Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/251724768
CITATIONS
READS
50
102
5 authors, including:
Zhenhua Min
Mohammad Asadullah
Curtin University
SEE PROFILE
SEE PROFILE
Shu Zhang
24 PUBLICATIONS 490 CITATIONS
SEE PROFILE
Fuel
journal homepage: www.elsevier.com/locate/fuel
Fuels and Energy Technology Institute, GPO Box U1987, Curtin University of Technology, Western Australia 6845, Australia
Department of Chemical Engineering, PO Box 36, Monash University, Victoria 3800, Australia
a r t i c l e
i n f o
Article history:
Received 6 February 2011
Received in revised form 17 March 2011
Accepted 17 March 2011
Available online 30 March 2011
Keywords:
Char-supported iron catalyst
Steam reforming
Tar reforming
Hot gas cleaning
a b s t r a c t
Char, char-supported catalysts and ilmenite were investigated for the steam reforming of biomass tar
derived from the pyrolysis of mallee wood in situ. Special attention was given to the reforming of
aromatic ring systems in tar. The results indicated that the char-supported iron/nickel catalysts exhibited
much higher activity for the reforming of tar than the char itself. Ilmenite and the char-supported iron
catalyst contained similar active phase but showed different tar reforming activities. Kinetic compensation effects demonstrated that the reaction pathways on the char-supported catalysts were similar but
were different from those on ilmenite. The proprieties of support could play important roles for the activities of the catalysts and the reaction pathways on the catalysts. Char would not only disperse the catalysts but also interact with the catalysts to enhance their activity for the steam reforming of tar.
2011 Elsevier Ltd. All rights reserved.
1. Introduction
Biomass gasication faces a number of technical challenges to
become a commercial renewable energy technology. The most
important one is the presence of tar in the gasication product
gas. Catalytic reforming is commonly accepted as the best option
that can convert tar into syngas efciently. Many types of catalysts
such as minerals [18] and Ni-based catalysts [915] have been
tried to reform tar. Iron-based catalysts [8,1619], having adequate
catalytic activities but less prone to coke formation and cheaper
than the Ni-based catalysts, show great promise for reforming tar
during biomass gasication.
The functions of support are important for the activity of the
catalyst. The support should be favourable for extending the life
of catalyst by inhibiting the agglomeration of the active phases
in the catalyst. Our recent review [20] has clearly indicated that
the chars from the pyrolysis and gasication of low-rank coal
and biomass could serve as support for a new class of cheap industrial catalysts with superior performance. Char itself also exhibits
some activity for tar reforming with its reforming activity being
inuenced by pore size, surface area and AAEM contents of char
[2123]. The coke deposited on the char can block the pores and
reduce the surface area to decrease its activity. However, the coke
can also be consumed by gasication/reforming agents and gas Corresponding author at: Fuels and Energy Technology Institute, GPO Box
U1987, Curtin University of Technology, Western Australia 6845, Australia. Tel.: +61
8 9266 1133; fax: +61 8 9266 1138.
E-mail address: chun-zhu.li@curtin.edu.au (C.-Z. Li).
0016-2361/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2011.03.027
eous products, such as steam, O2, and CO2, which would inhibit
its deactivation by producing new pores [22,23]. Furthermore, char
can act as a good support to disperse the active clusters at
nanoscale following pyrolysis or partial gasication [20,2426].
The char-supported catalysts would have low costs and be simply
burned/gasied to recover the energy of the char without the need
of often expensive regeneration after deactivation [20]. Hence, it is
worthy to understand the catalytic activities of the char-supported
catalysts on tar reforming and the functions of char during the catalytic steam reforming process.
In this study, char-supported iron/nickel and ilmenite catalysts
are used for the steam reforming of biomass tar derived from the
pyrolysis of mallee wood in situ. The performance of the catalysts
on the steam reforming of biomass tar was studied under a wide
range of operating conditions. The effects of support on the steam
reforming of tar were investigated. The kinetic parameters have
been calculated using a rst-order kinetic equation and are discussed in terms of changes in the reaction pathways.
2. Experimental
2.1. Preparation of catalysts
A Victorian brown coal was used to prepare the catalyst. The
main procedures for treating brown coal were acid washing and
ion-exchanging [2426]. Briey, Loy Yang brown coal (Victoria,
Australia) with a particle size range of 53150 lm (Table 1) was
washed with 0.2 M H2SO4 solution and de-ionised water to remove
inherent metallic species (mainly, Na, Mg, Ca, Fe, etc.). These
2546
Table 1
Proximate and ultimate analyses of coal and biomass.
Proximate analysis (wt.%)
a
Brown
coal
Mallee
wood
a
b
c
Ash
Volatile
matterb
Fixed
carbonb
Cb
Hb
Nb
Sb
Clb
Ob,c
3.2
51.9
48.1
68.1
4.9
0.6
0.6
0.1
25.7
0.9
81.6
18.4
48.2
6.1
0.15
0.01
0.04
45.5
Dry basis.
Dry and ash-free basis.
By difference.
inherent metallic species are removed only for the purpose of fundamental investigation to ascertain the catalytic activities of each
species although these species themselves could also act as tar
reforming catalysts. The air-dried acid-washed coal that contains
a negligible amount of inorganic species is termed as the H-form
coal because all carboxylates (COOM) have been turned into acids
(COOH). The H-form coal was ion-exchanged with 0.2 M FeCl3 or
Ni(NO3)2 solutions to load iron or nickel into the H-form coal,
respectively. The treated coals are thus termed as the iron-loaded
coal or the nickel-loaded coal.
The catalysts were prepared by pyrolysing and subsequently
gasifying the H-form coal, the iron-loaded coal and the nickelloaded coal at 800 C in a one-stage uidised-bed/xed-bed quartz
reactor [27]. Silica sand, ranging from 212 to 300 lm, was used as
the bed material. The ow rates of feeding gas and uidizing gas
were each 1 L min1. The coal particles were fed into the reactor
at a rate of around 300 mg min1 through a water-cooled probe
when the reactor had reached the target temperature. After the
feeding process, the steam was injected into the reactor to start
gasication that lasted for 10 min. An HPLC pump (Alltech 426)
was used to supply water directly into the reactor at the required
ow rate to make the steam concentration as 15 vol.% of the total
gas fed into the reactor. Finally, the char samples were collected
and used as the catalysts for the steam reforming of biomass tar
in subsequent experiments. The iron content of the char-supported
iron catalyst and the nickel content of the char-supported nickel
catalyst, measured using ICP, were 2.34 wt.% and 0.88 wt.% on
dry basis, respectively.
In some experiments, prior to being used for the steam reforming experiments, the char-supported iron catalyst was reduced at
600 C in situ for 1 h using Ultra High Purity (>99.999%) hydrogen
controlled by a mass ow controller (0.3 L min1).
In order to investigate the functions of the active phase and the
support of the catalysts, ilmenite (TiWest Joint Venture, WA, 106
150 lm) and iron oxide (SigmaAldrich, Fe3O4, <5 lm) were also
employed as catalysts for the steam reforming of biomass tar.
The main components in the raw ilmenite were Fe2O3 and TiO2 [8].
2.2. Steam reforming of biomass tar
A two-stage uidised-bed/xed-bed quartz reactor with three
frits, modied from our earlier design [28], was used for the
in situ catalytic steam reforming of biomass tar (Fig. 1a). The middle frit separated the reactor into two stages. Biomass was pyrolysed in the bottom stage and the product volatiles (tar) were
reformed with steam and the catalysts in the top stage.
In the three-frit reactor (Fig. 1a), the gas ow rates used were so
high for the char-supported catalyst particles (i.e. in the transport
reactor region) that the particles were elutriated to form a thin bed
underneath the top frit. The majority of the top stage was actually
available for the complete mixing of steam and volatiles before
they came into contact with the catalyst particles. However, when
ilmenite and Fe3O4 were used as catalysts, their high density
Fig. 1. A schematic diagram of the experimental set-ups for the steam reforming
experiments. (a) Three-frit reactor [modied from Ref. [24]], (b) four-frit reactor
[modied from Ref. [8]]. (1) water-cooled probe; (2) thermocouples; (3) biomass
particles and argon; (4) steam; (5) argon; (6) gas outlet (to condenser).
meant that these catalyst particles were only uidised. There was
concern that the volatiles and steam might not have well mixed
before contacting with the catalyst particles in the lower part of
the top stage. In order to ensure that steam and volatiles are well
mixed before entering the catalyst bed, the three-frit reactor was
modied into a four-frit two-stage uidised-bed/xed-bed quartz
reactor [8] (Fig. 1b). The distance between the catalyst bed and
the steam injection position of two reactors was the same
(30 mm) and the heights of the static catalyst bed were similar
(1.01.5 mm). These conditions ensured that the gas residence
time in the two reactors was comparable, allowing the direct comparison of the data obtained in these two reactors.
Mallee wood (WA, Australia, Table 1) with a particle size range
of 90180 lm was dried at 105 C for overnight prior to use. Silica
sand (212300 lm) was chosen as the uidised-bed material in the
bottom stage.
The catalysts were pre-loaded into the top stage before the
reactor was heated up. The temperature of the top frit and the temperature distribution of the bottom stage were measured using
two thermocouples for each reactor conguration (Fig. 1). The
feeding gas ow rate was 1 L min1. The uidising gas ow rate
was varied in the range of 0.721.50 L min1 to guarantee the same
residence time at different reforming temperatures. The steam
(15 vol.% of the total gas ow rate) was generated by feeding water
directly into the reactor with a HPLC pump (Alltech 426). The biomass particles were fed into the reactor (at approximately
100 mg min1) through a water-cooled probe after the reactor
had reached and stabilised at the target temperature. After the required amount of biomass had been fed into the reactor, the reactor was lifted out of the furnace immediately.
2.3. Sampling and analysis of tar
The details have been given in Part I of this series [8]. Briey, a
mixture of HPLC grade methanol and chloroform (1:4, vol) was
2547
35
used to capture the tar in three tar traps tted with thimbles and
placed in an ice-water bath and a dry ice bath, respectively. The
solvents improved heat transfer and dissolved the tar immediately,
avoiding the blockage of the thimble lters. Steam in the gas was
condensed in the rst trap with 50 ml solvent that was placed in
an ice-water bath (0 C), which avoided the blockage by ice of
the thimble lter in the second trap setting in a dry ice bath
(78 C). The second trap captured the tar efciently in experiments because negligible amounts of heavy tar (i.e. no colour)
were found in the third trap placed in a dry ice bath (78 C). After
experiments, the tar condensed at the exit of the reactor was
washed with the solution in the third trap. The recovered tar solutions from all three traps were mixed well and weighed accurately.
The tar yield was determined by evaporating the solvents and
water at 35 C for 4 h. The tar is hereby dened experimentally
as the material soluble in the mixture solvents (methanol and chloroform, 1:4) but would not evaporate (with the solvents) at 35 C
within 4 h. The residues in the solvents themselves (i.e. blank)
and the moisture content of biomass were considered in the calculation of the tar yields. The estimated reproducibility of tar yield
measurement is about 0.3 wt.%.
The aromatic ring systems in tar were characterised using a PerkinElmer LS50B luminescence spectrometer with a 1 cm light
path length. The slit widths were 2.5 nm and the scan speed was
200 nm min1. The synchronous spectra were recorded with a constant energy difference of 2800 cm1. The uorescence intensity
was multiplied by the tar yield to be displayed on the basis of per
gram of biomass (db) [29,30].
Pyrolysis
Reforming no catalyst
Reforming with H-char
Reforming with Fe-char
Reforming with Ni-char
30
25
20
15
10
5
0
400
500
600
700
800
900
Temperature ( C)
Fig. 2. The yields of tar after the reforming of mallee wood volatiles with steam
using char and char-supported catalysts as a function of reforming temperature.
Pyrolysis refers to the thermal decomposition of volatiles in the absence of
externally added steam or catalyst.
2000
60
1. 4 ppm
2. 2 ppm
1600
2548
1200
2
800
400
0
260
300
340
380
420
460
1. 4 ppm
2. 2 ppm
45
30
15
0
260
500
300
340
380
420
460
500
Wavelength (nm)
Wavelength (nm)
(b)
(a)
Fig. 3. Constant energy (2800 cm1) synchronous spectra of mallee wood tar solution. (a) The tar was produced from the pyrolysis of mallee wood at 500 C, (b) the tar was
produced from the steam reforming of mallee wood volatiles at 850 C with the char-supported iron catalyst.
2500
2000
4
3
1500
1
2
1000
2500
(a), 500C
500
(b), 600C
2000
1500
4
1
2
1000
5
500
0
0
260
300
340
380
420
460
260
500
300
Wavelength (nm)
Intensity (a.u., per g of
biomass (db))
2500
(c), 700C
2000
1
1500
2
3
1000
500
5
0
260
300
340
380
340
380
420
460
500
Wavelength (nm)
420
460
2500
(d), 800C
2000
1500
1000
2
3
500
0
260
500
300
5
340
380
420
460
500
Wavelength (nm)
Wavelength (nm)
2500
(e), 850C
2000
1500
1000
1
500
0
260
300
340
380
420
460
500
Wavelength (nm)
Fig. 4. Constant energy (2800 cm1) synchronous spectra of mallee wood tar solution (4 ppm) after the steam reforming of its volatiles with/without the char-supported
catalysts at the temperatures (1, pyrolysis; 2, reforming without catalyst; 3, reforming with H-char; 4, reforming with Fe-char; 5, reforming with Ni-char).
systems (mainly 23 rings, 280 nm < wavelength < 360 nm) after
the reforming with H-char or Fe-char were higher than those after
pyrolysis or non-catalytic reforming while the yields of the larger
aromatic ring systems (wavelength >360 nm) were similar at
500 C and 600 C (Figs. 4a and 4b). The increases in the yields of
smaller aromatic ring systems took place when the corresponding
changes in tar yields due to catalyst were small at this temperature
level. There are three possible reactions due to the catalysts. Firstly,
2549
(g)
300
0
300
M M
M M
M M
0
300
Intensity (a.u.)
(f)
Q
M
M
0
300
Q
M
0
300
Q
M
M
0
300
Q
M M
0
300
M M
(e)
M M M
(d)
MM
(c)
M M M
(b)
M M
(a)
0
10
20
30
40
50
2 (degree)
60
70
80
Fig. 5. XRD spectra of fresh and spent char-supported iron catalysts after the steam
reforming of mallee wood tar. (a) Raw Fe-loaded coal, (b) fresh char-supported iron
catalyst prepared at 800 C, (c) spent catalyst after reforming at 500 C, (d) spent
catalyst after reforming at 600 C, (e) spent catalyst after reforming at 700 C, (f)
spent catalyst after reforming at 800 C, and (g) spent catalyst after reforming at
850 C. (M: magnetite (Fe3O4); a: a-Fe; c: c-Fe; Q: quartz).
Table 2
Crystal sizes of iron-containing species in fresh and spent char-supported iron
catalysts calculated using the Scherrer equation.
Fresh catalyst
After being used
After being used
After being used
After being used
After being used
at
at
at
at
at
500 C
600 C
700 C
800 C
850 C
SiO2 (nm)
Fe3O4 (nm)
34.1
33.4
33.9
33.8
34.6
34.6
16.2
17.9
18.7
21.5
15.4
15.3
12.0
H-char
80000
60000
ln[-ln(1-X tar)/]
40000
20000
Fe-char
Ni-char
Ilmenite
8.0
4.0
r
Fe
-c
ha
ha
r
-c
0.8
0.9
1.0
1.2
1.3
1.4
(a)
R
ed
1.1
1000/T (1/K)
10
Fe
3O
0.0
uc
ed
Fe
Fe
-c
h
3O
ar
ha
r
H
-c
ta
ly
st
0
ca
2550
18
Fig. 6. Peak areas of constant energy (2800 cm1) synchronous spectra of mallee
wood tar solution (4 ppm) after the steam reforming with different catalysts at
800 C.
16
According to the tar yields, the kinetic parameters were calculated using a simple rst-order kinetic equation (Eqs. (1) and (2))
for the overall tar reforming reactions [3438].
dX tar
Eapp
kapp exp
1 X tar
dt
RT
ln1 X tar
Eapp
ln kapp
s
RT
where Xtar is the tar conversion after the steam reforming, on the
basis of tar yield after pyrolysis at 500 C, kapp is the apparent
pre-exponential factor, m3/(kg h), Eapp is the apparent activation energy, kJ/mol, s is the spacetime of gas the in the catalytic bed,
(kg h)/m3.
The spacetime for the catalytic bed is dened in Eq. (3).
H-char
14
Ilmenite
12
10
40
50
60
70
80
90
100
Eapp (kJ/mol)
(b)
Fig. 7. (a) Arrhenius plots and (b) CremerConstable relationship for the catalytic
steam reforming of mallee wood tar.
ln
lnk app
Fe-char
Ni-char
It should be noted that the steam supply was in excess and the
concentration of steam remained unchanged in all experiments.
Therefore, the term relating to the concentration of steam has been
lumped into the apparent pre-exponential factor kapp.
The Arrhenius plots and CremerConstable diagram of the catalytic steam reforming (500850 C) are shown in Fig. 7. The slopes
and intercepts of these lines were calculated by linear regression to
give the apparent activation energies and pre-exponential factors
(Table 3) for reforming with the H-form char catalyst, the charsupported iron catalyst, the char-supported nickel catalyst and
ilmenite.
As expected from their relative tar reforming activities, the
apparent activation energy and pre-exponential factor of the
Table 3
Apparent activation energies and pre-exponential factors for the overall tar removal.
Eapp (kJ/mol)
82.1
60.8
57.3
72.6
5.39 106
2.11 106
1.32 106
5.29 105
160000
No catalyst
Fe-char
120000
Ilmenite
80000
40000
0
500
600
700
800
Temperature (oC)
850
Fig. 8. Peak area of constant energy (2800 cm1) synchronous spectra of mallee
wood tar solution (4 ppm) after the steam reforming with/without catalysts at
different temperatures.
proprieties must have played important roles on the reaction pathways on the catalysts. In other words, the support would not only
disperse the iron-containing species but also interact with the
iron-containing species to become involved in the catalytic reactions during tar reforming.
In order to further understand the importance of support in
iron-based catalysts for the steam reforming of tar, the peak areas
of UV-uorescence spectra of aromatic ring systems in tar produced with ilmenite and the char-supported iron catalyst are compared in Fig. 8. These data conrm the drastic difference between
char-supported catalyst and ilmenite. At low temperatures
(<600 C), the aromatic ring systems increased after the steam
reforming with the char-supported iron catalyst, whereas they decreased after the steam reforming with ilmenite. As was mentioned above, the char-supported iron catalyst appears to be
particularly suitable for the dehydration and dehydrogenation
reactions to form additional aromatic ring systems.
In contrast, the char-supported iron catalyst showed higher
activity for reforming aromatic ring systems than ilmenite at high
temperature. Coke formation was still signicant on ilmenite at
high temperature (Fig. 9) [8], leading to slow deactivation of the
catalyst. However, in the case of char-supported catalysts, there
16
12
8
4
0
2551
-4
-8
-12
400
500
600
700
800
900
Temperature (oC)
Fig. 9. The yields of coke after the reforming of mallee wood volatiles with steam
using char and char-supported catalysts as a function of reforming temperature.
(The data of ilmenite were re-plotted from Ref. [8]. The total gas ow rates of argon
were 2.50 L min1 at 500 C, 2.21 L min1 at 600 C, 1.98 L min1 at 700 C,
1.80 L min1 at 800 C and 1.72 L min1 at 850 C, respectively. The concentration
of steam was always maintained at 15 vol.% in the gas stream.)
References
[1] Delgado J, Aznar MP, Corella J. Calcined dolomite, magnesite, and calcite for
cleaning hot gas from a uidised bed biomass gasier with steam: life and
usefulness. Ind Eng Chem Res 1996;35:363743.
[2] Prz P, Aznar MP, Caballero MA, Gil J, Martn JA, Corella J. Hot gas cleaning and
upgrading with a calcined dolomite located downstream a biomass uidised
bed gasier operating with steamoxygen mixtures. Energy Fuels 1997;11:
1194203.
2552
[3] Orivares A, Aznar MP, Caballero MA, Gil J, Francs E, Corella J. Biomass
gasication: produced gas upgrading by in-bed use of dolomite. Ind Eng Chem
Res 1997;36:52206.
[4] Rapagn S, Jand N, Foscolo PU. Catalytic gasication of biomass to produce
hydrogen rich gas. Int J Hydrogen Energy 1998;23:5517.
[5] Gil J, Caballero MA, Martn JA, Aznar MP, Corella J. Biomass gasication with air
in a uidised bed: effect of the in-bed use of dolomite under different
operation conditions. Ind Eng Chem Res 1999;38:422635.
[6] Corella J, Aznar MP, Gij J, Caballero MA. Biomass gasication in uidised bed:
where to locate the dolomite to improve gasication. Energy Fuels
1999;13:11227.
[7] Myrn C, Hrnell C, Bjrnbom E, Sjstrm K. Catalytic tar decomposition of
biomass pyrolysis gas with a combination of dolomite and silica. Biomass
Bioenergy 2002;23:21727.
[8] Min Z, Asadullah M, Yimsiri P, Zhang S, Wu H, Li C-Z. Catalytic reforming of tar
during gasication. Part I. Steam reforming of biomass tar using ilmenite as a
catalyst. Fuel 2011;90:184754.
[9] Arauzo J, Radlein D, Piskorz J, Scott DS. Catalytic pyrogasication of biomass.
Evaluation of modied nickel catalysts. Ind Eng Chem Res 1997;36:6775.
[10] Zhang R, Brown RC, Suby A, Cummer K. Catalytic destruction of tar in biomass
derived producer gas. Energy Convers Manage 2004;45:9951014.
[11] Courson C, Udron L, Swierczynski D, Petit C, Kiennemann A. Hydrogen
production from biomass gasication on nickel catalysts: tests for dry
reforming of methane. Catal Today 2002;76:7586.
[12] Furusawa T, Tsutsumi A. Comparison of Co/MgO and Ni/MgO catalysts for the
steam reforming of naphthalene as a model compound of tar derived from
biomass gasication. Appl Catal, A: Gen 2005;278:20712.
[13] Kinoshita CM, Wang Y, Zhou J. Effect of reformer conditions on catalytic
reforming of biomass-gasication tars. Ind Eng Chem Res 1995;34:294954.
[14] Martnez R, Romero E, Garca L, Bilbao R. The effect of lanthanum on NiAl
catalyst for catalytic steam gasication of pine sawdust. Fuel Process Technol
2004;85:20114.
[15] Miyazawa T, Kimura T, Nishikawa J, Kado S, Kunimori K, Tomishige K. Catalytic
performance of supported Ni catalysts in partial oxidation and steam
reforming of tar derived from the pyrolysis of wood biomass. Catal Today
2006;115:25462.
[16] Tamhankar SS, Tsuchiya K, Riggs JB. Catalytic reforming of benzene on iron
oxidesilica: catalyst activity and reaction mechanism. Appl Catal
1985;16:10321.
[17] Nordgreen T, Liliedahl T, Sjostrom K. Elemental iron as a tar breakdown
catalyst in conjunction with atmospheric uidised bed gasication of biomass:
a thermodynamic study. Energy Fuels 2006;20:8905.
[18] Nordgreen T, Liliedahl T, Sjostrom K. Metallic iron as a tar breakdown catalyst
related to atmospheric, uidised bed gasication of biomass. Fuel 2006;85:
68994.
[19] Matsuoka K, Shimbori T, Kuramoto K, Hatano H, Suzuki Y. Steam reforming of
woody biomass in a uidised bed of iron oxide-impregnated porous alumina.
Energy Fuels 2006;20:272731.
[20] Li C-Z. Some recent advances in the understanding of the pyrolysis and
gasication behaviour of Victorian brown coal. Fuel 2007;86:166483.