Catalytic steam gasification of cellulose using reactive flash volatilization

Article abstract of DOI:10.1002/cctc.201402434

Biomass gasification is considered to be one of the most promising technologies to deliver renewable energy. However, tar formation in the gasifier is one of the main challenges. Ni-based catalysts are one of the most effective transition-metal catalysts in biomass gasification for tar cracking and reforming. Alumina-supported Ni, Pt-Ni, Ru-Ni, Re-Ni and Rh-Ni catalysts were tested for their activity in the reactive flash volatilisation (RFV) of cellulose to produce synthesis gas in 50ms reaction time. RFV is a catalytic gasification process that utilises a high carbon space velocity and mass flow rate with oxygen and steam as gasification agents. Re-Ni, Rh-Ni and Ru-Ni supported catalysts showed a higher gasification efficiency than the other catalysts, which was because of their higher metal surface area, high reducibility and lower CO desorption temperature. The highest gasification efficiency was achieved at 750C with a carbon-to-oxygen ratio of 0.6 and a carbon-to-steam ratio of 1.0 without any oxygen breakthrough.

Full text of DOI:10.1002/cctc.201402434

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201402434
Catalytic Steam Gasification of Cellulose Using Reactive
Flash Volatilization
[a]
Fan Liang Chan and Akshat Tanksale*
Biomass gasification is considered to be one of the most prom-
ising technologies to deliver renewable energy. However, tar
formation in the gasifier is one of the main challenges. Ni-
based catalysts are one of the most effective transition-metal
catalysts in biomass gasification for tar cracking and reforming.
Alumina-supported Ni, Pt-Ni, Ru-Ni, Re-Ni and Rh-Ni catalysts
were tested for their activity in the reactive flash volatilisation
carbon space velocity and mass flow rate with oxygen and
steam as gasification agents. Re-Ni, Rh-Ni and Ru-Ni supported
catalysts showed a higher gasification efficiency than the other
catalysts, which was because of their higher metal surface
area, high reducibility and lower CO desorption temperature.
The highest gasification efficiency was achieved at 7508C with
a carbon-to-oxygen ratio of 0.6 and a carbon-to-steam ratio of
1.0 without any oxygen breakthrough.
(RFV) of cellulose to produce synthesis gas in 50 ms reaction
time. RFV is a catalytic gasification process that utilises a high
Introduction
[
6]
Lignocellulosic biomass is a renewable energy resource that
can be derived from organic sources such as energy crops, ag-
ricultural residues, forestry residues and recycled cardboard
purity. As Ni-based catalysts are used industrially for the
[7]
steam reforming of methane and naphtha, they are expected
to catalyse the steam reforming of tars and the subsequent
[
1]
and paper. Biomass utilisation for fuels and energy produc-
tion is recognised as one of the most promising solutions for
water–gas shift reaction to produce H . However, monometallic
2
Ni catalysts suffer from rapid deactivation because of coke for-
mation if they are used as primary catalysts in fluidised-bed
[2]
the energy crisis and anthropological CO emission problems.
2
[6c,8]
Thermochemical processes for biomass conversion, such as
combustion, gasification and pyrolysis, can be used for power
gasifiers.
The doping of Ni catalysts with a small amount of
noble metal increases its reforming activity, reducibility and
[
2]
[9]
generation and biofuels production. Among the biomass-
conversion technologies, gasification is one of the most inter-
esting technologies from both an industrial and academic re-
coke resistance.
Conventionally, biomass gasification is performed in one of
the following reactor setups:
[3]
search point of view because of its high conversion efficiency.
Biomass gasification can be achieved at temperatures in excess
of 7008C in the presence of oxygen or air with or without ad-
ditional steam. However, at this temperature, a significant
amount of condensable oxygenated hydrocarbons, commonly
referred to as tar, is produced. In the absence of catalysts, tar-
* Fluidised-bed gasifier with a downstream catalytic tar-clean-
ing reactor
* Fast pyrolysis reactor with a downstream catalytic steam re-
former
* Catalytic fluidised-bed gasifier
* Entrained-flow gasifier
[4]
free gasification requires higher temperatures (ꢀ10008C).
Tar removal is a major hurdle that hinders the commerciali-
[
5]
sation of biomass gasification. Many factors can affect the
amount and type of tar formed during gasification. These fac-
tors include gasifier type and design, operating parameters
Fluidised-bed reactors are capital-intensive and hence large-
scale reactors are required to make them economical. The cap-
ital cost of the gasifier and gas-cleaning system can account
[10]
(temperature, pressure, heating rate and residence time), type
for 66.7 to 85.5% of the total capital cost. Entrained-flow
gasifiers are more expensive because of their higher flow ve-
locities and temperatures in excess of 10008C, which require
Ni-based alloys (e.g., Hastelloy) to provide oxidation resistance
and high-temperature strength. In comparison, the capital cost
of fixed-bed reactors for biomass gasification is significantly
lower. Therefore, a new approach called reactive flash volatili-
of feedstock and the type of catalyst used. The optimisation of
these factors may maximise the efficiency of gasification with
minimum tar formation. Dolomite- and CeO /SiO -supported
2
2
Ni, Pt, Pd, Ru and alkali metal oxides have been used in the
past to catalyse gasification reactions, reduce tar formation, im-
prove conversion efficiency and improve the product gas
[11]
sation (RFV) was proposed recently for cellulose gasification.
[
a] F. L. Chan, Dr. A. Tanksale
Department of Chemical Engineering
Monash University
Clayton, VIC 3800 (Australia)
E-mail: akshat.tanksale@monash.edu
RFV uses a fixed-bed gasifier with a carbon space velocity
and carbon mass flow rate 10–100 times higher than that of
[11a]
fluidised-bed reactors.
As a result, RFV reactors require sig-
nificantly less catalyst and a smaller reactor volume to process
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2727

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
a unit mass of carbon feedstock compared to fluidised-bed
gasifiers. Therefore, RFV reactors can be economically viable in
small-scale operation, which is useful for distributed or decen-
tralised biomass-processing facilities. The advantage of decen-
tralised facilities is that the cost of transporting large volumes
of low-energy-density biomass is minimised by placing these
Results and Discussion
Catalyst characterisation
Nitrogen physisorption
The specific surface area, total pore volume and pore size of
the catalysts prepared in this project are summarised in
Table 2. The BET surface area of the commercial alumina sup-
[
12]
facilities closer to the source of biomass.
Similar to conventional gasification, the chemistry of RFV is
complex and yet to be understood fully. However, it is general-
ly accepted that the major reactions include pyrolysis, oxida-
tion, partial oxidation, reduction, steam reforming and water–
2
ꢂ1
port used in this study was 101.62 m g , which decreased by
varying amounts for the impregnated catalysts. This is expect-
ed to be because of pore blockage caused by impregnated
metals. Among the impregnated catalysts, the Ru-Ni catalyst
[
5,13]
gas shift reactions.
Side reactions such as methanation may
[
14]
2
ꢂ1
also occur but to a lesser extent. These reactions are sum-
marised in Table 1.
had the highest specific surface area of 95.74 m g , and the
2
ꢂ1
Re-Ni catalyst had the lowest surface area of 43.62 m g . This
is because the Re-Ni catalyst was not calcined before
the BET measurement was performed, therefore, the
metal precursors are expected to remain on the alu-
mina surface and block a large proportion of the
Table 1. Gasification reactions and their enthalpies for C
6
compounds.
^
0
Reaction
Stoichiometric equation
DHr
T=278C, x=6)
(
pores. After reduction, the BET surface area of the
2
ꢂ1
y
2
Re-Ni catalyst increased to 74.16 m g .
C
C
C
C
C
C
C
C
x
x
x
x
x
x
x
x
H
H
H
H
H
H
H
H
y
y
y
y
y
y
y
y
O
O
O
O
O
O
O
O
z
z
z
z
z
z
z
z
! (1ꢂx)CO+ H
2
+C
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
180
300
71
ꢂ213
ꢂ778
310
230
64
ꢂ41
pyrolysis
ðyꢂ4Þ
! (1ꢂx)CO+
H
2
+CH
4
2
1
2
y
2
+ O
2
! xCO+ H
2
y
2
partial oxidation
+O
+2O
+H
+nH
+(2xꢂz)H
O ! CO +H
CO+3H +H
! CH
2
! (1ꢂx)CO+CO
2
+ H
2
X-ray fluorescence spectroscopy
x
2
x
2
y
2
!
CO+ CO
2
+ H
2
2
The results of the X-ray fluorescence (XRF) elemental
analyses of the catalysts are presented in Table 2.
The NiO and Al O contents were measured directly
2
O ! xCO+yH
2
steam reforming
2
O ! aCO+(xꢂa)CO
2
+yH
2
y
2
2
O ! xCO
2
+(2n+ ꢂz)H
2
2
3
water–gas shift
methanation
CO+H
2
2
2
in the analyses, and the promoter metals content
was calculated based on mass balance. The XRF re-
sults and calculated promoter contents for Ni, Pt-Ni,
Ru-Ni and Rh-Ni are in agreement with the nominal
values, which indicates the effectiveness of the catalyst prepa-
ration procedure.
2
4
2
O
(10) ꢂ206
In RFV, the formation of char is avoided by oxidising the cel-
lulose pyrolytic products rapidly into gases by steam reforming
and partial oxidation reactions. The heat generated from the
oxidation reaction is used in the steam reforming reaction,
which makes the RFV process autothermal. A conversion of
H temperature-programmed reduction
2
9
9% of the feed at ꢁ70% hydrogen selectivity can be ach-
The reducibility of Ni catalysts is an important factor to deter-
[
11b]
0
ieved without catalyst deactivation.
However, so far the RFV
mine the activity of the catalyst. It is known that Ni is active
studies have been performed at temperatures >8008C and by
for steam reforming and the water–gas shift reaction, whereas
[
11]
2+
using supported Rh-Ce catalysts. There is a need to develop
a low-cost catalyst that is active and selective at low tempera-
tures to make this process economically attractive. In this
regard, promoted Ni catalysts may play a vital role.
Ni and NiO are not active. H temperature-programmed re-
2
duction (TPR) was conducted to test the reducibility of pre-
pared catalysts (Figure 1). The monometallic Ni catalyst had
the highest onset reduction temperature of ꢁ4208C, and the
Table 2. Comparison of SBET, total pore volume, amount of active metal, metal dispersion, NiO, Al
2
O
3
and promoter contents of the investigated catalysts.
XRF spectroscopy
Catalyst
Nitrogen physisorption
CO chemisorption
Irreversible CO Metal disper-
S
[
BET
2
V
pore (BJH)
NiO [wt%]
Al
2
O
3
[wt%]
Measured
Promoter [wt%]
ꢂ1
3
ꢂ1
ꢂ1
[a]
m g
]
(P/P
0
) [cm g
]
uptake [mmolg
]
sion [%]
Theoretical
Measured
Theoretical
Calculated
Al
Ni
Pt-Ni
Ru-Ni
Re-Ni
Rh-Ni
2
O
3
101.62
91.99
77.18
95.74
43.62
77.92
0.18
0.21
0.14
0.23
0.07
0.15
n.d.
n.d.
–
–
–
–
–
–
1.37
1.09
n.d.
1.29
12.32
67.88
41.96
78.92
37.67
0.66
3.87
2.33
4.49
2.09
13.59
12.37
12.36
12.36
12.36
15.13ꢃ0.02
13.25ꢃ0.02
11.96ꢃ0.07
10.25ꢃ0.08
14.99ꢃ0.08
86.41
86.53
86.44
86.43
86.41
84.09ꢃ0.07
84.30ꢃ0.05
86.95ꢃ0.05
50.42ꢃ0.18
83.72ꢃ0.10
[
b]
[a] Calculated based on mass balance. [b] Uncalcined catalyst.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2728

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
CO chemisorption
The results obtained from CO chemisorption are listed in
Table 2. The Re-Ni catalyst had the highest amount of irreversible
CO uptake, and therefore, the largest metal surface area and dis-
persion. Overall, all the promoted Ni catalysts exhibited a higher
metal surface area than the monometallic Ni catalyst. This is be-
cause of the higher reducibility of the promoted catalyst and
the higher dispersion of high-atomic-weight noble metals.
CO temperature-programmed desorption
The CO temperature-programmed desorption (TPD) profiles of
all the catalysts used in this study are shown in Figure 2. CO
chemisorption is not activated on Ni and noble metal active
Figure 1. Comparison of the TPR profiles of the investigated catalysts: ~: Ni,
*
: Pt-Ni, !: Ru-Ni,
&
: Re-Ni, ^: Rh-Ni.
[19]
sites, therefore, at room temperature (ꢁ218C) full coverage
can be achieved easily. CO desorption at elevated tempera-
[18]
tures is a complex process that is not well understood.
reduction was not complete at 8008C. The low reducibility of
Nonetheless, it is well known that the rate of desorption is
a function of heat and activation energy of adsorption, and
low activation energy sites exhibit a higher rate of CO desorp-
tion at lower temperatures. The profiles for CO desorption are
shown in Figure 2a, whereas Figure 2b shows profiles for CO2
the Ni catalyst is attributed to the formation of nickel alumi-
[
15]
nate (NiAl O ) spinel if Ni supported on alumina is calcined
2
4
[
16]
at or above 5008C.
All the promoted catalysts showed
a lower onset reduction temperature. The onset reduction
temperature increased in the order: Ru-Ni (1658C) <Pt-Ni
(
2408C)ꢁRe-Ni (2408C)<Rh-Ni (3628C)<Ni (4208C). It is
known from the literature that noble metals promote the re-
duction of NiO through the surface migration of a chemisorbed
hydrogen atom, a phenomenon known as the hydrogen spill-
[
17]
over effect. The most promising catalyst was Re-Ni, which
was reduced almost fully at 8008C.
The total hydrogen consumed by each catalyst in H -TPR
2
was calculated by integrating the area under the curve using
the trapezoidal method (Table 3). As the Re-Ni catalyst was not
Table 3. Comparison of the amount of H
and CO adsorbed on each catalyst.
2
consumed and amount of CO
2
Catalyst H
2
consumed in CO desorbed CO
2
desorbed CO+CO
2
desor-
ꢂ1
TPR [mmolgcat ] in TPD
in TPD
bed in TPD
Ni
124
163
306
344
107
31.98
33.45
29.01
22.41
31.49
1637
1025
1789
2124
2501
1669
1058
1818
2146
2533
Pt-Ni
Ru-Ni
Re-Ni
Rh-Ni
calcined before the H -TPR measurement, the precursors of Re
2
and Ni present on the surface of the catalyst react with hydro-
gen through the reactions shown in Equations (11) and (12)
[
18]
and lead to a high hydrogen consumption.
2
NH ReO þ7 H ! 2 Reþ8 H OþNH
ð11Þ
ð12Þ
4
4
2
2
3
NiðNO Þ þ6 H ! NiþN þ6 H O
3
2
2
2
2
Figure 2. a) CO-TPD profiles and b) CO
2
-TPD profiles of the investigated cata-
lysts. ~: Ni, *: Pt-Ni, !: Ru-Ni,
&
: Re-Ni, ^: Rh-Ni.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2729

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
evolution, which is formed from the dehydroxylation of alumi-
ratios constant at 0.5 and 1.0, respectively. The temperature
range for gasification was selected based on the thermody-
namic analysis of Colby et al. who used HSC chemistry soft-
ware to demonstrate that the CO selectivity increased with the
[
9a,20]
na and subsequent reaction with desorbed CO.
The low-temperature CO and CO2 desorption peaks that
appear at 1208C for the Ni, Ru-Ni and Re-Ni catalysts were at-
tributed to the weak multilayer chemisorption of CO over the
increasing temperature, whereas the H selectivity peaked at
2
[11a]
catalyst. A low-temperature CO peak was also observed for
around 7008C.
A significant amount of char was produced
2
the Rh-Ni catalyst but the corresponding CO peak was not ob-
served, which suggests that Rh-Ni promotes the rapid conver-
at 7008C with all the catalysts (Figure 3), except Re-Ni, which
led to the eventual clogging of the reactor (reduction in the
product gas flow rate). The Re-Ni catalyst showed a char selec-
tivity of only 1%, whereas all other catalysts showed a char se-
lectivity of 19–35%. The gasification efficiency increased with
the increasing temperature, whereas the gas selectivity was
the highest at 7508C. Re-Ni, Ru-Ni and Rh-Ni catalysts per-
formed best at this temperature in terms of a high gas yield
(in that order) and no char. The composition of synthesis gas
on a dry basis produced in these runs is shown in Figure 4. In
general, all the catalysts showed a high selectivity to CO over
sion of CO into CO at ꢁ1208C.
2
The second CO desorption peak observed at a higher tem-
perature of 275–3008C can be attributed to the strong mono-
layer adsorption of CO on the metal surface that requires
a higher temperature and energy to desorb. Broad CO and CO2
desorption peaks were observed for Ni, Pt-Ni and Rh-Ni cata-
lysts in this temperature range (275–3208C), whereas a should-
er peak was observed for the Re-Ni catalyst in this temperature
range. The CO uptake of each catalyst was quantified by inte-
grating the area under the curves for CO and CO desorption
methane and C compounds as a result of the high activity for
2
2
using the trapezoidal method. Re-Ni and Rh-Ni showed the
steam reforming reactions.
highest amount of CO+CO desorption in that order (Table 3).
A low char selectivity and high gasification efficiency were
observed at 750 and 7758C because steam reforming and par-
tial oxidation reactions are favourable at higher tempera-
2
In summary, from the results of the characterisation of the
catalysts used in this project, it can be observed that com-
pared to the monometallic Ni catalysts, the promoted Ni cata-
lysts had a significantly lower reduction temperature, higher
dispersion and higher amount of active metal sites. This is ex-
tremely important, because only a small amount of noble
metal was used in this study, which resulted in a significant im-
provement in the catalyst properties. Therefore, it is expected
that the promoted Ni catalysts will be more active for the RFV
run. Overall, the Re-Ni catalyst showed the highest metal sur-
[21]
tures. These reactions convert the char and tar into gaseous
products. It is known from the literature that the cellulose par-
ticles undergo oxidative pyrolysis on the hot surface of cata-
lysts to form a film of bio-oils that undergo catalytic steam re-
[11a]
forming and partial oxidation.
Therefore, high temperatures
led to the complete conversion of cellulose into gases with
small amount of unconverted volatile organics. However, for
the Ni and Ru-Ni catalysts, temperatures higher than 7508C
were not favourable because of the sintering of metal particles
face area from H -TPR, CO-TPD and CO chemisorption.
2
[22]
at this temperature. The sintering of active metal reduces
the metal surface available for reactions and, therefore, re-
duces the gasification efficiency. Ni catalysts supported on alu-
mina also suffered from the transformation of NiO/Al O to
Reactive flash volatilisation of cellulose
2
3
Effect of gasification temperature
[23]
NiAl O spinel at temperatures above 7508C. A noble-metal
2
4
promoter is able to suppress this phase transformation of Ni
catalyst into spinel by keeping the Ni in a reduced state by the
The catalysts were tested for their activity and stability in RFV
at three operating temperatures (700, 750 and 7758C) by keep-
ing the carbon-to-oxygen (C/O) and carbon-to-steam (C/S)
[9c,24]
hydrogen spillover mechanism.
Figure 3. Effect of RFV temperature and catalyst promoters on the selectivity of char (&), tar (&) and gas (&) based on carbon balance. The numbers on the
top of each bar represent the gasification efficiency, which is the percentage of carbon in the gas and tar combined. Reaction conditions: cellulose flow
ꢂ1
rate=15 gh , C/O=0.5, C/S=1.0 and residence time=50 ms. Here, tar is defined as water-soluble organics.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2730

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 4. Effect of RFV temperature and catalyst promoters on product gas composition. C
conditions: cellulose flow rate=15 gh , C/O=0.5, C/S=1.0 and residence time=50 ms.
2
compounds include ethane, ethylene and acetylene. Reaction
ꢂ1
The H /CO and CO/CO ratios are shown in Figure 5a and b,
this process [Eqs. (13)–(16)].
2
2
respectively. These ratios are important for the downstream
application in biofuels production. For example, for synthesis
2
2
ðCOþ2 H Ð CH OHÞ
ð13Þ
ð14Þ
ð15Þ
ð16Þ
2
3
of dimethyl ether a H /CO ratio close to 1 is desirable because
2
CH OH Ð CH OCH þH O
3
3
3
2
of the overall stoichiometry of DME synthesis from CO and H2
COþH O Ð H þCO
[
Eq. (16)]. Whereas, for methanol synthesis, a H /CO ratio of 2.0
2
2
2
2
is desirable [Eq. (13)]. Therefore, the ability to tune the H /CO
2
Overall : 3 COþ3 H Ð CH OCH þCO
2
3
3
2
ratio by changing the reaction parameters is an advantage in
A H /CO ratio of close to 1.0 is achieved at 7508C with most
2
the catalysts (Figure 5a), whereas a wider distribution of this
ratio is achieved at lower and higher temperatures. The results
ranged from a minimum of 0.66 with Re-Ni at 7008C to a maxi-
mum of 1.57 with Rh-Ni at 7508C. In general, the CO/CO ratio
2
showed an inverse relationship to temperature. The CO/CO2
ratio decreased at high temperatures because of the complete
oxidation of carbon and the higher water–gas shift activity on
[24]
promoted catalysts. A low CO/CO is not desirable as it re-
2
duces the calorific value of the synthesis gas. Therefore, further
tests were performed at 7508C, and the effect of feed ratios of
C/O and C/S on the product H /CO and CO/CO ratios were
2
2
tested to achieve the desired target.
Effect of carbon-to-oxygen ratio in the feed
The effect of the C/O ratio was tested on all the promoted cat-
alysts. The monometallic Ni catalyst was not considered for
these tests because of the high char selectivity observed with
this catalyst at all of the temperatures tested. The cellulose
ꢂ1
feed rate was kept constant at 15 gh , similar to the tests
above, and the oxygen flow rate in these tests was controlled
such that the C/O ratio was 0.5, 0.6 and 0.7. The reactor tem-
perature and C/S ratio for these runs were kept constant at
7508C and 1.0, respectively. The results, illustrated in Figure 6,
show that all the promoted Ni catalysts demonstrated a high
gasification efficiency of 88–100% at all the C/O ratios tested.
In general, the gasification efficiency decreased with the in-
creasing C/O ratio, which was expected because a high C/O
ratio means a low oxygen flow rate and, therefore, a low par-
tial oxidation. This result is consistent with previous observa-
tions in which higher C/O ratios in air-steam gasification led to
Figure 5. Effect of temperature on a) H
ed catalysts at a C/O ratio of 0.5 and C/S ratio of 1.0. ~: Ni, *: Pt-Ni, !: Ru-Ni,
: Re-Ni, ^: Rh-Ni.
2 2
/CO and b) CO/CO of the investigat-
[25]
&
a lower carbon conversion. Although a low C/O ratio may
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2731

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 6. Effect of C/O feed ratio and catalyst promoters on the selectivity of char (&), tar (&) and gas (&) based on carbon balance. The numbers on the
top of each bar represent gasification efficiency, which is the percentage of carbon in the gas and tar combined. Reaction conditions: cellulose flow
ꢂ1
rate=15 gh , 7508C, C/S=1.0 and residence time=50 ms. Here, tar is defined as water-soluble organics.
lead to high gasification temperatures (hot spots in the cata-
lyst bed), which is favourable for steam reforming, the higher
rate of oxidation reactions may also result in the oxidation of
CO to CO and H to H O. Therefore, low C/O ratios may de-
gas quality and the mole fractions of H and CO in the product
2
gas. In addition, the catalysts are less likely to be re-oxidised
under an oxygen deficit environment (high C/O), which can
improve the catalyst service life significantly.
2
2
2
grade the quality of synthesis by increasing the CO and H O
2
2
mole fraction. This is precisely what was observed in our ex-
Effect of the carbon-to-steam ratio in the feed
periments (Figure 7). A C/O ratio of 0.5 resulted in a CO mole
2
fraction in the dry synthesis gas in excess of 50%. If the C/O
The effect of the C/S ratio was tested on all the promoted cat-
alysts. The monometallic Ni catalyst was not considered for
these tests because of the high char selectivity observed with
this catalyst at all the temperatures tested in the previous sec-
ratio was increased to 0.6, the CO mole fraction decreased to
2
ꢁ
40%. However, a further increase of the C/O ratio resulted in
incomplete gasification in most cases, that is, the formation of
char. The Ru-Ni catalyst performed the best amongst all the
promoted catalysts. The char selectivity with the Ru-Ni catalyst
was zero at all the C/O ratios. Ru is known to exhibit a higher
activity for CꢂC bond cleavage than Re, Rh and Pt in that
ꢂ1
tion. The cellulose feed rate was kept constant at 15 gh , sim-
ilar to the tests above, and the steam flow rate in these tests
were controlled such that the C/S ratio was 1.0, 1.5 and 2.0.
The reactor temperature and C/O ratio for these runs were
kept constant at 7508C and 0.6, respectively. The results show
that all the promoted Ni catalysts showed a high gasification
efficiency of 82–100% at all the C/S ratios tested (Figure 9). In
general, the gasification efficiency reduced with the increasing
C/S ratio (i.e., a lower steam flow rate), which is expected be-
cause of the reduction in steam reforming and water–gas shift
activity at high C/S ratios. This result is consistent with the cat-
[
26]
order. Ru also shows a higher water–gas shift reaction activi-
[
27]
ty than Pt and Rh and it is nearly equal to that of Re. There-
fore, it is expected that Ru will perform well under gasification
conditions, which requires a combination of CꢂC bond cleav-
age and water–gas shift reaction. The water–gas shift activity
was high because of the high H /CO ratio and low CO/CO2
2
ratio observed for the promoted Ni catalysts (Figure 8).
In summary, a higher C/O ratio, especially C/O=0.6, was fa-
vourable as it improved the gasification efficiency, synthesis
[28]
[29]
alytic steam gasification and air-steam gasification of bio-
mass in previous studies. Although lower C/S ratios increase
Figure 7. Effect of C/O feed ratio and catalyst promoters on product gas composition. C
tions: cellulose flow rate=15 gh , 7508C, C/S=1.0 and residence time=50 ms.
2
compounds include ethane, ethylene and acetylene. Reaction condi-
ꢂ1
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2732

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
thermic. The required heat for steam reforming is provided by
partial oxidation and water–gas shift reactions. In this section
a high C/O ratio was used, therefore, high C/S ratios were not
favourable because the amounts of both the gasifying agents
(
O and steam) were low, which results in incomplete gasifica-
2
tion. The molar composition of the product gases are shown
in Figure 10, and it is clear that higher C/S ratios led to
a higher fraction of CO in the product gas because the oxida-
2
tion reactions dominated the gasification reaction if the steam
flow rate was reduced (higher C/S ratios). The fraction of CO2
increased from ꢁ40% to over 50% if the C/S ratio was in-
creased from 1.0 to 1.5. The H /CO and CO/CO ratios against
2
2
the changes in C/S ratio are shown in Figure 11. The H /CO
2
ratio recorded in this part of the study was the highest among
all the previous tests reported in the previous two sections. On
average the H /CO ratio ranged from 1.0 to 1.5, which is be-
2
lieved to be because a high C/S ratio leads to oxidation-domi-
nated gasification and conversion of CO into CO . This led to
2
a reduction in the CO/CO ratio as seen in Figure 11b.
2
Among the catalysts tested in this study, the Re-Ni and Ru-
Ni catalysts performed the best in terms of gasification efficien-
cy and hydrogen yield. As seen in the catalyst characterisation
section, the Re-Ni and Ru-Ni catalysts had a high metal surface
area, low reduction temperature and high CO uptake in CO-
TPD, which combined explains the superior performance of
these two catalysts in the gasification experiments.
Overall, the results from the catalyst characterisation and
RFV sections combined show that the Re-Ni, Rh-Ni and Ru-Ni
catalysts were the best catalysts among those tested in this
project. Among the conditions tested in this study, the best
conditions were determined to be a reaction temperature of
7
508C, C/O=0.6 and C/S=1.0. This conclusion is based on the
Figure 8. Effect of C/O on a) H
lysts at 7508C and a C/S ratio of 1.0. *: Pt-Ni, !: Ru-Ni, ^: Rh-Ni,
2
/CO and b) CO/CO
2
of the investigated cata-
: Re-Ni.
overall gasification efficiency, in particular the low char selectiv-
ity and the product gas composition, principally, the low CO2
mole fraction in the gas phase. These conditions resulted in
the sustained performance of the catalysts (>300 min) without
noticeable deactivation of the catalyst. The duration of the
&
the rate of steam reforming, they can reduce the temperature
of the catalyst bed because steam reforming is highly endo-
Figure 9. Effect of C/S feed ratio and catalyst promoters on the selectivity of char (&), tar (&) and gas (&) based on carbon balance. The numbers on the
top of each bar represent gasification efficiency, which is the percentage of carbon in the gas and tar combined. Reaction conditions: cellulose flow
ꢂ1
rate=15 gh , 7508C, C/O=0.6 and residence time=50 ms. Here, tar is defined as water-soluble organics.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2733

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 10. Effect of C/S feed ratio and catalyst promoters on product gas composition. C
ditions: cellulose flow rate=15 gh , 7508C, C/O=0.6 and residence time=50 ms.
2
compounds include ethane, ethylene and acetylene. Reaction con-
ꢂ1
Stability of the catalysts
The rapid deactivation of Ni catalysts because of coke forma-
tion is a well-known problem in gasification and hydrocarbon
[6c,8–9,30]
reforming reactions.
Commercial Ni catalysts used as
primary catalysts in biomass gasification are especially prone
to deactivation because of fouling, sintering and coke deposi-
[6c,8,31]
tion.
Methods to improve the activity and coke resistance
of supported Ni catalysts include doping with noble met-
[
17b,c]
[21]
[17b]
als,
als.
alkali metals, rare earth metals
or transition met-
[17c,21,24a]
In this project, the first strategy to promote the Ni
catalyst with noble metals was adopted. Coke deposition over
the spent catalysts used in these investigations was studied by
TEM, and the resulting images are shown in Figure 12. The
spent catalyst samples selected for imaging had a char selec-
tivity of 2–12%, which is believed to be in the range in which
the catalysts started to deactivate. Coke deposition of varying
degrees was found on all five catalysts. The highest coke depo-
sition was found on the monometallic Ni catalyst, which ex-
plains its rapid deactivation that led to a high char selectivity.
Modification of the Ni catalyst by the addition of noble metals
reduced the extent of coke deposition to a great extent. Re-Ni
and Rh-Ni showed the highest coke resistance among the cata-
lysts tested. This result concurs with previous reports in which
noble-metal-promoted Ni was used in biofuel steam reform-
[18]
[21]
ing and the partial oxidation of propane. Interestingly, the
monometallic catalyst, which showed the highest coke deposi-
tion, was used for the shortest time in the gasification run
(
180 min), whereas, Re-Ni and Rh-Ni, which showed negligible
coke deposition, were used in the gasification run for over
00 min. Therefore, the rate of coke deposition on the Ni cata-
3
lyst was several times higher than that on the Re- and Rh-pro-
moted catalysts. Carbon deposited on the catalysts can be cat-
egorised into three types: amorphous, filamentous and graphi-
Figure 11. Effect of C/S ratio on a) H
catalysts at 7508C and a C/O ratio of 0.6. ~: Ni, *: Pt-Ni, !: Ru-Ni,
: Rh-Ni.
2
/CO and b) CO/CO
2
of the investigated
&
: Re-Ni,
^
[24b]
tic.
Amorphous carbon is formed at the lowest temperature
range of Tꢄ5708C, followed by filamentous carbon at
5708C<T<10008C and finally graphitic carbon at the highest
temperature range of Tꢀ10008C. In our study, only filamen-
tous carbon deposits were found on the spent catalyst. This is
because the reaction temperatures used in our experiments
(700–7758C) were favourable for the formation of filamentous
carbon.
gasification study could not be extended because of health
and safety aspects. The gasification rig had to be operated
manually and monitored at all times, and it was considered
a potential hazard to leave the rig unattended overnight,
given the toxic and highly flammable nature of the product
gases.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2734

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
which had been used in gasifi-
cation runs for over 300 min,
are shown in Figure 13. Al-
though it was difficult to quan-
tify the sintering effect, it can
be said qualitatively that the
metal particle size of the spent
catalysts was slightly larger than
that of the fresh catalyst. How-
ever, the effect of sintering on
gasification is unclear because
the catalyst was stable for the
duration of the study. We could
not perform longer gasification
runs to determine the maxi-
mum duration for which the
catalyst can remain active.
In addition to testing the
coke deposition and metal sin-
tering during the gasification
runs, the stability of the alumina
support under hydrothermal
conditions was also studied.
Phase transformation from g-
Figure 12. TEM images of spent catalysts a) Ni, b) Pt-Ni, c) Ru-Ni, d) Rh-Ni and e) Re-Ni. The spent catalyst samples
selected for the imaging had a char selectivity of 2–12%. Scale bars=100 nm.
Catalyst deactivation may also be caused by the sintering of
metal crystallites to lead to the loss of active metal surface
alumina into d-alumina or q-alumina phases may take place at
high temperatures, which leads to the loss of support surface
[
22,25–26]
[28–29]
area.
High temperatures and the presence of steam, re-
area.
The loss of support surface area may lead to lower
ferred to jointly as hydrothermal conditions, may lead to the
catalytic activity by increasing the diffusional and film mass
transport resistances. The support surface area of the fresh
(calcined) catalysts was compared against spent catalysts in
the gasification runs. Results from the BET measurement of the
fresh and spent catalysts are listed in Table 4. The results indi-
cate clearly that the surface area of the spent catalyst had de-
creased, which was an indication of possible phase change.
TEM images (Figures 12 and 13) of the spent catalysts also
suggest that all five catalysts had gone through some phase
transformation. The TEM images of spent catalysts showed
that alumina particles had agglomerated, whereas the particles
in the fresh catalyst were much smaller and had a more
porous texture. Phase transformation of g- to d-alumina may
occur at temperatures as low as 7808C, and d-alumina may fur-
[
27]
sintering of Ni nanoparticles. To test the hydrothermal stabil-
ity of the catalysts, TEM images of the spent catalysts were re-
corded. The spent catalyst samples selected for the imaging
had a char selectivity of 2–12%. Images of the fresh (calcined)
Re-Ni and Rh-Ni catalysts compared to the spent catalysts,
[28–29]
ther transform into q-alumina at around 9508C.
To test the
hydrothermal stability of the commercial g-alumina support
used in this study, we heated the alumina samples to 750 and
Table 4. Comparison of SBET of fresh (calcined) and hydrothermally treated
catalysts.
Catalysts Fresh catalyst
Conditions
T [8C] C/O C/S
Spent catalyst Change in
2
ꢂ1
2
ꢂ1
SBET [m g
]
S
BET [m g
]
SBET [%]
Ni
91.99
77.18
95.74
77.92
74.16
74.16
750
750
775
750
750
750
0.5 1.0 25.42
0.5 1.0 31.25
0.5 1.0 34.78
0.6 1.5 40.64
0.6 1.5 31.83
0.7 1.0 28.56
ꢂ72.36
ꢂ59.51
ꢂ63.67
ꢂ47.84
ꢂ57.08
ꢂ61.49
Pt-Ni
Ru-Ni
Rh-Ni
Re-Ni
Re-Ni
[
[
a]
a]
Figure 13. TEM images of a) spent and b) fresh Rh-Ni, and c) spent and
d) fresh Re-Ni. Scale bars=100 nm.
[
a] Re-Ni was reduced. All other catalysts were calcined only.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2735

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
8
008C under air (calcination) and air+steam atmospheres and
tions, it can be said that the noble-metal-promoted Ni catalysts
were effective for RFV because of their superior metal surface
area, high steam reforming and water–gas shift activity and
good coke resistance. In particular, Re-Ni, Rh-Ni and Ru-Ni, in
that order, provided a much improved performance over the
monometallic Ni catalyst.
then analysed their morphology using powder XRD. The resul-
tant XRD patterns are shown in Figure 14. Pattern (a) corre-
sponds to the as-received commercial g-alumina sample. Pat-
Conclusion
Various reaction conditions were tested for reactive flash volati-
lisation, which include the effects of temperature, carbon-to-
oxygen (C/O) feed ratio, carbon-to-steam (C/S) feed ratio and
catalyst promoter. The highest gasification efficiency was ach-
ieved at 7508C, C/O=0.6 and C/S=1.0. It was observed that
Re-, Rh- and Ru-promoted Ni catalysts performed the best in
that order. This conclusion is based on the catalyst properties,
gasification efficiency and stability. Monometallic Ni was active
for reactive flash volatilisation; however, the catalyst suffered
rapid deactivation because of coke deposition, metal active-
site sintering and alumina support pore collapse because of
phase change. Coke deposition and sintering were reduced
significantly by using noble-metal promoters. The noble-metal
promoters increased the activity of the Ni catalysts by increas-
ing the metal surface area, reducibility and reducing the CO
desorption temperature. The noble-metal promoters also in-
creased coke resistance and reduced sintering. Therefore, the
promoted Ni catalysts were active for longer with little to no
deactivation.
Figure 14. Powder XRD patterns of g-alumina support: a) as received; cal-
cined at b) 7508C; c) 8008C; and treated in air+steam atmosphere at
d) 7508C and e) 8008C.
terns (b) and (c) correspond to the g-alumina support calcined
at 750 and 8008C, and patterns (d) and (e) correspond to the
g-alumina support treated under air+steam atmosphere at 750
and 8008C, respectively. XRD patterns (a) and (b) are nearly
identical, which indicates that calcination at 7508C does not
change the alumina phase. However, the increase of tempera-
ture and addition of steam has an effect on the alumina phase
change. The peaks at 2q=21.3 and 23.78, which correspond to
d-alumina, increase in intensity from patterns (b) to (e). If the
treatment temperature under air is increased from 750 to
Experimental Section
Catalyst preparation
Five Ni-based catalysts were developed using the impregnation
method. They were Ni/Al O , Pt-Ni/Al O , Ru-Ni/Al O , Rh-Ni/Al O
8008C, a small but measurable impact in the phase transforma-
2
3
2
3
2
3
2
3
tion of g- to d-alumina is observed. It is, however, under
air+steam atmosphere that the phase transformation is more
pronounced.
and Re-Ni/Al
2
O
3
. Nickel nitrate (Ni(NO
3
)
2
·6H O), alumina (Al O ) and
2 2 3
chemicals such as H PtCl , RuCl , RhCl and NH ReO obtained from
2
6
3
3
4
4
Sigma–Aldrich were used as the precursors for the catalyst synthe-
sis. First, the required amount of nickel nitrate was measured and
dissolved in distilled water. Then, the corresponding amount of
alumina and metal promoter precursor were added into the solu-
Therefore, from the results in Table 4 and Figure 14 it can be
concluded that the alumina support used in this study was not
completely stable during RFV. Although all the catalysts used
in this project were calcined at 6008C in air, except Re-Ni, the
reaction temperatures of 700–7758C in the presence of steam
caused the alumina pores to collapse, which led to a loss of
the BET surface area (SBET) and phase transformation of g- to d-
alumina. Although the phase transformation of g- to d-alumina
begins at ꢁ7808C, which is higher than all the RFV experi-
ments, the presence of steam in the reactor lowered the phase
transition temperature.
tion. The monometallic Ni/Al O3 catalyst had a Ni content of
2
11 wt%, and all bimetallic catalysts had Ni contents of 10 wt% and
a metal promoter content of 1 wt%. To ensure a homogeneous
mix of all precursors with alumina, the solution was heated up to
6
58C and maintained for 5 h under constant stirring. The solutions
were then dried overnight in a 1008C oven. Dry solid was recov-
ered and calcined in a muffle furnace at 6008C for 6 h under air at-
mosphere. Only the Re-Ni/Al O catalyst was not calcined at high
temperature because of the high volatility of rhenium(VI) and (VII)
oxides. The resultant calcined catalysts were reduced in situ before
2
3
ꢂ1
In summary, coke deposition, active metal sintering and alu-
mina phase transformation all contributed, in that order, to-
wards the deactivation of the catalysts used in this study. The
monometallic Ni catalyst had the worst performance because
of high coke deposition and Ni sintering. The addition of
a small amount of noble metal prevented coke deposition to
a great extent and reduced Ni sintering. If we combine the re-
sults presented in the catalyst characterisation and RFV sec-
the reaction studies under 60 mLmin of 5% H /N flow at 4008C
for 5 h.
2
2
Catalyst characterisation
Catalysts were characterised using techniques that include N2
physisorption, CO chemisorption, XRF spectroscopy, TEM, H -TPR
and CO-TPD.
2
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2736

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Nitrogen physisorption: The catalyst specific surface area, pore
partial pressure of gases with respect to time on stream. To report
the results, the partial pressures were converted into molar flows,
based on the volumetric flow rate used and the time on stream
was converted to temperature, based on the ramp rate of the fur-
nace.
size distribution and pore volume were measured using N physi-
2
sorption by using a Belsorp mini-II instrument. The surface area
was characterised using the BET method, whereas the pore size
distribution and pore volume were measured using the Barrett–
Joyner–Halenda (BJH) method on the adsorption–desorption iso-
therm curve. Catalyst (200 mg) was degassed under high vacuum
at 1208C for at least 5 h prior to the measurement.
CO-TPD: CO-TPD was performed to examine the interaction be-
tween Ni and the metal promoter and to measure the strength
and number of metal active sites available on the catalyst surface.
The same custom-built instrument described for the H -TPR study
was used in this investigation. In this study, fresh catalyst sample
was loaded into the quartz reactor and reduced in situ under
XRF spectroscopy: The elemental composition of Ni and alumina
of the prepared catalysts were determined by using an Ametek
Spectro iQ II XRF, and the promoter content was calculated based
on mass balance.
2
ꢂ1
6
0 mLmin of 5.22% H /N gas mixture at temperature of 4008C
2
2
ꢂ1
for 5 h followed by purging with 100 mLmin of He gas at 4008C
for 1 h. The catalyst was then cooled to RT under He gas. To dope
the catalyst surface with CO, 10% CO/He gas mixture was fed into
the reactor for 15 min at a flow rate of 100 mLmin followed by
purging with He gas at 100 mLmin for 2 h to remove any excess
CO chemisorption: The amount of active metal sites and the dis-
persion of the catalysts were measured using CO chemisorption by
using a Micromeritics ASAP 2020. Catalyst (200 mg) was loaded
into a flow-through quartz tube reactor in which an in situ reduc-
ꢂ1
ꢂ1
tion was performed. The sample was reduced with 5.22% H /N2
CO such that only a monolayer CO adsorbed was left on the sur-
face of the catalyst. Once the CO partial pressure, monitored by
using the RGA, was stable, the He flow was subsequently reduced
2
gas mixture at 4008C for 5 h and then cooled to RT under N flow.
2
The chemisorption measurement was started once the sample
reached 308C.
ꢂ1
to 50 mLmin , and the TPD data were recorded as the sample
ꢂ1
was heated from RT to 8008C at 108Cmin . Typical CO-TPD prod-
TPR: TPR measurements were performed to investigate the reduci-
bility of the fresh catalyst and examine the interaction between Ni,
metal promoter and the support. Measurements were performed
in a custom-built instrument that consisted of a vacuum compart-
ment fitted with an Agilent Technologies TPS Compact vacuum
pump and a Stanford Research Systems Residual Gas Analyser
ucts include CO and CO . It is believed that the formation of CO is
2
2
largely because the water–gas shift [Eq. (17)] and Boudouard
Eq. (18)] reactions occurred if the adsorption layer becomes
[
[9a]
mobile at high temperature. However, CO2 formation is largely
dominated by Equation (17). As a result, both CO and CO partial
2
pressures were monitored and reported in the CO-TPD analysis.
(
RGA) 300. A simplified schematic diagram of the setup is present-
ed in Figure 15. Catalyst (500 mg) was loaded into a custom-made
quartz reactor that was placed in a vertical tube furnace (Labec)
and heated from RT to 8008C at a heating rate of 108Cmin .
2
COþ2 OHAl2O2 ! 2 CO þH2
ð17Þ
ð18Þ
2
ꢂ1
During the heating process, a 5.22% H /N gas mixture was fed
2
2
2 CO Ð CO þC
2
ꢂ1
continuously into the reactor at a flow rate of 50 mLmin to
reduce the metal oxide to its pure metal state. The resultant gas
was introduced into the vacuum chamber through a capillary tube
and a leak valve. The RGA, which is a quadrupole mass spectrome-
ter, analysed the gases in the vacuum chamber and reported the
Powder XRD: Powder XRD measurements were performed by
using a Rigaku Miniflex powder diffractometer with monochroma-
tised CuK radiation (l=0.154 nm) at 40 kV and 15 mA.
a
Figure 15. Schematic diagram of the custom-made TPR-TPD-TPO instrument.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2737

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 16. Schematic diagram of the RFV reactor setup. MFC=Mass flow controller.
Catalytic activity evaluation
235 sccm, residence time 50 ms, C/O ratio 0.5, 0.6, 0.7, C/S ratio
.0, 1.5, 2.0.
1
Reactor setup: The catalytic activity was evaluated by using
The gas yield and composition results were computed by using
the trapezoid rule, which integrates the molar flow rate of each
gas species over total run time. Carbon balance was performed by
calculating the moles of carbon atoms in the gas, liquid and solid
products collected from the reactor. Carbon in the gas phase was
calculated by measuring the amount of carbonaceous molecules in
the gas phase using GC (described above). Carbon atoms in the
liquid phase (tar) were calculated by measuring the total organic
carbon by using a Shimadzu TOC-L series analyser. A liquid sample
(0.5 mL) was diluted with deionised water (200 mL) prior to the
analysis. The number of moles of carbon atoms in the solid prod-
ucts (char) was calculated by measuring elementary carbon by
using a PerkinElmer 2400 Series II CHNS/O system. Based on the
carbon balance, gas selectivity, tar and char are reported according
to Equations (19)–(21):
a bench-scale reactor setup that consisted of a 25 mm OD and
7
00 mm long quartz tube reactor, a K-Tron twin screw powder
feeder (K-MV-KT20), an Alltech HPLC pump (model 426), three Tele-
dyne Hasting mass flow controllers for nitrogen and oxygen,
a Labec vertical split tube furnace, a Brooks Instrument DLI Series
evaporator and a custom-built gas–liquid separator. A schematic
diagram of the reactor setup is shown in Figure 16.
Catalyst (1 g) was loaded into the quartz reactor prior to assem-
bling it in the furnace. The catalyst was held in the centre by
a porous quartz disk. Before the catalytic tests, in situ reduction
was performed with H /N gas mixture at 4008C for at least 4 h.
2
2
RFV was performed at 700, 750 and 7758C with various C/O and C/
ꢂ1
S ratios. During these experiments, 15 gh of cellulose was gravity
fed by the twin screw powder feeder and a mixture of O , N and
2
2
steam was fed into the reactor. The amounts of steam and O fed
2
into the reactor were controlled by the HPLC pump and the mass
moles of carbon atoms in the gas phase
moles of carbon atoms in the products
Sgas ½%ꢅ ¼
Star ½%ꢅ ¼
Schar ½%ꢅ ¼
ꢆ 100
ð19Þ
ð20Þ
ð21Þ
flow controller, respectively. N was used as a control parameter to
2
vary the reactor space velocity. Product gases from the reactor
were fed into a custom-built gas–liquid separator in which the gas-
eous products were sampled periodically for composition analysis.
Gases were analysed by using a Shimadzu gas chromatograph GC-
moles of carbon atoms in the liquid phase
moles of carbon atoms in the products
ꢆ 100
ꢆ 100
2
8
014 equipped with a molecular sieve 5 ꢁ column (60/80 mesh, 1/
inch diameter, 6 feet in length) using a thermal conductivity de-
moles of carbon atoms in the solid phase
moles of carbon atoms in the products
tector and a flame ionisation detector. The volumetric flow rate of
the gaseous product was determined after each GC analysis by
using a 50 mL bubble flow meter.
Biomass feedstock: High-purity microcrystalline cellulose, Cellets
200 (Pharmatrans Sanaq AG) was used in our studies. The particle
size of Cellets 200 was 200–355 mm.
Detailed operating parameters: temperature 700, 750, 7758C, cellu-
ꢂ1
lose type Cellets 200, cellulose feed rate 15 gh , N2 flow rate
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2738

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[
[
13] a) D. Sutton, B. Kelleher, J. R. H. Ross, Fuel Process. Technol. 2001, 73,
Acknowledgements
1
55–173; b) Y. Wang, C. M. Kinoshita, Solar Energy 1993, 51, 19–25.
14] a) A. Kumar, K. Eskridge, D. D. Jones, M. A. Hanna, Bioresour. Technol.
009, 100, 2062–2068; b) A. Haryanto, S. D. Fernando, L. O. Pordesimo,
The authors are grateful for the financial support from the Rural
Industries Research and Development Corporation (RIRDC) project
grant PRJ-004758 and the Department of Chemical Engineering,
Monash University.
2
S. Adhikari, Biomass Bioenergy 2009, 33, 882–889.
[
[
[
15] S. Wang, Appl. Catal. B 1998, 16, 269–277.
16] J. M. Rynkowski, T. Paryjczak, M. Lenik, Appl. Catal. A 1993, 106, 73–82.
17] a) A. Tanksale, J. N. Beltramini, G. M. Lu, Renewable Sustainable Energy
Rev. 2010, 14, 166–182; b) M. El Doukkali, A. Iriondo, P. L. Arias, J. Re-
quies, I. Gandarꢄas, L. Jalowiecki-Duhamel, F. Dumeignil, Appl. Catal. B
Keywords: biomass · nickel · platinum · rhenium · supported
catalysts
2
012, 125, 516–529; c) Z. Hou, P. Chen, H. Fang, X. Zheng, T. Yashima,
Int. J. Hydrogen Energy 2006, 31, 555–561.
18] L. P. R. Profeti, E. A. Ticianelli, E. M. Assaf, Int. J. Hydrogen Energy 2009,
34, 5049–5060.
[
[1] M. Ni, D. Y. C. Leung, M. K. H. Leung, K. Sumathy, Fuel Process. Technol.
2
006, 87, 461–472.
[19] D. O. Hayward, B. M. W. Trapnell, Chemisorption, 2ednd edButterworths,
London, 1964.
[
[
2] A. Kumar, D. D. Jones, M. A. Hanna, Energies 2009, 2, 556–581.
3] L. Devi, K. J. Ptasinski, F. J. J. G. Janssen, Biomass Bioenergy 2003, 24,
1
[20] S. D. Jackson, B. M. Glanville, J. Willis, G. D. McLellan, G. Webb, R. B.
Moyes, S. Simpson, P. B. Wells, R. Whyman, J. Catal. 1993, 139, 207–220.
[21] M. Shiraga, D. Li, I. Atake, T. Shishido, Y. Oumi, T. Sano, K. Takehira, Appl.
Catal. A 2007, 318, 143–154.
25–140.
4] a) J. J. Hernꢂndez, G. Aranda-Almansa, A. Bula, Fuel Process. Technol.
010, 91, 681–692; b) K. Qin, P. A. Jensen, W. Lin, A. D. Jensen, Energy
[
2
Fuels 2012, 26, 5992–6002.
5] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098.
6] a) S. Rapagnꢃ, N. Jand, A. Kiennemann, P. U. Foscolo, Biomass Bioenergy
[22] J. Sehested, J. A. P. Gelten, S. Helveg, Appl. Catal. A 2006, 309, 237–246.
[23] A. Shrotri, L. K. Lambert, A. Tanksale, J. Beltramini, Green Chem. 2013,
15, 2761–2768.
[
[
2
000, 19, 187–197; b) K. Tomishige, M. Asadullah, K. Kunimori, Catal.
[24] a) K. Omata, Y. Endo, H. Ishii, A. Masuda, M. Yamada, Appl. Catal. A
2008, 351, 54–58; b) M. Asadullah, T. Miyazawa, S. I. Ito, K. Kunimori, S.
Koyama, K. Tomishige, Biomass Bioenergy 2004, 26, 269–279.
[25] a) L. Wei, S. Xu, L. Zhang, C. Liu, H. Zhu, S. Liu, Int. J. Hydrogen Energy
2007, 32, 24–31; b) P. M. Lv, Z. H. Xiong, J. Chang, C. Z. Wu, Y. Chen,
J. X. Zhu, Bioresour. Technol. 2004, 95, 95–101.
[26] J. H. Sinfelt, Adv. Catal. 1973, 23, 91–119.
[27] D. C. Grenoble, M. M. Estadt, D. F. Ollis, J. Catal. 1981, 67, 90–102.
[28] B. Moghtaderi, Fuel 2007, 86, 2422–2430.
Today 2004, 89, 389–403; c) E. G. Baker, L. K. Mudge, M. D. Brown, Ind.
Eng. Chem. Res. 1987, 26, 1335–1339.
[
7] a) R. J. Farrauto, C. Bartholomew, Introduction to Industrial Catalytic Pro-
cesses, Chapman and Hall, London, 1997; b) J. R. Rostrup-Nielsen, Steam
Reforming Catalysts: An Investigation of Catalysts for Tubular Steam Re-
forming of Hydrocarbons, Taknisk Forlag, Copenhagen, 1975.
8] X. T. Li, J. R. Grace, C. J. Lim, A. P. Watkinson, H. P. Chen, J. R. Kim, Bio-
mass Bioenergy 2004, 26, 171–193.
[
[
9] a) A. Tanksale, J. N. Beltramini, J. A. Dumesic, G. Q. Lu, J. Catal. 2008,
[29] M. Campoy, A. Gꢅmez-Barea, F. B. Vidal, P. Ollero, Fuel Process. Technol.
2009, 90, 677–685.
258, 366–377; b) A. Tanksale, Y. Wong, J. N. Beltramini, G. Q. Lu, Int. J.
Hydrogen Energy 2007, 32, 717–724; c) A. Tanksale, C. H. Zhou, J. N. Bel-
tramini, G. Q. Lu, J. Inclusion Phenom. Macrocyclic Chem. 2009, 65, 83–
[30] R. T. K. Baker, M. A. Barber, P. S. Harris, F. S. Feates, R. J. Waite, J. Catal.
1972, 26, 51–62.
8
8.
10] D. Y. C. Leung, X. L. Yin, C. Z. Wu, Renewable Sustainable Energy Rev.
004, 8, 565–580.
11] a) J. L. Colby, P. J. Dauenhauer, L. D. Schmidt, Green Chem. 2008, 10,
73–783; b) J. R. Salge, Science 2006, 314, 801–804.
[31] A. Tanksale, J. N. Beltramini, G. Q. Lu, Devel. Chem. Eng. Mineral Process.
2006, 14, 9–18.
[
[
[
2
7
12] a) Y. N. Shastri, L. F. Rodriguez, A. C. Hansen, K. C. Ting, Biofuels, Bioprod.
Bioref. 2012, 6, 21–31; b) P. J. Dauenhauer, J. Colby, L. D. Schmidt, Salt
Lake City, 2007.
Received: June 11, 2014
Published online on August 21, 2014
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 2727 – 2739 2739