SNCR vs SCR: removing NOx from a small incinerator

NOx removal from incinerator flue gas: SNCR and SCR

A wet scrubber removes acid gases because they dissolve. Nitric oxide, which is the greater part of the nitrogen oxides a furnace produces, does not dissolve to any useful degree, so a train that removes hydrogen chloride efficiently removes almost none of the NOx. That is why the nitrogen oxide limit is met either in the furnace or in a reactor immediately after it, and why the two routes — selective non-catalytic reduction and selective catalytic reduction, SNCR and SCR — are the subject here rather than the scrubber.

The two share a chemistry. A nitrogen-bearing reagent, ammonia or urea, is injected into the gas and reacts with nitric oxide to give nitrogen and water, so the nitrogen leaves as an inert molecule instead of as an oxide.

• A reagent that supplies the reducing agent — ammonia water, anhydrous ammonia or urea solution.
• A temperature band in which the reaction is fast enough to compete with everything else happening in the gas.
• Mixing good enough for the reagent to meet the gas rather than pass beside it, which is the practical limit on NOx removal from incinerator flue gas.
• Residence time in that band, because the reaction is not instantaneous.
• A way of dealing with the reagent that does not react, because it leaves as ammonia rather than disappearing.

Selective non-catalytic reduction for incinerator flue gas

SNCR injects the reagent into the hot gas path and relies on temperature alone. The band usually quoted is roughly 850 to 1,100 °C, and that should be read as a starting method rather than as a specification: it describes where the reaction is fast on a typical gas, and it moves with the composition of the gas and the mixing the furnace provides. Below the band the reagent passes through unreacted and leaves as ammonia slip; above it, the ammonia is oxidised and can produce more nitrogen oxides than it removes.

The band overlaps the temperature the secondary chamber is already held at, and that is the attraction. A plant running at or above 850 °C already owns the conditions, so SNCR adds a lance, a reagent system and a control loop rather than a pressure vessel. The reagent is usually urea solution on a small plant, because it is safer to store and handle than ammonia water.

The limitation is that the band is a band in the gas, not a setting on a panel. Gas temperature moves with load, with the waste and with excess air, and the zone in which the reagent actually meets the gas at the right temperature is a slice of the furnace rather than the whole of it. Removal is therefore partial — the figures quoted for SNCR are of the order of a third to a half of the inlet nitrogen oxides — and what is not removed is paid for twice, once as reagent and once as the ammonia that leaves with the gas.

Selective catalytic reduction for waste incineration

SCR runs the same chemistry on a catalyst, which lets a much lower temperature do the work — broadly the 200 to 400 °C band, read again as a starting method rather than as a specification. The catalyst is typically vanadium oxide on a titanium support, formed into honeycomb or plate elements inside a reactor, and the gas passes through it.

The lower temperature is the advantage and the catalyst is the cost. On a waste plant the question that decides the arrangement is where the reactor sits. Put it upstream of the dust collector and the catalyst sees fly ash, heavy metals and acid gases, which is where deactivation comes from. Put it downstream of the filter and the gas is clean but cold, so it has to be reheated before the reactor and the energy for that has to be found somewhere in the plant. A small incinerator usually has no spare heat to give away, which is the practical reason the choice so often falls to SNCR.

The reactor also needs space, a support structure and a gas path that can be taken out of service for catalyst replacement. Which arrangement fits a given unit, and the catalyst volume that goes with it, are confirmed with the factory for your waste composition.

The temperature window each one needs

Both windows are stated at full load on a gas of a given composition, and both move when the plant moves. What matters is the window the gas actually passes through, over the range of loads the plant runs, not the window that appears on a data sheet.

• SNCR depends on a temperature the chamber is already holding, but the residence time in the band shortens as the load rises and the gas speeds up, and the band itself shifts with the waste.
• SCR depends on a temperature the plant has to produce, so a plant that cannot reheat its gas reliably has a reactor that is either cold and inactive or hot enough to damage the catalyst.
• At low load both lose performance first, which is why a guarantee is normally stated at more than one load rather than at the design point alone.
• The reagent that does not react is not lost quietly: it leaves as ammonia, and that slip is a parameter of its own.

Catalyst poisoning in waste incineration

A waste gas is not a power station gas. It carries alkali metals, heavy metals and sulphur, and all three attack a catalyst.

• Alkali metals in the fly ash — sodium and potassium — deposit on the surface and block the active sites, partly reversibly and partly for good.
• Arsenic binds to the active material and cannot be removed, so the catalyst has a finite life in proportion to what the waste contains.
• Sulphur dioxide with the ammonia forms ammonium bisulphate, which deposits in the cold end of the reactor and blinds the elements.
• Heavy metals and fine ash blind the channels physically if the reactor is placed upstream of the dust collector.
• Vanadium is volatilised above the working band, so the active material itself is lost if the reactor runs hot.

Two of those have consequences beyond the reactor. Ammonium bisulphate deposits foul the ductwork or the heat exchanger behind it, and a vanadium catalyst oxidises some of the sulphur dioxide in the gas to sulphur trioxide, which raises the acid dew point downstream and changes what the duct and the stack have to be made of. The catalyst and the reactor position are therefore decided together with the gas train, not separately from it.

Comparing SNCR and SCR on a small incinerator

• Reagent — both use ammonia or urea, and SNCR uses more of it per unit of nitrogen oxide removed because part of it never reacts.
• Temperature — SNCR works inside the chamber temperature the plant already holds; SCR needs the gas within its own window, which on a clean-gas arrangement usually means a reheater.
• Removal — SNCR reaches a part of the inlet nitrogen oxides and SCR reaches most of it. Where the limit leaves little margin, that difference is the whole decision.
• Consumables — SCR carries a catalyst that the waste poisons; SNCR has none.
• By-products — SNCR leaves ammonia slip; SCR leaves ammonia slip and, with a vanadium catalyst, some conversion of sulphur dioxide to sulphur trioxide.
• Footprint and structure — SNCR is a lance, a pump skid and a tank; SCR is a reactor, a support structure and often a reheater.
• Operating duty — a catalyst reactor holds a steadier figure and asks for scheduled inspection and replacement that a small site has to be able to carry out.

Working back from the NOx limit

The NOx emission limit for medical waste incinerators in the chart here is 300 mg/m³ as a one-hour mean and 250 mg/m³ as a 24-hour mean, referred to dry gas at 11 % oxygen. The two figures are alternative averaging periods for the same parameter, and they ask different things of a control system: the one-hour mean is met by a plant that responds quickly, and the 24-hour mean is met by a plant that does not wander. A catalyst reactor on a steady gas produces a flat figure; SNCR produces a figure that follows the furnace, which is why the 24-hour mean, rather than the one-hour mean, is usually the number that decides whether SNCR is enough on its own.

Two things follow for a small plant. Reagent consumption is a running cost that scales with the waste and with how much of the nitrogen oxide survives the furnace, so a furnace that is well mixed and consistently hot makes the problem downstream smaller. And whichever route is chosen, it is one stage in a train rather than a fix on its own: the nitrogen oxide limit sits in a set of limits that are met together, as the emissions chart shows.

Common mistakes

• Assuming a scrubber will take the nitrogen oxides. It will not: nitric oxide is close to insoluble.
• Quoting an SNCR removal figure without stating the load and the gas temperature at which it was achieved.
• Fitting SCR upstream of the dust collector to save a reheater, and then paying for it in catalyst replacements.
• Sizing the reagent system on the reaction alone, without allowing for the ammonia that does not react.
• Treating the catalyst as a permanent item. In a waste plant it is a consumable with a life set by what is in the ash.
• Forgetting the oxygen reference. A limit written at 11 % oxygen is not the same number as the concentration measured in the duct.

From a NOx limit to a reagent system

SNCR and SCR answer the same chemistry from two different places in the plant, and the choice follows from how hot the gas is where the plant can afford to put a reactor, and from what it can afford to keep clean. The emissions chart sets the nitrogen oxide limit with the rest of the limits the train has to meet, and the article on flue gas treatment covers the stages the nitrogen oxide system sits among. The reagent, the injection point and the catalyst arrangement for your plant are confirmed with the factory for your waste composition and the emission standard that applies.