Tar abatement for clean syngas production during biomass gasification in a dual fluidized bed

The tar removal efficiency of four different catalytic filter designs was evaluated with real biomass tar produced in situ in a dualfluidized bed gasifier DFBG.. Four catalytic filter designs

Research article

L.F de Diegoa,⁎ , F García-Labianoa, P Gayána, A Abada, T Mendiaraa, J Adáneza, M Nackenb, S Heidenreichb

a

Department of Energy and Environment, Instituto de Carboquímica (ICB-CSIC), Miguel Luesma Castán 4, 50018 Zaragoza, Spain

b Pall Filtersystems GmbH Production Site Schumacher, Zur Flügelau 70, 74564 Crailsheim, Germany

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 30 March 2016

Received in revised form 26 May 2016

Accepted 31 May 2016

Available online xxxx

Syngas obtained from biomass gasification needs to fulfil strong purity requirements before being used as raw material in power energy generation or chemicals manufacturing The use of hot catalyticfilter candles inside the freeboard offluidized bed gasifiers allows obtaining clean syngas without dust and low tar content The tar removal efficiency of four different catalytic filter designs was evaluated with real biomass tar produced in situ

in a dualfluidized bed gasifier (DFBG) The tar conversion reached at the outlet of the fluidized bed gasifier was larger for the candles with catalytically active layer design If a monolith is also incorporated, the tar conver-sion increases up to 95% which is one of the highest values obtained up to date In this case, the tar content at the outlet of the catalyticfilter was as low as 0.2 g/Nm3(N2free, d.b.)

© 2016 Elsevier B.V All rights reserved

Keywords:

Biomass gasification

Syngas cleaning

Dual fluidized-bed

Catalytic filter

1 Introduction

Biomass gasification represents a promising technology to produce

energy from a renewable source with zero CO2emissions Gasification

allows transforming biomass in a gas with high content of H2and CO

which account for more than 70% of the energy stored in the biomass

Among the available technologies for biomass gasification, dual

fluid-ized bed gasifiers (DFBG) allow reaching high gasification efficiencies

[1], as it has been shown in some operating gasification plants in Austria

[2]and Sweden[3] In a DFBG, steam gasification takes place in a

bub-blingfluidized bed (BFB) where biomass is converted to syngas

Follow-ing this, the residual char is transferred to a circulatFollow-ingfluidized bed

(CFB) which acts as a combustor, where the char is oxidized and

there-fore heat is generated to be used in the subsequent gasification process

Nevertheless, other gasification products also present in the

gasifica-tion gas can lead to operagasifica-tional problems in the further use of the syngas

generated as raw material in power energy generation or chemicals

manufacturing One of these products is the solid particles leaving the

fluidized bed In recent years, the use of ceramic and metallic filters

for particlefiltration at hot conditions has been investigated[4–6]

An-other product is tar, composed by those organic compounds with a

mo-lecular weight larger than benzene [7] In order to prevent tar

condensation and therefore fouling, it is desirable that the tar content

is decreased down to 30 mg/Nm3or even lower if the gasification gas

is to be used in downstream units such as gas engines or turbines[8]

If the gas is intended for syngas or methanol production or for use in a fuel cell, then more severe restrictions are applied and the tar content should be further reduced to values between 0.1 and 1 mg/Nm3[9]

In recent years, catalytic hot gasfilters for tar abatement have been developed as a cost-effective way to upgrade biomass gasification gas

[10–12] A catalyticfilter candle is normally placed in the freeboard of

afluidized bed where gasification takes place The incorporation of a catalyticfilter inside the gasifier presents several advantages On one hand, it contributes to maintain the thermal efficiency of the biomass conversion process and on the other hand, particle entrainment is avoided Therefore, a hot and clean gas is obtained at the outlet of the gasifier with reduced investment costs Three different types of manu-facture processes for catalyticfilters have been described in literature

[13,14]

i Incorporation of a catalytic component in the ceramic grain and binder mixture during the ceramicfilter manufacture process

ii Modification of the design of the ceramic filter by including a porous inner tubefixed at the head of the filter candle to allow the integra-tion of a catalystfixed bed

iii Catalytic coating on the porous support of a conventional hot gas ce-ramicfilter

Thefirst process was early discarded due to the low surface area of the catalyticfilters produced The high temperatures used in the manufacturing process led to grain sintering and therefore to losses in the active surface of the catalyst[13] Catalyticfilters produced under the other two processes have been optimized and tested under different

⁎ Corresponding author.

E-mail address: ldediego@icb.csic.es (L.F de Diego).

http://dx.doi.org/10.1016/j.fuproc.2016.05.042

Contents lists available atScienceDirect Fuel Processing Technology

j o u r n a l h o m e p a g e :w w w e l s e v i e r c o m / l o c a t e / f u p r o c

conditions.Fig 1presents a scheme of the different configurations of

these two catalyticfilters

Catalyticfilters with a catalyst fixed bed (FB) are shown inFig 1A

They present a high catalytic potential given theirflexibility to integrate

a custom-made tar reforming catalyst and their capability to integrate

high amount of this catalyst in the hollow cylindrical space inside the

fil-ter candle considering the limitations imposed by the total weight of the

candle and the price of the catalyst incorporated However, face

veloci-ties referred to the outer surface of the catalyticfilter candle should be

limited to allow enough residence time for the catalytic tar reforming

reaction[15]

Studies about the optimum composition of thefixed bed catalyst can

be found in literature Nacken et al.[13]tested several tar reforming

cat-alyst systems of different NiO loadings They evaluated the effect of the

variation of the catalyst support material, the preparation conditions,

the NiO loading and the effect of doping with ruthenium on the catalytic

activity of the tar reforming catalyst The catalytic activity tests were

conducted using naphthalene as model tar compound The highest

cat-alytic reforming activity was found for a MgO supported Ni catalyst with

a NiO loading of 6 wt% With this catalyst, complete naphthalene

con-version at 800 °C during 100 h operation even in the presence of H2S

was reached Therefore, this catalyticfilter was tested in a larger scale

A catalyticfilter candle of adequate dimensions was manufactured

and inserted in the freeboard of a bubblingfluidized bed gasifier

where crushed almond shells were used as feedstock[17] Gas and

hy-drogen yields were notably increased with the use of this catalyticfilter

and tar content at the outlet of the catalyticfilter was between 0.7 and

0.95 g/Nm3(N2free, d.b.) Besides, stable performance of thefilter was

observed after 22 h of gasification

The catalyticfilters with catalytic coating on the porous support of

the conventional hot gas ceramicfilter are denoted as catalytic layer

fil-ters (CL) (Fig 1B) This design of catalyticfilter had already been tested

for combined particle separation and NOxremoval from laboratory[18]

to pilot scale[19] The advantages of the catalytic layerfilters is that

higher face velocities can be used compared to thefixed bed catalytic

fil-ters and therefore the size and weight of the catalyticfilter could be

re-duced This also implies that for the same outer diameter and superficial

velocity higher residence time can be achieved when compared to the

fixed bed catalytic filters Besides a simplification of the manufacture

process compared to thefixed bed filters is also expected[15] The

pos-sibility of integration of a tar reforming catalyst as a catalytic layer by

catalytic activation of 10 mm thick alumina basedfilter disks was first

demonstrated[10,20,21] Then, several studies were carried out aiming

atfinding suitable catalytic systems for the pore walls of ceramic filters which combine high surface support materials and active catalysts[11, 15] In these studies, MgO and CaO-Al2O3were used as supports as well

as MgO-CaO and MgO-Al2O3 In some cases, they were doped with

La2O3, olivine or ZrO2 In all cases, the coatedfilters were catalytically ac-tivated by impregnation with the appropriate aqueous solution of nickel nitrate hexahydrate to adjust the NiO loading amounts of 6 and 60 wt% related to the amount of catalyst support[15] The catalytic activity was evaluated in all the cases using naphthalene as model tar compound Promising materials were selected to manufacturefilter candles to be tested in the freeboard of a bubblingfluidized bed[22] In these exper-iments, tar conversion extent obtained by means of the catalyticfilter was around 58% withfinal tar contents in the gas around 0.8 g/Nm3 (N2free, d.b.) Methane was also partially converted (28%) As a result,

a significant increase in the gas yield (15%) and in hydrogen concentra-tion was reported

Modifications and improvements of the first design of catalytic layer filters (CL) have been recently presented First, the replacement of SiC as filter material with another material which could withstand the high gasifier freeboard temperatures (between 800 and 850 °C) was accom-plished SiC was initially used due to the high heat conductivity and good thermal shock resistance for cyclic back pulse cleaning of the cat-alyticfilter[16] However, it was replaced by Al2O3which allowed long operating times at 850 °C[23] One of the new configurations for cata-lytic layerfilter candles included an additionally integrated catalyst: a catalytically activated Al2O3-based hollow-cylindrical monolith inte-grated in the hollow cylindrical space of the catalyticfilter candle[24] The incorporation of the monolith increased the Ni load of the catalytic filter Promising results were obtained with this new design of catalytic layerfilter candles (CL + M) Using the same experimental apparatus at the same operating conditions[23], tar conversion of 93.5% was reached with the use of the catalytically activated monolith in comparison with the 58% tar conversion of the catalytic layer SiC candle Results were also better than those obtained for SiC candle offixed bed design, where 79% tar conversion was obtained Thefinal tar content of the clean gas was around 0.25 g/Nm3(N2free, d.b.) A catalytically activated ceramic foam as additional reforming step for integration into the hollow-cylin-drical space of the catalytically activated filter candle was used (CL + Foam) The catalytic activity of this combination at different su-perficial velocities was first examined using naphthalene as model tar compound[25] Based on these results, a catalyticfilter of combined

Fig 1 Scheme of configurations of catalytic filters: (A) fixed bed and (B) catalytic layer.

design was developed allowing a reduction of the manufacturing costs

compared to those previously reported for catalyticfilter candles of

fixed bed design, because the catalyst grain filling procedure to realize

thefixed bed was avoided[24] The combined catalyticfilter was tested

with naphthalene as model tar compound and also in experiments in a

bubblingfluidized bed gasifier[12]showing good performance The

hy-drogen content was increased up to 56% (from 39% without catalytic

fil-ter) and the tar content was equal to 0.14 g/Nm3(N2free, d.b.)

In the present work, a comparison of the tar abatement performance

of catalyticfilters with different designs is presented Four catalytic filter

designs were tested:fixed bed (FB), fixed bed with catalytically active

inner tube (FB + CL), catalytic layer (CL) and catalytic layer with

addi-tional monolith (CL + M) The fact that not a biomass tar model

com-pound but real biomass tar produced in situ in a dualfluidized bed

gasifier was used in all the catalytic activity tests adds novelty and

appli-cability to the results presented

2 Experimental

2.1 Dualfluidized bed gasification plant

Biomass gasification was carried out in a bench-scale dual fluidized

bed gasification plant located at ICB-CSIC (Fig 2) and described in a

pre-vious work from the authors[26] The gasification plant consisted of

two interconnectedfluidized beds The gasifier was a bubbling fluidized

bed where biomass was fed in The biomass used was pine wood with

an average particle size of 0.5–2.0 mm The proximate and ultimate

analyses of biomass are shown inTable 1 Steam was used as gasifying

agent for biomass The gasifier bed consisted of Fe/olivine in the size

range 0.1–0.25 mm Fe/olivine material was prepared by impregnation

Iron nitrate (Fe(NO3)3·9H2O) was dissolved in heated water and olivine

was added to the iron aqueous solution The excess water was

eliminated and the sample dried before being calcined for 4 h at

1000 °C Thefinal content of Fe in the Fe/olivine used was 16% A more detailed description of the preparation method of the Fe/olivine can

be found elsewhere[26] After biomass gasification, the solids leaving the gasifier were transferred to the combustor through another bub-blingfluidized bed acting as a loop seal to avoid mixing gaseous atmo-spheres The char which was not gasified was burned in the combustor Hot particles were then returned to the gasifier through a riser

The tar produced in situ during biomass gasification was used in the catalytic activity tests of the differentfilter candles used in the present work The catalyticfilters were located downstream the gasifier prior

to the tar measurement as it is shown inFig 2 Thefilter was placed in-side a reactor and heated by a furnace to control the temperature inin-side thefilter Once the tar has been collected for measurement, several gas analysers were used to determine the composition of the gas product streams: CO, CO2and CH4concentration was measured in a non-disper-sive infrared (NDIR) analyser and H2using a thermal conductivity de-tector Moreover, the presence of C2-C3 hydrocarbons was also analysed off-line using a gas chromatograph (HP 5890) with a Poropack

fluidized bed gasifier.

Table 1 Proximate and ultimate analysis of pine wood (wt.%).

N column It was possible to by-pass the catalyticfilter in order to

deter-mine the composition of the gaseous stream at the inlet of the catalytic

filter

2.2 Tar sampling and analysis

Tar sampling and analysis was based on the European Tar Protocol

[27] Moisture and tar were collected in impingers filled with

isopropanol Water content of the tar was determined using the

Karl-Fi-scher titration method (CRISON Titromatic KF1S) A gas chromatograph

(Agilent 7890A) coupled with a mass spectrometer (Agilent 5975C) was

used in the determination of the concentration of the different tar

com-pounds in the samples collected from the impingers The GC wasfitted

with a capillary column (HP-5) and aflame ionization detector

2.3 Catalyticfilters

Four different catalyticfilter designs were tested in the biomass

gas-ification unit previously described Their scheme is shown inFig 3 All of

them were supplied by Pall Filtersystems GmbH with the following

specifications:

- DeTarCat FB (Fixed Bed)

- DeTarCat FB + CL (catalytically active inner tube)

- DeTarCat CL (Catalytically active Layer)

- DeTarCat CL + M (catalytically active monolith inside)

In all cases, a 21 mm-heightfilter segment taken from the

corre-sponding full-sizefilter candle was used for testing Results can be

con-sidered as representative on small-scale of the behaviour of the full size

filter candle The segment was covered with ceramic caps at the bottom

and upper part of thefilter

In the catalyticfilter fixed bed design (DeTarCat FB), MgO powder

with a BET surface of about 0.15 m2/g and catalytically impregnated

with NiO wasfilled as fixed catalyst bed into the cylindrical space

be-tween two porous silicon carbide tubes The silicon carbide tube has

an open pore volume of 38 vol% In a second embodiment of thefixed

bed design, the inner porous tube was made of silicon carbide and addi-tionally catalytically activated by a MgO-NiO coating (DeTarCat

FB + CL) For the DeTarCat CL design, a porous alumina basedfilter tube with an open pore volume of 45 vol% was catalytically

impregnat-ed with MgO-Al2O3supported NiO In the fourth design tested, the cat-alyst amount of the catalytic layer design was further increased by integration of an alumina foam tube into the interior of the alumina fil-ter element tube (DeTarCat CL + M) The alumina foam tube with an open pore volume of 71 vol% was also catalytically impregnated with MgO-Al2O3supported NiO.Table 2presents a summary of the main characteristics of the four catalyticfilter designs tested

2.4 Experimental plan

Table 3summarizes the tests performed with the different catalytic filters tested Once the steady state was reached, tars produced in the gasifier were measured bypassing the catalytic filter This measurement was considered as a reference test and corresponds to the tar composi-tion at the catalyticfilter inlet After the reference was set, the gasifica-tion gas was forced to pass through the catalytic filter Tar measurements at the outlet of the catalyticfilter were then performed and the clean gas was sent to the analysers To determine the tar conver-sion efficiency of the catalytic filter, the tar reference data were com-pared to the tar measurements after passing through the catalytic filter It was intended that the amount and composition of tar at the inlet of the catalyticfilter were similar for all the experiments per-formed Two gasification parameters were maintained roughly constant

in order to reach this condition First, the temperature in the biomass gasifier was set to 800 °C in all the cases The second parameter was the H2O/biomass ratio In our experiments, it varied between 0.52 and 0.68, which produced a variation in the characteristics of the tar pro-duced in the gasifier and it was considered in the treatment of the results

Regarding the catalyticfilter operating conditions, the influence of two parameters was evaluated On one hand, two temperatures of the catalyticfilter were tested, 800 and 850 °C, according to the recommen-dations made by the supplier On the other hand, the face velocity was varied in order to determine its influence on the performance of the

Fig 3 Gas path through the catalytic filter segments for the four different catalytic filter designs: (A) fixed bed (FB), (B) fixed bed with catalytically active inner tube (FB + CL), (C) catalytic layer (CL) and (D) catalytic layer with additional monolith (CL + M) Ceramic filter ; catalytic fixed bed ; catalytic layer ; catalytic foam ; ceramic caps (For interpretation of the

figure legend, the reader is referred to the web version of this article.)

filter for tar abatement The face velocity was defined as the ratio

be-tween the gasflow and the filter external area Finally, the cumulative

time of each type offilter was also presented inTable 3 The total

oper-ation time for each type of catalyticfilter is bold marked

3 Results and discussion

3.1 Comparison of tar conversion

Fig 4presents the comparison of the amount of tar and the

corre-sponding tar conversion obtained at the catalyticfilter outlet when

the four different catalyticfilters were used The values are represented

versus the corresponding values of face velocities used InFig 4, closed

symbols are used to represent tar reference values and open symbols for

the tar content at the catalyticfilter outlet The amount of tar in the gas

at catalyticfilter inlet oscillates between 2.5 and 4.5 g/Nm3for all the

ex-periments performed

In the experiments with the catalyticfilter with a fixed bed design

(FB), the temperature in thefilter was set to 800 °C This was the

max-imum operating temperature allowed by the manufacturer considering

that SiC is thefilter support material The effect of the face velocity on

the tar content and conversion is clearly seen The highest the face

ve-locity, the larger the tar content at the catalyticfilter outlet and

there-fore, the lower the tar conversion reached This fact can be attributed

to a decrease in the residence time of the gasification gas inside the filter

when the face velocity increases as it has been observed before by the

authors for this type of catalyticfilters[28] The tar conversion

de-creased from 85 to around 50% when the face velocity inde-creased from

40 to 87 m/h At the lowest face velocity tested (40 m/h) the tar content

was 0.7 g/Nm3

If an internal catalytically active inner tube is added to the catalytic filter with a fixed bed design (FB + CL), the resulting filter configuration improves tar removal Experiments were performed at the same tem-perature in thefilter as the experiments with the catalytic filter with fixed bed design (FB), i.e 800 °C In this case, the effect of face velocity

on the tar content at the catalyticfilter outlet is softened The tar content

in the experiment at the lowest face velocity (46 m/h) was 0.8 g/Nm3 Tar conversion values decreased with the increase in the face velocity, although the decrease was not as sharp as in the previous experiments with the FB catalyticfilter Tar conversion values were around 75% The use of catalyticfilters with catalytically active layer (CL) allowed

an improvement in tar conversion when compared to FB and FB + CL catalyticfilters at 800 °C Moreover, it is possible to operate at higher temperatures than with the other two catalyticfilters as Al2O3is used

asfilter support material in the CL catalytic filters At 850 °C, the tar con-tent at the outlet of the CL catalyticfilter was decreased to 0.3 g/Nm3 This corresponds to a tar removal efficiency of 88%, higher than that found with FB-based catalyticfilters It must be also mentioned that the difficulty in the tar removal process increases as tar content decreases

The results obtained with the CL catalyticfilter were further im-proved when a monolith was integrated in the hollow cylindrical space of the catalyticfilter (CL + M) Actually, the best results in this work were obtained using this catalyticfilter configuration Again, tem-peratures up to 850 °C could be reached with this type offilter, which also contributed to its better performance in tar removal For face veloc-ities around 70 m/h, the tar removal efficiency at 800 °C was 80% and it increased up to 95% at 850 °C The later efficiency value corresponds to a tar content at the outlet of the catalyticfilter 0.2 g/Nm3, which can be considered as excellent taking into account the very high tar content

at the inlet of the catalyticfilter (about 4.5 g/Nm3

) These results

Table 2

Characteristics of the catalytic filters used.

Catalytic filter configuration

DeTarCat Fixed bed design

DeTarCat Catalytic layer design

Catalyst support density (g/cm 3

0.0110 (CL)

0.0483 (M) NiO density (g/cm 3

0.0124 (CL)

0.0222 (M) Differential pressure (mbar)

(25 °C; face velocity = 90 m/h)

a

FB design: A: catalytic filter outer diameter B: FB outer diameter C: FB inner diameter D: inner tube inner diameter CL design: A: CL outer diameter B: CL inner diameter C: monolith outer diameter D: monolith inner diameter.

Table 3

Experimental tests with the four DeTarCat catalytic filter elements.

T gasifier

(°C)

H 2 O/biomass dry (g/g)

T filter

(°C)

Face velocity (m/h)

Cumulative time (h)

confirm previous findings by Rapagnà et al.[23]which also pointed to

the CL + M design for catalyticfilters as the most promising for efficient

biomass tar removal above the FB or CL catalyticfilter designs

In the comparison betweenfixed bed (FB) and catalytically active

layer (CL) catalyticfilters it should be born in mind that due to their

de-sign characteristics higher residence time can be expected for a CL

cata-lyticfilter than for the same fixed bed filter with the same candle outer

diameter and at the same superficial velocity Considering this, the

values of the gas hourly space velocity (GHSV) were calculated for the

experiments performed with the different types of catalyticfilters

test-ed with the aim of facilitating the comparison.Table 4summarizes the

values obtained In the case of FB catalyticfilters, larger GHSV values

were observed compared to the other three catalyticfilters Therefore,

shorter residence times for the gasification gas in the catalytic filter

can be expected Nevertheless, these values were close to those

report-ed by Nacken et al.[11]for similar FB catalyticfilters in experiments to

evaluate the catalytic activity using naphthalene as model tar

com-pound They performed 50 h long–term tests at 800 °C and observed

100% naphthalene conversion at GHSV of 3120 h−1and 99.3%

conver-sion at GHSV of 4160 h−1

Another aspect to take into account is the pressure drop in the

cata-lyticfilter Measurements were performed during operation with the

four types of catalyticfilters Results are shown inFig 5 As expected,

higher face velocities lead to a higher pressure drop through the

catalyt-icfilter in all types of filters Nevertheless, the pressure drop registered

for the FB + CL catalyticfilter is notably higher than for the rest of the

filters tested It oscillated between 30 and 43 mbar while for the rest

the values varied between 9 and 23 mbar in the tests This fact

repre-sents an additional disadvantage for the further use of this type of

cata-lyticfilter In comparison to that, the pressure drop measured for the

CL + M catalyticfilter is not especially higher when compared to that

of FB and CL catalyticfilters, because the catalytically activated monolith

creates no additional differential pressure under the applied superficial

flow conditions This result together with the high tar conversion

ob-tained in the experiments with this catalytic filter makes this

configuration the most promising for further development among all those studied in the present work

3.2 Comparison of syngas and tar composition The composition of syngas and tar at the inlet and outlet of the cat-alyticfilters was measured Selected operating conditions of 800 °C in the catalyticfilter and face velocities around 70–80 m/h were chosen

in all the cases in order to compare the results for the differentfilters

Table 5presents the operating conditions and experimental results for the different catalyticfilters including the syngas composition in dry and N2free basis From these values, the H2production and conversion

of the other gases was calculated as the ratio between the variation of moles through the catalyticfilter (outlet minus inlet) and the moles at the inlet In all the cases, an increase in H2at the outlet of the catalytic filter was observed, as it was reported before by other authors[12,22] The largest increment was observed for the FBfilter, probably due to a high catalyst amount present in thefixed bed Among the CL catalytic configuration, the largest H2production was observed for the CL + M filter This result agrees with the observed by other authors when using this type of catalyticfilter[23]

Tar composition is also plotted inFig 6for more clarity In all the cases, the major tar compounds at the inlet of the catalyticfilter were naphthalene, indene and biphenylen At the catalyticfilter outlet, naph-thalene was the major compound and in some cases almost the only tar compound that could be detected in significant level However, it suf-fered a significant drop during its passage through the catalytic filter

Fig 4 (A) Tar amount and (B) tar conversion as a function of the face velocity for the different designs of catalyticfilters tested at 800 or 850 °C (T g = 800 °C) Closed symbols = reference values; open symbols = values at the catalytic filter outlet.

Table 4

Gas hourly space velocity (GHSV).

Type of DeTarCat catalytic filter

T filter

(°C)

GHSV (h−1)

CL + M 800–850 1550–1915 Fig 5 Differential pressure in the catalyticfilters tested at 800 or 850 °C (Tfilter as a function of the face velocity for the

Naphthalene conversion was 42.8% for the FB catalyticfilter and 52.0%

for the FB + CLfilter In the case of the catalytic layer filters,

naphtha-lene conversion reached 50.6% for the CLfilter and 57.8% for the

CL + Mfilter

Considering all the results presented above, some guidelines for fu-ture optimization of the use of catalyticfilters in tar abatement during biomass gasification could be indicated Obviously, primary measures for tar reduction should be applied in the gasifier in order to decrease the tar concentration at the gasifier outlet (i.e the catalytic filter inlet)

as much as possible Temperature and gasflow to be treated are the most important variables affecting the design of a catalyticfilter Al-though high temperatures favor tar conversion, some materials used

in catalyticfilters manufacture can limit its use at temperatures below

800 °C, i.e SiC

In addition, the syngasflow to be cleaned would determine the number of catalyticfilters to be used Low face velocities, large filter di-ameters and wall thickness with increasing catalyst load lead to an in-crease in the residence time for tar and hydrocarbons in the catalytic filter, favoring tar abatement However, these parameters should be op-timized so that a compromise is reached between the tar abatement ef-ficiency and the weight and pressure drop through the filter, considering that the catalyticfilter would hang in the freeboard of the gasifier In this sense, the values of the parameters used in this work can be considered as normal in a future application of these catalytic fil-ters in an industrial gasifier

Regarding the results obtained in our experiments, the lowest values obtained were 0.2 g/Nm3 This value is still far from the limits set for ap-plications with high quality gas requirements (b1 mg/Nm3) such as methanol production or the combustion in a fuel cell However, it would be easier to reach the specifications for combustion in gas en-gines or turbines Although different values are given in literature, the maximum allowable concentration would be 100 mg/Nm3[9] Model-ing calculations based on experimental results for thefixed bed catalytic filters presented in this work showed that tar contents below 0.1 g/Nm3

Table 5

Operating conditions and experimental results for the different catalytic filters.

Catalytic filter

DeTarCat FB

DeTarCat

FB + CL

DeTarCat CL

DeTarCat

CL + M

(%)

Operating conditions

Gas composition (vol%)

Gasifier (N 2 free dry basis)

Combustor

Tar (g/Nm 3

Tar composition (g/Nm 3

dry)

a

H 2 production or gas conversion.

Fig 6 Tar composition at the inlet and outlet of the catalytic filter for the different catalytic

filters tested with T = 800 °C and face velocity = 70–80 m/h (T = 800 °C).

can be reached with catalyticfilter thickness of 25 mm However,

pres-sure drop and weight should be reduced in this configuration Catalytic

layerfilter design could reach similar tar abatement efficiency with less

restriction regarding pressure drop and weight

4 Conclusions

Four different design configurations of catalytic filter for hot gas

con-ditioning have been tested in biomass gasification experiments These

are based on different possibilities to include the catalyst: in afixed

bed or inside a catalytically active layer This novel tar abatement

tech-nology has been evaluated using real biomass tar produced in situ in a

dualfluidized bed gasifier (DFBG) The effect of temperature and face

velocity was evaluated in order to optimize tar abatement

The most promising catalyticfilter design is the catalytic layer

inte-grated with a catalytically activated alumina foam tube (CL + M

de-sign) High tar removal efficiencies up to 95% at 850 °C with

corresponding tar contents down to 0.2 g/Nm3have been achieved

This design provides a technically feasible solution for combined tar

and particulate removal with high performance at acceptable

differen-tial pressure under operating conditions

Further optimization of the catalyticfilter design would be needed in

order to use the syngas in gas engines or turbines However, for

applica-tions with more restricted requirements, such as methanol production

or the use in a fuel cell, additional cleaning downstream the gasifier

would be needed

Acknowledgments

This work was supported by the European Commission (EC Project

UNIQUE No 211517-ENERGY FP7-2008/2011) and the Spanish Ministry

MINECO (ENE2014-56857-R) T Mendiara thanks for the“Ramón y

Cajal” post-doctoral contract awarded by the Spanish Ministry of

Econ-omy and Competitiveness Cristina Igado is also acknowledged for her

contribution to the experimental work

References

[1] G Xu, T Murakami, T Suda, Y Matsuzawa, H Tani, The superior technical choice for

dual fluidized bed gasification, Ind Eng Chem Res 45 (2006) 2281–2286.

[2] M Bolhàr-Nordenkampf, R Rauch, K Bosch, C Aichernig, H Hofbauer, Biomass CHP

Plant Güssing-Using Gasification for Power Generation, International Conference on

Biomass Utilisation, 2002.

[3] M.C Seemann, H Thunman, The New Chalmers Research-gasifier, International

Conference on Polygeneration Strategies, 2009.

[4] R Ghidossi, J.P Bonnet, G Rebollar-Perez, E Carretier, J.H Ferrasse, J Vicente, F.

Topin, P Moulin, Separation of particles from hot gases using metallic foams, J.

Mater Process Technol 209 (2009) 3859–3868.

[5] P.J Kilgallon, N.J Simms, J.F Norton, J.E Oakey, Metallic components in

coal-gasifi-cation fuel gas paths, J Corros Sci Eng 6 (2003).

[6] S.D Sharma, M Dolan, D Park, L Morpeth, A Ilyushechkin, K McLennan, D.J Harris,

K.V Thambimuthu, A critical review of syngas cleaning technologies - fundamental

limitations and practical problems, Powder Technol 180 (2008) 115–121.

[7] K Maniatis, A.A.C.M Beenackers, Tar protocols IEA bioenergy gasification task: in-troduction, Biomass Bioenergy 18 (2000) 1–4.

[8] A.V Bridgwater, The technical and economic feasibility of biomass gasification for power generation, Fuel 74 (1995) (631–53).

[9] H.A.M Knoef, Handbook Biomass Gasification, BTG Biomass Technology Group, En-schede, The Netherlands, 2005.

[10] K Engelen, Y Zhang, D.J Draelants, G.V Baron, A novel catalytic filter for tar removal from biomass gasification gas: improvement of the catalytic activity in presence of H2S, Chem Eng Sci 58 (2003) 665–670.

[11] M Nacken, L Ma, S Heidenreich, G.V Baron, Catalytic activity in naphthalene reforming of two types of catalytic filters for hot gas cleaning of biomass-derived syngas, Ind Eng Chem Res 49 (2010) 5536–5542.

[12] A D'Orazio, S Rapagnà, P.U Foscolo, K Gallucci, M Nacken, S Heidenreich, A Di Carlo, A Dell'Era, Gas conditioning in H2 rich syngas production by biomass steam gasification: experimental comparison between three innovative ceramic filter can-dles, Int J Hydrog Energy 40 (2015) 7282–7290.

[13] M Nacken, M Lina, K Engelen, S Heidenreich, G.V Baron, Development of a tar reforming catalyst for integration in a ceramic filter element and use in hot gas cleaning, Ind Eng Chem Res 46 (2007) 1945–1951.

[14] M.M Yung, W.S Jablonski, K.A Magrini-Bair, Review of catalytic conditioning of bio-mass-derived syngas, Energy Fuel 23 (2009) 1874–1887.

[15] M Nacken, L Ma, S Heidenreich, G.V Baron, Performance of a catalytically activated ceramic hot gas filter for catalytic tar removal from biomass gasification gas, Appl Catal B Environ 88 (2009) 292–298.

[16] M Hackel, G Schaub, M Nacken, S Heidenreich, Kinetics of reduction and oxidation reactions for application in catalytic gas-particle-filters, Powder Technol 180 (2008) 239–244.

[17] S Rapagnà, K Gallucci, M di Marcello, M Matt, M Nacken, S Heidenreich, P.U Foscolo, Gas cleaning, gas conditioning and tar abatement by means of a catalyticfil-ter candle in a biomass fluidized-bed gasifier, Bioresour Technol 101 (2010) 7123–7130.

[18] M Nacken, S Heidenreich, M Hackel, G Schaub, Catalytic activation of ceramic filter elements for combined particle separation, NOx removal and VOC total oxidation, Appl Catal B Environ 70 (2007) 370–376.

[19] S Heidenreich, M Nacken, M Hackel, G Schaub, Catalytic filter elements for com-bined particle separation and nitrogen oxides removal from gas streams, Powder Technol 180 (2008) 86–90.

[20] L Ma, H Verelst, G.V Baron, Integrated high temperature gas cleaning: tar removal

in biomass gasification with a catalytic filter, Catal Today 105 (2005) 729–734 [21] L Ma, G.V Baron, Mixed zirconia-alumina supports for Ni/MgO based catalytic fil-ters for biomass fuel gas cleaning, Powder Technol 180 (2008) 21–29.

[22] S Rapagnà, K Gallucci, M.D Marcello, P.U Foscolo, M Nacken, S Heidenreich, In situ catalytic ceramic candle filtration for tar reforming and particulate abatement in a fluidized-bed biomass gasifier, Energy Fuel 23 (2009) 3804–3809.

[23] S Rapagnà, K Gallucci, M Di Marcello, P.U Foscolo, M Nacken, S Heidenreich, M Matt, First Al 2 O 3 based catalytic filter candles operating in the fluidized bed gasifier freeboard, Fuel 97 (2012) 718–724.

[24] M Nacken, G.V Baron, S Heidenreich, S Rapagnà, A D'Orazio, K Gallucci, J.F.M Denayer, P.U Foscolo, New DeTar catalytic filter with integrated catalytic ceramic foam: catalytic activity under model and real bio syngas conditions, Fuel Process Technol 134 (2015) 98–106.

[25] M Nacken, L Ma, S Heidenreich, F Verpoort, G.V Baron, Development of a catalytic ceramic foam for efficient tar reforming of a catalytic filter for hot gas cleaning of biomass-derived syngas, Appl Catal B Environ 125 (2012) 111–119.

[26] M Virginie, J Adánez, C Courson, L.F De Diego, F García-Labiano, D Niznansky, A Kiennemann, P Gayán, A Abad, Effect of Fe-olivine on the tar content during bio-mass gasification in a dual fluidized bed, Appl Catal B Environ 121–122 (2012) 214–222.

[27] P Simell, P Ståhlberg, E Kurkela, J Albrecht, S Deutsch, K Sjöström, Provisional protocol for the sampling and analysis of tar and particulates in the gas from large-scale biomass gasifiers Version 1998, Biomass Bioenergy 18 (2000) 19–38 [28] F García-Labiano, P Gayán, L.F De Diego, A Abad, T Mendiara, J Adánez, M Nacken,

S Heidenreich, Tar abatement in a fixed bed catalytic filter candle during biomass gasification in a dual fluidized bed, Appl Catal B Environ 188 (2016) 198–206.