Fuel 103 (2013) 42–52

Contents lists available at ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel

Optimization of char and tar conversion in fluidized bed biomass gasifiers

Alberto Gómez-Barea a,⇑, Pedro Ollero a, Bo Leckner b

a Chemical and Environmental Engineering Department, University of Seville, Seville 41092, Spainb Department of Energy and Environment, Chalmers University of Technology, Göteborg S-412 96, Sweden

a r t i c l e i n f o

Article history:Received 24 December 2010Received in revised form 31 March 2011Accepted 22 April 2011Available online 21 July 2011

Keywords:GasificationFluidized bedCharTarReview

0016-2361/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.fuel.2011.04.042

⇑ Corresponding author. Tel.: +34 95 4487223; fax:E-mail address: agomezbarea@esi.us.es (A. Gómez

a b s t r a c t

In fluidized-bed gasification (FBG) of biomass and waste the temperature is maintained relatively low toprevent agglomeration. This slows down carbon conversion in conventional FBG, and a gas with relativelyhigh concentration of tar is generated. Then the gasification efficiency is reduced and utilization of the gasis difficult in applications where the gas is cooled or compressed. In the present work the conversion ofchar and tar is studied to identify the main factors hindering complete conversion of the fuel into a prod-uct gas that is free from tar. It is concluded that char conversion can be increased by solids recirculation indirectly heated FBG (stand-alone units) or by burning the char in a separate chamber in indirectly heatedFBG. However, the tar content of the gas remains high, making gas cleaning necessary. Downstreamcleaning of gas by catalytic cracking and/or scrubbing is complex and/or expensive for small to mediumgasification plants, so conversion of tar within the gasifier is preferred. The optimization of conventionaldirectly heated FBG by use of in-bed catalyst and distribution of the gasification agent to various zones ofthe gasifier, although improving the process, is not sufficient to attain the gas purity required for cold gasapplications. Staged gasification is a suitable way to reach high char conversion, while yielding a gas withlow concentration of heavy tar. Most of the staged-gasification developments proposed up to date havebeen based on fixed-beds, thus having relatively small capacity. A recently proposed concept to achievealmost complete tar and char conversion in fluidized bed is presented.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Gasification is a thermochemical route for conversion of solidfuel into a gas, which can be used in a variety of applications [1].Gasification of biomass and waste is similar to that of coal in thesense that biomass can be regarded as a young coal [2]. Essentially,biomass contains more oxygen and volatiles (and sometimes, morewater) than coal, and the nature of the ash differs substantiallyfrom coal ash.

Gasification in a fluidized bed has several advantages over thatin a fixed/moving bed or an entrained-flow gasifier [3]. The fluid-ized bed provides high mixing and reaction rates, accommodatesvariation in fuel quality and allows scaling-up of the process. Var-ious concepts have been developed for gasification in FB. Stand-alone, air-blown, bubbling fluidized-bed gasification (FBG) is thesimplest, directly heated design, but it delivers a gas diluted bynitrogen, having low heating value (4–6 MJ/Nm3) and high tar con-tent (10–40 g/Nm3). Medium heating-value gas (12–15 MJ/Nm3)can be produced using steam as gasification agent. For this purposetwo approaches have been developed: directly heated gasifier, in

ll rights reserved.

+34 95 4461775.-Barea).

which a mixture of oxygen and steam is introduced in one singlereactor [4], and indirectly heated gasifier, consisting of two reac-tors using air in one and steam in the other [5]. In the latter case,heat for devolatilization is generated by burning char in the air-fed reactor and transferring the heat to the second reactor, wherethe fuel is devolatilized in steam. Highly purified oxygen is expen-sive, so gasification based on two reactors seems to be more prom-ising for medium-scale application than oxygen-blown gasification[1,5]. In both types of gasifier the tar concentration in the gas ishigh.

In FBG the bed temperature is limited to prevent agglomerationand sintering of bed material [1–3]. This usually results in incom-plete carbon conversion and production of gas with relatively highconcentration of tar. The first factor reduces the efficiency of theprocess, whereas the latter limits the application of the gas to caseswhere it can be used without cooling, like burning in kilns andboilers. Therefore, applications, such as gas engines, turbines, fuelcells, and synthesis of gas for fuels or chemicals, need extensiveand costly gas cleaning [1]. Effective secondary methods to capturetar are available [6–11]. Removal by washing with water is theleast complicated method, but the waste water is contaminatedby tar and needs expensive treatment before disposal [6,7]. Tarremoval using an organic solvent prior to the condensation ofwater, avoids contamination of the water stream and improves

A. Gómez-Barea et al. / Fuel 103 (2013) 42–52 43

the efficiency of the process by recirculating the tar to the gasifier[8]. Although the process seems to be efficient, it is complex andexpensive for small or medium-size plants [1]. Another possibilityis conversion of tar by catalytic reforming/cracking in a down-stream vessel, which is an effective way to convert tar at the ther-mal level of the gas leaving the gasifier, i.e. 800–900 �C [9–11].However, catalysts have technical shortcomings, such as inactiva-tion by carbon, soot and H2S. Novel catalysts can overcome suchdisadvantages, but they need demonstration prior to industrialimplementation, so they are not yet commercially available[12,13]. In summary, methods to reach high char and tar conver-sion within the gasifier are needed [14], especially for small tomedium scale plants where secondary cleaning has to be kept sim-ple and cheap [1].

In the present overview the fundamentals and methods to im-prove char and tar conversion inside the reactor applicable to con-ventional FBG are discussed. The objective is to assess theeffectiveness of the existing techniques. It is shown that thesemethods are not efficient enough for applications where the gasis cooled or compressed, so innovative designs based on stagedgasification are explored. Although the fundamentals discussedhere are of general applicability, the work is focussed on directlyheated FBG, processing biomass and waste, in small to mediumscale plants for electricity production where primary measuresare preferable. Modeling of the process is out of the scope of thiswork and can be found in [3].

2. Fundamentals of fuel conversion in FBG

2.1. Qualitative description

A biomass particle fed to a gasification reactor undergoes a ser-ies of conversion processes, listed in Table 1 and sketched in Fig. 1.The fluidization gas is, in general, a mixture of steam, oxygen,nitrogen and carbon dioxide (nitrogen is present in air-blown FBGs,whereas carbon dioxide could be part of the input gas in a gasifierintegrated in an oxy-fired plant, intended for carbon-capture). Ini-tially, the fuel particle is dried, devolatilized (R1), (primary pyroly-sis), yielding char and volatiles. Subsequently, volatiles (R7-9) andchar (R2-3) may be oxidized, and finally, char may be gasified bycarbon dioxide and steam (R4 and 5). Fuel particles shrink, and pri-mary fragmentation may occur immediately after the injection ofthe fuel into the bed. Secondary fragmentation and attrition of char

Table 1Main reactions in the biomass gasification process.

Stoichiometry S

Biomass! charþ tarþ H2Oþ light gasðCOþ CO2 þH2 þ CH4 þ C2þ þ N2 þ . . .Þ E

Char combustionCþ 1=2O2 ! CO �Cþ O2 ! CO2 �

Char gasificationCþ CO2 ! 2CO +Cþ H2O! COþH2 +Cþ 2H2 ! CH4 �Homogeneous volatile oxidationCOþ 1=2O2 ! CO2 �H2 þ 1=2O2 ! H2O �CH4 þ 2O2 ! CO2 þ 2H2O �COþ H2O$ CO2 þ H2 �

Tar reactions (tar assumed CnHm)CnHm þ ðn=2ÞO2 ! n COþ ðm=2ÞH2 ECnHm þ n CO2 ! ðm=2ÞH2 þ ð2nÞCOCnHm þ n H2O ! ðm=2þ nÞH2 þ n COCnHm þ ð2n�m=2ÞH2 ! n CH4

CnHm ! ðm=4ÞCH4 þ ðn�m=4ÞC

take place together with char conversion. The energy for heatingthe fuel to reactor temperature and for satisfying the needs ofthe endothermic reactions is provided by combustion of part ofthe fuel in autothermal gasification. The volatiles include non-con-densable gases, such as CO2, H2, . . . , condensable gases (tar), andwater (chemically bound and free water). After primary decompo-sition, a variety of gas–gas and gas–solid reactions take place: sec-ondary conversion, during which the tar may oxidize (R11), reform(R12 and R13), and further react by cracking (R15), dealkylation,deoxygenation, aromatization and formation of soot by polymeri-zation. Primary and secondary tar conversion processes can behomogeneous and heterogeneous, occurring inside as well as out-side of a particle. The tar conversion can be catalyzed by solidsadded to the bed (dolomite, olivine, etc.) or simply by the carbona-ceous surfaces in the devolatilizing particles.

The simplification in splitting up the process into primary (gen-eration) and secondary conversion in Fig. 1 is based on the differ-ent conversion time of the stages described above [3]. The ratesof char gasification with H2O and CO2 are orders of magnitude low-er than those of primary pyrolysis: it takes typically a few minutesto gasify the char at temperatures below 900 �C, while a fuel parti-cle is devolatilized in a few seconds. Conversion of volatiles is morerapid than char. Therefore, when pyrolysis takes place in an atmo-sphere containing steam and O2, O2 is consumed preferentially bycombustion of the volatiles [15]. Due to the fast release of volatiles,the gasification agent does not significantly penetrate into the par-ticle during devolatilization. Therefore, the rate and yield of devol-atilization are quite insensitive to the composition of thesurrounding gas; the bed simply provides the heat for the process[3]. Once the volatiles are emitted from the particle, they mix withthe surrounding gas and react by secondary reactions, mainlyreforming and oxidation, resulting in secondary gas, char and tar,indicated with ‘‘2’’ in Fig.1. The main processes of primary and sec-ondary transformation are listed at the bottom of Fig. 1.

2.2. Influence of fluid-dynamics of FB on char and tar conversion

The conversion of char and tar is related to the effective time forreaction with the gas and catalyst, which in turn depends on theresidence time of the fuel, char and gas, as well as on the local con-ditions of mixing in the reactor [16–18]. Of particular importanceare the contact of char and tar with oxygen and steam, and the po-sition where the fuel particle is devolatilized in the reactor [18].

tandard heat of reaction (kJ/mol) Name Number

ndothermic Biomass devolatilization R1

111 Partial combustion R2394 Complete combustion R3

173 Boudouard reaction R4131 Steam gasification R575 Hydrogen gasification R6

283 Carbon monox. oxidation R7242 Hydrogen oxidation R8283 Methane oxidation R941 Water gas-shift reaction R10

ndothermic (except R11) (200–300) Partial oxidation R11Dry reforming R12Steam reforming R13Hydrogenation R14Thermal cracking R15

Fig. 1. Scheme of reactions of the primary conversion process during devolatilization. The primary products have been indicated in Fig. 1 with ‘‘1’’, while subscript ‘‘2’’ is usedfor the secondary products.

AIR

t4

t1 t3t2

t=t 0FUEL

AIR

FUEL

FUELPARTICLE

D

L

t=t0

(a) High segregation

FUEL t=t0

(b) Low segregation

Fig. 2. Motion of a biomass particle during devolatilization in the dense bed of anFBG. (a) and (b) correspond, respectively, to high and low ratio of the rates ofvertical mixing and devolatilization. Times t1, t2, . . . , t1 represent successiveinstants during the devolatilization process.

44 A. Gómez-Barea et al. / Fuel 103 (2013) 42–52

Formation of bubbles, bypassing of gas, entrainment of materialand other factors influence the reaction time [3,18,19]. Factorssuch as fuel properties (density and size, volatile content) andtopology of the gasifier (bed diameter, aspect ratio, and numberand position of feeders) influence the conversion process. Thekey operation parameter in an FB is the superficial velocity of thegas, which affects mixing and entrainment [16–23].

Higher superficial velocity improves solids mixing, but biomassparticles with lower density and larger size than the bed particlescan be non-uniformly distributed [22,23]. At a given gas velocity,char particles more likely circulate within the bed, while devolatil-izing fuel particles tend to float on the bed’s surface as a result ofthe lift force caused by escaping volatile matter [3]. A key issueis whether the fuel particles remain on the bed’s surface or if theyare induced to descend due to the gross movement of the bed, the‘‘gulf-stream’’. The competition between mixing and segregationhas been studied experimentally in various particle systems andfor a wide range of superficial velocities [22,23]. In these works,measurements were made on cold rigs, keeping the particle den-sity and size constant during a test. In a real gasifier, the changeof particle properties during the release of volatiles and generationof fines add further complexity [3,18,24]. Fig. 2 shows two (limit-ing) cases, illustrating the influence of mixing of a devolatilizingparticle on the tar content in the outlet gas and the way of conver-sion of the char in the reactor. Fig. 2a corresponds to an under-fedFBG, when the vertical mixing of fuel particles is rapid compared tothe time of devolatilization. It also includes the case when the bio-mass is fed at the top and keeps there. The devolatilization takesplace on the top of the bed, where the fuel particles remain float-ing. The gas obtained from a case like this has a pyrolytic nature,i.e. it has a similar composition to that obtained in pyrolysis tests.At the bottom of the bed, the amount of volatile matter is expectedto be small, and hence, the fluidization gas meets the hot char,which is oxidized to some extent, depending on the local temper-ature. The remaining char is converted by gasification with steamand CO2. In Fig. 2b the devolatilization is rapid, and most of vola-tiles are released in the bottom zone of the bed. The presence ofcombustible matter i.e. H2, CO and CH4 at the bottom consumesthe oxygen in a short height [24]. Then, due to the absence of O2

in most reactor zones, the char can only be converted by H2Oand CO2. These gasification reactions are slow at the usual

temperature of a biomass FBG (800–900 �C), and the char conver-sion may be limited unless the residence time is not long enough(larger bed volume) [3,15]. Finally, there is a key feature regardingthe yield and nature of the tar produced: the concentration of tar inthe gas in Case a (Fig. 2) is higher than in Case b; however, the tarcompounds produced in Case a are more reactive than those inCase b; The reason is that the tar compounds are thermally trans-formed and cracked into a more stable form in Case b, due to thelonger time in contact with the bed material. This may have a sig-nificant impact on the method selected to remove the tars as willbe discussed below.

Increasing gas velocity enhances entrainment of material anddecreases the residence time of a particle in the reactor [20,21].A qualitative judgment of the tendency of a particle to be carriedaway can be made by the terminal velocity of a single particle ut

0 1 2 3 4 5Particle diameter, mm

Term

inal

vel

ocity

, m/s

Silicasand

Calcined dolomite

Raw dolomite

Wood chips

Char from wood chips

0

5

10

Fig. 3. Terminal velocities of single particles of various materials in a CFB(calculated for air, at 850 �C and at a pressure of 1.1 bar) as a function of particlesize. Properties of solids: Silica sand (2600 kg/m3, sphericity 0.9); Raw dolomite(2400 kg/m3, sphericity 0.8); Calcined dolomite (1300 kg/m3, sphericity 0.9); Woodchips (850 kg/m3, sphericity 0.4); Char from wood chips (350 kg/m3, sphericity 0.4).

A. Gómez-Barea et al. / Fuel 103 (2013) 42–52 45

[19]. Fig. 3 presents ut versus particle diameter for different mate-rials used in a circulating FBG [3]. At a superficial gas velocity of5 m/s, the diagram indicates that, while silica sand of 0.5 mmwould stay in the bottom bed, dolomite of 0.6–1.0 mm would re-main in the bed initially, but once calcined, it could be rapidly elu-triated. Biomass like wood chips with particle sizes smaller than3 mm would be carried away from the bottom bed and particlesizes up to 10 mm of the char generated from this fuel could betransported. Solids recirculation yields higher carbon conversionbecause it increases the residence time of particles, a generic solu-tion that can be applied for both bubbling and circulating FBG.However, there are circumstances when recirculation is not effi-cient because the reactivity of the char falls due to deactivationby interaction with volatiles [25]. In circulating FBG, large devola-tilizing fuel particles move towards the top of the riser, increasingthe tar yield of the gas as discussed above with regard Fig.2. In abubbling bed, in contrast, the fuel particles are likely to remainin the bed (or on its surface) most of the devolatilization timedue to lower superficial velocity (lower entrainment) [17,18].Entrainment of biomass particles during devolatilization does notplay an important role in a bubbling unit, but entrainment of finerchar particles may be severe [21].

The size of the gasifier may influence the conversion process. Inlarge commercial devices, characterized by bed aspect ratios ofunity or lower (wide beds), the mixing of the volatiles generatedaround the feed port may not be fast enough and changes ofconcentration of gas and tar in the horizontal direction are found,affecting the tar and volatile conversion through the bed [3,16,18].Ascending plumes with highly concentrated pyrolysis gas are likelyto be formed, yielding gas with high tar concentration [18]. Theseaspects may significantly change the expected performance of

0 200 400 600 800 1000Temperature (ºC)

Cha

r yie

ld

Gas

yie

ld

0.0

0.2

0.4

0.6

0.8

1.0

0 200 400Temper

0.0

0.2

0.4

0.6

0.8

1.0

Fig. 4. Yields of char, light gas and total pyrolytic liquids (tar + water) as a function ofempirical model Adapted from Neves et al. [39].

large units. For instance, if a catalyst has been found efficient inthe lab, it may be insufficient for tar conversion in a commercialunit. The reason is that in the latter unit the bypassing of tar inthe plumes of volatiles and the mass transport resistance betweenthe bubble and emulsion may slow down the rate of transport oftar compounds over the catalyst surfaces.

2.3. Fundamentals of tar and char conversion in an FBG

The rate of devolatilization and the yields of primary char, tar,and light gases vary mainly with the temperature and the heatingrate at which the fuel is transformed, which in turn are determinedby the fuel particle size and operating temperature of the reactor[26]. The rate of devolatilization of mm-size particles in an FBGis usually controlled by internal heating [3,16,26]. Then, the reac-tion zone within the particle is narrow compared to the particlesize, and the fuel material is devolatilized at a temperature intervalof say 400–600 �C, significantly lower than the bed temperature(800–900 �C) [3]. The fuel emits a variety of organic compounds,from aliphatic chains to parent fuel structures, such as levogluco-san or glucose. Vapor-phase secondary pyrolysis in the internalpores of the fuel particle involves complex parallel and consecutivereactions, which are included within the primary pyrolysis [26].The compounds released are thermally unstable at temperaturesabove 600 �C due to the multiple aliphatic bonds and heteroatomsin its structure. Most tars are emitted before the particle is heatedabove 600 �C [3], but outside of the fuel particle they meet the bedtemperature and their chemical nature changes by breaking of thealiphatic structures into lower hydrocarbon molecules and lightgas, as well as by rearrangement of lower hydrocarbon chains tomore stable aromatic structures. In general, the secondary reac-tions decrease the tar yield but enhance generation of refractory,non-substituted, stable, polyaromatic hydrocarbons (PAHs) [27–30]. At temperatures below 900 �C, PAHs are not significantly con-verted into gas via non-catalytic steam reforming in the time rangeavailable for reaction in a FBG (seconds) [15].

Prediction of the pyrolysis has been made in various ways, rep-resenting the process by a single reaction (global model), by com-bination of series and parallel reactions, and by assuming adistribution of activation energy [3,19,26]. The models predictthe yields of tar, gas, and char released during pyrolysis, but theydo not predict the yield of the individual gas and tar species inthe gas or the properties of the char generated. Instead, empiricalcorrelations developed for coal [31] and experimental data for avariety of biomasses using different laboratory equipment havebeen obtained under different process conditions [32–39]. Datafor FBG should be obtained at high temperature and heating ratewith particle sizes to be used in an industrial application [3].

General trends of devolatilization can be determined bycomparison of published data, as it is shown in Fig. 4, where theyields of char, light gas, and liquids (tar + water) as a function of

Liqu

ids

yiel

d

600 800 1000ature (ºC)

0 200 400 600 800 1000Temperature (ºC)

0.0

0.2

0.4

0.6

0.8

1.0

the peak pyrolysis temperature. d -‘‘fast heating rate’’; s -‘‘slow heating rate’’; �

46 A. Gómez-Barea et al. / Fuel 103 (2013) 42–52

temperature at various heating rates are presented for a variety ofbiomass materials [39]. Despite the scatter in the data, due to thelarge amount of biomass fuels and types of research rigs includedin the analysis, the main trends can be observed. At the lowesttemperatures (<300 �C) char is the main product. At middle-rangetemperatures (450–550 �C) a maximum is observed for liquids.Qualitatively, these general trends in product yield as a functionof temperature are the same for slow and fast heating rates. Onheating up to around 450–550 �C, slow heating rates give morechar and less tar than fast heating rates due to intraparticle char-ring of the primary tars and the low activity of secondary reactionsof volatiles. The observed decrease in the yield of char as temper-ature increases indicates that the major mass loss of fuel occurs inthe range of 200–600 �C, i.e. most gas is released from the solidfuel. At these low temperatures, the heating rate has a small influ-ence. As temperature increases above 450–550 �C the variation inthe yield of char is small, with low heating rates generally associ-ated with higher yield of char. Temperature and heating rate (par-ticle size) also determine the physical properties of the primarychar, which in turn influence the chemical reactivity of the charand its fragmentation properties. The influence of the heating rateon the tar yield becomes less important at the highest tempera-tures analyzed (>800 �C). The low temperature range where thethermal decomposition of biomass takes place can be referred toas the primary pyrolysis stage, which is the likely process to applyto mm-sized particles in FBG.

Primary fragmentation has not usually taken place in the testsincluded in Fig. 4, but it can occur in real FBG, and this aspect shouldbe considered because of the change in the particle size to bedevolatilized (primary fragmentation can be considered almostinstantaneous) [40–43]. The probability for large fuel particles tobreak up during devolatilization is high, whereas small particlesdo not undergo primary fragmentation [41]. Recent advancedmodels attempt to quantify the variables producing fragmentation,but the process is very complex and the fragmentation patterns arestill determined by measurements [3,40–43].

To estimate the heterogeneous conversion of char after devola-tilization (secondary conversion) one needs to estimate the reac-tion time of the char particles in the vessel. This is governed bythe chemical reactivity, secondary fragmentation and elutriationof char fines by attrition. In a directly heated FBG the char is notmuch affected by oxygen, since the oxygen is preferentially con-sumed by the volatiles, and the contribution of combustion tothe overall char conversion is small [24,44]. Then the char ismainly converted by CO2 and H2O whose reaction rates are slow[45]. The rate of reaction of char with different gases can be repre-sented by global nth order kinetics, valid under certain operationconditions [3,45]. More complex Langmuir–Hinshelwood kineticsare necessary for analyzing a broader range of conditions [45]. Dur-ing char conversion by gasification in FBG, the reaction zone ex-tends over most of the particle volume, and the interior surfacechanges significantly [46]. Heat transfer plays a secondary role inchar conversion, since thermal gradients at the particle scale aresmall. The overall or effective reactivity in the bed depends onthe distribution of char particles in the bed [47], which, besidesthe char properties, is mainly governed by solids mixing, and so,by the gas velocity.

2.4. Concluding remarks

From the fundamentals discussed and from further literatureresults, the following guidelines for optimization of char and tarconversion in FBG can be outlined:

� A directly heated FBG for high char conversion is operated at themaximum temperature permissible to avoid agglomeration and

with a long residence time of the char particles. The latter isachieved by lowering the gas velocity to prevent high entrain-ment of fuel and char fines in bubbling units; however, thismeasure slows down solids mixing, and increases the segrega-tion of fuel, which results in higher tar yield. Recirculation ofmaterial may be a solution for improving char conversion,although there are circumstances when recirculation is noteffective because volatiles interact with the char and decreaseits reactivity. Conversion of high-ash fuels makes it necessaryto discharge bed material, and this reduces the carbon conver-sion further. In practise, it is difficult to achieve higher conver-sion than 97% in conventional directly heated FBG. In indirectlyheated gasification using dual FB, the char is burnt separately inthe combustion unit, and char conversion up to 99% can beachieved, but the tar concentration remains high.� The conversion of tar by oxygen is limited in directly heated

FBG, because tar competes with the light gases and char forthe oxygen (the light gases are more reactive than tar). In indi-rectly heated FBG, the devolatilization of fuel is conducted in asteam atmosphere, and the tar yield is high because steamreforming of tar is not efficient at the usual process temperature(<900 �C). When there is no catalyst in the bed, thermal crack-ing only partially converts the tar in the reactor. External solidcatalysts may significantly enhance cracking and reformingreactions, reducing the tar concentration. The optimization ofthe contact of steam and oxygen with freshly generated tarand with an external catalyst is necessary if tar reduction isessential. Feeding the fuel at the bottom increases the timefor tar conversion if the fuel remains in the bed during devola-tilization, reducing the tar concentration; however this alsoleads to more stable and heavy (aromatic) tar compounds inthe gas, making secondary conversion difficult.

3. Improving tar and char conversion in conventional FBG

Primary methods are preferred over secondary methods [14,48].An ideal primary method should eliminate the need for further gastreatment, and the char should be completely converted in the gas-ification reactor. For a given type and flow rate of biomass (thermalinput), the optimization can be made by adjusting the flow rate ofoxygen, steam (and catalyst or additive, if any) to give sufficientlyhigh temperature, residence time of the char and tar (gas mixingand effective time in contact with catalyst and char), under the con-dition of a safe operation (without sintering) [1,3].

In the following, various ways to optimize the gasifier byadjusting the manipulated variables in a given gasifier (size andtopology) are discussed qualitatively, and to a certain extent, quan-titatively, based on literature and own data. The methods dis-cussed include: composition and flow rate of the gasificationagent, staging of the gasification agent and addition of in-bed cat-alyst. Temperature and residence times of char and tar are the ma-jor parameters affecting the performance of the gasifier, but theyare indirect variables, which are not known a priori, i.e. they can-not be directly controlled by the above manipulating variables. Inpractice, temperature is measured during operation and can be ad-justed on-line by trial and error. Comprehensive modeling, how-ever, allows understanding of the process, and therefore it shouldbe used to support optimization of design and operation [3]. Theeffects of temperature and residence time of char and tar on con-version are therefore discussed first by theoretical simulations.

3.1. Effects of temperature and residence time of char and tar in thegasifier

Fig. 5 summarizes the effect of temperature in a gasifier on thekey output variables: char conversion, tar concentration, and

Higher Gas heating value Lower

Higher Tar content Lower

Lower Char conversion Higher

Decreasing risk Sintering Increasing risk

700 °C 800 °C 900 °C 1000 °C

Agrofuels

Refusedderived

fuel

Woodybiomass Coal

Fig. 5. Effect of temperature on parameters and processes during gasification. Adapted from Devi et al. [14] who quoted Hallgren [49].

A. Gómez-Barea et al. / Fuel 103 (2013) 42–52 47

heating value of the product gas, as well as the propensity for sin-tering. The temperature range of a biomass gasifier is between 800and 900 �C, considering the balance between benefits and draw-backs associated with the thermal level. It is clear that raisingthe temperature increases tar and char conversion. However, thedanger for sintering of ash and bed material also increases and setsthe maximum temperature. Even at the highest temperatureacceptable, it is difficult to convert PAH into gas via steam reform-ing by non-catalytic reactions or by contact with bed material. Asdiscussed previously, fluid-dynamic effects may limit the effectivecontact of tar and catalyst.

Fig. 6 shows the effect of residence time on the char conversionattained in an FBG at various temperatures. The residence time ofthe char is calculated as the ratio of the mass of char in the gasifierto the fixed carbon added with the input fuel-stream. The resultcorresponds to an atmospheric air-blown bubbling FBG with aninternal diameter of 1 m [50], processing wood pellets, as obtainedby a fluid-dynamic and fuel-conversion model [3]. This modeltakes into account most of the processes discussed in this work:entrainment, fragmentation, elutriation, chemical conversion ofchar particles, considering their internal surface, density and size.It is seen that at 800 �C the residence time of the char has to belonger than 0.5 h to reach 80% conversion of char. In contrast, at900 �C, the same conversion is achieved 10 times faster. It mustbe underlined that all the points in the graph do not correspondto the same thermal input of the gasifier. For instance, to changetemperature of the given gasifier from 800 to 900 �C, the air to

Fig. 6. Effect of temperature and char residence time on char conversion in an FBG.Calculated by modeling with the model descript in [3].

biomass ratio has to be increased. This can be done by increasingthe air flow-rate or reducing the biomass flow-rate. In the formercase the superficial gas velocity and entrainment are increased,reducing the residence time of the char. If the biomass flow rateis reduced, for instance, to such extent that the superficial gasvelocity is kept constant and equal to that at 800 �C, the char res-idence-time is increased and then, the conversion of char rises sig-nificantly, since it is converted at much higher rate at the highertemperature. In summary, it is clear how the char conversion is af-fected by different operating variables.

A similar figure could be made showing the conversion of tar inthe bed as a function of residence time and temperature. However,there is no model taking into account key phenomena, such as by-passing of tars with the plumes of volatiles and interaction be-tween one tar and another during conversion. A simplified picturecan be displayed considering tar as a single or a few lumps, whichreact in a single global reaction and without considering the com-plex fluid-dynamic processes in the bed (see Fig. 24 in [3]).Although the figure would display the expected trend (tar conver-sion increases with residence time and temperature), the resultsare not realistic, and such a prediction is not useful for optimiza-tion. Developments in this area are needed.

3.2. Effect of composition of gasification agent

Fig. 7 shows the distribution of the fuel energy into sensible andchemical energy of the gas during conversion of a solid fuel as afunction of the oxygen added (quantified by the stoichiometric ra-tio, SR). The temperature of the system (not shown) increases withSR, and so does the sensible heat of the gas up to a certain level be-cause more volatiles are emitted and more char is converted withraising temperature. This occurs from a SR of 0 up to 0.3–0.4,meanwhile the fuel is increasingly converted: the more the solidsare converted, the more chemical energy from the char is trans-ferred into gas. However, when the oxygen supply exceeds thepoint where the char is already completely converted into gas,more fuel is burned to CO2 and H2O and the heat release increasesat the cost of product gas, lowering the chemical energy in the gas.An optimum (from a thermal efficient point of view) is foundwhere the chemical energy in the gas is maximized with significanttar reduction. The optimum from the tar point of view, however, isnot clear, since the reduction of tar is not the only issue to be con-sidered. In fact the optimum is found once the gas application isestablished: for instance, for power generation in an alternativeengine the dew point of the gas is a more correct index of gas qual-ity [8,48]. The calculations in Fig.7 have been made at equilibrium.

Fig. 7. Energy conversion from solid fuel to gas, illustrating the distribution of theenergy in the gas between sensible and chemical energy as a function ofstoichiometric ratio, SR (SR = kg of O2 fed/kg of stoichiometric O2). (At any SR,Total = Sensible + Chemical).

48 A. Gómez-Barea et al. / Fuel 103 (2013) 42–52

In practise, equilibrium is not attained because the conversion ofchar by CO2 and H2O is slow at the usual process temperature,and the char is removed from the system before reaching equilib-rium. In other words, the conversion of char controls the processfor a typical biomass and residence time (see Fig. 6 showing a casefor wood pellets) and the optimum (complete char conversion) isobtained by setting an actual SR Higher than that calculated byequilibrium, where the temperature is higher, and hence also theutilization of the fuel (higher char conversion). In summary, anequilibrium simulation such as presented in Fig. 7 is useful tounderstand the process qualitatively; however, complex modelssuch as that used to calculate the results shown on Fig. 6 are nec-essary for optimization.

Addition of steam to a directly heated gasifier enhances tarreforming and char gasification, improving the quality of the gasand reducing its tar content [51,52]. However, steam addition re-duces the temperature of the gasifier and more oxygen has to beadded to maintain the temperature level, lowering the heating va-lue of the fuel gas produced [51]. There is an optimal steam to oxy-gen ratio where steam addition positively compensates for theburnt fuel gas, further converting char to CO and H2 [51,52]. Steamaddition at high temperature is an effective measure [53], but thetemperature of the input steam is limited to that achievable byheat integration (for instance by heat exchange with the producedgas) if the gasification process is conducted autothermically (noexternal heat is added). Low-cost oxygen, such as produced bymembranes with a purity of 40–50%, together with steam, pre-heated by the hot produced gas, has been shown to improve theprocess significantly, achieving a carbon conversion of up to 97%[52]. However, the tar content in the gas is not reduced to the limitrequired, and the dew point of the gas is above 100 �C [52].

3.3. Effect of distribution of the gasification agent

Staging the gasification agent makes it possible to create vari-ous thermal levels in the gasifier. The principle is illustrated inFig. 8a. Staging by injection of secondary air has been tested in con-ventional FBG at pilot scale [54–57]. A portion of the inlet oxygen(air) is conducted to a port situated in the upper part of the bed orthe freeboard. Significant reduction of tar has been reported (below1 g/Nm3) [54,57]. These data, however, should be interpreted withcaution: in [54] the air ratio is increased simultaneously with there-distribution of the air, i.e. more gas is burnt, lowering the heat-ing value of the gas; in [57] a great reduction of tar is achieved, but

the operation is conducted in an external oven, so it is not possibleto make any conclusion about the tar reduction under autothermaloperation. Injection of secondary air at constant air ratio, i.e. toanalyze the effect of injection in an isolated way, has been shownto be effective for reduction of phenols and other light tar com-pounds [56], but the total tar concentration is still high (a fewgrams per Nm3) and the proportion of stable aromatic tar com-pounds in the gas increases significantly. Staging of the gasificationagent, oxygen and steam, despite being potentially interesting, hasnot been reported. It seems that a more drastic division of zones inthe gasifier is necessary for further tar reduction; however thismakes it necessary to leave the conventional design as we willsee in Section 4.

3.4. Effect of in-bed catalyst

The presence of an in-bed catalyst and the local rate of gas andsolids mixing influence greatly the cracking and/or reforming reac-tions [14,58]. Typical in-bed catalysts are calcined limestone anddolomite [56,59,60], olivine [59,61], and, less frequently, Ni-basedor other metallic catalysts. Reduction of tar by cheap catalysts,such as natural mineral (dolomite, olivine, etc.) reaches tar concen-trations down to 0.5 g/Nm3. However, the remaining small amountof heavy tar leads to a gas with a dew point of around 100 �C, andthis will cause problems in applications where the gas has to becooled to the ambient temperature, so this method has to be com-plemented by secondary treatment [56]. More effective catalysts,such as those based on nickel, improve tar reduction in the bed,but the rapid degradation of the catalyst makes this option unfea-sible. Ni-based catalysts are preferred in downstream vessels, butdeactivation is a problem still to be solved [9–13].

Char generated in the process acts as a catalyst enhanced by thealkali and alkaline earth-metals remaining in its structure, espe-cially potassium in the form of carbonates, hydroxides or oxides,which has an effect on steam reforming of the nascent tar[62,63]. The main mechanisms of tar conversion on char surfacesare still not well understood [64]. The char structure undergoessignificant transformations during the conversion process, and itis simultaneously gasified by steam in the fluidization gas. Poly-merization with coke formation seems to be the main decomposi-tion mechanism of PAH tar compounds at temperatures above700 �C [64], lowering the rate of char gasification with steam. Attemperatures below 600 �C, in contrast, the char acts as a tar filterwith little or no gas emission, but reducing significantly the tarconcentration. The effect of the light gas compounds of the pro-ducer gas on tar conversion is not well understood. Hydrogenseems to have a negative effect on the conversion rate of tar[64], similar to the processes leading to soot formation.

Much work has been done towards the determination of thekinetics of tar generation and conversion, but the process is notyet fully understood. Therefore, existing kinetics data are globaland none of them represents elementary steps. Therefore, specificmeasurements are still the only data source available to study tarconversion [3,26]. To give a rough idea on the effectiveness of dis-tinct catalysts, experimental findings [65] have been illustrated inFig. 9. The measurements comprised two model compounds: phe-nol and naphthalene. The figure shows a ranking of the activities ofthe catalysts (and also thermal cracking) for conversion of naph-thalene at 900 �C (a) and for phenol at 700 �C (b). Naphthalene isthermally stable at 900 �C (only 2% was converted over silica sand),whereas it is more converted with dolomite and olivine, and al-most fully converted in the presence of nickel-based catalyst andchar. It is seen that the results are quite insensitive to the tar con-centration in the inlet gas (the inlet concentration is that typicallyfound in FBG). Phenol conversion is represented in the figure at700 �C because at 900 �C all catalysts gave 100% phenol conversion,

Devolatilizationzone

(Firststage)

Gasification zone(Second stage)

Pyrolysis unit

Gas

ifica

tion

unit

Char bed

Pyrolysis gas,tar

Pyrolysis gas,tar and char

BiomassssamoiB

Gas with lowtar content

Gas with lowtar content

Primaryoxygen

Secondaryoxygen

Steam+Oxygen

Heat

Fig. 8. Staging of the gasification process: (a) secondary air injection. (b) Two-stage gasifier.

A. Gómez-Barea et al. / Fuel 103 (2013) 42–52 49

and more than 98% of the phenol was already thermally cracked.Although the figures are obtained in specific conditions [65], theresults lead to important conclusions: the conversion of phenol isnot a problem in FBG at temperatures above 800–850 �C, whereasthe reduction of naphthalene, or a mixture of tar compounds withsimilar reactivity, down to 0.5–1 mg, is difficult in an FBG. Chareffectively converts the heavy tar compounds, so optimization ofthe contact of tar with in situ generated char might be targeted.However, this it is difficult to achieve in a single vessel due tofluid-dynamic effects, as shown in the previous sections. Stagedgasification could make this concept possible, as we will see below.

3.5. Concluding remarks

The combined use of in-bed catalyst, injection of secondary airand optimization of the composition of the gasification agent inconventional FBG designs, despite improving the process, wereshown to be insufficient to attain the gas purity required for coldgas applications [56]. It is concluded that further measures involv-ing redesign of the gasifier/process are necessary, i.e. it is desirableto develop innovative designs. One rational option to do this isdealt with in the following section.

4. Staged gasification

4.1. Existing developments

It has been shown that optimization of conventional directlyheated FBG is not effective enough to sufficiently reduce tar (dewpoint) to the required level. Indirectly heated gasifiers consistingof dual FB convert most of the char, but the tar problem remains,so it is also not a sufficiently good solution as long as expensivesecondary cleaning of the gas is needed. Therefore, staged gasifica-tion in directly heated gasifiers has been proposed, searching forhigh conversion of both tar and char inside the reactor [66–75].

Staged gasification creates zones in the gasifier, allowing opti-mization of the process by increasing simultaneously the conver-sion of char and tar. The essential idea is similar to thatconceptualized in Fig. 8a but here more drastic zone division thanin the case of secondary air injection is achieved by dedicated de-signs. Various thermal levels are created in the gasifier by stagingthe oxidant: fuel is devolatilized at relatively low temperature by

decreasing the oxygen supply to this zone; the temperature is in-creased in a second zone by addition of the remaining oxygen to-gether with steam. This two-stage procedure favors theconversion of tar because it creates a gas with high-reactive tarcompounds at high temperature in the presence of steam (non-cat-alytic gas-phase reforming). If the char generated is used to form athird stage, the char can be gasified with steam and further tar canbe converted. An example of this three-stage concept is shown inFig. 8b. In this figure the devolatilization of the fuel is conductedwithout addition of oxygen [70]. The heat for pyrolysis comes fromthat generated in the second stage, conducted by successive heatexchanges (not shown in the figure).

A few innovative processes have been proposed based on stagedgasification. Examples are processes like CASST, developed at En-ergy research Centre of the Netherlands [71], the ‘‘Viking’’ and‘‘Low-Tar BIG’’ developed at Danish Technical University for fixedbed [70] and fluidized bed [72], respectively, STAR-MEET at TokyoInstitute of Technology [66], CleanStgGas at ITE Graz University ofTechnology [73], and other [74,75]. In most gasifiers of this typethe char is converted by gasification (with steam or CO2), so theefficiency of the process depends on how the conversion is ’’orga-nized’’. Since char gasification reactions are slow, it is necessary toprovide long residence time to achieve significant char conversion.Therefore, most such processes are based on fixed bed designswhere long residence time easier can be achieved. A process com-bining fluidized and moving beds [68,75] has been suggested re-cently, oriented to the conversion of difficult waste with highfuel utilization, but the tar content in the gas is still high.

4.2. FLexible Three-stage GASification (FLETGAS) of biomass and waste

All mentioned staged gasification designs where high conver-sion of tar and char has been achieved are fixed or moving beds.In order to carry out staged gasification in FB, enabling highthroughputs, a new staged-gasification concept is under develop-ment by the Bioenergy Group at the University of Seville [76].The system is focused on processing of difficult wastes, whoseash content is high. The nature of the ash limits the temperatureof the gasifier because of the risk for agglomeration.

The gasification concept consists of three stages: FB devolatil-ization (first stage), non-catalytic air/steam reforming of the gascoming from the devolatilizer (second stage), and chemical filter-ing of gas in a moving bed supplied with the char generated in

(a) (b)

2

55 61

100

60.373.7

94.4 99.6

Nap

htha

lene

co

nver

sion

(%)

6

34.542.7

81.6 87.1 90 91

0

20

40

60

80

100

Phen

ol

conv

ersi

on (%

)

020406080

100

Fig. 9. (a) Effect of catalytic bed material on naphthalene conversion. T = 900 �C, s = 0.3 s, feed gas composition: 6% v/v CO2, 10% v/v H2O, and balance N2. The bars indicate theinitial naphthalene concentration: 40 g/Nm3 (white bars) and 90 g/Nm3 (black bars) (b) Effect of catalyst bed material on phenol conversion. T = 700 �C, s = 0.3 s, feed gascomposition: 6% v/v CO2, 10% v/v H2O, and balance N2, initial phenol concentration 8–12 g/Nm3 (Adapted from El Rub et al. [65]).

DEVOLATILIZER

LOOP-SEAL

VALVE

Gas Flow Solids Flow

CHAR CONVERTER

Fuel

Ash

Fuel Gas

NON-CATALYTICREFORMER

Steam + Enriched Air

Steam + Enriched Air

Steam + Enriched Air

Fig. 10. Conceptual performance diagram of FLETGAS process.

50 A. Gómez-Barea et al. / Fuel 103 (2013) 42–52

the devolatilizer (third stage). The operational principle of the sys-tem is outlined in Fig.10, where the directions of the flows of solids(biomass, bed material and char) and gas (inlet flow rates of fluid-ization agent) are indicated. A control valve is included to adjustthe residence time of the solids through control of the solids inven-tory in the system.

Air and steam can be injected at various points (in the devola-tilizer, steam reformer and seal) with different proportions of thetwo reactants. Enriched air, with an oxygen concentration of upto 40% (to keep the price reasonably low, for instance, producedby membranes) can be used instead of air. The fuel is fed nearthe bed’s surface and has to circulate down to the bottom beforeleaving the bed. The devolatilizer, where a high yield of fresh taris generated, is operated at relatively low temperature (700–750 �C). The fresh tar compounds are drastically reduced in the re-former, where a temperature of up to 1200 �C is created. The injec-tion of steam into the reformer avoids coking and polymerizationof tar. The gas is filtered in a moving bed made of char coming fromthe loop seal. The loop seal can be operated as an oxidiser (fed withenriched air) or as a light reformer (fed with H2O) depending onthe fuel’s reactivity and ash properties. The char filter also coolsdown the gas (chemical quench) by the endothermic char gasifica-tion reaction with steam, while it acts as a catalytic filter promot-ing tar decomposition reactions with steam.

It is challenging to achieve the appropriate movement of solidsand gas, so a cold model (scaled-down from an imaginary 2MWe

gasification plant, operating with dried sewage sludge) was built[76]. A fluid-dynamic model was developed to understand themovement of gas and solids, predicting the distribution of massof solids between the bed and loop seal [76]. Fuel conversion tests(devolatilization, char gasification, and tar conversion) in a bench-scale FBG have also been conducted [77]. The data from the coldand fluid-dynamic models and from the fuel and tar conversiontests have been a support for the design of a plant that is currentlyunder construction to demonstrate this new gasification concept atpilot scale (30 kg/h).

5. Summary and conclusions

Measures to optimize fuel conversion in an FBG to simulta-neously achieve high tar and char conversion have been reviewed.Factors limiting the complete conversion of tar and char were iden-tified. Char conversion can be enhanced by indirectly heated gasi-fication in twin-bed units or, to some extent, by solids recirculationin conventional directly heated FBG. However, the tar content re-mains high, so the gas has to be further cleaned from tar by sec-ondary measures. Treatment of tarry gas downstream of thegasifier is complex and/or expensive. Therefore, small to mediumsize gasification plants need methods to improve both tar and charconversion within the gasifier. The optimization of conventionalFBG by in-bed catalysts and distribution of the gasification agent,although improving the process, is shown to be insufficient to at-tain the gas purity required for cold gas application. Staged gasifi-cation creates zones in the gasifier, which promote high conversionof char and tar, and, therefore, this is an effective and feasible wayfor reduction of tar during gasification. Various developments ofthis concept have been proposed for small-scale gasification duringlast years. These developments seem to have solved the technicalchallenges, but they have still not reached the commercial status.In addition, all developments involving staged gasification arefixed/moving bed designs, thus having relatively small capacityand limited fuel flexibility. A new fluid-bed three-stage gasifier(FLETGAS) has been briefly described here. The new process ispromising because it enables to increase the flexibility and capac-ity of existing staged gasification developments. The FLETGAS pro-cess is still under development at pilot scale.

Acknowledgments

This work was financed by the Junta de Andalucía under theProject FLETGAS. The first author acknowledges the invitation of

A. Gómez-Barea et al. / Fuel 103 (2013) 42–52 51

Prof. Jun-Ichiro Hayashi to ISGA in Fukuoka, where the material ofthis Paper was presented and discussed.

References

[1] Gómez-Barea A, Leckner B. Gasification of biomass and waste. In: Lackner M,Winter F, Agarwal AK, editors. Handbook of combustion, vol.4. Weinheim: Wiley-Vch; 2009. p. 365–97.

[2] Higman C, Van der Burgt M. Gasification. 2nd edition. Amsterdam: GulfProfessional Publishers/Elsevier Science; 2008.

[3] Gómez-Barea A, Leckner B. Modeling of biomass gasification in fluidized bed.Progr Energy Combust Sci 2010;36(4):444–509.

[4] Rauch R, Hofbauer H, Bosch K, Siefer I, Aichernig C, Tremmel H et al. Steamgasification of biomass at CHP plant in Güssing – status of the demonstrationplant. In: Proc 2nd world conf on biomass for energy, industry and climateprotection, vol. II. Rome; 2004. p. 1687–90.

[5] Paisley MA, Overend RP. The Sylvagas process from future energy resources – acommercialization success. In: Palz W, Spitzer J, Maniatis K, Kwant K, Helm P,Grassi A, editors. Proc 12th European conference on biomass for energy,industry and climate protection. Amsterdam; 2002. p. 975�8.

[6] Stevens DJ. Hot gas conditioning: recent progress with larger-scale biomassgasification systems. NREL, Golden (CO), USA, Report No. NREL/SR-510-29952;2001.

[7] Hasler P, Nussbaumer T. Gas cleaning for IC engine applications from fixed bedbiomass gasification. Biomass Bioenergy 1999;16:385–95.

[8] Boerrigter H. ‘‘OLGA’’ tar removal technology. The Energy research Centre ofthe Netherlands, ECN-C-05-009; 2005.

[9] Sutton D, Kelleher B, Ross JRH. Review of literature on catalysts for biomassgasification. Fuel Process Technol 2001;73:155–73.

[10] Simell P. Catalytic hot gas cleaning of gasification gas. VTT Publication No. 330;1997.

[11] Dayton D. A review of the literature on catalytic biomass tar destruction.National Renewable Energy Laboratory, NREL/TP-510-32815; 2002.

[12] Salo K. Biomass gasification plant in Skive. Presented at IV internationalseminar of biomass gasification 2010. Göteborg. Sweden.

[13] Hannula I, Lappi K, Simell P, Kurkela E, Luoma P, Haavisto I. High efficiencybiomass to power operation experiences and economical aspects of the novelgasification process. In: 15th European Biomass Conference & Exhibition, 7–11May 2007, Berlin, Germany 2252–5.

[14] Devi L, Ptasinski KJ, Janssen FJJG. A review of the primary measures for tarelimination in biomass gasification processes. Biomass Bioenergy2002;24:125–40.

[15] Hayashi JI, Hosokai S, Sonoyama N. Gasification of low-rank solid fuels withthermochemical energy recuperation for hydrogen production and powergeneration. Process Safe Environ Protect 2006;84(B6):409–19.

[16] Leckner B. Fluidized bed combustion: mixing and pollutant limitation. ProgEnergy Combust Sci 1998;24:31–61.

[17] Kersten SRA, Moonen RHW, Prins W, van Swaaij WPM. Gas mixing in pilotscale (500 kWth) air blown circulating fluidised bed biomass gasifier. In:Bridgwater AV, editor. Progress in thermochemical biomassconversion. Cambridge MA: Blackwell Science; 2001. p. 452–64.

[18] Petersen I, Werther J. Three-dimensional modeling of a circulating fluidizedbed gasifier for sewage sludge. Chem Eng Sci 2005;60:4469–84.

[19] Souza-Santos ML. Solid fuels combustion and gasification. New York: MarcelDekker; 2004.

[20] Geldart D. Particle entrainment and carryover. In: Geldart D, editor. Gasfluidization technology. New York: Wiley; 1986. p. 123.

[21] Miccio F, Moersch O, Spliethoff H, Hein KRG. Generation and conversion ofcarbonaceous fine particles during bubbling fluidized bed gasification of abiomass fuel. Fuel 1999;78:1473–81.

[22] Wirsum M, Fett F, Iwanowa N, Lukjanow G. Particle mixing in bubblingfluidized beds of binary particle systems. Powder Technol 2001;120:63–9.

[23] Zhang Y, Jin B, Zhong W. Experimental investigation on mixing andsegregation behavior of biomass particle in fluidized bed. Chem Eng Process2009;48:745–54.

[24] Yan HM, Heidenreich C, Zhang DK. Modelling of fluidised–bed coal gasifiers:elimination of the combustion product distribution coefficient by consideringhomogeneous combustion. Chem Eng Process 1999;39:229–37.

[25] Kajita M, Kimura T, Norinaga K, Li CZ, Hayashi JI. Inhibition of steamgasification of char by volatiles in a fluidized bed under continuous feedingof a brown coal. Fuel 2006;85:340–9.

[26] Di Blasi C. Modeling chemical and physical processes of wood and biomasspyrolysis. Prog Energy Combust Sci 2008;34:47–90.

[27] Boroson ML. Secondary reactions of tars from pyrolysis of sweet gumhardwood. PhD. thesis; 1987.

[28] Kinoshita CM, Wang Y, Zhou J. Tar formation under different biomassgasification conditions. J Anal Appl Pyrol 1994;29:169–81.

[29] Brage C, Yu Q, Chen G, Sjöström K. Tar evaluation profiles obtained fromgasification of biomass and coal. Biomass Bioenergy 2000;18:87–91.

[30] Morf P. Secondary reactions of tar during thermochemical biomass conversion.PhD Thesis. Zurich: Swiss Federal Institute of Technology; 2002.

[31] Loison R, Chauvin R. Pyrolyse rapide du charbon. Chem Ind 1964;91:269–75.[32] Hajaligol MR, Howard JB, Longwell JP, Peters WA. Product compositions and

kinetics for rapid pyrolysis of cellulose. Ind Eng Chem Process Des Dev1982;21:457–65.

[33] Nunn TR, Howard JB, Longwell JP, Peters WA. Product compositions andkinetics in the rapid pyrolysis of sweet gum hardwood. Ind Eng Chem ProcDesign Develop 1985;24:836–44.

[34] Boroson ML, Howard JB, Longwell JP, Peters WA. Product yields and kineticsfrom the vapor phase cracking of wood pyrolysis tars. AIChE J 1989;35:120–8.

[35] Jand N, Foscolo PU. Decomposition of wood particles in fluidized beds. Ind EngChem Res 2005;44:5079–89.

[36] Thunman H, Niklasson F, Johnsson F, Leckner B. Composition of volatile gasesand thermochemical properties of wood for modeling of fixed or fluidizedbeds. Energy Fuels 2001;15:1488–97.

[37] Kersten SRA, Wang X, Prins W, Van Swaaij WPM. Biomass pyrolysis in afluidized bed reactor Part1: Literature review and model simulations. Ind EngChem Res 2005;44:8773–85.

[38] Font R, Marcilla A, Devesa J, Verdu E. Kinetic study of the flash pyrolysis ofalmond shells in a fluidized bed reactor at high temperatures. J Anal Appl Pyrol1993;27:245–73.

[39] Neves D, Thunman H, Matosa A, Tarelho L, Gómez-Barea A. Characterizationand prediction of biomass pyrolysis products. Prog Energy Combust Sci2011;37:611–30.

[40] Chirone R, Massimilla L, Salatino P. Comminution of carbons in fluidized bedcombustion. Prog Energy Combust Sci 1991;17:297–326.

[41] Scala F, Salatino P, Chirone R. Fluidized bed combustion of a biomass char(Robinia pseudoacacia). Energy Fuels 2000;14:781.

[42] Scala F, Chirone R, Salatino P. Combustion and attrition of biomass chars in afluidized bed. Energy Fuels 2006;20:91–102.

[43] Chirone R, Salatino P, Scala F. The relevance of attrition on the fate of ashesduring fluidized-bed combustion of a biomass. Proc Combust Inst2000;28:2279–86.

[44] van den Aarsen FG. Fluidised bed wood gasifier. Performance and modelling.Ph.D. dissertation, University of Twente: Twente University Publication; 1985.

[45] Di Blasi C. Combustion and gasification rates of lignocellulosic chars. ProgEnergy Combust Sci 2009;35:121–40.

[46] Gómez-Barea A, Ollero P, Leckner B. Mass transport effects during measurementsof gas-solid kinetics in fluidised bed. Chem Eng Sci 2007;62:1477–93.

[47] Gómez-Barea A, Leckner B, Santana D, Ollero P. Gas-solid conversion influidised beds. Chem Eng J 2008;141:151–68.

[48] Kiel JHA, van Paasen SVB, Neeft JPA, Devi L, Ptasinski KJ, Janssen FJJG et al.Primary measures to reduce tar formation in fluidized-bed biomass gasifiers.Report ECN-C–04-014, The Netherlands; 2004.

[49] Hallgren A. Improved technologies for the gasification of energy crops.Publishable Final Report (TPS AB), European Commission JOULE IIIProgramme. Project No. JOR3-CT97-0125; 1998.

[50] Gómez-Barea A, Vicente-Enamorado, JM, Cruz, JL, Rodríguez P, Campoy M,Ollero P. Commissioning of a 3 MWth demonstration unit for biomassgasification in fluidised bed. In: Proceedings of 16th European biomassconference. Valencia; 2008, p. 2333–6.

[51] Campoy M, Gomez-Barea A, Villanueva AL, Ollero P. Air-steam gasification ofbiomass in a fluidized bed under simulated autothermal and adiabaticconditions. Ind Eng Chem Res 2008;47:5957–65.

[52] Campoy M, Gomez-Barea A, Villanueva A, Ollero P. Air-steam gasification ofbiomass in a fluidised bed: process optimisation by enriched air. Fuel ProcessTechnol 2009;90:677–85.

[53] Ponzio A, Kalisz S, Blasiak W. Effect of operating conditions on tar and gascomposition in high temperature air/steam gasification (HTAG) of plasticcontaining waste. Fuel Process Technol 2006;87:223–33.

[54] Narváez I, Orío A, Aznar MP, Corella J. Biomass gasification with air in anatmospheric bubbling fluidized bed. Effect of six operational variables on thequality of produced raw gas. Ind Eng Chem Res 1996;35:2110–20.

[55] Pan YG, Roca X, Velo E, Puigjaner L. Removal of tar by secondary air injection influidized bed gasification of residual biomass and coal. Fuel 1999;78:1703–9.

[56] Campoy M, Gómez-Barea A, Fuentes-Cano D, Ollero P. Tar reduction byprimary measures in an autothermal air–blown fluidized bed biomass gasifier.Ind Eng Chem Res 2010;49(11):294–301.

[57] Cao Y, Wang Y, Riley JT, Pan WP. A novel biomass air gasification process forproducing tar-free higher heating value fuel gas. Fuel Process Technol2006;87(4):343–53.

[58] Milne TA, Abatzoglou N, Evance RJ. Biomass gasifire ‘‘tars’’: their nature,formation and conversion. NREL/TP-570-25357; 1998. p. 1–68.

[59] Corella J, Toledo JM, Padilla R. Olivine or dolomite as in-bed additive inbiomass gasification with air in a fluidized bed: which is better? Energy Fuels2004;18:713–20.

[60] Olivares A, Aznar MP, Caballero MA, Gil J, Frances E, Corella J. Biomassgasification: produced gas upgrading by in-bed use of dolomite. Ind Eng ChemRes 1997;36:5220–6.

[61] Devi L. Catalytic removal of biomass tars; olivine as prospective in-bed catalystfor fluidized-bed biomass gasifiers. PhD Thesis. Eindhoven, The Netherlands:Technical University of Eindhoven; 2005.

[62] Hayashi Ji, Takahashi H, Iwatsuki M, Essaki K, Tsutsumi A, Chiba T. Rapidconversion of tar and char from pyrolysis of a brown coal by reactions withsteam in a drop-tube reactor. Fuel 2000;79:439–47.

[63] Hayashi Ji, Iwatsuki M, Morishita K, Tsutsumi A, Li CZ, Chiba T. Roles ofinherent metallic species in secondary reactions of tar and char during rapidpyrolysis of brown coals in a drop-tube reactor. Fuel 2002;81:1977–87.

[64] Hosokai S, Kumabe K, Ohshita M, Norinaga K, Li C-Z, Hayashi J-i. Mechanism ofdecomposition of aromatics over charcoal and necessary condition formaintaining its activity. Fuel 2008;87:2914–22.

52 A. Gómez-Barea et al. / Fuel 103 (2013) 42–52

[65] Abu El-Rub Z, Bramer EA, Brem G. Experimental comparison of biomass charswith other catalysts for tar reduction. Fuel 2008;87:2243–52.

[66] Wang Y, Yoshikawa K, Namioka T, Hashimoto Y. Performance optimization oftwo-staged gasification system for woody biomass. Fuel Process Technol2007;88:243–50.

[67] Bui T, Loof R, Bhattacharya SC. Multi-stage reactor for thermal gasification ofwood. Energy 1994;19:397–404.

[68] Susanto H, Beenackers AACM. A moving bed gasifier with internal recycle ofpyrolysis gas. Fuel 1996;75(11):1339–47.

[69] Van der Drift A. An overview of innovative biomass gasification concepts. In:12th European conference and exhibition on biomass for energy and climateprotection, 17–21 June 2002, Amsterdam.

[70] Henriksen U, Ahrenfeldt J, Jensen TK, Gobel B, Bentzen JD, Hindsgaul C.Sorensen LH The design, construction and operation of a 75 kW two-stagegasifier. Energy 2006;31:1542–53.

[71] den Uil H. CASST: a new and advanced process for biomass gasification. In:Developments in Thermochemical Biomass Conversion, 17–22 September,Tyrol, Austria; 2000.

[72] Houmøller S, Hansen MW, Henriksen U. Two-stage fluid bed pyrolysis andgasification unit. In: Proceedings of biomass for energy and industry 9thEuropean conference and technology exhibition 24–27 June 1996,Copenhagen, Denmark.

[73] Lettner F, Haselbacher P, Timmerer H, Leitner P, Suyitno, Rasch B. Latest resultsof CLEASTGAS-Clean staged biomass gasification CHP. In: 15th Europeanbiomass conference & exhibition. Berlin, Germany; 2007.

[74] Schmid C., Mühlen HJ. Staged reforming—a thermal conversion processspecifically developed for biomass. In: Proceedings of the second OlleLindström symposium; 1999. p. 75–82.

[75] Hamel S, Hasselbach H, Weil S, Krumm W. Autothermal two-stage gasificationof low-density waste-derived fuels. Energy 2007;32:95–107.

[76] Gómez-Barea A, Fuentes D, Nilsson S, Tirado J, Ollero P. Fluid-dynamics of acold model of a fluidized bed gasification system with reduced tar content. In:Proceedings of 18th European biomass conference, May 2010, Lyon (France).Paper VP2.1.13.

[77] Gomez-Barea A, Nilsson S, Vidal-Barrero F, Campoy M. Devolatilization ofbiomass and waste in fluidized bed. Fuel Process Technol 2010;91:1624–33.