Dry Scrubber

What a Dry Scrubber Is and What Is Being Decided

The machine trades water for solid. A wet scrubber absorbs acid gases into a liquid slurry and pays with a wastewater stream, scrubber blowdown that needs treatment and disposal. A dry scrubber arranges the same acid-base chemistry to happen on dry or drying surfaces: sorbent particles, dry powder or atomizing slurry droplets, capture the acid gas, and the reaction product stays solid, collected with the fly ash in the fabric filter or the electrostatic precipitator that the train already needs. What leaves the system is a dry residue, salts plus unreacted sorbent plus ash, handled by conveyor and truck. That single difference, no liquid effluent, is why the dry family owns waste incineration, why it dominates mid-size industrial boiler retrofits, and why buyers accept sorbent consumptions that a wet system would not tolerate.

What is being decided in this purchase is an architecture plus a sorbent plus a collector strategy, all three interlocked. The architecture, DSI, SDA, or CDS, sets removal ceiling and capital: injection is cheap and modest, the spray dryer is mid-capital and high removal, the circulating bed is high removal at higher fan power. The sorbent, hydrated lime, sodium bicarbonate, or trona, sets reagent cost per ton of acid gas, reaction temperature window, and residue chemistry, sodium salts dissolve and can leach, which some residue permits care about. The collector, a bag filter in almost every modern train, is not downstream of the scrubber but part of it: the filter cake on the bags is a second reaction stage where unreacted sorbent meets acid gas for seconds, not milliseconds, and fabric filter design, cleaning intensity, and cake management are scrubber design variables. An RFQ that buys the reactor and ignores the collector has bought half a machine.

The competitive frame against the wet family deserves one paragraph here and depth in the selection chapter. The wet scrubber, covered by its own leaf, takes acid gas into liquor at saturation, achieves the highest removals at the lowest stoichiometric ratios, handles SO2 loads that would bury a dry system in residue, and pays for it all with a wastewater plant, a saturated wet plume, and wet-grade materials. The dry family answers where water is scarce or unwelcome, where the residue is marketable or landfillable as dry solid, where loads are moderate, and where retrofit simplicity matters: DSI installs in weeks into existing duct. Between them, the semi-dry SDA occupies the middle ground, using just enough water to humidify the reaction before it all evaporates. The choice is site economics, and the chapters below give it numbers.

Two equipment neighbors need explicit disentangling because trains interleave them. The bag filter is the collector this page's systems feed; its leaf covers fabric filtration on its own terms, and here it appears as the integrated second stage, its cake as reactor surface. The electrostatic precipitator can serve some DSI trains as collector, and a companion leaf covers that machine; the chemistry in this page applies regardless of which collector follows, though the filter cake benefit belongs to fabric. The cyclone separator appears as a pre-cleaner and, in CDS designs, as the internal recirculation separator. And the wet scrubber is the alternate universe: same pollutants, different phase discipline, its own leaf, and the selection chapter runs the honest comparison.

Scale and market shape follow architecture. DSI systems are packaged skids, silo, blower, injector lance array, controls, installed into existing ducts on units from a few to hundreds of megawatts equivalent, typically where a modest acid gas trim, HCl compliance on boilers, SO2 polish on top of other controls, is the target. SDA trains are engineered vessels, 8 to 12 m tall absorber towers with rotary or dual-fluid atomizers, serving waste incinerators and mid-size utilities at 30,000 to 400,000 cubic meters per hour. CDS systems are engineered reactor-plus-recirculation packages at similar scales with higher fan duty. The purchase pattern follows: DSI as a component retrofit quoted fast, SDA and CDS as process trains with guarantees, performance tests, and the full RFQ discipline of chapter 5.

How Sorbent Chemistry Sets the Size and the Bill

Three sorbents carry the market and each has a temperature window. Hydrated lime, calcium hydroxide, reacts with SO2 and HCl to form calcium sulfite and sulfate and calcium chloride; it works both in the humid near-saturation environment that SDA and CDS create and, less efficiently, in dry duct injection, and it is the cheapest reagent per ton almost everywhere. Sodium bicarbonate decomposes above about 140 deg C into porous sodium carbonate, which reacts fast with acid gases in the dry state, reaching high removals in plain DSI at boiler air heater outlet temperatures, 140 to 280 deg C; it costs more per ton but its stoichiometry is far better, so its cost per ton of acid gas removed is competitive. Trona, sodium sesquicarbonate ore, ground and injected, behaves similarly at a lower reagent price with slightly worse kinetics, and it dominates DSI on utility coal duty in North America for exactly that arithmetic.

Stoichiometry is the number that prices everything. Defined as moles of reagent injected per mole of acid gas removed, real systems run: lime SDA and CDS at 1.5 to 2.5 on SO2 and lower on HCl, which reacts faster; sodium DSI at 1.1 to 1.6 near optimum; and the vendor's guaranteed removal at guaranteed stoichiometric ratio is the core of every bid. The unreacted fraction, a quarter to half of the lime in some duty, leaves in the residue as purchasing waste, which is why reagent price per ton times stoichiometric ratio, not price per ton, compares sorbents honestly. The RFQ demands both numbers together, at the design acid gas load and at the permit point, because removal and stoichiometry trade along a curve and a bid quoting 95 percent at 1.4 and a bid quoting 95 percent at 2.0 are different machines with different residue streams.

Water sets the reaction quality in the semi-dry family. SDA and CDS performance lives at the approach to adiabatic saturation temperature: the amount of water evaporated cools and humidifies the gas toward its saturation point, and the closer the operating point sits to saturation without wetting, typically 10 to 20 deg C approach, the better the lime reacts and the less reagent the guarantee needs. Too close, and the vessel, ducts, and downstream bags paste up with wet residue; too far, and stoichiometry climbs. The absorber is therefore designed around a water balance: slurry or humidification flow, gas flow and inlet temperature, and a outlet temperature that the controls hold as the primary operating variable. Spray dryer absorbers run outlet temperatures of 140 to 160 deg C on incinerators and lower approaches on tolerant duties; DSI runs on whatever humidity the gas already has, which is why lime DSI without humidification underperforms and why sodium sorbents, which need temperature rather than humidity, dominate dry injection.

Residence time and contact geometry decide whether the chemistry finishes. SDA atomizers, rotary cups spinning at thousands of revolutions per minute or dual-fluid nozzles, make a droplet cloud that dries in one to two seconds while reacting; the tower is sized for that drying flight, and gas distribution at the atomizer zone is the make-or-break detail. CDS fluidizes a deep bed, a meter or more of hydrated lime and recycled product, with the gas itself, giving seconds of contact and fresh surface continuously as particles collide and abrade; the recirculation separator returns the bed and only lets spent fines leave. DSI has only the duct run and the filter cake: seconds of total time, of which the bag cake is the productive part, which is why DSI removal claims are quietly fabric filter claims and why collector condition is scrubber condition.

The residue is a purchase line, not an afterthought. Dry scrubber residue is fly ash plus calcium or sodium salts plus unreacted reagent: lime residues are generally benign and often landfilled or used in cement, sodium residues carry soluble chlorides and sulfates that leach, and some permits and some ash markets treat them differently, which can flip the sorbent choice. Quantities scale with stoichiometry and inlet load: a 100,000 cubic meter per hour train removing 500 milligrams per normal cubic meter of SO2 at ratio 2.0 produces tonnes per day of extra residue, every day, and the silo, discharge, and truckout scope belongs in the scrubber purchase, not in a later surprise. The weigh feeder and the rotary airlock under the silo, the same hardware family that dust collection systems use, are part of the reagent train's integrity.

Monitoring closes the guarantee loop. Inlet and outlet acid gas measurement, SO2 and HCl by CEMS with proper sample conditioning, sets the removal arithmetic the permit audits; gas temperature and humidity define the approach operation; and the fabric filter's pressure drop trace, read across the collector, is the cake management signal. On trains with oxygen trim requirements, a zirconia oxygen analyzer feeds the correction basis. The same discipline that makes an electrostatic precipitator guarantee enforceable applies here: a removal promise without named instruments, reference methods, and averaging periods is arithmetic, and the RFQ writes the measurement basis into the guarantee.

Types, Construction, and Duty Boundaries

Dry sorbent injection is the retrofit machine. A silo of trona or bicarbonate, a weighing feeder, a blower, and lance arrays injecting into the duct or the economizer hopper ahead of the air heater; the sorbent entrains, reacts in flight, and the bag filter or ESP collects it with the ash. DSI removes 50 to 80 percent of HCl and 50 to 70 percent of SO2 at practical ratios on moderate loads, enough for many boiler HCl rules and MATS-style compliance trims, installs in a shutdown window rather than an outage year, and carries the lowest capital in the whole air pollution control catalog. Its limits are equally plain: high SO2 loads bury it in sodium residue, cold dry gas starves lime, and its performance is chained to collector condition.

The spray dryer absorber is the incineration standard. A tall vessel with rotary atomizer at the top, or dual-fluid lances on smaller units, sprays lime slurry into the hot gas; droplets dry in flight while the lime captures acid gases; the dry product drops to the vessel cone and exits to the collector with the fly ash. SDA removes 90 percent-plus of HCl and SO2 at moderate ratios, tolerates load swings well because the water balance absorbs them, and its residue, dry calcium salts, suits the waste industry's residue management. Its discipline is the atomizer and the water balance: rotary atomizer maintenance is a real cost line, the approach temperature is the operating covenant, and pasting the downstream equipment by running too close to saturation is the classic failure.

The circulating dry scrubber buys its performance with recirculation and fan power. Gas accelerates through a venturi or fluidizing grid into a riser where it entrains a deep circulating bed of hydrated lime and recycled reaction product; the bed is separated at the top, returned at the bottom, and only a spent fraction leaves to the collector. The long cumulative contact and the constantly abraded fresh surface drive removals of 95 to 98-plus percent at ratios competitive with SDA, with a dry vessel that never atomizes slurry, humidification water only rather than slurry handling. The price is pressure drop, 1,500 to 2,500 pascals across the reactor against the SDA's few hundred, and the fan energy that buys; on high sulfur duty the CDS plus fabric filter train is often the best dry-family answer.

Construction materials are mild by wet standards and specific at the wet spots. Vessels and ducts are carbon steel, insulated to keep the gas above dew point everywhere, because the dry family's whole premise is staying out of condensation; the humidification and slurry circuits on SDA are the exceptions, alloy or lined piping and nozzles on the wet path; and the downstream collector follows its own construction practice, fabric filters with membrane or treated bags chosen for the residue's chemistry, or ESPs in some DSI retrofits. Bag selection is scrubber chemistry: sodium residues and fine salts drive needle felt or membrane choices, temperature excursions matter more than in plain ash duty, and the bag vendor becomes a member of the scrubber project team in every serious train.

Duty boundaries, stated bluntly. A dry scrubber is not a particulate machine: the collector behind it does that work, and the collector has its own leaf. It is not a wet scrubber: no liquid effluent, no saturation plume, no scrubber liquor chemistry, and lower top-end removal on brutal SO2 loads, where wet flue gas desulfurization owns the utility megastation duty. It is not a nitrogen oxides machine: NOx belongs to SCR and SNCR, and though trains sit side by side, the chemistries do not mix. It is not, in itself, a dioxin or mercury answer: activated carbon injection handles mercury as a parallel sorbent stream, and dioxin control rides the fabric filter and temperature discipline; both are frequent companions of this equipment, both are separate purchase lines. The table below consolidates the type map.

Key Specification Parameters

The datasheet rows below decide whether a quoted dry scrubber matches the gas and the permit. Each row is a purchase defining parameter; the paragraphs after the table explain what to demand and what a weak answer looks like.

The removal and stoichiometry rows must be read as a pair and demanded as a curve. A single guarantee point hides the trade: ask for removal versus ratio at the design gas conditions, at two or three ratios, so the operating cost of chasing the last few percent is visible in the bid, not discovered in the first summer of compliance. The same paragraph applies to approach temperature on the semi-dry machines: the guarantee's outlet temperature, its tolerance band, and the controls that hold it are the difference between a train that holds removal through load swings and one that pastes its bags in a downpour of bad tuning.

The sorbent row prices the decade. Reagent cost per ton removed equals price per ton times stoichiometric ratio times purity corrections, and the parallel table below compresses the comparison; the rule under it is one sentence: compare sorbents at guaranteed ratio and delivered price, including the residue chemistry consequence, because the cheapest reagent per ton is rarely the cheapest compliance. Sodium sorbents at 1.3 beat lime at 2.2 whenever sodium price is under twice lime's, and lime wins on residue leach rules that sodium fails.

The collector and residue rows close the system. The fabric filter is the second reaction stage: its design air-to-cloth ratio, bag media, and cleaning regime belong in the scrubber specification, and the pressure transmitter across it is the scrubber's most informative instrument once the train runs. Residue handling, silo capacity in days, discharge hardware, and the truck-out interface, is purchased with the train or owned deliberately by the plant; leaving it to the general contractor is how scrubbers meet their residue silo full in month two.

Selection Decision Factors and RFQ Language

The decision sequence runs as a checklist, and each step eliminates a class of bids before price comparison begins.

1. Write the gas basis honestly: flow and temperature at the reactor inlet across load, acid gas loads by species with their fuel-driven range, and moisture.
2. Write the permit line with its measurement basis: percent removal or outlet concentration, reference method, averaging period, and oxygen correction.
3. Bracket the architecture: if the target is a 60 percent HCl trim on an existing baghouse, DSI answers at a fraction of vessel capital; if it is 95 percent SO2 on a waste train, only SDA or CDS bid; in between, invite both.
4. Price the sorbents at delivered price and guaranteed ratio, including residue class consequences, and let the arithmetic pick the reagent.
5. Fix the downstream collector as part of the system: new fabric filter or existing, its condition, media, and the cake management strategy.
6. Fix the water and approach basis on semi-dry machines: available water quality and quantity, outlet temperature covenant, and paste-up protection.
7. Fix residue handling end to end: tonnage at ratio, silo days, discharge, transport, and the disposal or sale path with its permit.
8. Fix the verification evidence: removal versus ratio curve, performance test protocol, monitoring instruments with sample conditioning, and the correction basis.

Against the neighboring control machines, the comparison a specifier actually runs is the following, and it belongs here rather than in the page title. Against the wet scrubber, the dry family trades top-end removal and low reagent ratios for the absence of wastewater, a dry residue, and a simpler retrofit story; high sulfur utility loads stay wet, moderate loads and water-sensitive sites go dry, and the semi-dry SDA splits the difference with the water it evaporates. Against the bag filter, no rivalry, only integration: the collector is this system's second stage, and DSI without a healthy fabric filter is half a promise. Against the electrostatic precipitator, the same integration question in collector choice, with the cake-stage benefit lost and higher DSI removals therefore harder. Against the cyclone separator, upstream pre-collection protects the reactor and recirculation hardware on ash-heavy gas, a line item decision, not a rivalry. Choose the train the load, water economics, and residue rules demand, not the machine a supplier has on promotion.

The application map below turns those choices into a table the project file can carry into bid evaluation.

The RFQ block should be written once and reused. A complete block reads: gas flow, temperature, moisture, and pressure at inlet with load range; acid gas loads by species with fuel range and design basis; required removal or outlet concentration with reference method, averaging period, and oxygen basis; architecture invited, DSI, SDA, CDS, or combinations; sorbent proposed with delivered price basis and guaranteed stoichiometric ratio at the permit point; removal versus ratio curve at design conditions; approach temperature and outlet temperature covenant with control description on semi-dry machines; reactor residence and geometry basis; downstream collector scope, new or existing, media, air-to-cloth, and cleaning; pressure drop allowance and fan scope; reagent storage and feed system, silo capacity, feeder type, lances or atomizers; residue quantity, silo days, discharge, and interface to disposal; monitoring scope, CEMS for acid gases with sample conditioning system, temperature, draft, oxygen trim; materials and insulation scope with dew point margin; performance test protocol with penalty structure; and documentation, process flow and heat and material balance, control narrative, and the residue data sheet. A quotation returned without the ratio curve or the residue quantity is incomplete at any price.

Installation, Standards, Failure Modes, and Maintenance

Installation is duct and vessel work around a hot gas path that must stay dry. Reactors and ducts are insulated to hold the gas above dew point everywhere, including the idle branches and the collector, because the dry family's corrosion and pasting risks are all condensation risks; the reagent silo sets on foundations with the feed skid below it, and the weigh feeder and blower train are aligned for metering accuracy; lances and atomizer penetrations are installed with their access platforms, because these are the wear parts that will be serviced; and the residue path, vessel cone, collector hoppers, conveyors, silo, is proven empty-to-full before first fire. On SDA trains the slurry plant, lime slaking or delivery, slurry tanks, and the atomizer pump loop, is commissioned as its own chemical system, and the atomizer is run on water before it ever sees slurry.

Commissioning proves the curve, not just the point. Baseline gas measurements first, inlet and outlet acid gas and flow; then sorbent feed at low ratio and climb, recording removal versus ratio on the way to the guarantee point, which is the curve the plant will operate against for its life; on semi-dry machines the water balance is walked down toward the approach covenant with paste-up watched at every step, bag differential inspected, vessel cones checked for buildup. The performance test then runs at stated conditions with the reference methods the permit names, and the residue stream is sampled and characterized for its disposal class, because a removal guarantee on a residue the landfill rejects is a compliance failure wearing a different hat.

The standards frame is regulatory with engineering practice behind it. In the United States, EPA control cost manual practice frames wet and dry acid gas scrubber economics, and the agency's flue gas desulfurization fact sheets describe wet, spray dry, and dry injection systems for the permit audience; European waste incineration runs the BAT framework with its emission and monitoring expectations; and the monitoring regime, CEMS with reference method tie, follows the jurisdiction's rules with sample conditioning discipline throughout. Reagent quality, lime reactivity or trona grind, is a purchase specification line because it feeds the guarantee arithmetic. There is no single dry scrubber hardware standard: the machine is a system of vessels, feeders, and collectors each answering to its own codes, and the specification names them per part.

Failure modes follow a short, predictable list. Paste-up on semi-dry trains: outlet temperature control drifts, water wins locally, and vessel walls, ducts, or bags cake with damp residue; the guards are the approach covenant, good atomizer condition, and the differential pressure trend that catches the cake early. Atomizer wear on SDA: droplet size grows, drying flight lengthens, and removal sags between maintenances; rotary cup and nozzle overhaul is a scheduled scope with spares on site. Sorbent feed drift: weigh feeder calibration or milling quality slips, ratio silently falls, and the first symptom is at the stack; the feed system is instrumented and audited like any metering system. Bag blinding on the collector: wrong media for the residue chemistry, or pasting from upstream, shows as climbing differential and falling fan margin; media selection and cake management are the cures. Bed collapse on CDS: fluidizing velocity loss or reagent quality change drops circulation, and removal falls with it; the fan and grid design margins cover the normal range. And ash-related erosion at vessel inlets and lance tips on hot dirty gas: wear allowance and replaceable tips are the honest design answers.

Maintenance economics are metering and wear parts. Daily: read the ratio actually fed against gas load, the outlet temperature, and the collector differential; these three trends are the train's health chart. Weekly: verify feeder calibration by inventory drop, inspect lances or atomizer on schedule, and blow down any humidification nozzles. Monthly: CEMS calibration checks with the sample conditioning system maintained, residue silo level audit against trucking, and a vessel cone inspection on semi-dry duty. Annually: full atomizer or lance overhaul, bag filter row-by-row inspection with media condition documented, feeder and airlock rebuild kits, and a reconfirmation run of the removal versus ratio curve, which is the acceptance test the plant gives itself. A train with its curve history and its residue data current is a system whose next RFQ is written from evidence, which is how a purchasing file gets smarter instead of just older.