne removed and should come first, but they have limits. A cooler flame means more carbon monoxide and unburnt carbon in the ash, reducing zones can corrode boiler tubes, and in cement, glass and metals the flame shape is part of product quality. On their own they rarely meet today's limits.

Oxidation and scrubbing: the problem moves into water

There is also a wet route. Nitric oxide, most of the NOx in flue gas, hardly dissolves in water, so you oxidise it first. Ozone, hydrogen peroxide, chlorine dioxide or sodium chlorite turn NO into nitrogen dioxide and higher oxides, which dissolve easily; a caustic or lime scrubber then washes them out. The oxidant is the cost: ozone takes about ten kilowatt-hours per kilogram to make, and the first step alone, NO to NO2, needs about 1.6 kilograms per kilogram of NO. Oxidising further to N2O5, which is what makes the washing easy, needs about 2.4 kilograms, and plants dose above stoichiometry on top.

The larger point is where the nitrogen ends up: in the scrubber water as nitrate and nitrite. That water then needs a discharge permit, often biological treatment or evaporation, and carries a salt load. Nitrite is toxic to fish. The NOx is out of the air, but it has been moved rather than removed.

A third route: holding the NOx on a solid

Adsorption takes a different path. Instead of converting NOx with a reagent or washing it into water, you let it stick to a solid surface. When the solid is full you regenerate it, driving off what it holds, and put it back to work. The historic barriers were water competing for the same sites and regeneration energy.

Krajete, an Austrian gas-purification company, builds regenerative zeolite adsorbers for this duty. They run at ambient temperature, so there is no reheating and no temperature window to hold. They use no ammonia, no catalyst, no process water and no reagents. The adsorbent is regenerated with gentle heat and goes back into service, cycle after cycle.

Published results sit in the 1 ppm class at the outlet in industrial service. In CO2 matrices, 99.9 percent capture has been demonstrated, and CO2 polishing has reached below 10 ppb. On a cement kiln, a slipstream of about 100 Nm3/h of hot, dusty gas gave roughly 10 ppm NOx at the outlet: day one went into conditioning the bed, and uptake stabilised above 85 percent on day two. The pilot envelope runs from 50 to 18,000 m3/h, containerised.

What separates this route from the others is regeneration. The NOx comes off concentrated, so it can be recovered as feedstock for nitric acid or fertiliser instead of being destroyed or flushed away.

The scale needs stating plainly: Krajete delivers to pilot scale, and a full-scale plant is built by the client's own engineering partner, with Krajete supplying the design basis, the media and the licence. So this is not yet a like-for-like swap for an SCR reactor, and it does not make SCR or SNCR obsolete. Those remain the right answer in many plants, and in-furnace measures remain the cheapest first step. What is new is a third option to test, one that works cold and gives the nitrogen back.