Part II Wastegas Engineering 9 Control of Primary Particles 9.1 Wall Collection Devices The first three types of control devices we consider--gravity settlers, cyclone separators, and electrostatic precipitators--all function by driving the particles to a solid wall, where they adhere to each other to form agglomerates that can be removed from the collection device and disposed of. Although these devices look different from one another, they all use the same general idea and are described by the same general design equations. 9.1.1 Gravity Settlers A gravity settler is simply a long chamber through which the contaminated gas passes slowly, allowing time for the particles to settle by gravity to the bottom. It is an old, unsophisticated device that must be cleaned manually at regular intervals. But it is simple to construct, requires little maintenance, and has some use in industries treating very dirty gases, e.g., some smelters and metallurgical processes. Furthermore, the mathematical analysis for gravity settlers is very easy; it will reappear in modified form for cyclones and electrostatic precipitators. 9.1.2 Centrifugal Separators We have spent considerable time on gravity settlers because it is easy to see what all their mathematics mean. But they have little practical industrial use because they are ineffective for small particles. If we are to use them or devices like them, we must find a substitute that is more powerful than the gravity force they use to drive the particles to the collection surface. Physics and mechanics books usually show that centrifugal force is a pseudoforce that is really the result of the body's inertia carrying it straight while some other force makes it move in a curved path. It is convenient to use this pseudoforce for calculational purposes. At even modest velocities and common radii, the centrifugal forces acting on particles can be two orders of magnitude larger than the gravity forces. For this mason centrifugal particle separators are much more useful than gravity settlers. How does one construct a practical centrifugal particle collector? There are many types, but the most successful is sketched in Fig 9.1. It is universally called a cyclone separator, or simply a cyclone. It is probably the most widely used particle collection device in the world. In any industrial district of any city, a sharp-eyed student can find at least a dozen of these outside various industrial plants. A cyclone consists of a vertical cylindrical body, with a dust outlet at the conical bottom. The gas enters through a rectangular inlet, normally twice as high as it is wide, arranged tangentially to the circular body of the cyclone, so that the entering gas flows around the circumference of the cylindrical body, not radially inward. The gas spirals around the outer part of the cylindrical body with a downward component, then turns and spirals upward, leaving through the outlet at the top of the device. During the outer spiral of the gas the particles are driven to the wall by centrifugal force, where they collect, attach to each other, and form larger agglomerates that slide down the wall by gravity and collect in the dust hopper in the bottom. There are many other variants on the centrifugal collector idea, but none approaches the cyclone in breadth of application. These devices are simple and almost maintenance-free. Because any medium-sized welding shop can make one, the big suppliers of pollution control equipment, who have test data on the effects of small changes in the internal geometry, have been unwilling to make these data public. The same basic device as the cyclone separator is used in other industrial settings where the goal is not air pollution control, but some other kind of separation. When it is used to separate solids from liquids it is generally called a hydroclone. A cyclone called an air-swept classifier is attached to many industrial grinders. It passes those particles ground fine enough, and collects those that are too coarse, returning them to the grinder. 9.1.3 Electrostatic Precipitators (ESP) If gravity settlers and centrifugal separators are devices that drive particles against a solid wail, and if neither can function effectively (at an industrial scale) for particles below about 5 Ix in diameter, then for wall collection devices to work on smaller particles, they must exert forces that are more powerful than gravity or centrifugal force. The electrostatic precipitator (ESP) is like a gravity settler or centrifugal separator, but electrostatic force drives the particles to the wall. It is effective on much smaller particles than the previous two devices. The basic idea of all ESPs is to give the particles an electrostatic charge and then put them in an electrostatic field that drives them to a collecting wall. This is an inherently two-step process. In one type of ESP, called a two-stage precipitator, charging and collecting are carried out in separate parts of the ESR This type, widely used in building air conditioners, is sometimes called an electronic air filter. However, for most industrial applications the two separate steps are carried out simultaneously in the same part of the ESP. The charging function is done much more quickly than the collecting function, and the size of the ESP is largely determined by the collecting function. Fig 9.2 shows in simplified form a wire-and-plate ESP with two plates. The gas passes between the plates, which are electrically grounded (i.e., voltage = 0). Between the plates are rows of wires, held at a voltage of typically -40 000 volts. The power is obtained by transforming ordinary alternating current to a high voltage and then rectifying it through some kind of solid-state rectifier. This combination of charged wires and grounded plates produces both the free electrons to charge the particles and the field to drive them against the plates. On the plates the particles lose their charge and adhere to each other and the plate, forming a "cake." The cleaned gas then passes out the far side of the precipitator as shown in Fig. 9.3. Solid cakes are removed by rapping the plates at regular time intervals with a mechanical or electromagnetic rapper that strikes a vertical or horizontal blow on the edge of the plate. Through science, art, and experience designers have learned to make rappers that cause most of the collected cake to fall into hoppers below the plates. Some of the cake is always re-entrained, thereby lowering the efficiency of the system. If the collected particles are liquid, e.g., sulfuric acid mist, they run down the plate and drip off. For liquid droplets the plate is often replaced by a circular pipe with the wire down its center. Some ESPs (mostly the circular pipe variety) have a film of water flowing down the collecting surface, to carry the collected particles to the bottom without rapping. There are many types of ESPs; Fig. 9.3 shows one of the most common in current use in the United States. Gas flow is from right to left. The gas enters at the right through an inlet diffuser (not shown) in which the flow spreads out from the much narrower duct to the perforated gas distribution plate that distributes the gas evenly across the entrance face of the precipitator. A similar plate and converging nozzle on the left side (not shown) maintain a uniform flow at the outlet and then reduce the cross-sectional flow area to that of the outlet duct. The whole interior of the structure is filled with discharge electrodes and collecting plates; the cutaway shows only one set of plates and discharge electrodes. The discharge electrodes consist of rigid frames with many short, pointed stubs, which serve the same function as the wires in Fig.9.2. The collecting surfaces are made of sheet metal sections with vertical joints that tend to trap the particles. Each pair of plates, along with the discharge electrode between them, acts like the single channel in the simplified version of an ESP shown in Fig. 9.2. The rappers strike the supports for the discharge electrodes and the collecting plates at regular time intervals to dislodge the cake of collected particles. The multiple power supply transformer-rectifier sets supply DC current at m -40,000 V to the discharge electrodes. The collected particles, dislodged from the plates by the rappers, fall into the particle collecting hoppers, from which they are automatically removed to storage. The drawing shows some of the structural steel frame and enclosure of the ESP and the handrail on its top, but not the internal seals that hinder the gas from flowing around the area of the collecting plates. Each point in space has some electrical potential V. If the electrical potential changes from place to place, then there is an electrical field, E = V/x, in that space. If we connect two such points with a conductor, then a current will flow. This V is the voltage we are all familiar with, and E is its gradient in any direction; the units of E are V/m. In a typical wire-and-plate precipitator, the distance from the wire to the plate is about 4 to 6 in., or 0.1 to 0.15 m. With a voltage difference of 40 kV and 4-in. spacing, one would assume a field strength of 40 kV/0.1 m = 400 kV/m. This is indeed the field strength near the plate. However, all of the electrical flow that reaches the plate comes from the wires, and the surface area of the wires ii much lower than that of the plate; thus, by conservation of charge, the driving potential near the wires must be much larger. Typically it is 5 to 10 MV/m. (The first person to utilize this fact was presumably Benjamin Franklin, who invented the sharp, pointed lightning rod.) When a stray electron from any of a variety of sources encounters this strong a field, it is accelerated rapidly and attains a high velocity. If it then collides with a gas molecule, it has enough energy to knock one or more electrons loose, thus ionizing the gas molecule. These electrons are likewise accelerated by the field and knock more electrons loose, until there are enough free electrons to form a steady corona discharge. In a dark room this discharge appears as a dim glow that forms a circular sheath about the wire. The positive ions formed in the corona migrate to the wire and are discharged. The electrons migrate away from the wire, toward the plate. Once they get far enough away from the wire for the field strength to be too low to accelerate them fast enough to ionize gas molecules, the visible corona ceases and they simply flow as free electrons. As the electrons flow toward the plate, they encounter particles and can be captured by them, thus charging the particles. Then the same electric field that created the electrons and that is driving them toward the plate also drives the charged particles toward the plate. The typical linear velocity of the gas inside an ESP is 3 to 5 ft/s, much lower than that in a cyclone. The typical pressure drop is 0.1 to 0.5 in. H20, again much less than in a cyclone. The pressure drop in the ducts leading to and from the precipitator is generally more than in the ESP itself. The ESP industry is now well established. Standard package units are available for small flows (down to the size of home air conditioners), and large power plants have precipitators costing up to $30 million. Wet ESPs are more complex, and the collected particles are not in the convenient form of a dry powder. But for the final 5% cleanup these problems seem a modest price to pay for the greatly improved collection efficiency. Another approach is to make the final 5% collection in a filter, as described next. Sometimes the ESP-filter combination is more economical than an equivalent-performance ESP or filter. 9.2 Dividing Collection Devices Gravity settlers, cyclones, and ESPs collect particles by driving them against a solid wall. Filters and scrubbers do not drive the particles to a wall, but rather divide the flow into smaller parts where they can collect the particles. In this section we shall first consider the two types of filters used in air pollution control, surface filters and depth filters. Then we shall discuss scrubbers. The public often refers to any kind of pollution control device as a filter, giving the word filter the meaning "cleaning device." Technically, a filter is one of the devices described in this section. Other devices (e.g., the "biofilters" described in Chapter 10) are not truly filters. Engineers must live with the difference between the technical meaning and that used by nonprofessionals. 9.2.1 Surface Filters Most of us have personal experience with surface filters, as exemplified by those in a coffee percolator or a kitchen sieve. The principle of operation is simple enough; the filter is a membrane (sheet steel, cloth, wire mesh, or filter paper) with holes smaller than the dimensions of the particles to be retained. Although this kind of filter is sometimes used for air pollution control purposes, it is not common because constructing a filter with holes as small as many of the particles we wish to collect is very difficult. One only needs to ponder the mechanical problem of drilling holes of 0.1μ diameter or of weaving a fabric with threads separated by 0.1 μ to see that such filters are not easy to produce. It can be done on a laboratory scale by irradiating plastic sheets with neutrons and then dissolving away the neutron-damaged area. The resulting filters have analytical uses but are not used for industrial air pollution control (although they are used industrially to filter some beers and other products, removing trace amounts of bacteria). Although industrial air filters rarely have holes smaller than the smallest particles captured, they often act as if they did. The reason is that, as fine particles are caught on the sides of the holes of a filter, they tend to bridge over the holes and make them smaller. Thus as the amount of collected particles increases, the cake of collected material becomes the filter, and the filter medium (usually a cloth) that originally served as a filter to collect the cake now serves only to support the cake, and no longer as a filter. This cake of collected particles will have average pore sizes smaller than the diameter of the particles in the oncoming gas stream, and thus will act as a sieve for them. The particles collect on the front surface of the growing cake. For that reason this is called a surface filter One may visualize this situation with a screen having holes 0.75 in. (1,91 cm) in diameter. We could collect a layer of Ping-Pong balls easily on this screen. Once we had such a layer, we could then collect cherries, which, by themselves, could pass through the holes in the screen but cannot pass through the spaces between the Ping-Pong balls. Once we have a layer of cherries, we could put on a layer of peas, then of rice, then of sand. In that way we could collect sand on a screen with holes 0.75 inch in diameter. In typical industrial filters the particles are of a wide variety of sizes, so they do not go onto the screen in layers, but all at once. The effect is the same; very small particles are collected by the previously collected cake on a support whose holes are much larger than the smallest particles collected. The theory of cake accumulation and pressure drop for this type of device is well-known from industrial filtration. The two most widely used designs of industrial surface filters are shown in Fig 9.4 and 9.5 on pages 285 and 286. Because the enclosing sheet metal structure in both figures is normally the size and roughly the shape of a house, this type of gas filter is generally called a baghouse. The design in Fig. 9.4, most often called a shake-deflate filter, consists of a large number of cylindrical cloth bags that are closed at the top like a giant stocking, toe upward. These are hung from a support. Their lower ends slip over and are clamped onto cylindrical sleeves that project upward from a plate at the bottom. The dirty gas flows into the space below this plate and up inside the bags. The gas flows outward through the bags, leaving its solids behind. The clean gas then flows into the space outside the bags and is ducted to the exhaust stack or to some further processing. For the baghouse in Fig. 9.4 there must be some way of removing the cake of particles that accumulates on the filters. Normally this is not done during gas-cleaning operations. Instead the baghouse is taken out of the gas stream for cleaning. When the gas flow has been switched off, the bags are shaken by the support to loosen the collected cake. A weak flow of gas in the reverse direction may also be added to help dislodge the cake, thus deflating the bags. The cake falls into the hopper at the bottom of the baghouse and is collected or disposed of in some way. Often metal rings are sewn into filter bags at regular intervals so that the bag will only partly collapse when the flow is reversed, and a path will remain open for the dust to fall to the hopper. Because it cannot filter gas while it is being cleaned, a shake-deflate baghouse cannot serve as the sole pollution control device for a source that produces a continuous flow of dirty gas. For this reason, one either uses a large enough baghouse so that it can be cleaned during periodic shutdowns of the source of contaminated gas or installs several baghouses in parallel. Typically, for a major continuous source like a power plant, about five baghouses will be used in parallel, with four operating as gas cleaners during the time that the other one is being shaken and cleaned. Each baghouse might operate for two hours and then be cleaned for 10 minutes; at all times one baghouse would be out of service for cleaning or waiting to be put back into service. Thus the baghouse must be sized so that four of them operating together provide adequate capacity for the expected gas flow rate. The other widely used baghouse design, called a pulse-jet filter, is shown in Fig. 9.5. In it the flow during filtration is inward through the bags, which are similar to the bags in Fig. 9.4 except their ends open at the top. The bags are supported by internal wire cages to prevent their collapse. The bags are cleaned by intermittent jets of compressed air that flow into the inside of the bag to blow the cake off. Often these baghouses are cleaned while they are in service; the internal pulse causes much of the collected solids to fall to the hopper, but some are drawn back to the filter cloth. Just after the cleaning the control efficiency will be less than just before the next cleaning, but the average efficiency meets the legal control requirements. The flow velocities through such filters are very low, typically a few feet per minute. In contrast, in devices like cyclones the flow is about 60 feet per second. A wind velocity equal to the typical flow through such a filter is so low that someone standing in it could not tell in which direction it was blowing and would report that there was no wind at all. This calculation shows that the collected cake is about 0.07 in. thick, the average increase during one cycle. If the cleaning were perfect, this would be the cake thickness. However, it is hard or impossible to clean the bags completely, and in power plant operation it is common for the average cake thickness on the bags to be up to 10 times this amount. During each cleaning cycle some part of the cake falls completely away, leaving bare patches on the bag; and most of the cake does not come off at all. If one could examine a bag after a cleaning, one would probably see nine-tenths of the surface covered with a cake perhaps 0.7 in. thick, and one-tenth of the bag with a bare surface. The operators would like to clean the bags more thoroughly, but more vigorous cleaning procedures (harder shaking, faster reverse gas flow) tend to wear out the bags faster and lead to more frequent maintenance shutdowns. Most operators have used mild cleaning cycles, leading to long bag lives and low maintenance costs but higher pressure drops than would be needed if all the cake came off the bag at each cleaning cycle. One of the advantages of the pulse-jet design is that it cleans the bags more thoroughly, allowing a higher Vs, at the cost of a somewhat shortened bag life. If surface or cake-forming filters are operated at low superficial velocities, they can have very high efficiencies, and they generally collect fine particles as efficiently as coarse ones. For these two reasons they have found increasing application, particularly in electric power plants, as particle emission regulations have become steadily more stringent, making it necessary to collect particles in the size range from 0.1 to 0.5μ, which are difficult for ESPs to collect. 9 .2.2 Depth Filters Another class of filters, widely used for air pollution control, does not form a coherent cake on the surface, but instead collects particles throughout the entire filter body. The examples with which the student is probably familiar are the filters on filter-tipped cigarettes and the lint filters on many home furnaces. In both of these a mass of randomly oriented fibers (not woven to form a single surface) collects particles as the gas passes through it. Such filters are often used where the particles to be caught are fine drops of liquids that are only moderately viscous. Such drops will coalesce on the fibers and then run off as larger drops, leaving the fibers ready to catch more fine drops. If the particles were solid, then this type of filter would require regular cleaning; for the liquid application it does not. The most widespread air pollution control use of depth filters is in the collection of very fine liquid drops, sulfuric acid mist, produced in sulfuric acid plants. Similar devices are used in many gas-liquid contacting devices to catch fine droplets; one brand uses the trade name Demister. This kind of device is also used for removing solid particles from gas streams that contain 'few of them, e.g., for cleaning the air of industrial clean rooms or hospital surgical suites and in personal protection dust masks. The filters are thrown away when they have collected enough particles that their pressure drop begins to increase. The depth filters used in those applications are normally called high-efficiency, particle-arresting (HEPA) or absolute filters. The air filters on household furnaces operate this way as well; typically the fibers are coated with a sticky substance to improve the retention of the collected dust and lint. Depth filters collect particles mostly by impaction. Some older types of particle collectors also used impaction, to catch particles on solid walls, but they are seldom used now. Some size-specific particle analyzers (impactors or cascade impactors) use impaction on collecting surfaces to collect specific sizes of particles. In liquid scrubbers (discussed later), one of the principal collection mechanisms is the collision between the particle and a moving drop of liquid (usually water). It has also been observed that if the particles are charged before they enter the filter, they will be collected with a higher efficiency than if they am not. This has led to the ESP-baghouse combination, in which an old ESP that does not meet new emission standards has a baghouse attached to its downstream side. The particles passing from the ESP to the baghouse are mostly the smallest of the particles that entered the ESP, and many of them are charged. The measured performance of this combination is often better than one would predict for an ESP plus a baghouse treating uncharged particles. 9.2.3 Filter Media Whether a filter behaves as a surface or a depth filter depends on the type of filter medium used. For shake-deflate baghouses (Fig. 9.4) the filter bags are made of tightly woven fibers, much like those in a pair of jeans. (The reader is invited to look at the sun through a single layer of such fabric, seeing that it has some pinholes, allowing light to come through, and to blow into such a fabric, observing that one can breathe in and out through one.) Pulse-jet baghouses (Fig. 9.5) use high-strength felted fabrics, so that they act partly as depth filters and partly as surface filters. This allows them to operate at superficial velocities (air-to-cloth ratios) two to four times those of shake-deflate baghouses; in recent years this higher capacity per unit size has allowed them to take market share away from the previously dominant shake-deflate type baghouses. Filter fabrics are made of cotton, wool, glass fibers, and a variety of synthetic fibers. The choice depends on price and suitability for the expected service. Cotton and wool cannot be used above 180 and 200~F, respectively, without rapid deterioration, whereas glass can be used to 500~F (and short-term excursions to 550。F). The synthetics have intermediate service temperatures. In addition the fibers must be resistant to acids or alkalis if these are present in the gas stream or the particles as well as to flexing wear caused by the repeated cleaning. Typical bag service life is 3 to 5 years. Generally fibers that have many small microfibers sticking out their sides form better cakes than those that do not. The student should examine under a microscope a thread of cotton, which has such microfibers, and one of monofilament fishing line, which does not. 9.2.4 Scrubbers for Particulate Control Just as filters work by separating the flow of particle-laden gas into many small streams, so also scrubbers effectively divide the flow of particle-laden gas by sending many small drops through it. In air pollution control engineering, the term scrubber originally meant a device for collecting fine particles on liquid drops. Then when liquid drops were used to collect sulfur dioxide, the devices that did that were also called scrubbers. Recently, alas, some other types of devices have been marketed as dry scrubbers. In this chapter, we will use the original meaning of the term: a scrubber is a device that collects particles by contacting the dirty gas stream with liquid drops. Most fine particles will adhere to a liquid drop if they contact it. So if we can make the drop and the particle touch each other, the particle will be caught on the drop. Particles 50 μ and larger are easily collected in cyclones. If our problem is to collect a set of 0.5-μ particles, cyclones will not work at all. However, if we were to introduce a large number of 50-μ diameter drops of a liquid (normally water) into the gas stream to collect the fine particles, then we could pass the stream through a cheap, simple cyclone and collect the drops and the fine particles stuck on them. This idea is the basis of almost all scrubbers for particulate control. A complete scrubber has several parts, as sketched in Fig. 9.6. Most often, the gas-liquid separator is a simple cyclone of the type discussed in Sec. 9.1.2; water drops of the size encountered in most scrubbers pose few difficulties for such cyclones. The liquid-solid separator can be of many kinds although gravity settlers seem to be the most common. If possible, the engineer should try to save money by finding a place where the contaminated water stream can be recycled inside the plant without first removing the solids. There are many examples where that has been done successfully. Obviously, if there is no good way to deal with the contaminated water stream, then the scrubber has merely changed an air pollution problem into a water pollution problem. 9.3 Choosing the Collectors In choosing a primary particle collection device one must consider the size of the particles to be collected, the required collection efficiency, the size of the gas flow, the allowed time between cleanings, and details of the nature of the particles. The following rules of thumb may be helpful: 1. Small or occasional flows can be treated by throwaway devices, e.g., cigarette and motor oil filters, in which the collected particles remain in the device. Large and steady flows require collection devices that operate continuously or semicontinously, and from which the collected particles can be removed continuously or semicontinuously. A throwaway device may be used as a final cleanup device, e.g., a high-efficiency filter may remove the last few particles from the air flowing to a microchip production clean room. 2. Sticky particles (e.g., tars) must be collected either on throwaway devices or into a liquid, as in a scrubber or cyclone, filter, or wet ESP whose collecting surfaces are continually coated with a film of flowing liquid. There must be some way to process the contaminated liquid thus produced. 3. Particles that adhere well to each other but not to solid surfaces are easy to collect. Those that do the reverse often need special surfaces, e.g., Teflon-coated fibers in filters that release collected particles well during cleaning. 4. Electrical properties of the particles are of paramount importance in ESPs, and they are often significant in other control devices where friction-induced electro-static charges on the particles can aid or hinder collection. 5. For nonsticky particles larger than about 5 μ, a cyclone separator is probably the only device to consider. 6. For particles much smaller than 5 ix one normally considers ESPs, filters, and scrubbers. Each of these can collect particles as small as a fraction of a micron. 7. For large flows the pumping cost makes scrubbers very expensive; other devices are chosen if possible. 8. Corrosion resistance and acid dew point (Sec. 7.12) must always be considered. 9.4 Summary 1. Gravity settling chambers, cyclones, and ESPs work by driving the particles to a solid wall where they form agglomerates that can be collected. These three devices have similar design equations. 2. Filters and scrubbers divide the flow. They have different design equations from wall collection devices and from each other. 3. Both surface and depth filters are used for particle collection. Surface filters are used to collect most of the particles in a heavily laden gas stream. Depth filters are mostly used for the final cleanup of air or gas that must be very clean or for fine liquid drops, which coalesce on them and then drop off. 4. To collect small particles, a scrubber must have a very large relative velocity between the gas being cleaned and the liquid drops. For this reason co-flow scrubbers are most often used. The venturi scrubber is the most widely used type of co-flow scrubber.