Chemical Engineering Science 90 (2013) 284–298

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Science

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

Gasification of wood in a dual fluidized bed gasifier: Influence of fuel

feeding on process performance
Stefan Kern n, Christoph Pfeifer, Hermann Hofbauer
Vienna University of Technology, Institute of Chemical Engineering, Getreidemarkt 9/166, 1060 Vienna, Austria

H I G H L I G H T S G R A P H I C A L A B S T R A C T

c Gasification of wood pellets in a

dual fluidized bed steam gasifier.
c Variation of feeding position of the
solid feedstock.
c Improved gas yield but higher tar
content by feeding of solid feed-
stock onto bed.
c Gas measurement at different levels
in gasification reactor to study gas
formation.

a r t i c l e i n f o a b s t r a c t

Article history: Gasification of wood pellets at a gasification temperature of 850 1C and a fuel power of 90 kW was
Received 29 October 2012 performed in a dual fluidized bed gasifier. Pure steam was used as a gasification agent at a steam-to-
Received in revised form fuel ratio of 0.6 kgH2O/kgfuel,daf. One focus of the investigation was the influence of the position at which
22 December 2012
the solid fuel was fed into the gasifier. In one case the fuel was fed directly into the bubbling fluidized
Accepted 25 December 2012
bed while in the second case the fuel was fed from the top onto the bubbling bed. With in-bed feeding
Available online 4 January 2013
much lower tar contents and a higher H2 content were observed, while with on-bed feeding the amount
Keywords: of product gas generated was significantly higher. A second focus of this test series was the gas
Dual fluidized bed composition inside the reactor. To investigate this, gas measurements were carried out at different
Gasification
height levels in the reactor for both fuel feeding options. It was observed that the gas composition
Renewable energy
changed drastically along the height of the gasifier. A decrease in the H2 content of about 12 vol%db was
Fuel
Fluidization measured from the lowest to the highest sampling point in the gasifier while the contents of higher
Thermodynamics process hydrocarbons, CH4, C2H4, and C2H6, increased towards the gasifier outlet.
& 2013 Elsevier Ltd. All rights reserved.

1. Introduction years. In particular, the major part of natural gas is imported into
the European Union from only three countries (Green Paper,
The worldwide demand for energy is growing and the largest 2006). As carbon dioxide emissions from biomass are perceived
part of this growth is currently covered by oil and gas. as being neutral, the utilization of biomass as a substitute for
In particular the developing countries show high rates of growth fossil fuels is attractive. Consequently, finding new, more effective
connected to their growing economies (World Energy Outlook, and wide-ranging applications for low grade and cheap biogenic
2010). The European Union’s dependency on energy imports is and fossil fuels is essential. Gasification of biomass or other solid
expected to rise from 50% at present to around 70% in 20–30 fuels enlarges the range of its conventional application.
In many applications for gasification of solid feedstock, air is
used as gasification agent, as it is cheap and drives the process
n
Corresponding author. Tel.: þ43 1 58801 166382; fax: þ 43 1 58801 16699. autothermally. The drawback of using air is that a lot of nitrogen
E-mail address: [email protected] (S. Kern). is introduced, dilutes the product gas, and lowers the heating

0009-2509/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ces.2012.12.044
 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 285

value of the product gas down to about 3–6 MJ/Nm3db(Zainal in the gasification reactor at different height levels of the free-
et al., 2002). When steam or CO2 is used as a gasification agent, board as well as in the bubbling bed will be carried out.
the product gas is also free of nitrogen and the calorific value of
the gas is higher: for steam, gasification values between 10 and
18 MJ/Nm3db can be reached (Rapagna et al., 2000; Schuster et al., 2. Fundamentals
2001). The advantage using steam instead of CO2 is that the
reactivity of steam is on average about four times higher than that 2.1. Gasification of solid feedstock with steam
of CO2 (Molina and Mondragón, 1998), so residence times of the
char in the gasification section would have to be longer and the The conversion of solid fuels into gaseous products by gasifi-
gasification efficiency would suffer. With H2O or CO2 as a cation occurs in several steps. The first step is drying of the fuel,
gasification agent, the process becomes allothermal, so the heat followed by the release of volatile components and thermal
for the endothermic gasification reactions has to be provided degradation in the absence of oxygen, called pyrolysis. The
externally. residual char can be converted into product gas by a gasifying
Dual fluidized bed (DFB) gasification solves this problem as the agent like air, oxygen, steam, carbon dioxide, or a mixture of
combustion reactor, which provides the energy for gasification, is oxygen and steam. In fluidized beds, gasification takes place at
separated from the gasification reactor and pure steam is used as temperatures between 800 and 900 1C. For DFB gasification, pure
gasification agent. Circulating bed material between these two steam will be used as a gasification agent, so this is the type that
reactors carries the heat from the combustion reactor to the will be considered here.
gasification reactor. This gasification technology (Hofbauer et al., The main gasification reactions are shown in Table 1. These
2002) has been developed at Vienna University of Technology and reactions are considered as equilibrium reactions with changing
has been successfully demonstrated in Güssing (Hofbauer et al., equilibrium conditions depending on gas concentrations, tem-
2003) and Oberwart (Kotik, 2010), Austria, on the 8 and 10 MWth perature, and pressure. For the applied temperature range of
scales, respectively, since 2001 and 2008. Another plant in Villach, fluidized bed gasification, equilibrium will normally not be
Austria with a fuel power of 15 MW is in operation since 2011 reached. In the gasification reactor these reactions can take place
(Klotz, 2010). Further plants in Klagenfurt, Austria, Gothenburg, at the same time and place and some reactions can be forced by
Sweden (Gunnarsson, 2010), and Senden, Germany, are currently operating parameters and by the utilization of catalytically active
in planning (Klagenfurt), under construction (Gothenburg), or in bed material. When applying steam for gasification, the water–
the startup period (Senden), and will achieve fuel power of gas shift reaction is forced, so carbon is converted into H2 and CO
15 MW (Senden), 25 MW (Klagenfurt), and 32 MW (Gothenburg). (Eq. (1)), and carbon which is also present in the gas in the form of
The Institute of Chemical Engineering at Vienna University of CO can be converted to H2 and CO2 (Eq. (2)) already in the
Technology operates a 100 kW DFB pilot plant for research gasification reactor. This leads to a product gas with a high
purposes, where research concerning different types of biomass content of hydrogen. Based on the gasification reactions above
(Pfeifer et al., 2011a; Kitzler et al., 2011), coal (Kern et al., 2011; and including the aspect of present oxygen in the solid fuel, the
Kern et al., 2012a), sewage sludge (Schmid et al., 2011), or plastics overall reaction for gasification of solid fuels can be expressed as
(Wilk et al., 2011) as well as investigations regarding operating  y 
parameters (Kern et al., 2012b) and bed materials (Pfeifer et al., Cx Hy Oz þ ðxzÞH2 O-xCO þ x þ z H2 for ðx 4 zÞ DHR,850 4 0
2
2011b; Koppatz et al., 2011) have been carried out over the last ð9Þ
decade.
Gasification of solid fuels in a fluidized bed reactor is an From Eq. (9) the minimum amount of steam that must be
extensive matter of interest worldwide for the time being. In present during the process for a known composition of the fuel
most of the cases of investigations there is a focus on the products can be defined. The stoichiometric steam demand can be
(product gas composition) at the outlet of the gasification reactor, extracted as
but, in particular, previous research by Miccio et al. (1999) fH2 O ¼ ðxzÞ ð10Þ
showed that in the freeboard of a fluidized bed gasifier that uses
air as a gasification agent quite a lot of reactions change the For the tests in this publication the feedstock can be expressed
composition and yield of the product gas. For biomass gasification as the system CxHyOz free of sulfur and free of nitrogen. From the
in fluidized bed reactors, the location where the fuel is introduced fuel analysis (see Table 2), the molarities of C, H, and O are found
into the gasification reactor is critical (Corella et al., 2008). to be CH1.44O0.66. With the knowledge of the fuel composition, the
Gómez-Barea et al. (2013) concluded that for the case of auto- stoichiometric steam demand can be calculated as
thermal gasification with oxygen and steam the location of fuel fH2 O ¼ 14:76 molH2 O =kgdaf,N,S,Cl free
feeding influences the gas composition since the gas obtained if
the feedstock is introduced at the top of the reactor has a fH2 O ¼ 0:266 kgH2 O =kgdaf, N,S,Cl free
pyrolytic nature compared to the case where the feedstock is
fed into the middle of the bubbling bed.
The present work deals with gasification of wood pellets with Table 1
Equilibrium reactions in biomass gasification according to Kaltschmitt et al.
pure steam as gasification agent in a dual fluidized bed gasifier (2009).
operated with a fuel power of 90 kWth at a gasification tempera-
ture of 850 1C. To determine the influence of the fuel feeding Name of reaction Chemical equation DH[kJ/mol] Equation
position, two operating points will be investigated. The first
Water–gas (i) C þH2O2CO þH2 þ 118.5 (1)
operating point deals with the fuel feeding position that intro-
Water–gas (ii) C þ2H2O2CO2 þ 2H2 þ 103 (2)
duces the feedstock directly into the bubbling bed, which is Boudouard C þCO222 CO þ 159.9 (3)
typically applied for most applications, while the second operat- Methanation C þ2H22CH4  87.5 (4)
ing point outlines the influence on the system if the fuel is fed Oxidation (i) C þO22CO2  393.5 (5)
from the top onto the bubbling bed. In addition to the general Oxidation (ii) C þ0.5 O22CO  123.1 (6)
Water–gas shift CO þH2O2CO2 þ H2  40.9 (7)
reactor performance, the product gas quality, the energy flows in Methane reforming CH4 þH2O2CO þ3H2 þ 225 (8)
the reactor, and the efficiencies, gas composition measurements
286 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

Table 2 to wood, the amount of coal required to reach the same fuel
Proximate and ultimate analyses of the feedstock. power is smaller than the amount of wood. This would result in
very different steam-to-fuel ratios, but a constant steam-to-
Component Unit Wood pellets
carbon ratio is essential for maintaining comparable gasification
Dry basis As used conditions

Water content – 6.11 _ steam þ nH2 O Um

m _
jSF,wt ¼   fuel ð12Þ
Ash content 0.29 0.27 1nH2 O nash Um _ fuel
C 50.23 47.16
H wt% 6.04 5.67
N 0.05 0.05 _ steam þ nH2 O Um
m _ fuel
jSC,wt ¼ ð13Þ
O 43.38 40.73 nC Um_ fuel
S 0.005 0.005
Cl 0.003 0.003 In steam gasification, the amount of water introduced that is
Volatile matter wt% 86.45 81.17 consumed for the gasification and steam reforming reactions is an
Fixed carbon 13.55 12.72
LHV [MJ/kg] 18.75 17.46
indicator for the whole process. This value is called water
conversion. The relative water conversion is defined as the
amount of water consumed per mass unit of converted fuel

m_ H2 O,con:
In real gasifiers the water introduced for conversion of the X H2 O,rel ¼   ð14Þ
1nH2 O nash Um _ fuel
feedstock will not be converted completely, as a result of several
facts. On the one hand, to maintain a good fluidization, a steam The absolute water conversion simply expresses the relation
flow higher than that required stoichiometrically is typically between the amount of water consumed and the amount of water
chosen. On the other hand the aspects discussed before only introduced as
consider the conversion of the feedstock to H2 and CO. In a real  
m_ H2 O,con:
case many simultaneous reactions take place, such as pyrolysis, X H2 O ¼ 100 ð15Þ
gasification, reforming, and cracking as well as recombination nH2 O Um
_ fuel þ m_ H2 O,steam
reactions for tar compounds that lead to the actual product gas
The conversion of carbon in the gasifier to gaseous products
composition.
can also be used as a key figure for the performance of the
From the facts mentioned above it can be assumed that the
gasification process. This value is the carbon conversion. For a
water–gas shift reaction takes place in the reactor. In general, this
dual fluidized bed gasifier, one has to distinguish between the
reaction is desired, as a high hydrogen content is welcome in
conversion of carbon to product gas in the gasification reactor
many cases. Moreover, materials that promote this reaction also
itself and the conversion of carbon in the whole system to product
force the decomposition of tar compounds (Kirnbauer et al., 2012)
gas and flue gas. In the first case, the carbon conversion is the
and reduce the energy required for the process due to its
ratio of carbon leaving the gasification reactor in the form of
exothermic character (Eq. (7)). To quantify the distance to
gaseous products in the product gas stream to the amount of
equilibrium, a model parameter is introduced which is defined
carbon introduced by the feedstock
as the logarithm of the ratio of the actual partial pressure product
 
to the equilibrium constant (Eq. (11)). If pdeq,CO  shift o0, the m_ C PG
XC ¼ 100 ð16Þ
actual state is still on the side of the reactants, so further reaction nC Um_ fuel
is thermodynamically possible. If pdeq,CO  shift ¼0, the water–gas
shift equilibrium is fulfilled by the product gas composition, and if XC can be used as a kind of parameter for determination of the
pdeq,CO  shift 40, the actual state is on the side of the products. The amount of char that leaves the gasification reactor and enters the
latter case cannot be reached thermodynamically by the water– combustion reactor, neglecting char present in the product gas
gas shift reaction, but is a result of the products of devolatilization stream.
and the other gasification reactions and the water content in the
product gas. If, for example, the gasifier is operated with a low
amount of steam for gasification, the water content will be lower 2.2. Olivine as bed material
in the product gas. This moves the distance to water–gas shift
equilibrium beyond 0 (pdeq,CO  shift 40). The water–gas shift During the last few years, olivine has become a widely known
reaction will then move towards its reactants and consume and used bed material and in-bed catalyst for fluidized bed
energy gasification. It is a naturally occurring mineral formed of silicate
" Q ni # tetrahedra which contain iron and magnesium (Mg1  X, FeX)SiO2.
  i pi The content of iron and magnesium usually differs depending on
pdeq,COshift pi ,T ¼ log10 ð11Þ
K p,COshift ðTÞ where the olivine has been mined. The effect of catalytic tar
reduction through the use of olivine as a bed material was
KP,CO  shift is the equilibrium constant of the water–gas shift reported by Koppatz et al. (2011). Calcination of the olivine before
reaction and can be determined via several sources (HSC, 2002). using it can greatly improve its catalytic activity (Devi et al.,
For practical operation of gasifiers an easier expression of the 2005). For long-term utilization of olivine in biomass gasification
amount of steam is used. The steam-to-fuel ratio expresses the systems, Kirnbauer, Hofbauer (2011) investigated the interaction
sum of water present in the system in relation to the total mass of of the fuel ash with the olivine particles to form layers, rich in
dry and ash free fuel introduced (Eq. (12)). As the steam is initially calcium, around the particles. The effect on the gasification
required for gasification of carbon particles (Eqs. (1) and (2)) the process is that the catalytic effect is considerably improved, so
formulation of the so-called steam-to-carbon ratio will also be the GC/MS tar values were found to be around 80% lower and the
used here (Eq. (13)). This formulation makes it easier to compare gravimetric tar produced was approximately 65% lower in the
the gasification of fuels from with different origins, like coal and operation with coated olivine particles compared to the operation
biomass, because due to the high heating value of coal compared using fresh, uncoated olivine particles (Kirnbauer et al., 2012).
 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 287

3. Materials and methods particle separation from the flue gas at the exit of the combustion
reactor, the bed material flows back to the gasification reactor via
3.1. The dual fluidized bed pilot plant at Vienna University of the upper loop seal. Both the lower and upper loop seals are
Technology fluidized with steam to ensure a high throughput of bed material
and to avoid any leakage of gas between the reactors. In practical
For pilot scale experiments, Vienna University of Technology operations, the gasification temperature is normally controlled by
operates a 100 kW dual fluidized bed (DFB) gasification reactor the addition of fuel (e.g. recycled producer gas, part of the
(Pfeifer et al., 2011a). A schematic drawing of the pilot rig is feedstock, etc.; Fig. 1, fuel to comb.) into the combustion reactor.
shown in Fig. 1. The fuel, usually biomass, enters the gasification In the case of the 100 kW pilot plant, light heating oil is used as
reactor, a bubbling bed fluidized with steam, where drying, fuel in the combustion reactor as it is easy to handle in pilot scale
pyrolysis, and heterogeneous char gasification takes place. The processes. The main basic geometry data of the DFB reactor
remaining residual char leaves the gasification reactor at the system are summarized in Koppatz et al. (2011).The process
bottom together with bed material, which circulates between the yields two separate gas streams at high temperatures: a high
two reactors, through the lower loop seal to the combustion quality producer gas and a conventional flue gas.
reactor. This reactor is designed as a fast fluidized bed that is The pilot plant is equipped with three different hoppers to
fluidized with air to maintain combustion of the residual char and enable fuel to be fed into the gasification reactor at different
additional fuel, if required. By burning char and additional fuel in locations as well as to give the possibility of co-gasification of
the combustion reactor, the bed material is heated up, and after several feedstock at any mixing ratio. The three hoppers are used
for the following feeding locations or fuel requirements:

 Hopper 1: for feeding of solid fuels into the bubbling bed of

olivine particles. The screw conveyor introduces the fuel about
0.3 m below the splash zone of the bubbling bed. In most cases
this hopper is used.
 Hopper 2: for feeding of solid fuels from the side into the
freeboard of the gasifier. The screw conveyor introduces the
fuel about 0.3 m above the splash zone of the bubbling bed.
 Hopper 3: for feeding of solid fuels from the top of the gasifier
(top-down feeding). This feeding position is designed for fuels
with a low melting point like plastics and ensures that the
fuels do not come into contact with hot surfaces (free-falling
into the reactor) before making contact with the hot bed
material.

For the experiments discussed later in this article, only

hopper 1 and hopper 2 were used.

3.2. Analytics

3.2.1. Measurement of main product gas composition

The composition of product gas is measured after the gasifica-
tion reactor. The permanent gas components CH4, H2, CO, CO2,
and O2 are measured by a Rosemount NGA 2000. The components
N2, C2H4, and C2H6 are measured using an online gas chromato-
graph (PerkinElmer Clarus 500). To avoid any contamination and
damage of the gas analyzer and the column of the online gas
chromatograph, the product gas has to be cleaned in terms of
particulate matter and condensable components like higher
Fig. 1. Scheme of the dual fluidized bed gasification pilot plant at Vienna hydrocarbons (tars) and water, as both gas measurement devices
University of Technology. require dry gas. To maintain this, the gas cleaning line drawn in

Fig. 2. Gas cleaning line for online product gas measurement.

288 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

Fig. 2 is used. Particles like dust and char are removed by a glass- the tar composition. This measurement method gave the follow-
wool-stuffed filter. After this particle removal the gas is led ing results:
though six impinger bottles. Water and hydrocarbons that con-
dense at temperatures higher than 4 1C are collected in the first  Gravimetric tar content
two bottles. They are followed by three further impinger bottles.  GC–MS tar content
Those are filled with rapeseed oil methyl ester where tars are  GC–MS tar composition
washed out of the gas. The last bottle ensures that no rapeseed oil  Water content
methyl ester can leave the gas cleaning line accidentally and  Char load
cause damage to the online gas analyzer or the online gas  Dust load (inorganic matter).
chromatograph.

3.2.3. NH3 and H2S measurement

For ammonia measurement, gas is sampled in a similar way to
3.2.2. Tar measurement
the sampling for the tar measurements, using impinger bottles.
Tar is sampled isokinetically with impinger bottles and, after-
The solvent used in this procedure is diluted sulfuric acid at a
wards, gravimetric as well as GC/MS tars are determined. The tar
temperature of about  2 1C. The impinger bottles are placed in a
sampling is applied discontinuously by condensing and solving of
glycol/ethanol mixture whose temperature is cooled down by a
the tar components out of the product gas. The measurement
cryostatic temperature regulator. To avoid tar condensation in the
method is based on the tar protocol according to CEN/TS 15439
pump, a bottle with toluene is added after the solvent for NH3.
(Neft et al., 1999) focusing on tars originating from biomass
After this procedure the concentration of ammonium ions in the
gasification. The method applied here differs with regard to the
sulfuric acid can be detected by a photometric method according
solvent used, as CEN/TS 15439 proposes the use of isopropanol
to DIN 38 406 Part 5 and ISO 7150. Hydrogen sulfide is sampled
(IPA), but toluene is used here. This allows simultaneous detec-
using impinger bottles filled with an aqueous potassium hydro-
tion of the water content in the product gas because water can be
xide solution at a temperature of about  2 1C. Subsequently, the
measured as a separate phase in the impinger bottles. But unlike
H2S values were determined by potentiometry.
IPA, the use of toluene as a solvent does not enable the tar
components benzene, toluene, and xylene (BTX) to be detected.
However, the separation performance of toluene for tar compo- 3.2.4. Flue gas measurement
nents larger than BTX is higher than that of IPA. A schematic of The composition of the flue gas was measured with a
the arrangement of the tar sampling line is shown in Fig. 3. Rosemount NGA 2000 (CO, CO2, and O2).
The gas enters the heated sampling line, which consists of a
cyclone and a glass-wool-stuffed filled filter cartridge, where dust 3.2.5. Fuel analysis
as well as condensed tar components is deposited. Afterwards, The feedstock for the gasification tests is analyzed by the
the gas is led through six impinger bottles, five of which are filled Testing Laboratory for Combustion Systems at Vienna University
with toluene. The impinger bottles are located in a cooling bath of Technology. The sampling and preparation of the fuels are done
cooled down to  8 1C by a cryostat. There the tars and steam according to DIN 51701. After determination of the water content,
condense. The liquid phases in the impinger bottles are unified described in DIN 51718 (drying at 30 1C to constant mass,
and the aqueous phase is separated from the toluene phase. grinding of the dried sample to a maximum particle size of
Afterwards, the amount of water is determined to calculate the 1 mm, and drying of this sample at 10672 1C in an inert
water content in the gas stream. The amount of toluene is also atmosphere to constant mass), the ash content is determined
marked down and a GC–MS sample is taken. Then the main part according to DIN 51719 by burning the sample to constant mass.
of the toluene is evaporated from the sample in a petri dish. To C, H, N, and S are measured by an EA 1108 CHNS-O elementary
analyze the oil in the filter cartridge it is necessary to carry out a analyzer made by Carlo Erba. This is carried out by burning the
Soxhlet extraction with IPA. Again a GC–MS sample of the IPA sample under oxygen atmosphere. To make sure that CO forma-
phase is taken. Afterwards the IPA phase is handled like the tion is avoided, the gas passes a tungsten catalyst that ensures
toluene phase. The results of the toluene phase and the IPA phase complete oxidation. Afterwards the gas passes a layer of copper at
are added, which gives the amount of gravimetric tar in the a temperature of 860 1C where free oxygen is bound. Here,
product gas. The filter cartridge is reduced to its ashes by nitrogen oxides are also reduced to N2. As a result the gas consists
oxidizing the organic matter in a furnace. By weighing the only of the components CO2, H2O, N2, and SO2 that can be
cartridge before and after the muffle furnace treatment, the detected. Chlorine present in the gas is absorbed in an aqueous
amount of entrained char and dust can be calculated. Finally, perhydrol solution (H2O2) which is analyzed afterwards by
the GC–MS sample is analyzed in a GC–MS device to determine capillary electrophoresis.

Fig. 3. Tar sampling line.

 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 289

3.3. Balance of the pilot plant of wood pellets, a further objective of these test series is to study the
influence of the position where the fuel is fed into the gasification
All of the online measured values obtained during a gasifica- reactor on the performance of the system and the reactions in the
tion test are recorded by process-control software. Using the data freeboard of the gasifier. Both gasification tests were accomplished
collected, mass and energy balances of the DFB system can be with a fuel power of 90 kW, which corresponds to a mass flow rate
calculated. For this purpose, the balance tool IPSEpro was used. of wood pellets of 18.6 kg/h. For gasification test OP1, the wood
IPSEpro is a stationary, equation-oriented flow sheet simulation pellets were fed into the bubbling bed, whereas for OP2 the wood
tool that has been developed for power systems. Detailed infor- pellets were introduced into the freeboard of the gasifier where the
mation about IPSEpro, its mode of operation, and its utilization pellets drop onto the surface of the bubbling fluidized bed of
for biomass-based energy systems are published by Pröll and hot olivine particles. Both test runs were operated with the same
Hofbauer (2008). key data for gasification, that is, the same gasification temperature,
bed material size and type,and steam-to-fuel ratio (jSF ¼0.6 kgH2O/
3.4. Used bed material kgdaf),which represents the steam present in the gasifier that will act
as the gasification agent. For each test a new batch of 100 kg of
The bed material that is used for the tests is olivine, provided olivine was used. Tar was sampled at each operating point whereas
by the Austrian manufacturer Magnolithe GmbH. As discussed H2S and NH3 were only sampled at OP1 as their release into the
before it has been observed that calcination of olivine before product gas is relatively constant if the same feedstock is used and
using it in the gasifier increases its catalytic potential. This the gasification temperature is not affected.
thermal preparation is done by the manufacturer in a rotary kiln The steady state phase was held between 7 and 9 h for both
reactor at a temperature of up to 1600 1C for about four hours. operating points. This long time period was necessary as the
Due to this the material is sintered. The results of the XRF analysis discontinuous sampling methods had to be done one after
as can be found in Kirnbauer, Hofbauer (2011). Based on the sieve another, to get the opportunity to observe any interactions in
analysis of the bed material the particle diameter by this meth- the system that produce a certain trend in the gas composition
odology was calculated to be dp ¼ 411 mm. The shape of the and the process performance and to get a sufficient amount of
olivine particles was identified as approximately spherical, so data to close all balances. For start-up of the system, both reactors
the way to specify the mean sauter diameter of the used bed are operated in combustion mode (fluidization with air). When
material is dsv ¼ fdp . With the sphericity f ¼0.9 the mean sauter the temperatures for operation are reached, the fluidization agent
diameter results to dsv ¼ 370 mm. This particle size is classified for the gasification reactor is changed to steam.
according to Geldart (1973) as particle group B.
4.2. Key data and general gasification results
3.5. Feedstock
A first insight into the gasification behavior of wood pellets in
Wood pellets produced according to the Austrian standard the DFB system can be made by highlighting the temperatures in
ÖNORM M 7135 are usually used as a standard fuel to represent the gasifier at the different operating points and fluidization
wood in the gasifier. For the processing of biomass in a power conditions in the system. An overview of the fluidization regime
plant, wood chips are mostly the designated fuel, but in the pilot is presented in Table 4. The velocities Umf and Ut and their ratios
plant the pieces have to be smaller and the quality of the fuel has with the actual superficial velocity in the gasification and com-
to be held constant for the entire test campaign. Therefore, bustion reactors can be used for characterization of the fluidized
instead of wood chips, wood pellets are normally used for the bed system. The generalized regime map of gas–solid fluidized
tests. Furthermore, it was found during previous tests that wood beds proposed by Grace (1986), Haider and Levenspiel (1989),
pellets behave similarly to wood chips in the DFB gasifier
(Kirnbaueret al., 2012). Proximate and ultimate analyses of the Table 4
used wood pellets are listed in Table 2. Fluidization velocities and their ratios to minimal required values.

Value Unit OP1 OP2

4. Results and discussion
Minimum fluidization velocity, Umf m/s 0.07 0.07
Terminal velocity, Ut m/s 3.36 3.36
4.1. Overview Superficial velocity gasification reactor, Ug m/s 0.41 0.47
Fluidization number gasif. reactor, Ug/Umf – 5.69 6.53
Gasification of wood was carried out at two different operating Transport number gasif. reactor, Ug/Ut – 0.12 0.14
Superficial velocity comb. reactor, UC m/s 8.9 9.0
points (OP1 and OP2). The general parameters of these tests are
Fluidization number comb. reactor, UC/Umf – 124.1 124.5
summarized in Table 3. In addition, to investigate the general Transport number comb. reactor, UC/Ut – 2.66 2.67
performance and the product gas composition of steam gasification

Table 3
General parameters for the gasification tests.

Value Unit OP1 OP2

Fuel feeding position – In bubbling bed On bubbling bed

Fuel power wood pellets kW 90
Gasification temperature bubbling bed 1C 850 72
Mean temperature combustion reactor 1C 890 912
Fuel mass flow (wood pellets) kg/h 18.60
jSC kgH2O/kgcarbon 1.3
jSF kgH2O/kgfuel,daf 0.6
Initial bed material inventory kg 100
290 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

and Abba et al.(2003) is used to illustrate the position of the Rep

Un ¼ ð19Þ
operating points (Fig. 4).The Y-axis in this diagram represents the Ar 1=3
dimensionless velocity Un and the X-axis plots the dimensionless
particle size dPn. The values for Fig. 4 can be calculated by rg UU S Udsv
ReP ¼ ð20Þ
m
n 1=3
dP ¼ Ar ð17Þ
The values for the gasification and combustion reactors are
  calculated for the mean sauter diameter of the bed material
rg Ud3sv U rP rg Ug perticles dsv. In the gasification reactor the bubbling fluidized
Ar ¼ ð18Þ bed is operated with a value of Un that is relatively low but still
m2
around six times higher than Umf. The fluidization regime in the
combustion reactor has to be above the terminal velocity for a
single particle (Ut), but for practical operation of the riser it
should be above the superficial velocity where significant entrain-
ment of solids occurs (Use), and this criterion is fulfilled with the
chosen operating points. The temperature profile in the gasifica-
tion reactor for both operating points is shown in Fig. 5. This
diagram shows that the temperatures over the height of the
reactor differ from each other due to several impacts on
the system. The hot bed material which transports the heat from
the combustion reactor to the gasification reactor is circulated
to the gasification reactor in the lower freeboard section. This
causes a local hot spot where the highest temperature in the
reactor is measured. Where the superheated steam for fluidiza-
tion is introduced into the bubbling bed, a cold spot is produced,
because the steam temperature is about 300 1C, which is sig-
nificantly lower than the bed material temperature (  850 1C).
But due to good intermixing of steam, bed material, and fuel
particles and the high heat capacity of the bed material, the
temperature decrease is limited. The set point for the gasification
temperature (850 1C) is chosen at the height in the middle of the
bubbling bed. In this area the fuel is fed into the gasifier in case of
OP1. In the middle and upper sections of the freeboard, the
temperature drops slightly in a constant way due to heat losses
and energy-consuming reactions like steam reforming and crack-
ing reactions as well as pyrolysis and gasification of any particles
entrained in the freeboard.
With the temperature profiles of the gasification test, a first
estimation of the behavior of the system can be made. The
gasification temperature was controlled to 850 1C. To reach this,
the recirculated bed material had to provide a much higher
Fig. 4. Generalized fluidization regime map. temperature in the case of OP2, where on-bed feeding of the

Fig. 5. Temperature profile in the gasification reactor.

 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 291

solid feedstock was applied. As a result, the highest temperature Table 5

in the freeboard was reached for OP2 and was about 12 1C higher Specific production of the main gas components.
compared to feeding the feedstock into the bubbling bed. The cold
Value Unit OP1 OP2
spot in the surroundings of the steam fluidization was affected
more during OP1, where the temperature was cooled down to Specific H2 yield Nm3H2/kgfuel,daf 0.408 0.374
810 1C, while for OP2 a temperature of 820 1C was recorded at this Specific CO2 yield Nm3CO2/kgfuel,daf 0.168 0.167
point. The more intense cooling was a result of the fuel particles Specific CO yield Nm3CO/kgfuel,daf 0.316 0.396
Specific CH4 yield Nm3CH4/kgfuel,daf 0.096 0.117
being heated up, dried, and pyrolyzed; some of them were also Specific C2H4 yield Nm3C2H4/kgfuel,daf 0.019 0.031
gasified in this zone: all of these four steps of fuel conversion are Specific C2H6 yield Nm3C2H6/kgfuel,daf 0.002 0.002
energy consuming. For the case of OP2, where on-bed feeding of
the pellets was applied, these steps took place to a large extent
while floating in the area of the splash zone of the fluidized bed Table 6
and cooled the particles only in a small part of the vertical height. GC–MS tar components that were found only for OP2.

Component Unit OP2

4.3. Product gas composition
3
Benzo[a]pyrene mg/Nm 99.80
The DFB gasification process yields two separate gas streams: a Dibenz[a,h]anthracene mg/Nm3 99.46
Carbazole mg/Nm3 27.38
product gas stream (gasification reactor) and a conventional flue
4-Methylphenol mg/Nm3 23.68
gas stream (combustion reactor). The product gas primarily Isoquinoline mg/Nm3 13.83
consists of the gas components hydrogen (H2), carbon monoxide 1-Benzothiophene mg/Nm3 11.10
(CO), carbon dioxide (CO2), methane (CH4), ethylene (C2H4), Indole mg/Nm3 1.90
ethane (C2H6), and unconverted water (H2O). The difference
between the sum of these main components and 100% of the
gas components is made up by the following components or determination of the change in the components produced can
impurities: be provided. The specific amount of H2 decreased by about 8.2%
forOP2 compared to OP1. The highest increase in the specific
 A nitrogen (N2) content o1.5 vol%db can be present in the amounts produced was ascribed to ethylene (C2H4),which showed
product gas as N2 is used in the process as gas for inertization an increase of 58% for OP2 compared to OP1.The production of
of the fuel feeding system. carbon monoxide (CO) increased by about 25% and that of
 Gaseous C2–C5 hydrocarbons that are not detected with the methane (CH4) by about 21%. The relatively high increase in
applied measurement devices. C2H4 is caused by the very low total amount produced forOP1.
 Hydrogen sulfide (H2S) in the product originating from the C2H6 was almost unaffected as its share in the product gas was
sulfur content in the fuel. very low.
 Ammonia (NH3) in the product originating from the nitrogen The higher production of methane and higher hydrocarbons
in the fuel. also has an influence on the heating value of the product gas
(Table 7) as this value increased in the gas produced by feeding
the fuel above the bubbling bed (OP2). From the results of the
Mean values of the main permanent gas components are
specific production of the permanent gas components, it can be
summarized in Fig. 6,which provides a good impression of the
estimated that the total product gas amount increased with on-
influence of the feeding position on the gas composition. As
bed feeding, as shown in Fig. 7. An explanation for this could be
mentioned before, Gómez-Barea et al. (2013) concluded that a
that the residence time of the char particles in the gasification
fuel feeding position above a bubbling bed produces a gas that
reactor is increased in OP2 as the possibility that a char particle
shows more characteristics of a pyrolysis gas, namely a lower H2
will move together with the bed material to the combustion
content and higher contents of higher hydrocarbons and tar. This
reactor through the lower loop seal is higher when the fuel is fed
behavior can also be seen here. With fuel feeding into the bed
into the bubbling bed (OP1). This also results in a higher conver-
(OP1) the product gas consists of 39.3 vol%db H2, 30.5 vol%db CO,
sion rate of carbon in the gasification reactor to the gaseous
16.2 vol%db CO2, 9.3 vol%db CH4, and 1.9 vol%db C2H4. In contrast to
products shown in Fig. 8. The higher conversion of water
this, forOP2 only 32.8 vol%db H2 and 14.6 vol%db CO2 were
(gasification agent) with in-bed feeding (Fig. 8) is due to the fact
detected but CH4 increased up to 10.3 vol%db and C2H4 was found
that the gasification reactions are faster in the presence of a
to be 2.7 vol%db in the product gas. By the specific production of
catalyst like the hot solid olivine particles. With in-bed feeding
the gas components presented in Table 5, a more exact
(OP1) the olivine–char–steam system can drive the gasification
reactions more easily and especially steam reforming reactions
take place primarily in the bed zone. In the OP2 gasification test
the pyrolysis gases are only able to react with the olivine particles
and the steam in the small splash zone or in the freeboard
without any catalytic action of the particles. With in-bed feeding
much of the higher hydrocarbons was reduced by steam reform-
ing in the bubbling bed. This is also a reason why the hydrogen
content is higher and higher hydrocarbons are reduced with OP1.
The most interesting aspect concerning the permanent gas
composition of the product gas is the shift between H2 and CO.
The steam-to-fuel ratio (jSF) chosen for the test runs here was
low (jSF ¼0.6 kgH2O/kgfuel,daf) compared with previous research
activities at the pilot plant at Vienna University of Technology. In
a previous research by Koppatz et al. (2011) the influence of the
Fig. 6. Measured permanent gas composition of the product gas. steam-to-fuel ratio (jSF) in steam gasification of wood pellets was
292 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

Table 7
Specific data of the accomplished tests.

Value Unit OP1 OP2

Total product gas yield Nm3/h 27.52 30.68

H2O content in product gas vol% 32.62 35.76
Dry product gas yield Nm3db/h 18.09 19.88
Specific product gas yield Nm3db/kgfuel,daf 1.04 1.15
LHV MJ/Nm3db 13.62 14.23
Syngas power excl. tar kW 68.40 78.60
Cold gas efficiency, Zc,IP % 68.97 70.06
Logarithmic deviation from CO-shift equilibrium, pdeq,CO-shift –  0.45  0.49
Carbon conversion in the gasification reactor, XC % 66.85 82.31
Overall carbon conversion of the DFB system, XC,DFB % 96.62 97.19
Specific tar content, GC–MS g/kgfuel,daf 7.43 19.34
Specific tar content, grav. g/kgfuel,daf 1.54 11.18
Specific tar content, GC–MS g/kgcarbon 14.75 38.39
Specific tar content, grav. g/kgcarbon 3.06 22.19
Tar intensity per kWh of syngas, GC–MS g/kWhsyngas 1.89 4.26
Tar intensity per kWh of syngas, grav. g/kWhsyngas 0.39 2.46
Mean gas residence time freeboard gasifier, tF s 3.77 4.15
Mean gas res. time combustion reactor, tC s 0.86 0.85

Fig. 7. Total and specific gas production. Fig. 9. GC–MS and gravimetric tar content in the product gas.

to 0.74 kgH20/kgfuel,daf caused an increase of CO and a decrease of

CO2 in the product gas. Only a minor effect was found for H2
(a decrease when jSF was lowered) and CH4 was not influenced
significantly. In the case here the value of jSF is even lower
so this effect is already even more distinctive forOP1. The sum of
the effects explained before causes this difference in the gas
composition.
The amounts of H2S and NH3 were measured only for operation
of OP1 where 21.5 ppmv of H2S and 856.5 ppmv of NH3 were found.
These impurities are formed by the sulfur and nitrogen content of
the biomass. Previous research showed that nearly all of the
introduced nitrogen is released in the gasification reactor [20, 24].
In a DFB gasifier it has been found that the release of H2S in the
product gas is between 50 and 90% of the amount of sulfur
introduced while the rest leaves the system as SO2 in the combus-
tion reactor (Kern et al., 2011). Due to the higher residence time of
the fuel particles in the gasification reactor with on-bed feeding
of OP2 (explained before), it can be assumed that this percentage of
Fig. 8. Carbon- and water conversion in the gasification reactor. released sulfur in the product gas will increase slightly, but in total
volumetric numbers the H2S content does not have to be higher as
studied and it was found that especially CO and CO2 are influenced the amount of syngas produced is also higher and may dilute this
significantly if jSF drops below a certain value. For wood pellets sulfur content in the gas. Furthermore the measured H2S content in
this was when jSF o0.84 kgH20/kgfuel,daf. The reduction of jSF down the gas was already at a low level.
 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 293

Fig. 10. GC–MS tar classification according to ECN (2009), Milne et al. (1998) and Wolfesberger et al. (2011).

4.4. Tar content fluoranthene and benzo[g,h,i]perylene increased for in-bed feed-
ing. On the other hand there were also substances found in the tar
Especially for the downstream installations of utilization of the with on-bed feeding of the feedstock that were not present with
product gas, the tar content is a most crucial value. For each in bed-feeding. These are listed in Table 6.
operating point, various tar samples were taken and analyzed
according to their amounts, as shown in Fig. 9 (gravimetric tar
4.5. Conversion performance
and GC–MS tar), as well as the composition of the GC–MS tar
presented in Fig. 10 and Table 6. As expected, GC–MS tar as well
Summarizing the tests, some characteristic values are listed in
as the gravimetric tar content increased massively in the product
Table 7. As mentioned before, the water conversion was better for
gas when the pellets were fed onto the bubbling bed. Compared
OP1 where the fuel was fed into the bed. This caused a reduced
to in-bed feeding, the GC–MS tar increased from 7.2 g/Nm3db up
amount of water in the product gas for OP1 compared with OP2.
to 16.8 g/Nm3db, which corresponds to an increase of 135%. The
To provide a value for the efficiency of the gasification system, the
effect on the gravimetric tar was even more drastic because these
cold gas efficiency is used here. For calculation of this value in the
compounds in the product gas increased by about 555% (from
case of tests on the pilot plant it has to be kept in mind that a pilot
1.5 g/Nm3db for in-bed feeding up to 9.7 g/Nm3db for on-bed
plant usually does not reach a low values of heat losses like an
feeding). This massive increase in the gravimetric tars, consisting
industrial large scale plant does. In the case of the DFB pilot plant
mainly of large compounds with a high molecular weight,
the heat losses are nearly 20% of the fuel input. A large scale plant
strengthens the fact that the product gas of on-bed feeding shifts
can be operated with significantly lower heat losses. Stidl (2012)
toward the character of a pyrolysis gas (Gómez-Barea et al., 2013).
calculated the heat losses by radiation for the main parts of the 10
The individual tar components can be detected by GC–MS
MWth DFB gasification plant in Oberwart, Austria (Kotik, 2010).
analysis. A classification of these tar components can be made
Based on the reported heat losses, the heat loss by radiation for an
according to several aspects. Milne et al. (1998) categorized the
industrial plant can be assumed to be 2% of the input fuel power.
components into primary, secondary, and tertiary tar components
To compare the cold gas efficiency to other plants, it is calculated
depending on their temperature of formation. Another classifica-
for an industrial plant size according to
tion can be made when the individual components are dedicated
to their super-ordinated groups (Wolfesberger et al., 2011). These V_ PG ULHV PG
groups are phenolic compounds, furans, aromatic compounds, ZC,IP ¼ ð21Þ
ðPfuel,G þP fuel,C Q_ PP Q_ IP Þ3600
and polyaromatic hydrocarbons (PAH). Naphthalene, as the most
dominant and most stable tar component, would belong to the The distance to equilibrium of the water–gas shift reaction is
group of PAH. To make the content of naphthalene visible it will on the side of the reactants for both operating points because
be removed from the group of PAH and treated separately. A third pdeq,CO  shift o0. For on-bed feeding the reaction is even more on
and widely used classification is the characterization according to the side of CO and H2O. Basically this is caused by the relatively
ECN (2009). Here, tars are classified into five classes. Class I low content of H2 and relatively high content of CO in the
collects tars that are non-detectable by GC–MS as they condense gas, especially for OP2. Due to the missing gas–solid contact
at high temperatures (high molecular weight fraction of gravi- that enhances the reaction, the residence time of the gas in the
metric tar). The other tar classes (II–V) are lighter hydrocarbons freeboard (tF) is too short to favora better conversion to
that are present in the tar. The complete classification of the H2 and CO2.
individual GC–MS tar components into the classes discussed here To give further information about the possible result of the
can be found in Koppatz et al. (2011). The classification of the influence of the water–gas shift reaction on the gas composition,
GC–MS tar components into these groups is shown in Fig. 10. the gas composition for fulfilled water–gas shift equilibrium is
From these data it can be seen that the relative contribution of displayed in Fig. 11. As discussed before, OP2 yields a higher
naphthalene, as the most dominant GC–MS tar component, to the product gas output with a higher heating value. Therefore more
GC–MS tar decreased for OP2 compared to OP1 (Fig. 10). Further carbon from the fuel is converted to product gas. This can be
components whose shares decrease are indene and phenols. In explained by the fact that the possibility of a char particle leaving
contrast to this, the share of components like anthracene, the gasification reactor at the bottom together with the bed
294 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

material is lower when the feedstock is fed onto the bed. combustion reactor, this part has to be substituted with some
A graphic illustration of these energy flows is shown in Fig. 12 more fuel (light heating oil in the case of the pilot plant) for
for OP1 and Fig. 13 for OP2. In both figures the energy streams of combustion to close the balance and steadily provide the heat for
the input and output streams as well as the heat loss of the the gasification process.
gasification reactor (without the combustion reactor) are shown.
The input stream of wood pellets and the stream of char that 4.6. Product gas composition in the gasification reactor
leaves the gasification reactor to the combustion reactor are
assumed to be chemical power. Power in terms of sensible heat As shown in Table 7, the mean gas residence time in the
is displayed in the stream of the fluidization agent (steam), the freeboard (tF) is around 4 s. tF is calculated considering a plug
heat provided by the hot bed material from the combustion flow of the gas in the freeboard. To find out if and how the gas
reactor, and the heat losses of the pilot plant. The hot product composition is affected in the freeboard of the gasifier, the
gas that leaves the reactor is divided into a fraction made of permanent gas components H2, CO, CO2, CH4, C2H4, and C2H6
chemical power and a fraction where sensible heat, entrained were detected at three different levels in the gasifier in addition
char, and tar are summed up, because the product gas is usually to the standard sampling point at the outlet of the gasifier. The
cleaned and cooled for further utilization and therefore only the gas analysis was carried out similarly to the standard permanent
chemical power is used as a product. Therefore often only the gas measurement methods (Rosemount NGA2000 and Perkin
chemical energy can be used for the downstream processes. The Elmer Clarus 500). The gas was sampled at two heights in the
heat from the combustion reactor DQ_ DFB is calculated as the freeboard and one sampling point was chosen in the bubbling
difference in the sensible heat flows with the bed material from bed. The sampling points and their vertical heights, relative to the
and to the gasification reactor. The most important aspect when surface of the bubbling bed in the gasifier, are displayed in Fig. 5.
comparing those two operating points is that, for OP1,24.6 kW of The caption of these points is also shown. L1 (level 1) is the
chemical power in the form of residual char leaves the gasifica- sampling point in the bubbling bed which is 0.14 m below the bed
tion reactor and enters the combustion reactor whereas only surface. L2 (level 2) is located 0.34 m above the bubbling bed and
about 10.9 kW of char are transported to the combustion reactor L3 (level 3) is the sampling point at a vertical height of 0.8 m
in case of OP2. This is in accordance with the increased power in above the bubbling bed. L4 (level 4) represents the gas composi-
the product gas. As nearly 15 kW of char are missing for the tion at the outlet of the reactor at a vertical height 1.65 m above
the bubbling bed. For L4 the gas composition of the standard
sampling (these values were discussed before) is used. Next to
each sampling point (except L3) a thermocouple is also installed
to precisely determine the temperature at the sampling point. The
temperature profile of the gasification reactor is shown in Fig. 5.
To avoid any reactions of the gas in the sampling pipe, the
product gas was cooled down as fast as possible by a water
cooled pipe. The gas was cooled down from the high temperature
in the reactor to ambient temperature in less than 0.12 s. This
cooling rate (41000 K/sec) should be sufficient to avoid post-
sampling reactions. Unfortunately L2 was not ready for measure-
ments for the gasification test of OP2, where the fuel was fed onto
the bubbling bed, so this sampling point was used only for OP1.
With the data obtained, each gas component can be plotted
depending on the sampling level. Figs. 14–16 show these mea-
sured values for each gas component. On the Y-axis of these
Fig. 11. Calculated gas composition for fulfilled water–gas shift equilibrium. diagrams the height of the gasification reactor relative to the

Fig. 12. Energy balance of the gasification reactor for in bed feeding (OP1), values in kW.
 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 295

Fig. 13. Energy balance of the gasification reactor for on-bed feeding (OP2), values in kW.

Fig. 14. H2, CO and CO2 profile in the gasification reactor. Fig. 16. C2H6 profile in the gasification reactor.

increased. This can be caused by the fact that the larger part of
devolatilization of the fuel particles takes place in the upper area
of the bubbling bed. Primary char and hot bed material are
present in the bed, so a high H2 content can be detected there,
formed by the gasification reactions. If the pyrolysis gases are
added to this gas amount, the H2 content is reduced mostly in
terms of dilution. This is the global aspect of these trends, which
is much more distinctive in the case of on-bed feeding during OP2
than during in-bed feeding (OP1). This is because heating, drying,
and devolatilization occur with on-bed feeding while most of the
fuel particles float near the surface and the release of the gas is
abrupt in a very small vertical height difference.
For the higher hydrocarbons C2H4 and C2H6 the profiles of the
gas composition over the vertical height of the reactor showed a
quite constant trend where C2H4 increased from 0.23 to 1.87 vol%db
Fig. 15. CH4 and C2H4 profile in the gasification reactor. for OP1 and from 0.5 up to 2.75 vol%db for OP2. C2H6 was found in
an even lower concentration but it also increased for higher
sampling points in the freeboard. For both operating points, a
surface of the bubbling bed is listed. The black horizontal line in concentration of 0.03 vol%db was detected at the lowest level (L1)
the diagram represents the height where the bubbling bed ends, while a mean C2H6 concentration of 0.15 vol%db was found at the
so below this line is the bubbling fluidized bed of olivine particles top of the freeboard. The production of methane is influenced more
and above it is the freeboard of the gasifier. Big differences in the than any other species in the freeboard by changing the feedstock
gas composition can be observed depending on the vertical feeding position. In Figure 22, it can be seen that in contrast to OP1,
height. The most important aspect is that hydrogen increased the concentration of CH4 in the bubbling bed is marginal for OP2.
up to 51 and 63 vol%db for OP1 and OP2 respectively at the lowest When the wood pellets were fed onto the bed, only 15% of the final
level. So due to an increase in the height, the H2 content reactor outlet methane concentration was found at L1, while for the
was reduced whereas the content of other gas components case where the pellets were fed into the bed, at L1 already 63%
296 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

of the final methane concentration was found. This strengthens the feeding onto the bubbling bed (OP2) causes a higher efficiency of the
assumption that the larger part of the gas production process with whole process chain as the cold gas efficiency is higher in this case
woody biomass takes place in the bed when the fuel is fed into the and the higher tar content does not cause problems if the gas does
bed whereas with on-bed feeding most of the gaseous products are not have to be cooled below the dew point of the tars and is instantly
generated in the upper level and the splash zone of the bubbling burned in a boiler. If the product gas is used for synthesis processes
bed. The change in the CO2 concentrations (Fig. 14) along the or fuel cells, the product gas quality is an essential factor. Most of the
vertical height of the gasifier is, in relative numbers, not particularly synthesis processes like Fischer Tropsch-Synthesis or the production
noticeable; even the CO2 content is nearly the same for both of CH4 (SNG) require a H2:CO ratio higher than 2 and a gas with a
operating points at L1. The difference in the gasifier outlet concen- low amount of tars. The best fuel feeding design for this case is to
tration can be ascribed to the enhanced water–gas shift reaction at feed the fuel directly into the bubbling bed as the gas composition is
OP1. The content of carbon monoxide stayed more constant with in- closer by the requirements (high H2 content) of the downstream
bed feeding over the reactor height as shown in Figure 20. In utilization of the product gas.
addition to the missing contribution of the devolatilization products The feedstock feeding position also influences the profiles of
in the case of OP2, there is also a lower content of char present that the gas species in the gasification reactor. In general, the hydro-
drives the Boudouard reaction for the production of CO with char gen content decreases as it passes along the freeboard towards
and the nearly constant content of CO2 (Fig. 14) available. As was the exit of the gasifier. This can be explained by the fact that in
already discussed in Section 4.2, the temperature gradient also plays the bed some residual char forms primary H2 and CO with the
a role in influencing the gas composition. From L3 to the exit of the steam for gasification. The devolatilization products are released
reactor the temperature decreases from the local hot spot of bed especially in the case of on-bed feeding at the top of the bubbling
material recirculation. Above L3, the majority of the heterogeneous bed and are to a large part responsible for the dilution of
reactions and devolatilization processes of the feedstock particles hydrogen. The reactions in the freeboard are more distinctive
should be finished and only homogeneous reactions should prevail when the solid feedstock is placed on the surface of the bed rather
with the exception of the gas–char reactions with entrained char. than in the splash zone, as drying and pyrolysis of these particles
Entrained char was found in both cases in the range of 12–16 g/ mainly occur in a very small vertical height range compared to
Nm3db. The increase of the gas components CH4, C2H4, and C2H6 in the case when the fuel enters the bed in the middle of its height.
the freeboard cannot be neglected (Figs. 15 and 16) from L3 to L4. In the latter case the pyrolysis products can already use the very
The mean rise of the C2H6 content in relative numbers was about intense gas–solid contact in the bubbling bed and react with the
50%. Also C2H4 increased by 19–23% in relative numbers in the tests steam. These forced steam reforming reactions can also be seen as
and even the rise in CH4 was around 5%. The water–gas shift a part of the tar abatement and the lower content of higher
reaction will not be dominant in the freeboard of the gasification hydrocarbons in this case.
reactor, as it was found before that only in the presence of a solid
catalyst like olivine can influence the gas composition in a signifi-
cant way. Some of these findings also agree with the measurements Nomenclature
done by Miccio et al. (1999) for biomass gasification with air in a
bubbling fluidized bed. He also documented an increase of the Symbols
amount of CH4 and higher hydrocarbons (C2H6–C2H8) found at
higher levels in the freeboard and a higher total gas yield as the Ar Archimedes number, dimensionless
n
N2 content decreased as height increased in the gasifier. The lower dP dimensionless particle diameter, dimensionless
amount of H2O found in the gas also strengthened this fact. dP general particle diameter, mm
dsv sauter diameter, mm
g apparent gravity, m/s2
5. Conclusion DHR,850 heat of reaction at 850 1C, kJ/mol
K p, CO shift equilibrium constant of CO-shift, dimensionless
In this paper investigations on steam gasification of biomass LHVPG lower heating value of the product gas (dry), MJ/Nm3db
was carried out at the Institute of Chemical Engineering at Vienna m_ H2 O,actual actual mass flux of steam in the gasification reactor, kg/h
University of Technology with wood pellets, a fuel power of m_ H2 O,stoich: actual mass flux of steam in the gasification reactor, kg/h
90 kW, a gasification temperature of 850 1C, and a steam-to-fuel m_ H2 O,steam mass flux of steam in the gasification reactor, kg/h
ratio (jSF) of 0.6 kgH2O/kgfuel,daf are presented. It was found that m_ fuel mass flux of solid fuel into the gasification reactor, kg/h
the position of feedstock admission has a significant impact on m_ H2 O,con: amount of water that is converted to product gas, kg/h
the product gas composition, amount, and quality. Also, the m_ C PG carbon flux in the product gas stream, kg/h
energy flow and the temperature profiles in the gasifier were Pi actual measured gas phase partial pressure of the
affected. The major differences found in the case of feeding the species i, Pa
feedstock onto the bubbling bed compared to in-bed feeding are pdeq,CO  shift logarithmic deviation from CO-shift equilibrium,
dimensionless
 Significantly higher CO content and lower H2 content in the Pfuel,G input fuel power of solid fuel into gasification reactor, kW
product gas. Pfuel,C input fuel power of fuel for combustion reactor, kW
 Water–gas shift equilibrium closer to reactants CO and H2O. Q_ PP heat loss of the pilot plant, kW
 Higher amount of product gas. Q_ IP heat loss of an industrial size plant with the same power
 Higher carbon conversion in the gasification reactor. as the pilot plant, kW
 Lower water conversion. Re Reynolds number, dimensionless
 Higher tar yields (Seven new GC–MS tar species were found). T temperature, 1C
Un dimensionless gas velocity, dimensionless
For the evaluation of the product gas quality and process Umf minimum fluidization velocity for a single particle, m/s
performance the downstream utilization of the product gas has to Ut terminal velocity for a single particle, m/s
be distinguished between heat and power production or synthesis Use superficial velocity where significant entrainment of
processes and fuel cells. For heat and power production fuel solids occurs, m/s
 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298 297

Ug, Uc superficial gas velocity in gasification reactor (g) and Corella, J., Toledo, J.M., Molina, G., 2008. Biomass gasification with pure steam in
combustion reactor (c), m/s fluidized bed: 12 variables that affect the effectiveness of the biomass gasifier.
Int. J. Oil Gas Coal Technol 1, 194–207.
V_ PG volumetric flow rate of product gas (dry), Nm3db/h Devi, L., Craje, M., Thüne, P., Ptasinski, K.J., Janssen, F.J.J.G., 2005. Olivine as tar
X H2 O,rel water conversion in the gasifier, related to the fuel removal catalyst for biomass gasifiers: Catalyst characterization. Appl. Catal.
input, kgH2O/kgfuel,daf A: Gen 294, 68–79.
ECN, 2009. Tar classification system. /http://www.thersites.nlS.
x molarity of carbon in the fuel (dry, ash, N, Cl, and S free
Geldart, D., 1973. Types of gas fluidization. Powder Technol. 7, 285–292.
basis), mol/kgC,H,O Gómez-Barea, A., Ollero, P., Leckner, B., 2013. Optimization of char and tar
y molarity of hydrogen in the fuel (dry, ash, N, Cl, and S conversion in fluidized bed biomass gasifiers. Fuel, 103, 42–52.
free basis), mol/kgC,H,O Grace, J.R., 1986. Contacting modes and behavior classification of gas–solid and
other two-phase suspensions. Can. J. Chem. Eng. 64, 353–363.
z molarity of oxygen in the fuel (dry, ash, N, Cl, and S free Green Paper, 2006. A European Strategy for Sustainable. Competitive and Secure
basis), mol/kgC,H,O Energy. Commission of the European Communities, Brussels.
Gunnarsson, I., 2010. The GoBiGas project—efficient transfer of biomass to
biofuels. In: Proceedings of the International Seminar on Gasification, Gothen-
Greek letters burg, Sweden, 28–29 October.
Haider, A., Levenspiel, O., 1989. Drag coefficient and terminal velocity of spherical
and nonspherical particles. Powder Technol 58, 63–70.
d deviation, dimensionless Hofbauer, H., Rauch, R., Loeffler, G., Kaiser, S., Fercher, E., Tremmel, H., 2002. Six
f sphericity, dimensionless years experience with the FICFB-gasification process. In: Palz, W., Spitzer, J.,
fH2 O stoichiometric H2O demand, molH2 O =kgdaf, N,S,Cl free , kgH2 O = Maniatis, K., Kwant, K., Helm, P., Grassi, A. (Eds.), Proceedings of the 12th
kgdaf, N,S,Cl free European Biomass Conference, ETA Florence, Italy, pp. 982–985.
Hofbauer, H., Rauch, R., Bosch, K., Koch, R., Aichernig, C., 2003. Biomass CHP plant
ZC cold gas efficiency, dimensionless Güssing—A success story. In: Bridgewater, A.V. (Ed.), Pyrolysis and Gasifica-
jSF,wt steam-to-fuel ratio, kgH2O/kgfuel,daf,, dimensionless tion of Biomass and Waste. CPL Press, Newbury, Berks., UK, pp. 527–536.
jSC,wt steam-to-carbon ratio, kgH2O/kgC, dimensionless HSC, HSC Chemistry 5.1, 2002. Outokumpu Research Oy, Pori, Finland.
Kaltschmitt, M., Hartmann, H., Hofbauer, H., 2009. Energie aus Biomasse. Springer,
jSC,mol molar steam-to-carbon ratio, molH2O/molC, dimensionless Berlin, Heidelberg.
lH2 O stoichiometric H2O ratio mol/mol, kg/kg Kern, S., Pfeifer, C., Hofbauer, H., 2011. Dual fluidized-bed steam gasification of solid
m absolute/dynamic viscosity, kg=ðmsÞ feedstock: Matching syngas requirements with fuel mixtures. In: Luckos, T., den
Hoed, P. (Eds.), Proceedings of the Industrial Fluidization South Africa (IFSA 2011),
rg gas density, kg/m3
November 2011, Johannesburg, South Africa, 2011, pp. 67–78.
rp density of a solid particle, kg/m3 Kern, S., Pfeifer, C., Hofbauer, H., 2012a. Synergetic utilization of renewable and
tF product gas residence time in the freeboard of the fossil fuels: dual fluidized bed steam co-gasification of coal and wood. APCBEE
gasification reactor, s Procedia 1, 136–140.
Kern, S., Pfeifer, C., Hofbauer, H., 2012b. Co-gasification of wood and hard coal in a
tC gas residence time in the combustion reactor, s dual fluidized bed steam gasifier: process efficiency vs. gasification tempera-
nash ash mass fraction in the fuel, dimensionless ture. In: Proceedings of the 21st International Conference on Fluidized Bed
nC carbon mass fraction in the fuel, dimensionless Combustion (FBC), June 2012, Naples, Italy.
Kirnbauer, F., Hofbauer, H., 2011. Investigations on bed material changes in a dual
nH2 O water mass fraction in the fuel, dimensionless
fluidized bed steam gasification plant in Güssing, Austria. Energy Fuels 25,
3793–3798.
Abbreviations and subscripts Kirnbauer, F., Wilk, V., Kitzler, H., Kern, S., Hofbauer, H., 2012. The positive effects
of bed material coating on tar reduction in a dual fluidized bed gasifier. Fuel
95, 553–562.
BTX benzene, toluene, xylene Kitzler, H., Pfeifer, C., Hofbauer, H., 2011. Gasification of reed in a 100 kW dual
c carbon, cold gas (efficiency), combustion reactor fluidized bed steam gasifier. In: Proceedings of the 19th European Biomass
Conference and Exhibition, June 2011, Berlin, Germany, pp. 1101–1105.
CHP combined heat and power plant Klotz, T., 2010. A regional energy-supply-showcase – the 15 MW fuel-power
daf dry and ash free basis biomass gasification plant Villach. In: Proceedings of the International Seminar
db dry basis on Gasification, Gothenburg, Sweden, October 28–29.
Koppatz, S., Pfeifer, C., Hofbauer, H., 2011. Comparison of the performance
DFB dual fluidized bed
behaviour of silica sand and olivine in a dual fluidised bed reactor system for
F freeboard steam gasification of biomass at pilot plant scale. Chem. Eng. J. 175, 468–483.
g gasification reactor Kotik, J., 2010. Über den Einsatz von Kraft-Wärme-Kopplungsanlagen auf Basis der
GC–MS gas chromatography mass spectrometry Wirbelschicht-Dampfvergasung fester Biomasse am Beispiel des Biomassek-
raftwerks Oberwart. Ph.D. Thesis, Vienna University of Technology (written in
IPA isopropanol German).
OP operating point Miccio, F., Moersch, O., Spliethoff, H., Hein, K.R.G., 1999. Generation and conver-
PAH polycyclic aromatic hydrocarbons sion of carbonaceous fine particles during bubbling fluidised bed gasification
of a biomass fuel. Fuel 78, 1473–1481.
PG product gas
Milne, T.A., Abatzoglou, N., Evans, R.J., 1998. Biomass Gasifier ‘Tars’: Their nature,
SNG substitute natural gas formation, and conversion. National Renewable Energy Lab, Golden, CO, USA.
sv surface volume Molina, A., Mondragón, F., 1998. Reactivity of coal gasification with steam and CO2.
Fuel 77, 1831–1839.
Neft, J.P.A., Knoef, H.A.M., Zielke, U., Sjöström, K., Hasler, P., Simell, P.A., et al., 1999.
Guideline for Sampling and Analysis of Tar and Particles in Biomass Producer
Gases (Tar Protocol), ERK6-CT1999-20002 Version 3.1, ECN.
Acknowledgments Pfeifer, C., Koppatz, S., Hofbauer, H., 2011a. Steam gasification of various feed-
stocks at a dual fluidised bed gasifier: Impacts of operation conditions and bed
materials. Biomass Convers. Biorefin 1, 39–53.
The authors gratefully acknowledge the financial support Pfeifer, C., Koppatz, S., Hofbauer, H., 2011b. Catalysts for dual fluidised bed
granted by the European Commission as this study was carried biomass gasification – an experimental study at the pilot plant scale. Biomass
Convers. Bioref 1, 1–12.
out within the framework of the Fecundus project, funded by the Pröll, T., Hofbauer, H., 2008. Development and application of a simulation tool for
Research Fund for Coal and Steel of the European Union (Contract biomass gasification based processes. Int. J. Chem. React. Eng. 6, 89.
no. RFCR-CT-2010-00009). Rapagna, S., Jand, N., Kiennemann, A., Foscolo, P.U., 2000. Steam-gasification
of biomass in a fluidized-bed of olivine particles. Biomass Bioenerg 19,
187–197.
References Schmid, J.C., Wolfesberger, U., Koppatz, S., Pfeifer, C., Hofbauer, H., 2011. Variation
of feedstock in a dual fluidized bed steam gasifier – influence on product gas,
tar content and composition. Environ. Prog. Sustain. Energy 31, 205–215.
Abba, I.A., Grace, J.R., Bi, H.T., Thompson, M.L., 2003. Spanning the flow regimes: Schuster, G., Löffler, G., Weigl, K., Hofbauer, H., 2001. Biomass steam gasification –
Generic fluidized-bed reactor model. AIChE J. 49, 1838–1848. an extensive parametric modeling study. Bioresour. Technol 77, 71–79.
298 S. Kern et al. / Chemical Engineering Science 90 (2013) 284–298

Stidl, M., 2012. Prozesssimulation von spezifischen Anwendungsfällen der Wolfesberger, U., Koppatz, S., Pfeifer, C., Hofbauer, H., 2011. Effect of iron
Zweibett-Wirbelschicht-Dampfvergasungs-Technologie für die Papier- und supported olivine on the distribution of tar compounds derived by steam
Zellstoffindustrie. Ph.D. Thesis, Vienna University of Technology, p. 112. gasification of biomass. In: Proceedings of the International Conference on
Wilk, V., Kern, S., Kitzler, H., Koppatz, S., Schmid, J.C., Hofbauer, H., 2011. Gasification of Polygeneration Strategies (ICPS11), 30 August–1 September, Vienna, Austria.
plastic residues in a dual fluidized bed gasifier – characteristics and performance World Energy Outlook, 2010. International Energy Agency.
compared to biomass. In: Proceedings of the International Conference on Poly- Zainal, Z.A., Rifau, A., Quadir, G.A., Seetharamu, K.N., 2002. Experimental investi-
generation Strategies (ICPS11), Vienna, Austria. gation of a downdraft biomass gasifier. Biomass Bioenerg 23, 283–289.