Biomass and Bioenergy 23 (2002) 113 128

Review

Ancillary equipment for biomass gasication

of Mechanical Engineering, Iowa State University, 2025 Black Engineering Building, Ames, IA 50011, USA of Mechanical Engineering and Chemical Engineering, Center for Sustainable Environmental Technologies, Iowa State University, 286 Metals Development Building, Ames, IA 50011, USA Received 18 August 2000; received in revised form 19 February 2002; accepted 13 March 2002

Abstract Considerable research has been conducted in the past 20 years to advance gasication technology and adapt this technology for biomass applications. The gasication process itself is relatively well developed, and producer gas can be generated at fairly large scales from a variety of feedstocks. The largest impediments now facing the application of biomass gasication for energy production are the inabilities of the current ancillary systems to allow for economical production of a clean producer gas, free of contaminants. Ancillary systems are operations other than the actual gasication process and generally fall into two categories: fuel preparation and feeding of the feedstock (prior to gasication) and gas-cleaning systems (subsequent to gasication). Fuel preparation systems ensure proper feeding of biomass into a gasication vessel and the obtainment of good-quality gas, while gas-cleaning systems remove contaminants from the producer gas to allow reliable and e cient operation of internal combustion engines, gas turbines, or fuel cells. These ancillary systems, particularly the fuel preparation and feeding systems, generally designed specically for a single gasication system, are often not mentioned when gasication systems are discussed. This paper attempts to collect this information together for a better understanding of overall gasication systems. ? 2002 Published by Elsevier Science Ltd.
Keywords: Biomass gasication; Fuel preparation; Feeding systems; Hot gas cleanup

1. Introduction Thermal gasication, which is the generation of gaseous fuels by thermal processes, is a particularly attractive means for converting biomass into useful energy. Biomass, such as wood and switchgrass, are renewable energy resources that are much less prone to polluting the environment than fossil fuels and, therefore, have an excellent potential for displacing fossil fuels in some applications. The conversion of solid biomass to gaseous fuel provides opportunities
Corresponding author. Tel.: +1-515-294-7934; +1-515-294-3091. E-mail address: rcbrown@iastate.edu (R.C. Brown).

fax:

for retrotting coal-red boilers, displacing natural gas in process heating, and developing distributed power systems using biomass that are based on internal combustion engines, gas turbines, or fuel cells. Considerable research has been devoted to improving the design of gasier reactors. However, just as important to the success of biomass energy systems is the performance of ancillary equipment, especially biomass feeding systems and devices to reduce gaseous and particulate contaminants from the producer gas. Unfortunately, less attention has been paid to the improvement of this ancillary equipment. While gasication of biomass occurs in a relatively simple reactor, two other processes are essential for successful energy production: a fuel preparation and

0961-9534/02/$ - see front matter ? 2002 Published by Elsevier Science Ltd. PII: S 0 9 6 1 - 9 5 3 4 ( 0 2 ) 0 0 0 3 8 - 7

114

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

feeding system and a gas cleanup system. Fuel preparation and feeding systems encompass the drying and sizing of the feedstock, as well as feeding of fuel into the gasier. Gas cleanup technologies include pollution control, elimination of particulate matter, and the reduction of tars and alkali present in the raw producer gas. Fuel preparation is crucial in order to obtain the highest-energy producer gas from a given feedstock. Gas cleanup is paramount in achieving the stringent gas quality standards required for devices such as turbines and fuel cells. The development of e ective ancillary systems is an important hurdle to overcome in order to begin the transition toward renewable energy from biomass. This paper is a review of fuel-feeding systems and gas-cleaning systems relevant to biomass gasication systems. 2. Fuel preparation and handling Feedstock arriving at the site of gasication can rarely be directly fed into the gasier. The feedstock must be properly prepared and delivered to the gasier, a process termed fuel preparation and handling. This process entails drying the feedstock to remove excess moisture, reducing the feedstock to the proper size, and entering the feedstock into the reactor, which may or may not be pressurized. The following section outlines these processes and describes some of the equipment commonly used for these tasks. 2.1. Drying In order to ensure reliable, consistent feeding and optimize gasication products, the feedstock must be properly dried and sized. Freshly harvested biomass may have moisture content as high as 60%. Most gasication systems prefer moisture in the range of 10 20%. Moist fuel is likely to clog the feeding system and can lower the heating value of the gas, while improperly sized fuel will also result in poor feeding and inhibit bed uidization [1]. The proper size and moisture content are dependent on requirements of the gasication and fuel feeding systems and are often determined by practical experience with a particular gasier. The drying and sizing sequence depends on the drying and sizing equipment used. Some dryers may require the feedstock to be sized prior to

drying, while some sizing equipment may require that the feedstock be dry. These constraints may require stages of drying and sizing in order to reach the desired size and moisture content. Drying processes require fairly large energy inputs (to produce the necessary heat), which reduces overall plant e ciency. This ine ciency can be mitigated by directing waste heat from the gasication process to the dryers. Sources of heat from within a biomass plant include hot producer gas, turbine exhaust, heat exchanger exhausts and combustion of by-products [2]. The exhaust from drying systems must be monitored for volatile organic compounds, which arise from either vaporization of volatile components in the biomass or from thermal degradation of the biomass in the dryer. When these volatile components are released, they give rise to a slightly smoky exhaust plume called blue haze, which can be hazardous. These emissions are usually released when the tem perature of the feedstock is greater than 100 C; therefore, an e ort should be made to prevent the feedstock from reaching this temperature. Cleanup equipment such as cyclones and adsorption beds may be necessary to lter the exhaust plume prior to its release from the drying system [2]. Another area of concern for drying systems is the risk of re and=or explosion. Fires and explosions can result from the ignition of a feedstock dust cloud in the dryer or from the ignition of combustible gases released from the feedstock during drying. If a sufcient amount of oxygen is present in the drying medium and a su ciently high temperature is reached in the dryer, then ignition may occur. Oxygen concentrations greater than 10% are considered potentially dangerous [2]. Many types of dryers are commercially available; the selection of an appropriate dryer depends on several factors. The size of the particle to be dried, the type of fuel (whether woody or herbaceous), and the necessary drying capacity of the system are all important considerations. Perforated oor bin dryers o er the simplest drying system for biomass. Primarily used for drying grain, perforated oor bins dry feedstocks in batches and are suitable for small biomass plants ( 1 MW). This system consists of a large bin through which hot gases are allowed to pass, a process known as

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

115

Fig. 1. Perforated oor bin dryer. Reproduced with permission [2]. Fig. 3. Rotary cascade dryer. Reproduced with permission [2].

Fig. 2. Band conveyor dryer. Reproduced with permission [2].

through-circulation (see Fig. 1). Wet feedstock forms a xed bed above the perforated oor. A relatively shallow bed depth of 0.4 0:6 m is recommended. Even with this shallow bed depth, vertical moisture gradients arise in the bed of material. The dried material needs to be mixed thoroughly before use, and often needs to be stored for a period of time for moisture to reach an equilibrium state. Since the lowest levels of the feedstock bed reach the gas temperature, inlet gas temperatures should be less than 100 C to avoid gaseous emissions [2]. Band conveyor dryers o er the simplicity of the perforated oor dryer while allowing a continuous stream of feedstock to be dried. The feedstock is carried through the drier on a permeable band as the drying medium is blown by fans through the band and feedstock (see Fig. 2). Drying is uniform for band conveyors due to the shallow depth of the feedstock on the band (215 cm). Gas may ow upward or downward, and the feedstock may be transported through the dryer several times. Bands can be arranged in series, or they can be arranged one above the other, with each band discharging onto the one below it. Band conveyors are divided into sections, where the drying medium can be continually added. Residence time can be controlled by adjusting the speed of the bands.

Temperatures in band conveyors are limited to 350 C due to problems lubricating the conveyor, chain and roller drives at higher temperatures. The drying medium may need to be ltered prior to exhausting due to gaseous emissions [2]. Rotary cascade dryers are used widely in industry to handle a wide range of materials and are the most common device used for drying in large-scale biomass gasication projects and large wood-chip combustion plants. The device is made up of a cylindrical shell, which slowly rotates (110 rpm). The shell diameter is as small as 1 m for smaller capacities and as large as 6 m for high-throughput applications. Longitudinal ights inside the shell lift the feedstock and cascade it through the drier as drying medium ows horizontally through the cylinder (see Fig. 3). The dryer is slightly inclined so that the feedstock progresses through the dryer as the material rotates [2]. One of the drawbacks of the rotary cascade dryer is its requirement for much larger gas volumes than those in through-circulation systems due to its lower e ciency of heat transfer. Temperatures up to 1000 C can be attained, but temperatures greater than 600 C require expensive steel alloys or refractory lining. Operating at such temperatures requires close monitoring to prevent re or explosion. Rotary cascade dryers also require cyclones and=or bag lters due to the unavoidable entrainment of feedstock particles in the drying medium [2]. At larger scales ( 10 MWe), more advanced drying systems may become practical. Two such systems are the uidized bed steam dryer developed by the Danish company, Niro, and the pneumatic conveying steam dryer developed by Stork Engineering. Both systems are relatively expensive but are cost e ective at large scales. The Niro steam dryer holds biomass

116

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

Fig. 4. Fluid bed steam dryer. Reproduced with permission [2].

Fig. 5. Pneumatic conveying steam dryer. Reproduced with permission [2].

in 16 cells surrounding a central heat exchanger, as shown in Fig. 4. Superheated steam leaves the heat exchanger and enters the cells, uidizing a bed of moist feedstock. The superheated steam vaporizes the moisture in the feedstock, and the vaporized moisture is entrained. The steam passes through an internal cyclone to remove entrained particles. After the cyclone, the excess steam from evaporation is discharged. The steam is reheated in a large heat exchanger to 200 C and then returns to the uidized bed. The excess steam ( 150 C) may be used as process steam elsewhere, and the Niro system exhausts no gaseous emissions to the atmosphere, eliminating the need for external cyclones or bag lters [2]. The Stork Engineering system, commercially available as the Exergy system, operates similarly. The Exergy system, illustrated in Fig. 5, operates with a series of heat exchangers rather than a uidized bed.

Moisture is vaporized as it is entrained by superheated steam, passes through a cyclone and then is discharged. Like the Niro system, the Exergy system has no gaseous emissions and can provide process steam [2]. For systems in which gasication occurs within a circulating uidized bed (CFB) or bubbling uidized bed (BFB) with a sand circulation system, the drying process may be integrated with the fuel-feeding and bed recirculation systems. This arrangement, developed by Imatran Voima Oy (IVO, and now known as Fortum), is termed a bed-mixing dryer. The innovative system takes advantage of the sensible heat of the hot bed material circulating through the gasication system. Bed material and superheated steam at atmospheric pressure are mixed together prior to the introduction of the wet fuel. When the fuel is injected into the ow, the sensible heat of the hot bed material vaporizes much of the moisture from the fuel. The bed material and fuel are entrained and transported by the superheated steam to a cyclone, where the bed material and fuel are separated from the steam and fed together into the bed. The excess moisture, now in the form of steam, is removed from the steam circuit and condensed at atmospheric pressure [3]. There are two great advantages of the bed-mixing dryer: increased e ciency of the overall gasication system and absence of both moving parts and heat exchangers in the drying circuit. The increase in e ciency is due to the recovery of the latent heat of vaporization during the condensation of excess steam. In a demonstration conducted by IVO, the e ciency of a plant incorporating a bubbling uidized bed boiler increased by 10 15% when the original fuel moisture content was 50% on a wet basis. One additional advantage is that long bers and agglomerates in the fuel can be broken up during the interaction with the bed material [3]. A simpler method of drying biomass is a mechanical or hydraulic press. Mechanical presses are used primarily in pulp-and-paper mills to dry bark. These devices squeeze water from extremely moist feedstocks, but they can reduce moisture contents only to approximately 55%. Another disadvantage of these machines is that they consume large amounts of energy and require a great deal of maintenance. Mechanical presses are recommended for removing some of the water from wet feedstocks before thermal drying [4].

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

117

One novel approach to biomass drying has been proposed by a group of Italian researchers [5]. This group suggests decentralizing the biomass-drying task and assembling a network of regional gas turbinedryer units. Each gas turbinedryer would be composed of a gas turbine power plant and a rotary cascade dryer. Inclusion of a heat recovery steam generator with such a system would allow steam reinjection to improve the e ciencies of the turbine and dryers. Each unit in the network would then send its dried biomass to a centrally located commercial power plant. The researchers have reported e ciencies of 35.4% for the network, compared to 31.5% for the power plant alone and 25.6% for a stand-alone gas turbinedryer unit. Decentralizing the drying process reduces transportation and storage costs, which can be large. However, it is not protable to supply a thermal power plant for each distributed turbinedryer unit owing to the high specic cost of small power plants. It is more e ective to feed a single, large power plant from a network of small gas turbinedryer units [5]. 2.2. Sizing The two most common devices for comminuting biomass to sizes appropriate for gasication are knife chippers and hammermills. Chippers are high-speed rotary devices, operating at speeds up to 1800 rpm, and are better suited for comminuting wood. The wood enters a cavity around a rotating cylinder where cutting blades attached to the cylinder and stationary blades break the wood into smaller pieces. Care must be taken to remove any metal that may be mixed with the wood, as this can severely damage the knives; however, this problem is usually limited to waste wood residues and is not a large concern for dedicated feedstocks. Hammermills are also rotary devices. Rather than being cut by blades, as in chippers, biomass is crushed by large metal hammers. As biomass falls into the hammermill, larger pieces are crushed by the spinning hammers against a breaker plate and then pulverized between the hammers and a screen at the bottom of the mill, as illustrated in Fig. 6. The screen determines the size of the comminuted particles. Fuel should be dried before size reduction in a hammermill [4]. Hammermills are suited to process wood as well as herbaceous energy crops such as switchgrass. Many hammermills can grind biomass small enough to pass a No. 30 sieve.

Fig. 6. Hammermill. Reproduced with permission [1].

Hammermills operate best when initial feedstock size is less than 4 cm (1:5 in); an auxiliary crusher may be necessary to meet this requirement [6]. Tub grinders are becoming a viable alternative to chippers and traditional hammermills, particularly for the sizing of forestry residues. Tub grinders are small, mobile hammermills, often designed as pull-behind units for agricultural uses or mounted on tractor-trailers for larger waste-removal uses. Tub grinders consist of a rotating tub, which feeds material into a hammermill. The mill discharges the comminuted material onto a conveyor that exits the tub grinder. The advantages of tub grinders over stationary hammermills are their mobility and ability to deal with large objects. Some tub grinders are able to process whole logs and stumps at a rate of 100 ton=h. The mobility of these grinders also allows grinding to occur on-site, potentially reducing transportation costs of the feedstock. It is conceivable to envision a system of on-site tub grinding at dedicated-feedstock wood farms, with further sizing performed with larger hammermills located within biomass power plants. In order to ensure that feedstocks have been properly sized, screens may be used. Screens may be used at the inlet of comminution equipment to divert undersized material while screens at the exit recirculate large pieces that require further size reduction. Other methods of ensuring proper size are by otation and air classication, using buoyancy and pneumatic

118

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

principles, respectively, to separate the di erent sizes. These two methods are best used for woody feedstocks, although otation methods saturate the fuel with water, and air classication is more complex than simple screen systems [1]. 2.3. Fuel-feeding systems The fuel-feeding systems convey the fuel from storage bins and hoppers to the gasier, as illustrated in Fig. 7. Since feeding systems have been developed in an ad hoc fashion, only a limited amount of information has been published on their design and operation. Systems are typically custom-designed for the specic application in mind. Feeding systems may or may not incorporate the drying and sizing systems mentioned above. An ideal feeding system provides smooth and continuous feeding and allows for accurate control of the feed rate. The system should be relatively insensitive to variations of fuel size and must maintain su cient pressurization to prevent the back ow of gases from the gasier to the feeding system [7]. Typically, the fuel-feeding system consists of two parts: fuel transport from storage to the gasier and injection into the gasier.

Fuel transport is best accomplished by one of the three methods: pneumatic transport, screw conveyance, or belt conveyance. Pneumatic transport is e ective for long distance transport of sized fuel. Capital costs are fairly low, although pneumatic transport requires high-energy consumption in order to generate the high-pressure transport air necessary for its operation. Screw conveyance is more generally suitable for transporting sized fuels over relatively short distances. Screw conveyors are inexpensive and require little energy consumption but are generally limited to transport distances less than 6 m. Belt conveyors are an e ective means of transport for unsized fuels (bales and logs). Like screw conveyors, they are relatively inexpensive and consume little energy [8]. Gasier feeding systems range from the simplest to the most complex part of a gasication energy system, depending upon the fuel and the kind of gasier. Fluidized beds, for example, present many options for injecting biomass into the reactor. The feedstock can be fed above, below, or directly into the uidized bed. Underbed and in-bed feeding provides a longer residence time for the fuel, ensuring complete gasication. Underbed and in-bed systems are often pneumatic but cooled augers can also be used. In either case, pressure seals must be used to prevent back ow of

Fig. 7. Illustrative feeding system incorporating drying and sizing equipment.

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

119

producer gas or bed material into the feeding system. This increases the complexity of the system and the potential for failure. Overbed feeding, often accomplished by gravity feeding from elevated hoppers, reduces the complexity of the system, but may allow the feedstock to become entrained by the gasication medium without being gasied [1]. When the gasier is pressurized, the feeding system must also be pressurized, regardless of where the feedstock is fed to the reactor. As mentioned previously, the degree of pressurization varies greatly among gasiers, from a few inches of water (virtually atmospheric) to pressures greater than 100 atm ( 10 MPa). The following feeding devices are common designs for pressurized gasication systems, but they function well for gasiers operating at atmospheric pressures with little or no modication. In addition, all the feeders discussed below have the ability to feed both woody and herbaceous feedstocks that have been properly sized and dried. The most common device for feeding solid fuel into a pressurized reactor is a lock hopper system. Lock hoppers work best for systems operating at less than 3:5 MPa (standard atmospheric pressure is 0:1 MPa). Above this pressure, excessive quantities of inert gas are required to back pressurize the system, and the strain on the lock hoppers increases the maintenance requirements [7]. One prominent design is the lock hopper system, shown in Fig. 8, developed by Thomas R. Miles Consulting Engineers, which can feed 5 ton=h of biomass into 10 25 bar pressure (1.0 2:5 MPa). The lock hopper system functions by pressurizing a lock hopper after the biomass has been loaded into the hopper. Biomass is then discharged into a metering bin, and metering screws transport feedstock into the injector screw, which feeds the biomass into the reactor [7]. The rotary-valve feeder shown in Fig. 9 is another simple system for feeding pressurized gasication systems ( 1:5 MPa) and functions much like a revolving door. The fuel is conveyed from an unpressurized area into a pressurized area between the blades of a rotor and the frame of the feeder. Pressure sealing is secured by an accurate adjustment of the rotor blades in the frame of the feeder, and the total discharge of the feeder pockets is ensured by blowing out the pockets with high-pressure steam. However, the rotor may stick as deposits build up on the frame of the feeder,

Fig. 8. Lock Hopper feeding system. (Reproduced with permission: Thomas R. Miles Consulting Engineers.)

and these feeders are susceptible to pressure losses and the back ow of producer gases [7]. Piston and screw feeders are more-advanced biomass gasication feeders. Piston feeders can operate in gasication systems pressurized up to 4.5 15:0 MPa, with capacities ranging from 11 to 115 m3 h1 . Metering screws feed biomass into a long, narrow channel where a piston compresses the biomass into a solid fuel plug, which is pushed into the gasier. This plug must be su ciently compressed to prevent back ow of gases into the feed line [7]. This system is sometimes referred to as a cigar feeder or a cigar burner as the plug burns much like a cigar. Screw feeders of the type shown in Fig. 10 operate similar to piston feeders but have a much lower pressurization range (0.5 1:5 MPa). In a screw feeder, an auger, rather than a piston, compresses the feedstock into a compact plug. The compression of the plug is aided by the tapering of the feed channel as it nears

120

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

the volume of the uidized bed is relatively small, the uid action of the bed is su cient to distribute the feedstock properly. However, large uidized beds of the size expected for commercial power production may not achieve good mixing of fuel within the bed. Poor mixing of fuel yields uneven gasication. In such cases, multiple inlets are required to eliminate such gradients. 3. Hot gas cleanup processes Raw producer gas contains tar, char, mineral matter, and other impurities. These contaminants must be removed before the producer gas can be used for chemical synthesis or red in power generation equipment. Tars can foul heat exchange surfaces and engine valves or they may not burn adequately in combustion chambers of gas turbines. Alkali forms deposits on cold heat exchange surfaces and turbine blades where it promotes corrosion [9,10]. Removal of tar and alkali is essential to the long-term success of biomass gasication processes. One of the challenges of hot gas cleanup in gas turbine applications is to maintain the producer gas at the temperature at which it leaves the gasier. Although gas cooling can simplify gas cleaning, it reduces the overall e ciency of the power cycle. Therefore, most development e orts today focus on hot gas cleanup techniques. 3.1. Particulate control Particles smaller than 1 m may need to be removed from producer gas. Cyclones, wet scrubbers, ceramic lters, and electrostatic precipitators have all been

Fig. 9. Rotary-valve feeder. (Reproduced with permission: A. Ahlstrohm Corp.)

the gasier. As in piston feeders, the feed plug forms a barrier to prevent the back ow of gases and bed material from the gasier [7]. With any feeding system, feed line temperature and pressure should be monitored in order to detect plugs in the line. Fuel that is too large or too moist is generally the cause of feed line plugging, although backpressure may also prevent the fuel from moving forward. Another problem that can occur is gasication of the fuel in the feed line. Measures should be taken to insulate the line and prevent the system from reaching temperatures at which pyrolysis commences. Another point that must be considered is the distribution of the feedstock within the gasier. When

Fig. 10. Screw feeder. (Reproduced with permission: Kamyr., Inc.)

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

121

Fig. 11. Cyclone particle collector. Reproduced with permission [11].

employed for particulate removal with varying degrees of success. Cyclone collectors, while among the simplest particulate control device, are very e ective for removing all but the nest particles. A cyclone is a device without moving parts in which the velocity of an inlet gas stream is transformed into a conned vortex from which centrifugal forces tend to drive the suspended particles against the wall of the cyclone body [11]. The gas stream exits through the top of the cyclone, while particles settle at the bottom of the cyclone, as illustrated in Fig. 11. Cyclones can be designed to capture particles as small as 5 m. Cyclones are frequently arranged in series, with the rst cyclone designed to capture the largest particles while subsequent cyclones are designed to capture progressively ner particles. A comprehensive discussion on cyclone design can be found in [11]. Unfortunately, particles smaller than 5 m, which escape capture in cyclones, can degrade performance of many kinds of power systems. Ceramic lters, illustrated in Fig. 12, consist of arrays of candle-shaped elements, which are e ective in removing ne particles not captured by cyclones. A typical candle element is 1:5 m long and 60 mm in diameter. Filter arrays are employed in series to

ensure complete particulate removal. The ceramic candle lters that are currently available are made of clay-bonded silicon carbide (SiC). The main body is composed of course-ground silicon carbide, while a thin outer layer made of ne-ground silicon carbide or aluminosilicate surrounds the main body. It is this outer layer that performs the ltering duties [12]. Westinghouse Electrical Corporation has been testing these lters since 1993. Ceramic candles have collection e ciencies approaching 100% [12]. There is concern, however, about the reliability of the ceramic lters when operated at extremely high temperatures ( 815 C) [13], as well as the possibility of chemical degradation, creep, and static fatigue failure over time [14]. Concerns about reliability of ceramic lters have led to the recent development of sintered metal lters, which, like ceramic lters, are barrier lters capable of high collection e ciencies for micron-sized particles. Metal lters are produced by lling molds with powdered metal and heating them to slightly below the melting point of the metal. Under these conditions, the metal particles sinter together to form a porous metal matrix. The pore size of metal lters is determined by the type of alloy, the sintering temperature and time, and the gaseous atmosphere in which the sintering process occurs [15]. The advantages of porous metal lters with respect to ceramic lters lie predominantly in the material qualities of the porous metals. Porous metals have demonstrated good mechanical strength under constant and transient loads. They have performed well at high temperatures with special alloys operating at temperatures up to 1000 C. Very importantly, they stand up well to thermal shock, which occurs during periodic cleaning of candle-barrier lters. In addition to these physical characteristics, these lters are resistant to the corrosion under typical gasication conditions. Porous metal lters are easy to machine and weld [16]. Granular bed lters provide an alternative to ceramic candle and porous metal lters. Granular bed lters are basically large bins containing a quantity of a granular material, often limestone or alumina, through which a stream of gas is forced. Granular bed lters may employ either static or moving beds of granular material. In static bed lters, a dirty gas stream ows down through the bed, where particles impact and adhere to the granules. As particles accumulate in

122

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

Fig. 12. Candle lter system. Reproduced with permission [12].

the bed, the void spaces in the bed become clogged. When this occurs, pressure drop increases, and the bed must be cleaned, requiring an interruption in the ltering process, which occurs with candle-barrier lters. Cleaning is accomplished by a reverse ow of gas through the bed. Moving bed lters, shown in Fig. 13, alleviate the need to periodically interrupt ltering to clean the granular bed. In these lters, the lter media ows downward through the lter, while dirty gas ows upward through the bed. Collection e ciency for a moving bed lter of this type developed by Combustion Power Company was greater than 99% for particles greater than 4 m in size, while 93% of particles less than 4 m were retained [17]. In another series of short duration tests ( 55 h), the collection e ciency for particles less than 10 m in size was 93% [18]. 3.2. Tar control Removal of tar is among the greatest technical challenges to the wide application of gasication systems. As mentioned previously, tars can foul heat exchanger and engine parts and are slower to burn than the gases and vapors in producer gas. Much of the current research on biomass gasication systems focuses on possible remedies for this problem. However, the lack of uniform sampling methods and consistent denitions hinder the search for a solution [19]. A tar-sampling

procedure was outlined in June 1998 in an attempt to standardize biomass gasication tar research [20]. In a review of tar research, Milne et al. [19] dened tars as the organics produced under thermal or partial-oxidation regimes (gasication) of any organic material: : :and are generally assumed to be largely aromatic. This denition is adopted for the following discussion. Tars can be classied into four categories [21]: primary products, secondary products, alkyl tertiary products, and condensed tertiary products. Primary products comprise relatively simple organic compounds derived from cellulose, hemicellulose, and lignin. Secondary products are characterized as phenolics, which are alcohols with the hydroxyl group bonded directly to an aromatic (benzene) ring, or as olens, which are hydrocarbons with carbon carbon double bonds. Alkyl tertiary products consist mainly of methyl derivatives of aromatics. Condensed tertiary products include benzene, napthalene, acenaphthylene, anthracene=phenathrene, and pyrene. Aromatic compounds contain a benzene ring (six carbon atoms alternately single- and double-bonded), a stable molecular structure. The types of tars produced is a function of both the time and temperature over which reaction occurs, sometimes known as reaction severity [21]. Higher temperatures and longer residence times correspond to a higher reaction severity. As reaction severity

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

123

Fig. 13. Moving granular bed lter. Reproduced with permission [18]. Table 1 Maturation scheme for tars [22] Temperature ( C) 400 500 600 700 800 900

Products Mixed oxygenates Phenolic ethers Alkyl phenolics Heterocyclic ethers Polynucleic aromatic hydrocarbons (PAH) Larger PAH

increases, the tars produced are more likely to be secondary and tertiary products. Primary and tertiary tars are mutually exclusive, indicating that primary tars are destroyed before the formation of tertiary products [21]. The presence of both primary and tertiary products indicates process upsets. Tar composition as a function of reactor operating temperature is shown in Table 1 [22]. Currently, methods to remove tars from producer gases fall into one of the three categories: physical

removal, thermal conversion of tars, and catalytic destruction of tars. Removal of tars via physical processes is a separation process. Thermal conversion, also called cracking, is the use of extremely high temperatures to disintegrate complex organic compounds into more benign forms. Catalytic destruction of tars involves the use of catalysts to promote disintegration of tars. Physical processes for tar removal include wet scrubbers, demisters, wet granular bed lters and wet electrostatic precipitators. These processes are only e ective for tar removal when the producer gas has been cooled to less than 100 C, which is thermodynamically ine cient for power systems. A second problem with these methods of tar removal is that the water used in the applications must be treated to remove the tars before it can be disposed. The costs of water treatment may prohibit the use of these methods. Dry applications such as fabric and ceramic lters may be used if the producer gas temperature can be reduced to less than 150 C [19]. Thermal conversion of tars is achieved by passing the producer gas through a second, high-temperature reactor where tars decompose or reform to carbon monoxide (CO), hydrogen gas (H2), and other light gases. Primary tar products readily decompose, with CO yields reaching 50 wt%, while condensed tertiary products are more di cult to crack. Milne et al. [19] concluded that temperatures in excess of 1000 C and reasonable residence times are necessary to destroy aromatic tars without the use of a catalyst. This temperature presents materials problems, requiring expensive alloys. An even greater concern is that the high-temperature reaction conditions produce soot, which may be even more problematic than tars. Addition of steam and=or oxygen is e ective in increasing cracking rates in the temperature range of 950 1250 C. Oxygen at 600 700 C accelerates the destruction of primary products, and inhibits the formation of aromatics; however, once benzene, the primary component of aromatic compounds in tar, is formed it cannot be destroyed by oxygen. The addition of steam has been reported to produce fewer refractory tars, enhance phenol formation, reduce the concentration of other oxygenates, have only a small e ect on the conversion of aromatics, and produce tars that are easier to reform catalytically [19].

124

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

The addition of steam also facilitates the water=gas shift reaction (CO + H2 O CO2 + H2 ). As greater amounts of steam are introduced into the system, H2 and CO2 concentrations increase, while CO and H2 O concentrations decrease [23]. This reaction is extremely benecial for methanol production applications, in which the ideal H2 =CO ratio is 2. The H2 =CO ratio for raw producer gas is usually less than 1; steam addition to a catalytic tar-cracking system can increase the H2 =CO ratio high as 13 [24]. Catalytic tar destruction is accomplished by passing the producer gas through a bed of catalytic material. The use of catalysts mixed with the feedstock prior to gasication has been attempted but has not been as successful as the use of separate catalytic reactors [19]. The catalysts may be metallic, such as nickel or alumina, or non-metallic, such as dolomite or limestone. Dolomite beds have been reported to cut tar levels in half at temperatures around 820 C [25]. Metallic catalysts have been more successful, reducing tar levels by more than 90% [19]. Two metallic catalysts operated in tandem have demonstrated as successful tar reduction. In research at Battelle Laboratories, the combination of DN-34, a proprietary catalyst made of essentially alumina, and ICI 46-1 was successfully used to catalytically crack tars [26]. The rst bed of DN-34 catalytically destroyed almost 100% of the tars present in the gas stream. DN-34 showed little deactivation during the 50 h of operation. This catalyst also showed the potential to destroy polynuclear aromatic hydrocarbons (PAH), which are the most di cult tars to control. The second bed of ICI 46-1, operated with steam addition, demonstrated the ability to decompose virtually all methane [24]. Thus, this combination of catalysts is ideally suited for methanol synthesis. Recently, Abatzoglou et al. [18] evaluated a proprietary catalyst for tar cracking and found cracking e ciencies greater than 95% during the rst 60 h of testing. Deactivation was signicant after this time, but the catalyst could be regenerated to its original activity. These researchers estimated that the useful life of the catalyst was 4000 5000 h, depending on the gasication feedstock. The key design parameters for the tar cracker are the linear gas velocity, the residence time of the producer gas, and the size of the catalyst particles [18].

In bench-scale testing, researchers at the University of Hawaii at Manoa found that steam injection into the catalytic tar cracker did not greatly a ect tar cracking, but a ected producer gas yield. Prior to steam injection, producer gas yields of 1:49 m3 kg1 of biomass were observed, while the yield improved to 1:69 m3 kg1 of biomass when steam (steam-to-biomass ratio of 1.2) was injected into the cracker. The gas yield of the raw producer gas was 0:98 m3 kg1 of biomass. The yield of the catalytic steam reactor injection system was also higher than gasier steam injection system yields [23]. It is unclear whether the point of steam injection ultimately a ects the nal yield, but catalytic tar cracking with steam injection noticeably improved producer gas yields. Researchers at the Technical Research Centre of Finland (VTT) have explored the e cacy of nickel catalysts for destruction of tar. In addition to pelletized and crushed catalysts, tests were conducted with catalyst monoliths. The monoliths contained several square channels, which allowed the producer gases to pass without plugging even if char and ash particles had not been removed from the producer gas. Tests conducted on producer gas from xed-bed gasiers showed that carbonate rocks and nickel catalysts were e cient for tar decomposition. Carbon deposition, or coking, was most likely to occur if the gas did not contain signicant H2 O or CO2 , or if the catalyst temperature was too low. Poisoning of the nickel catalyst by hydrogen sulde (H2 S) was most likely when reactors were operated with high pressures and low temperatures. Temperatures greater than 900 C helped to mitigate against catalyst poisoning. Although monoliths may be benecial in reducing plugging of catalyst by particulate matter, monoliths have substantially less surface area for heat and mass transfer between the bulk gas and the catalyst than do pelletized or crushed catalysts [27]. 3.3. Alkali control Alkali metals, especially potassium, are prevalent in many biomass materials. Alkali readily vaporizes during gasication after which it condenses at a temper ature below about 600 C. The resulting alkali aerosol has several detrimental e ects. Deposited on metal surfaces, it forms a sticky lm that causes impacting

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

125

particulate matter to adhere to the surface forming ash deposits that interfere with heat transfer through boiler tubes or with aerodynamics of turbine blades. Alkali metals are also thought to be erosive to turbomachinery and corrosive to metal surfaces. In a study of biomass combustion, alkali species were released prominently following the release of volatile species, during the combustion of the remaining char. The release of alkali species was largely una ected by the changes in temperature and oxygen content, while the addition of steam resulted in altered speciation of the alkali, transforming some of the potassium chloride (KCl) produced into potassium hydroxide (KOH). Regardless of steam addition, KCl is the predominant alkali species released during combustion [28]. In a study of the release of alkali from biomass gasication [29], tests were performed on producer gas exiting the gasier at 700 C and main tained above 600 C. Researchers concluded that alkali species were largely in particulate form, possibly deposited on y ash, rather than vapor form. This conclusion was based on the low levels of alkali present downstream of a heated particulate lter. Potassium was measured at a concentration of 400 parts per billion (ppb), while sodium was measured at a concentration of 30 ppb. When the sample was taken from an unltered line, the concentrations rose quickly to levels exceeding the calibration of the analyzer (in excess of 1800 ppb). Concentration returned to the initial levels when the sample was taken from the ltered sampling line. If alkali exceeds limits recommended by turbine manufacturers, then some method for controlling alkali emissions is required. One recommendation is for total alkali concentration at the turbine inlet so as not to exceed 24 ppb [30]. When alkali concentrations must be reduced, there are two processes commonly used to attain the desired alkali levels in producer gas: adsorption and leaching. Adsorption is a process by which molecules become adhered to appropriate materials, or sorbents. Adsorption can be categorized as physi- or chemi-sorption. In physi-sorption, molecules are attracted to the sorbent by van der Waals forces, which are weak intermolecular forces caused by dipoledipole attractions. Molecules adsorbed via physi-sorption can form several layers on the surface but they can also be desorbed, or removed from the sorbent, relatively

easily. Chemi-sorption forms a chemical bond between the molecule and the surface of the sorbent. Researchers believe that because of this surface bond, chemi-sorption cannot adsorb multiple layers of molecules, reducing the amount of molecules that can be captured by a given amount of sorbent. The forces that adhere the molecule to the sorbent are much stronger for chemi-sorption than physi-sorption; thus, desorption is much more di cult [31]. When adsorption is used to capture alkali, the sorbent is known as an alkali getter, and the adsorption process is sometimes referred to as gettering. The ideal alkali getter has a high-temperature capacity, a rapid rate of adsorption, and a high loading capacity. Irreversible chemi-sorption is sometimes considered preferable to reversible physi-sorption since it is less likely to release alkali if a gasier upset produces transients in temperature or pressure [31]. On the other hand, chemi-sorption makes sorbent regeneration problematic. Adsorption can occur in the bed itself if sorbents are mixed into the uidized bed or it can take place in a separate, post-gasier sorbent bed. A number of materials have been tested as potential alkali getters including emathlite, bauxite, diatomaceous earth, kaolinite, and a variety of clays [31]. While all the materials displayed some ability to adsorb alkali, bauxite showed the most promise. Bauxite is composed of 80% Al2 O3 , 10% SiO2 along with a few impurities. At a residence time of 0:2 s, the alkali removal e ciency for bauxite was 99% [31]. Bauxite removes alkali via physi-sorption, while emathlite, kaolinite, and diatomaceous earth adsorb alkali via chemi-sorption, inhibiting sorbent regeneration. Due to the physical nature of adsorption via bauxite, regeneration is a relatively simple process, using boiling water to remove the alkali. Physi-sorption also has the advantage of exhibiting higher rates of alkali loading with decreasing temperatures, which should improve performance at process temperatures typical of gasication systems [31]. A bench-scale experiment was conducted in order to determine the ability of activated bauxite to remove alkali from actual producer gas [32]. An initial test was conducted without any sorbent in the alkali getter reactor, and the concentrations of potassium, sodium, and chlorine in the producer gas in this case were 28, 11, and 1309 ppm, respectively. When activated bauxite was inserted into the alkali getter reactor, the

126

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

Table 2 Alkali indices for unleached and leached biomass fuels [33] Fuel Rice straw Wheat straw Switchgrass Wood fuels Banagrass Unleached 2.05 2.15 0.24 0.18 0.71 Leached 0.20 0.33 0.02 0.06 0.11 0.45

concentrations of alkali downstream of the getter reactor were reduced to 0:07 ppm for potassium and 0:6 ppm for sodium, while chlorine concentration was relatively unchanged. The activated bauxite was much more e ective in capturing potassium than sodium. This is benecial for alkali removal from biomass gasication producer gas, which in general contains a much greater amount of potassium than sodium. An alternative approach to alkali control is the elimination of alkali in the feedstock. This is accomplished by selecting fuels with extremely low alkali. The pre-selection of alkali-free fuels drastically reduces the number of potential biomass feedstocks, especially herbaceous plant materials, which are generally much higher in alkali content than woody feedstocks. A second method of reducing the alkali content of the feedstock is the pre-treatment of the feedstock prior to its injection into the gasier. Since alkali in biomass occurs as water-soluble compounds, washing the biomass readily dissolves most of the alkali. In a process known as leaching, the biomass is washed in water followed by mechanical dewatering. Simple leaching removes more than 80% of the potassium and sodium, as well as more than 90% of the chlorine in herbaceous biomass [33]. Alkali index is a measure of the ash-fouling potential of a given feedstock when used as boiler fuel. It also gives a relative measure of the di culties presented by the fuel when gasied. Alkali index is dened as the ratio of a fuels alkali content (kg kg1 ) to heating value (GJ kg1 ). The alkali index of several feedstocks tests (rice straw, wheat straw, switchgrass, banagrass, bagasse, and wood fuel) is listed in Table 2. Fuels with an alkali index of less than 0:17 kg GJ1 of energy potential are considered to have a low severity fouling potential, while fuels with

an alkali index greater than 0:34 kg GJ1 are considered to have a high severity ash fouling potential in boilers. On average, simple leaching reduced the alkali index of biomass fuels by 82% [33]. Turn and co-workers have evaluated the e ectiveness of various leaching processes on the removal of alkali from banagrass. The feedstock was forage-chopped, and the severity of the leaching process was varied: unpressed, pressed (mechanical dewatering), pressedrinsedpressed (dewatering followed by a leaching cycle), and sizedpressed rinsedpressed (sizing of the feedstock followed by dewatering and a leaching cycle). Removal of potassium increased from 40% when only pressed to 70% when the leaching cycle was added. When the feedstock was sized prior to treatment, 90% of the potassium was removed from the feedstock. This treatment also removed 54% of the sulfur, 70% of magnesium, sodium, and phosphorous and 98% of chlorine [34]. 4. Conclusions Ancillary equipment for biomass gasication systems ranges from commercially available drying systems to alkali getter systems still under development. Upstream equipment (drying, sizing and feeding systems), although commercially available, are not well documented in the published literature. Downstream equipment (particulate, tar, and contaminant removal systems) has been the subject of many published studies but some of the equipment has yet to reach technical maturity. A number of specic development activities are recommended to advance gasication technology. On the upstream side, more complete information on the performance of various fuel handling and feeding systems would help energy system designers and planners evaluate the feasibility of biomass gasication. In terms of new technologies on the upstream side, feeding systems that are able to operate under positive pressure but without elaborate arrangements of lock hoppers and rotary valves would be an important development in pressurized gasication. On the downstream side, control technology is still sought for high-temperature removal of particles, tar, and alkali without the use of high energy or water inputs. Multi-contaminant control systems that simultaneously remove particles, tar,

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128

127

and alkali would be an important advancement for the use of gasiers for power production.
[13]

Acknowledgements This project was funded by the Iowa Department of Natural Resources, Des Moines, IA. The authors would like to specically thank Alison Kovac for her help in the development of this article. References
[1] FBT, Inc. Fluidized bed combustion and gasication: a guide for biomass waste generators. Prepared for Southeastern Regional Biomass Energy Program, 1994. [2] Brammer JG, Bridgwater AV. Drying in a biomass gasication plant for power or cogeneration. Proceedings of the Fourth Biomass Conference of the Americas, Oakland, CA, USA, 1999. p. 2817. [3] Hulkkonen S, Parvio E, Raiko, M. Advanced fuel drying technology for uidized bed boilers. Proceedings of the 13th International Conference on Fluidized Bed Combustion, Orlando, FL, USA, 1995. p. 399 403. [4] Elliott TC. Standard handbook of powerplant engineering. New York: McGraw-Hill, 1989. [5] Cipollone R, Cocco D, Bontto E. Integration between gas turbine plants and biomass drying processes. Proceedings of the Fourth Biomass Conference of the Americas, Oakland, CA, USA, 1999. p. 2918. [6] Snow RL, Allen T, Ennis BJ, Litster JD. Size reduction and enlargement. In: Perry RH, Green DW, editors. Perrys chemical engineers handbook. 7th ed. New York: McGraw-Hill, 1997. [7] Wilen C, Rautalin A. Handling and feeding of biomass to pressurized reactors: safety engineering. Bioresources Technology 1993;46(12):7785. [8] Koch T, Winter E, Christensen H. Feed preparation of straw. Prepared by ELKRAFT and the Danish Energy Agency for the IEA Biomass Thermal Gasication Activity, 1996. [9] Miles TR, Miles TR, Baxter LL, Bryers RW, Jenkins BM, Oden LL. Alkali deposits found in biomass power plants: a preliminary investigation of their extent and nature. Summary Report for the National Renewable Energy Laboratory, NREL Subcontract TZ-2-11226-1, 1995. [10] Zakkay V, Gbordzoe EAM, Sellakumar KM, Lu CQ. Performance of hot gas clean-up devices tested at the NYU DOE-PFBC facility. Proceedings of the Joint ASME=IEEE Power Generation Conference, Dallas, TX, USA, ASME 89-JPGC=GT-8, 1989. [11] Dullien FAL. Introduction to industrial gas cleaning. San Diego, CA: Academic Press, 1989. p. 55 90. [12] Newby RA, Lippert TE, Alvin MA, Bruck GJ, Sanjana ZN, Smeltzer EE. Hot gas lters for coal and biomass power [14] [15]

[16] [17] [18]

[19] [20]

[21] [22]

[23]

[24] [25] [26] [27] [28]

[29]

systems. Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Indianapolis, IN, ASME 99-GT-267, 1999. Wiant BC, Bachovchin DM, Onischak M, Carty RH, Ratcli M. Biomass gasication hot gas lter testing results. Proceedings of the International Gas Turbine and Aeroengine Congress and Exhibition, Birmingham, UK, ASME 96-GT-336, 1996. DOE. State-of-the-art-technology: hot gas particulate ltration. Washington, DC: US Department of Energy, 1994. Terpstra, RL, Anderson IE, Gleeson B. Development of metallic hot gas lters. Proceedings of the 2001 International Conference on Powder Metallurgy and Particulate Materials, MPIF, Princeton, NJ, USA, 2001. p. 8.84 8.97. Neuman P, Rohlig R, Kohstoss A, Semrau A. Metallic membranes. Filtration and Separation 1998;35(12):402. Saxena SC, Henry RF, Podolski WF. Particulate removal from high-temperature, high-pressure combustion gases. Progress in Energy and Combustion Science 1985;11:193251. Abatzoglou N, Bangala D, Chornet E. An integrated modular hot gas cleaning conditioning technology. Proceedings of the Fourth Biomass Conference of the Americas, Oakland, CA, USA, 1999. p. 9539. Milne TA, Abatzoglou N, Evans, RJ. Biomass gasier tars: their nature, formation, and conversion. NREL=TP-57025357, Golden, CO, 1998. Albrecht J, Deutsch S, Kurkela E, Simell P, Sjostrom K. Provisional protocol for the sampling and analysis of tar and particulates in the gas from large-scale biomass gasiers, Version 1998. Prepared for the IEA, 1998. Evans RJ, Milne T. Molecular characterization of the pyrolysis of biomass. 1. Fundamentals. Energy and Fuels 1987;1:12338. Elliott, DC. Relation of reaction time and temperature to chemical composition of pyrolysis oils. In: Soltes EJ, Milne TA, editors. Pyrolysis oils from biomass, ACS Symposium Series, vol. 376, Washington, D.C.: American Chemical Society, 1988. Zhou J, Ishimura DM, Kinoshita CM. E ects of injecting steam on catalytic reforming of gasied biomass. Proceedings of the Fourth Biomass Conference of the Americas, Oakland, CA, USA, 1999. p. 9917. Gebhard SC, Wang D, Overend R, Paisley MA. Catalytic conditioning of synthetic gas produced by biomass gasication. Biomass and Bioenergy 1994;7:30713. Leppalahti J, Kurkela S. Behaviour of nitrogen compounds and tars in uidized bed air gasication of peat. Fuel 1991;70:4917. Paisley MA. Method for hot gas conditioning. Columbus, OH: Battelle Memorial Institute, US Patent 5, 1996. p. 494,653. Simell P, Kurkela E, Stahlberg P, Hepola J. Catalytic hot gas cleaning of gasication gas. Catalysis Today 1996;27:5562. Dayton DC, French RJ, Milne TA. Direct observation of alkali vapor release during biomass combustion. 1. Application of molecular beam=mass spectroscopy to switchgrass combustion. Energy and Fuels 1995;9:85565. Smeenk J, Brown RC, Eckels D. Determination of vapor phase alkali content during biomass gasication. Proceedings

128

K.R. Cummer, R.C. Brown / Biomass and Bioenergy 23 (2002) 113 128 removal from biomass producer gas using a xed bed of solid sorbent. Proceedings of the Fourth Biomass Conference of the Americas, Oakland, CA, 1999. p. 939 46. [33] Dayton DC, Jenkins BM, Turn SQ, Bakker RR, Williams RB, Belle-Oudry D, Hill LM. Release of inorganic constituents from leached biomass during thermal conversion. Energy and Fuels 1998;13:86070. [34] Turn SQ, Kinoshita CM, Ishimura DM. Removal of inorganic constituents of biomass feedstocks by mechanical dewatering and leaching. Biomass and Bioenergy 1997;12: 24152.

of the Fourth Biomass Conference of the Americas, Oakland, CA, USA, 1999. p. 9617. [30] General Electric. PFB coal-red combined cycle development program. FE-2357-66, 1980. [31] Turn SQ, Kinoshita CM, Ishimura DM, Zhou J, Hiraki TT, Masutani SM. A review of sorbent materials for xed bed alkali getter systems in biomass gasier combined cycle power generation applications. Journal of the Institute of Energy 1998;71:16377. [32] Turn SQ, Kinoshita CM, Ishimura DM, Hiraki TT, Zhou J, Masutani SM. An experimental investigation of alkali