Screening of nickel catalysts for selective hydrogen production using supercritical water gasification of glucose

Article abstract of DOI:10.1039/c2gc16378k

We report the activity and the selectivity of several heterogeneous nickel catalysts for the supercritical water gasification (SCWG) of biomass. The effects of catalyst support on the carbon conversion and hydrogen selectivity were demonstrated using 44 d

Full text of DOI:10.1039/c2gc16378k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2012, 14, 1766
www.rsc.org/greenchem
PAPER
Screening of nickel catalysts for selective hydrogen production using
supercritical water gasification of glucose
Pooya Azadi, Elie Afif, Faraz Azadi and Ramin Farnood*
Received 2nd November 2011, Accepted 29th March 2012
DOI: 10.1039/c2gc16378k
We report the activity and the selectivity of several heterogeneous nickel catalysts for the supercritical
water gasification (SCWG) of biomass. The effects of catalyst support on the carbon conversion and
hydrogen selectivity were demonstrated using 44 different materials, covering a wide range of chemical
and physical properties. At 5% nickel loading, α-Al O , carbon nanotubes (CNTs), and MgO supports
2
3
resulted in high carbon conversions, while SiO , Y O , hydrotalcite, yttria-stabilized zirconia (YSZ), and
2
2 3
TiO showed modest activities. Utilization of different γ-Al O supports resulted in a wide range of
2
2 3
catalytic activities from almost inactive to highly active. Other catalyst carriers such as zeolites, molecular
sieves, CeO , and ZrO had insignificant activity under the conditions tested (i.e., 380 °C, 2 wt% feed).
2
2
Aside from the catalytic activity, the stable metal oxide supports under the experimental conditions of this
work, as identified by XRD, were α-Al O , boehmite, YSZ, and TiO . Given the high hydrogen yield and
2
3
2
carbon conversion as well as its superior stability in supercritical water, α-Al O was chosen for a more
2
3
elaborate investigation. It was found that when using the same amount of nickel, the methane yield
significantly decreased by increasing the nickel to support ratio whereas the carbon conversion was only
slightly affected. At a given nickel to support ratio, a threefold increase in methane yield was observed by
increasing the temperature from 350 to 410 °C. The catalyst activation conditions (e.g., calcination and
reduction) had a small impact on its catalytic performance. The catalyst activity increased with the
addition of alkali promoters (i.e., K, Na, Cs) and decreased with the addition of tin. The highest catalytic
activity was obtained with the addition of 0.5% potassium. In summary, nickel loading and alkali
promoters improved the hydrogen selectivity and the carbon conversion of the Ni/α-Al O catalyst.
2
3
gasification processes (e.g., >800 °C). Considering the amount
and the high specific heat of water as well as the high costs
associated with the construction and operation of high pressure/
high temperature equipment, it is highly desirable to operate the
supercritical water gasifiers at milder reaction conditions. To
achieve this goal, solid catalysts are employed to enhance the
gasification rates and obtain a high conversion at temperatures
Introduction
Catalytic hydrogen production is among the most important
industrial processes for the production of high quality fuels and
foods. Hydrogen is used in ammonia and methanol syntheses,
fuel cells, hydrotreating and hydrogenation processes. Currently,
hydrogen is mainly produced by steam reforming of natural gas
and naphtha due to their low costs. However, it is anticipated
that utilization of biomass as a feedstock for hydrogen pro-
duction will gain a more significant share in the near future due
to the increasing concerns over the adverse environmental
impacts of the fossil fuels. Since lignocellulosic materials are the
most abundant biomass species, it is highly desirable to produce
renewable hydrogen from such resources.
Supercritical water gasification (SCWG) is a promising tech-
nology for the conversion of wet biomass into a gas mixture
composed of hydrogen, methane, carbon dioxide and carbon
monoxide without the need for drying the feedstock. The gasifi-
cation of organics in supercritical water typically proceeds at
lower temperatures (e.g., <600 °C) compared to the conventional
1
–14
below 500 °C and more preferably, below 400 °C.
Nickel
and ruthenium have been found to be active for catalyzing the
SCWG reactions which require C–C bond cleavage at such low
temperatures. For a review of the recent advances in catalytic
SCWG please refer to ref. 15 and 16.
The reaction pathway for the nickel-catalyzed supercritical
water gasification of cellulosic biomass is depicted in Fig. 1.
According to this figure, the polysaccharides are initially hydro-
lyzed to sugar monomers and small oligomers, followed by
chemical transformation to other intermediates such as alcohols
and acids. Decomposition of these chemical intermediates on the
surface of a metal catalyst through C–C bond cleavage results in
the formation of syngas. The syngas is further upgraded to
hydrogen and carbon dioxide by water-gas shift reaction on the
metal. On the other hand, if the catalyst has a high activity for
the cleavage of C–O bonds, the produced hydrogen will be used
Department of Chemical Engineering and Applied Chemistry, University
of Toronto, M5S 3E5, Canada. E-mail: ramin.farnood@utoronto.ca
1766 | Green Chem., 2012, 14, 1766–1777
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 1 Reaction pathways for production of hydrogen by reactions of oxygenated carbohydrates with water adapted from ref. 17 (*represents a
surface metal site).
for hydrogenation of the adsorbed species and results in the for-
mation of methane and other alkanes. From a thermodynamic
standpoint, reforming of the oxygenated compounds (e.g.,
biomass) at lower temperatures and higher pressures favors
greater alkane yields and thus, results in poorer hydrogen selec-
tivity. Consequently, water near its critical point (374 °C and
catalyst performance in terms of the activity and selectivity. The
catalyst supports included various materials covering a wide
range of chemical and physical characteristics. Among the cata-
lysts tested, the Ni/α-alumina catalyst was selected for a more
detailed investigation. The effects of the nickel to carrier ratio,
reaction temperature, catalyst activation conditions and promo-
ters were studied.
2
21 bar) is a highly effective medium for the formation of
alkanes, especially methane. Fortunately, methane formation in
SCWG of biomass does not have a pyrolytic origin and it is
almost solely formed through the catalytic cleavage of C–O
bonds in the presence of hydrogen. Due to this reason, at low
conversions where the H , CO and CO partial pressures are still
Experimental
2
2
The catalysts used in this study were prepared by impregnation
of the supports with a nickel nitrate hexahydrate (Sigma-Aldrich,
Canada) precursor using the incipient wetness method. The
characteristics of these supports are given in Table 1. A-1 to A-4,
A-12, Ca-1, HT-1, Mg-1, Mg-2, MS-1, MS-2, Si-1, Si-2, Yt-1,
YZ-1, Ze-4, and Zr-1 were obtained from Sigma Aldrich,
Canada. A-5 to A-11, SA-1 to SA-4, Si-3, Si-4, Ze-2, and Ze-3
were obtained from Grace Davision, USA. Ca-2 was obtained
from Cheap Tubes Inc. (Brattleboro, USA), Ti-1 to Ti-4 were
obtained from Degussa AG (Germany), and Ze-1 was obtained
from United Catalyst Inc. (USA). In some experiments, as indi-
cated by “*” in Tables 3 and 4, the support pellets were crushed
to fine powder prior to the impregnation. Following the impreg-
nation of the support materials with the desired amount of nickel
nitrate solution, catalysts were dried at 110 °C and calcined at
350 °C for 2 h (except for Ca-1 and Ca-2). Then, catalysts were
low, the methane yield is negligible. However, when the reaction
proceeds further, the partial pressures of these gases increase and
as a result of that, the methanation rate increases. In spite of the
high activity of nickel for C–C bond scission and the water-gas
shift reaction, the undesirable high activity of this metal for C–O
bond cleavage makes hydrogen production in SCW challenging.
Dumesic et al. showed that the addition of tin to RANEY nickel
can considerably suppress the ability of the catalyst for cleaving
the C–O bonds while retaining its original activity for the
1
8–21
cleavage of the C–C bonds.
Although there is a large body of literature on the catalytic
1
5
supercritical water gasification of biomass, little is known
about the effects of catalyst support and structure on its catalytic
performance and there is a need for a systematic study to address
this issue. Hence, in this study, we systematically examined the
performance of several supported nickel catalysts for SCWG of
glucose in terms of carbon conversion and hydrogen selectivity.
The reaction conditions (i.e., 380 °C, 15 min, 2 wt% feed, 1 g
catalyst) resulted in a wide range of carbon conversions and
hydrogen selectivities, allowing for a fair distinction between the
reduced in flowing hydrogen (50 ml min− STP, 30% H , 70%
1
2
N ) for 2 h. The appropriate reduction temperature for each cata-
2
lyst was identified in preliminary experiments and it ranged from
400 to 950 °C, depending on the severity of interaction between
nickel oxide and the catalyst carrier. The desired amount of
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1766–1777 | 1767

View Article Online
Table 1 Characteristics of the supports used in this study
BET [m g−1]
2
Particle size [μm]
Pore vol. [cc g−1
]
Comment
Entry
Support
a
A-1
A-2
A-3
A-6
A-7
A-8
A-9
γ-Al
γ-Al
γ-Al
γ-Al
γ-Al
γ-Al
γ-Al
2
2
2
2
2
2
2
O
O
O
O
O
O
O
3
3
3
3
3
3
3
200
200
200
147
225
170
210
200
100
100
100
4100
4100
15
1900
78
70
118
10
35
1
110
15
6500
20
5
0.02
1
40
Acid, pH 5
Basic, pH 9
Neutral, pH 7
Spheres
Spheres
Powder
Minispheres
4% Lanthana
Microspheres
Corundum
a
a
a
1.2
1.1
0.7
1.1
0.8
0.5
0.12
a
a
a
a
A-10
A-11
A-4
A-12
A-13
A-14
A-15
A-5
Ca-1
Ca-2
Ce-1
Ce-2
HT-1
Mg-1
Mg-2
MS-1
MS-2
Si-1
Si-2
Si-3
Si-4
SA-1
SA-2
SA-3
SA-4
SA-5
Ti-1
La
2
O
3
-γ-Al
2
O
3
a
γ-Al
2
O
3
165
8.2
0.6
α-Al
α-Al
α-Al
α-Al
α-Al
AlO(OH)
Activated carbon
Carbon nanotube
CeO
CeO
Hydrotalcite
MgO
2
O
2
O
2
O
2
O
2
O
3
3
3
3
3
a
a
5.5
Powder
Powder
Crystalline
Pseudo-boehmite
a
10
7.5
312
600
495
3
39
12
14
0.05
0.9
a
a
1.1
Multiwalled, OD < 8 nm
2
2
Nano powder
Synthetic
0.06
0.13
MgO
0.5
338
800
70
4
Molecular sieve
Aluminosilicate MCM41
Organophilic
1.0
3% Al, 5 nm pore
Nano powder
Pore diameter 6 nm
Large pore size
Granules
a
SiO
Silica gel
2
500
0.01
13
50
2000
74
59
74
73
78
357
0.7
1.0
0.4
1.1
0.6
1.1
0.8
0.8
0.9
a
SiO
SiO
2
40
a
2
687
424
327
364
573
552
49
57
50
50
7
144
298
414
760
748
7
a
Silica-alumina
Silica-alumina
Silica-alumina
Silica-alumina
Silica-alumina
2 3
13% Al O powder
a
SLA pyridine catalyst
a
13% Al
13% Al
25% Al
7710
2
O
2
O
2
O
3
3
3
powder
powder
powder
a
a
TiO
TiO
TiO
TiO
2
2
2
2
Ti-2
Ti-3
Ti-4
YT-1
YZ-1
Ze-1
Ze-2
Ze-3
Ze-4
ZR-1
PF2
1600
0.4
7711, 75% Anatase
PF25
Y
2
O
3
6
Yttria-stabilized zirconia
Zeolite pentasil
Zeolite H-ZSM5
Zeolite USY
0.1
2500
14
5
YSZ, 8% yttria
Extrudate
a
0.2
0.3
a
Zeolite Y
ZrO
2
5
a
Reported by the manufacturer.
potassium nitrate, sodium nitrate, cesium carbonate and tri-n-
butyltin acetate (Sigma Aldrich) were dissolved in water and
added together with the nickel nitrate to synthesize the promoted
catalysts.
A schematic of the experimental setup is shown in Fig. 2. A
5
0 mL stainless steel batch reactor was used in this study to
evaluate the activity of the supported nickel catalysts for super-
critical water gasification of biomass. The reactor was used
several times prior to the experiments to ensure a minimal cataly-
tic effect of the walls. The heating medium was a molten salt
bath containing sodium nitrate, potassium nitrate, and sodium
nitrite by which an initial heat-up rate of greater than 300 °C
−1
min was achieved. Using the explained reaction system, 90%
of the temperature change occurred within the first 100 s of the
experiments and the temperature reached its final value in about
Fig. 2 Schematic diagram of the experimental set-up: (1) molten salt
bath, (2) reactor, (3) electrical heater, (4) thermocouple, (5) PID tempera-
ture controller, (6) first valve, (7) low-pressure gauge, (8) second valve.
3
min. One should refer to ref. 7 for typical variation of the
reactor temperature and pressure using the salt bath systems and
1768 | Green Chem., 2012, 14, 1766–1777
This journal is © The Royal Society of Chemistry 2012

View Article Online
to ref. 4 for the heat transfer issues from molten salts to SCW
reactors. Generally, the pressure inside the reactor closely
matched with the corresponding water vapor pressure at any
given time during the heat-up. The temperature of the salt bath
was measured using a Type K thermocouple (Omega Engineer-
ing, Canada) and was maintained at the desired value using a
temperature controller (Hanyoung Electronic Co. Ltd., Korea). In
all experiments, 0.2 g of glucose (Sigma-Aldrich, Canada) was
added to 9.8 g of deionized water to obtain a 2 wt% solution.
This solution along with the desired amount of catalyst was then
introduced into the reactor and immersed in the molten salt bath.
After 15 min, the reactor was taken out from the bath and
immediately quenched in a cold water bath. Then, the pressure
in the reactor was determined using a digital pressure gauge
Results and discussion
In the first part of this section, the calculated equilibrium gas
composition is given. Then, the effects of the supports on the
catalytic activity are discussed. In addition to that, the effects of
the nickel to support ratio, the reaction temperature, the catalyst
activation conditions and the addition of promoters have been
studied for one of the promising alumina catalysts found in the
course of catalyst screening.
Thermodynamic considerations
Table 2 shows the equilibrium gas composition calculated at
350, 380, and 410 °C (and at their corresponding pressures). It is
known that the hydrogen mole fraction at equilibrium is posi-
tively correlated with temperature and inversely correlated with
the pressure. Since the pressure inside a closed vessel (e.g., batch
reactor) dramatically changes with temperature near the critical
point of water, the hydrogen mole fraction at equilibrium
decreases with increasing temperature under the conditions
studied in this work. However, the variation of the gas yields
with temperature within the investigated temperature range was
rather small. The calculation of the gas composition at equili-
brium revealed that only about 6 mmol of hydrogen would be
formed per gram of glucose, and a larger fraction of the hydro-
gen will be consumed to hydrogenate carbon dioxide. The corre-
sponding equilibrium hydrogen selectivity and hydrogen
gasification ratio (HGR) at the conditions relevant to this work
was found to be approximately 0.2 and 108%, respectively.
(Cecomp Electronics, USA) and was used to calculate the gas
yield using the ideal gas law. The gaseous product was then col-
lected in a gas bag for composition analysis. In all experiments,
the amount of nickel metal was set at 50 mg regardless of the
catalyst formulation. In the support screening section, 1 g of
5
wt% Ni catalyst was used to gasify the feed at 380 °C. The
corresponding pressure in the reactor at 380 °C was estimated to
be approximately 230 bar. A gas chromatograph (Hewlett-
Packard 5890 series) equipped with a thermal conductivity
detector and argon as the carrier gas was used to determine the
product gas composition. The BET surface areas of the catalyst
supports were obtained using Quantachrome Autosorb catalyst
characterization system (Quantachrome, USA). The equilibrium
gas compositions were calculated using Aspen Plus software
(
AspenTech, Burlington, USA).
The results are reported in terms of gas yields (mmol g−1
glucose), carbon gasification ratio (CGR), hydrogen gasification
ratio (HGR) and hydrogen selectivity (HS). The CGR and HGR
are defined as the ratio of carbon and hydrogen in the gas pro-
ducts to the carbon and hydrogen in the initial feed, respectively.
We note that the maximum amount of hydrogen that can be gen-
erated from glucose is 66.6 mmol g that is twice the amount of
hydrogen initially presented in the glucose molecules; therefore,
the HGR as defined here can reach a maximum value of 200%.
Alumina-supported catalysts
Table 3 demonstrates the gas yields and the carbon conversion of
glucose using the alumina-supported catalysts at 380 °C. For
comparison, results from the SCWG of glucose under the cata-
lyst-free conditions are also given in Table 3. The CGR obtained
from these Ni/Al O catalysts ranged from 21% (the same as cat-
−
1
2
3
alyst-free) to 84%, implying that the chemistry of the support
material (Al O in this case) was not solely responsible for cata-
Hydrogen selectivity is defined as the number of moles of H to
2
2 3
the number of moles of hydrogen in methane. The CGR, HGR
and hydrogen selectivity have been calculated throughout this
lyst activity, and other factors significantly influenced the cataly-
tic performance as well. It should be noted that the variation in
the catalytic activity among the Ni/α-Al O was less pronounced
study as follows:
2
3
YCO þ YCH þ YCO
2
4
CGR ¼
ꢀ 100
(typically CGR > 70%) compared to that of Ni/γ-Al
O (i.e.,
2 3
mmol C=g feed
2
1% < CGR < 79%). Fig. 3 shows the variation of hydrogen
yield and selectivity with carbon conversion using alumina-sup-
ported catalysts. It was observed that up to carbon conversions
of about 60–70%, the hydrogen yield increased, after which it
levelled off. The initial increase in hydrogen yield, however,
appeared to be due to the increase in conversion, such that the
hydrogen selectivity remained constant at low conversions. At
higher conversions (i.e., CGR > 60%), hydrogen selectivity
YH þ 2YCH
2
4
HGR ¼
ꢀ 100
mmol H2=g feed
YH2
H2 selectivity ¼
2
ꢀ YCH4
Table 2 The equilibrium gas composition calculated based on SCWG of 2 wt% glucose and water density of 200 kg m− using Aspen Plus software
3
[mmol g−
1
]
CO [mmol g−1
]
CH
4
[mmol g−1
]
CO
[mmol g−1
]
Hydrogen selectivity
HGR [%]
T [°C]
P [bar]
H
2
2
3
3
4
50
80
10
165
230
320
6.7
5.5
4.7
0.005
0.008
0.012
14.9
15.2
15.4
18.4
18.1
17.9
0.22
0.18
0.15
110
108
107
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1766–1777 | 1769

View Article Online
showed a clear declining trend with CGR. This observation can
be justified considering the partial pressures of the hydrogen and
carbon dioxide (and carbon monoxide to a lesser extent). At low
carbon conversions, the partial pressure of the hydrogen and
carbon dioxide are low and hence, the rates of the formation of
these two compounds (from the remaining feed and the inter-
mediates) are much higher than their consumption rate by metha-
nation reactions. At moderate conversions, hydrogen formation
and consumption rates are comparable and therefore the hydro-
gen yield remains almost constant. At higher conversions, hydro-
gen is formed at a lower rate due to the progressive consumption
of the feed and intermediates, while the rate of hydrogen con-
sumption is significantly higher due to the higher partial press-
ures of hydrogen and carbon dioxide. As a result, at high carbon
conversions, the net rate of change in hydrogen yield is expected
to decrease. It is worthwhile to mention that the amount of
hydrogen at the highest carbon conversion was almost threefold
higher than the equilibrium hydrogen yield (see Table 2), indi-
cating that the gas phase composition was far from equilibrium.
Given that all catalysts in Table 3 (except A-15) have the same
chemical formula and were prepared using the same procedure,
the large differences among the performances of these catalysts
could be attributed to one or more of the following reasons: (i)
nickel dispersion, (ii) severity of the nickel–support interaction,
(iii) support stability, (iv) mass transfer issues, and (v) the sup-
port’s acidity. Here, we discuss the possibility of how and to
what extent each of these parameters could affect the catalytic
performance in SCWG.
(
i) Nickel dispersion. As a general rule, the activity of a cata-
lyst is proportional to the number of the available active sites. At
a given metal loading, the number of active sites in a catalyst is a
function of the metal dispersion, which is defined as the fraction
of the total atoms of the active element that is exposed at the
surface. The dispersion is, in turn, inversely proportional to the
crystallite size. In the case of commercial Ni/Al O catalysts, the
Fig. 3 Hydrogen yield and selectivity vs. carbon conversion using 5%
Ni/alumina catalysts, α-Al (■), γ-Al (◯), La –γ-Al (▲)
380 °C, 15 min, 2 wt% glucose, 1 g catalyst).
2
O
3
2
O
3
2
O
3
2 3
O
(
2
3
metal dispersion is typically below 15%, corresponding to an
Table 3 Performance of alumina-supported catalysts (380 °C, 15 min, 2 wt% glucose, 1g catalyst)
[mmol g−
1
]
CO [mmol g−1
]
CH
[mmol g−1
]
CO
6.5
[mmol g−1
]
CGR [%]
HGR [%]
Support
Ni %
H
2
4
2
No catalyst
A-1
A-2
—
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
2.7
1.0
0.3
0.3
0.6
0.5
0.2
0.5
0.4
1.1
0.5
0.4
0.5
0.5
0.3
0.5
0.4
0.4
0.2
1.3
0.1
8.1
5.0
1.6
4.1
4.9
4.1
5.0
0.0
3.6
4.8
2.1
2.3
4.6
2.9
7.4
9.5
9.9
0.0
22
79
72
39
65
62
65
68
21
63
67
45
48
67
51
73
84
84
23
9
20.5
22.3
8.8
20.8
16.2
19.0
18.6
2.5
19.5
18.8
11.6
14.5
24.3
16.8
17.9
18.3
17.6
1.4
17.9
18.6
10.9
17.1
15.7
17.1
17.3
6.1
17.1
17.3
12.4
13.1
17.4
13.8
16.8
18.0
17.8
6.4
110
97
36
87
78
81
85
7
80
85
47
57
101
68
98
112
112
4
A-3
A-6
A-6
a
A-7
A-7
a
A-8
A-9
A-9
a
A-10
A-11
A-4
A-12
A-13
A-14
A-15
A-5
a
Catalyst pellets were crushed to fine powders before impregnation.
1770 | Green Chem., 2012, 14, 1766–1777
This journal is © The Royal Society of Chemistry 2012

View Article Online
average crystallite size of about 6.5 nm (and larger). The main
factors that govern the dispersion of Ni on alumina include the
support surface area, the metal loading, and the activation con-
ditions. Since the metal loading and the activation conditions
were identical for all catalysts shown in Table 3, the support
surface area was the only parameter which could possibly
change the metal dispersion. Using the Scherrer equation, it was
revealed that the size of NiO (before reduction, XRD not shown
here) and Ni (XRD shown in Fig. 5) on α-Al O were approxi-
conversion (CGR), no correlation between hydrogen yield and
selectivity and the support surface area was observed.
(
ii) Nickel–support interaction. The physicochemical prop-
erties of Ni/alumina are affected by the thermal treatment
process prior and after the metal deposition. The amorphous
transition alumina phases (i.e., γ, δ, θ) undergo phase transform-
ation to form crystalline α phase upon calcination at high temp-
eratures (e.g., 1300 °C). This phase change causes a significant
dehydroxylation, loss of surface area and pore volume and an
increase in the average pore diameter. Three distinguishable
species of nickel oxide can present upon calcination of a nickel
2
3
mately 20 nm and 32 nm, respectively. These crystallite sizes
correspond to 5% and 3% dispersion, respectively. These values
are consistent with the previously reported data for the dispersion
2
2
23
of nickel on α-Al O at low metal loadings. Therefore, consid-
salt deposited on alumina: bulk NiO (reducible at ∼400 °C),
nickel oxide with strong interaction with support (reducible at
∼400–800 °C), and surface and bulk nickel aluminate spinel
2
3
ering the metal dispersion in α-Al O and γ-Al O -supported
2
3
2 3
catalysts, no clear relationship was observed between the metal
dispersion and the activity of the catalyst for the gasification of
glucose in SCW. However, Ni dispersion effects might have
been masked by other factors (discussed below) that had more
pronounced impacts on the activity of the alumina-supported
catalysts. Fig. 4 demonstrates the hydrogen yield and selectivity
as a function of support surface area. Similar to the carbon
(
hardly reducible at >800 °C). Nickel aluminate (NiAl O ) is
2 4
2+
formed through strong bond between Ni and the divalent
vacancy in transition alumina. The formation of nickel aluminate
on α-Al O only occurs at high calcination temperatures (e.g.,
2
3
>600 °C), whereas nickel aluminate is the most dominant
species found in unreduced NiO/γ-Al O catalysts regardless of
2
3
22,24
the calcination temperature,
particularly at low metal load-
ings where the surface cation vacancies of the γ-Al O are yet to
2
3
2
+
be fully occupied by the Ni ions. Also, it has been shown that
the reducibility of NiO/Al O is typically facilitated by increas-
2
3
ing the metal loadings and decreasing the calcination tempera-
22
ture. It has also been reported that the addition of a few percent
La O to γ-Al O had little influence on the nickel–support inter-
2
3
2 3
25
action. In summary, at low metal loadings (e.g., 5 wt%), NiO
and NiAl O are the more dominant species in α-Al O and
2
4
2 3
γ-Al O -supported catalysts, respectively. Therefore, the greater
2
3
catalytic activity of Ni/α-Al O compared to that of Ni/γ-Al O
2 3
2
3
2+
can be partly attributed to the better reducibility of Ni depos-
ited on α-Al O and consequently, the higher extent of reduction
2
3
(
EOR). It is worth mentioning that reduction of 5% Ni/γ-Al O3
2
(A-8) catalyst at 500, 700 and 950 °C failed to activate the cata-
lyst for the reaction of interest. The XRD pattern of this catalyst
Fig. 5 XRD patterns of 5% Ni on (a) α-Al
γ-Al (A-6) reduced at 500 °C, (c) γ-Al
and (d) γ-Al (A-8) reduced at 700 °C. The vertical dashed lines rep-
resent nickel peaks.
2
O
3
reduced at 500 °C, (b)
Fig. 4 Hydrogen yield and selectivity vs. support surface area using
2
O
3
2 3
O (A-6) reduced at 800 °C,
5
% Ni/alumina catalysts, α-Al
2
O
3
(■), γ-Al
2
O
3
(◯), La
2
O
3
–γ-Al
2
O
3
2 3
O
(▲) (380 °C, 15 min, 2 wt% glucose, 1 g catalyst).
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1766–1777 | 1771

View Article Online
Fig. 6 XRD patterns of different types of alumina before (bottom) and after (top) exposure to the supercritical water at 380 °C for 1 h.
Table 4 Performance of other supported catalysts (380 °C, 15 min, 2 wt% glucose, 1 g catalyst)
[mmol g−
1
]
CO [mmol g−1
]
CH
[mmol g−1
]
CO
2
[mmol g−1
]
CGR [%]
HGR [%]
Support
Ni [%]
H
2
4
No catalyst
Ca-1
Ca-2
Ce-1
—
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
2.7
3.1
17.3
6.0
14.1
24.5
26.6
26.9
13.0
10.0
3.4
4.5
2.5
3.4
2.6
20.5
13.6
16.8
4.3
1.0
0.7
0.4
1.0
0.7
0.15
0.3
0.6
0.7
1.8
2.5
1.9
2.4
2.4
2.7
0.2
1.2
1.0
1.2
1.2
0.6
0.4
0.3
0.3
0.3
0.6
0.2
3.8
1.8
1.4
1.1
0.1
0.1
9.8
0.1
0.7
1.9
6.2
4.6
2.6
1.2
0.2
0.04
0.07
0.4
0.1
3.7
4.5
3.7
1.0
1.4
3.5
1.8
2.8
2.1
7.3
2.4
2.7
0.3
0.3
0.14
0.1
6.5
8.3
18.1
7.0
10.7
15.9
17.3
18.9
10.9
4.7
4.8
4.5
4.1
4.7
22
27
85
24
36
54
71
72
43
20
22
19
20
22
21
53
58
55
25
28
59
46
53
40
59
51
30
29
24
26
26
9
10
111
19
46
85
117
108
55
37
11
14
8
Ce-2
HT-1
Mg-1
Mg-2
MS-1
MS-2
SA-1
SA-2
SA-3
SA-4
SA-5
Si-1
13
8
4.3
14.4
13.6
13.7
6.3
84
68
73
19
24
71
58
61
51
123
49
108
16
21
2
Si-2
Si-3
Si-4
Si-4
a
5.2
6.8
Ti-1
Ti-2
Ti-3
Ti-4
YT-1
YZ-1
Ze-1
Ze-2
Ze-3
Ze-4
ZR-1
16.7
15.6
14.6
12.8
26.5
11.5
30.7
4.9
15.6
13.1
14.4
11.0
12.0
14.0
7.2
5.6
6.0
7.0
7.3
6.5
0.5
2.9
9
a
Catalyst pellets were crushed to fine powders before impregnation.
1772 | Green Chem., 2012, 14, 1766–1777
This journal is © The Royal Society of Chemistry 2012

View Article Online
(
Fig. 5) exhibited no nickel peak, implying the formation of very
higher alkanes. The acid sites also promote the polymerization
reactions, which in turn, lead to the formation of tarry materials.
Among the various alumina phases, α-Al O is the least acidic
well-dispersed nickel aluminate phase at the surface and perhaps
in the bulk of the γ-Al O support.
2
3
2 3
26
one. Therefore, the higher gasification efficiency obtained by
Ni/α-Al O can be partly due to the weaker acidic strength of
(iii) Support stability. Chemical degradation and/or phase
2
3
transformation of the catalyst support in SCW may diminish the
catalytic activity even during a short batch experiment. Fig. 6
shows the XRD of patterns of α-Al O (A-4), γ-Al O (A-6),
α-Al O compared to that of γ-Al O .
2
3
2 3
In summary, the following factors were proposed to influence
the catalytic activity of alumina-supported nickel catalysts for
SCWG in this study: (a) reducibility and the form of nickel
oxide prior to the reduction (NiO vs. NiAl O ), (b) support stab-
2
3
2 3
La O –γ-Al O (A-10), and AlO(OH) (A-5) supports before and
2
3
2 3
after exposure to SCW at 380 °C for 1 h. Thanks to its high
surface area and superior thermal stability over a wide range of
temperatures, γ-Al O is the most widely used industrial catalyst
2
4
ility, and (c) support acidity.
2
3
support. However, as previously reported in the literature,
γ-Al O undergoes phase transition to boehmite upon exposure
Other catalysts
2
3
10
to SCW. The phase transformation of γ-Al O resulted in a sig-
2
3
26
The gas yields and carbon conversion for the SCWG of glucose
using other catalyst support materials are shown in Table 4 and
the corresponding hydrogen yields and selectivities versus
carbon conversion are plotted in Fig. 7. The relative activity for
these catalysts reduced depending on the support material in the
nificant loss of the mesopores and thus, the surface area. This
would, in turn, lead to the occlusion of the catalytically active
crystallites (i.e., nickel) and cause a fast catalyst deactivation. In
contrast, α-Al O was found to be fairly stable under similar
2
3
conditions. We note that the BET surface area of α-Al O also
2
3
following orders: carbon nanotube (CNT) > MgO > Y O , TiO ,
remained unchanged after exposure to SCW. Exposure of the
high-surface area pseudo-boehmite to SCW for 1 h increased the
crystallinity of the material but did not change its chemistry.
Consequently, one can expect that crystalline pseudo-boehmite
could potentially be a good candidate as a catalyst carrier in
SCW. Unfortunately, the Ni/pseudo-boehmite catalyst exhibited
negligible activity for SCWG of glucose (Table 3). Addition of
small quantities of lanthanum oxide did not improve the hydro-
thermal stability of γ-Al O . Considering the above discussions,
2
3
2
2
3
instability of γ-Al O as well as the negligible activity of the
2
3
boehmite-supported nickel catalysts can be partly responsible for
the higher activity of Ni/α-Al O compared to that of γ-Al O .
2
3
2 3
Therefore, except for Ni/α-Al O , no practical use is foreseen for
2
3
other alumina supported catalysts in SCWG process. It should be
also emphasized that the extremely small solubility of α-Al O
2
3
2
7
in supercritical water along with its superior mechanical
26
strength render α-Al O an appropriate candidate for catalysis
2
3
in SCW.
(iv) Mass transfer issues. Under similar operating con-
ditions, the rate of mass transfer towards the active sites of a cat-
alyst depends on the morphological and textural characteristics
of the catalyst particulate (pellet, extrudate, granule, sphere,
powder). The SCWG of glucose with finely crushed spheres
(A-6*, A-7*, and A-9*) revealed that, expectedly, the inter-
particle pores had no measurable impact on the catalytic activity.
Moreover, considering the data presented in Tables 1 and 3 for
original and crushed catalyst, it becomes apparent that there was
no relationship between the catalyst particle size and activity.
This implies that the mass transfer in the mesopores was not the
rate determining factor under the experimental conditions used.
Therefore, based on these evidences and also considering the
2
8,29
high diffusion coefficients in supercritical state,
the chemical
reaction on the catalytic active site was most likely the rate deter-
mining step under the conditions used in this study.
(
v) Support acidity. In SCW, sugars undergo dehydration on
Fig. 7 Hydrogen yield and selectivity vs. carbon conversion using 5%
Ni/support catalysts, (380 °C, 15 min, 2 wt% glucose, 1 g catalyst). Mol-
ecular sieve (▲), YSZ (◯), hydrotalcite (■), silica (◆), titania (4),
yttria (◇), magnesia (□), and CNT (●).
the catalyst acidic sites. Utilization of bifunctional catalysts, with
active hydrogenation metal and strong acidic sites, results in suc-
cessive dehydration/hydrogenation and therefore, produces
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1766–1777 | 1773

View Article Online
Fig. 8 XRD patterns of MgO, Y
2
O
3
, TiO
2
, hydrotalcite, CeO
2
and YSZ before (bottom) and after (top) exposure to the supercritical water at 380 °C
for 1 h.
Fig. 9 Carbon gasification ratio and methane yield vs. Ni loading (on α-Al
2
O
3
(A-4)) at 410 °C (▲), 380 °C (■), and 350 °C (◆) (15 min, 2 wt%
glucose, 0.05 g Ni).
1774 | Green Chem., 2012, 14, 1766–1777
This journal is © The Royal Society of Chemistry 2012

View Article Online
SiO > hydrotalcite, YSZ > CeO , activated carbon, ZrO , mol-
2 3
Optimization of Ni/α-Al O catalyst
2
2
2
ecular sieve, silica-alumina and zeolites. The results obtained
from gasification of glucose in the presence of carbon nanotubes
Based on the above findings, among the active catalysts, Ni/
α-Al O and Ni/CNT had high CGR values (∼85%) and high
2
3
(and activated carbon) should be interpreted with caution when
−1
hydrogen yields (∼17–24 mmol g ) while Ni/MgO exhibited
an excellent hydrogen yield (∼26 mmol g− ) and a CGR of
using a batch reactor as the catalyst support may be also partially
1
gasified. Zeolite pentasil (Ze-1) showed an exceptionally high
−
1
hydrogen yield (∼30 mmol g ). Similarly, MgO and Y O per-
2
3
formed better than alumina supports (hydrogen yield:
24–26 mmol g ). Hydrotalcite (HT) also demonstrated a
−1
∼
extraordinarily high hydrogen selectivity of ∼6.5 but with a
modest CGR of 54%. The XRD spectra of the studied catalyst
supports before and after 1 h exposure to SCW at 380 °C are
given in Fig. 8. Based on these plots, TiO and YSZ were the
2
only stable supports in such conditions and other supports
materials; including MgO, hydrotalcite, Y O , and CeO under-
2
3
2
went chemical reaction with water and/or phase transformation
and thereby, were not useful for the process of interest. As dis-
cussed in the previous section for the Ni/Al O catalysts, the
2
3
acidic supports (e.g., zeolites, silica–alumina) led to very poor
gasification yields. Once again, no clear relationship was
observed between the support surface area and particle size with
the catalyst activity for SCWG of glucose.
Fig. 11 Effects of calcination temperature (for 2 h) and time (at
350 °C) on product yields for SCWG using 5% wt Ni/α-Al O (A-4)
2
3
catalyst (380 °C, 15 min, 2 wt% glucose, 1 g catalyst). Carbon monox-
ide yield remained less than 0.3 mmol g− (not shown).
1
Fig. 12 Effects of reduction temperature (for 2 h) and time (at 500 °C)
on the gas yields using 5% Ni/α-Al O catalyst (380 °C, 15 min, 2 wt%
2 3
Fig. 10 Hydrogen selectivity vs. Ni loading (on α-Al
2
O
3
(A-4)) at
glucose, 1 g catalyst). Carbon monoxide yield remained less than
0.3 mmol g− (not shown).
1
3
50 °C, 380 °C, and 410 °C (15 min, 2 wt% glucose, 0.05 g Ni).
2 3
Table 5 Effects of the addition of promoters to Ni/α-Al O catalysts on gasification of 2 wt% glucose solution. (380 °C, 15 min, 0.05 g Ni)
Promoter
[mmol g−
1
]
CO [mmol g−1
]
CH
[mmol g−1
]
CO
6.5
[mmol g−1
]
CGR [%]
HGR [%]
Ni %
Element
[%]
H
2
4
2
No catalyst
2.7
24.3
9.8
1.0
0.3
0.3
0.4
0.3
0.4
0.3
0.3
0.2
0.3
0.4
0.3
0.3
0.3
0.1
4.6
1.6
1.9
6.7
6.2
5.7
5.7
5.2
6.1
6.8
5.9
9.4
22
67
40
43
77
74
70
71
67
76
80
71
92
99
9
101
39
5
5
5
5
5
5
5
5
5
5
5
—
Sn
Sn
Cs
Cs
Na
Na
Na
K
K
K
K
K
—
17.4
11.4
12.2
18.7
18.1
17.4
17.6
16.8
18.9
19.6
17.6
20.7
22.6
0.50
1.00
0.50
1.00
0.25
0.50
1.00
0.25
0.50
1.00
0.25
0.50
11.1
21.7
22.1
20.3
21.3
21.5
22.7
21.2
20.6
21.2
21.3
45
105
103
95
98
96
105
104
97
120
127
1
1
0
0
10.5
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1766–1777 | 1775

View Article Online
about 72%. Considering its low cost and high stability, α-Al O3
2
was chosen as support for further study to enhance catalyst per-
formance by optimizing synthesis conditions and by introducing
promoters.
Effects of nickel loading at different reaction temperatures.
Fig. 9 shows the relationship between the nickel loading and the
catalyst activity for SCWG of glucose at 350, 380 and 410 °C. It
should be noted that regardless of the level of nickel loading on
the support, the total nickel weight was kept constant in the
experiments (i.e., 0.05 g). It can be seen that carbon conversion
slightly decreased while the methane yield substantially dropped
by increasing the nickel to the support ratio. As discussed
earlier, nickel loading affects the nickel crystalline size and the
nickel–support interactions. Therefore, it may be concluded that
under the experimental conditions of this work, the methanation
reaction was more sensitive to the nickel crystallite size and the
extent of Ni–support interaction than other reactions such as
C–C bond rupture. For RANEY nickel-catalyzed SCWG it has
been reported that the methane formation rate considerably
decreased over time (due to substantial sintering and hence, crys-
tallite growth) while the carbon conversion did not drop at the
3
same pace. Moreover, according to Fig. 9, the methanation rate
was heavily dependent on the reaction temperature. A threefold
increase in methane yield was observed upon increase of temp-
erature from 350 to 410 °C at any level of the nickel loading
which in turn, dramatically decreased the hydrogen selectivity
(Fig. 10).
Effects of catalyst activation. Fig. 11 and 12 demonstrate the
effects of calcination and reduction on the activity of the catalyst,
respectively. Based on the results presented in Fig. 11, the calci-
nation temperature and duration had a relatively small impact on
the performance of the catalyst for SCWG. In fact, the highest
methane yield was obtained with the catalyst that was prepared
without calcination. Furthermore, the rate of methane formation
decreased with increasing reduction temperature. This obser-
vation may be justified due to the possible decrease in the metal
dispersion caused by sintering of the crystallites at higher
reduction temperatures.
Effects of promoters. Due to the proven usefulness of alkali
3
0–32
metals for facilitating the SCWG reactions,
these metals
have been added to the Ni/α-Al O catalyst as promoters.
2
3
Additionally, when added to the surface of RANEY nickel, tin
has been found to dramatically decrease the methane formation
1
8–21
rate and improve the hydrogen selectivity.
Hence, the
addition of tin as a promoter was also considered to potentially
enhance the hydrogen selectivity of Ni/α-Al O .
2
3
The corresponding results obtained from SCWG of glucose
using promoted α-Al O are presented in Table 5. For compari-
2
3
son, performances of the same catalyst without a promoter is
also given in Table 5. It can be seen that the addition of tin,
while it enhanced hydrogen selectivity, also considerably
decreased the activity of the Ni/α-Al O catalyst for gasification
Fig. 13 Hydrogen yield, methane yield and hydrogen selectivity vs.
2
3
carbon conversion using promoted Ni/α-Al
2 3
O catalysts. No promoter
of glucose. On the other hand, the addition of small quantities of
alkali metals improved the CGR and methane yield but decreases
the hydrogen selectivity. A near complete carbon conversion was
achieved upon the addition of 0.5% K to 10% Ni/α-Al O .
(×), Sn (□), Na (◇), Cs (▲), and K (●) (380 °C, 15 min, 1 g catalyst).
corresponding hydrogen selectivity using alkali and tin-pro-
moted nickel catalysts. It was found that the hydrogen yield
2
3
Fig. 13 illustrated the hydrogen and methane yields and the
1776 | Green Chem., 2012, 14, 1766–1777
This journal is © The Royal Society of Chemistry 2012

View Article Online
remained almost constant while the methane yield increased
upon doping of the catalyst with alkali metals, particularly with
potassium. In all cases, the hydrogen selectivity was inversely
correlated with the carbon conversion.
5 P. Azadi, J. Otomo, H. Hatano, Y. Oshima and R. Farnood,
Int. J. Hydrogen Energy, 2010, 35, 3406–3414.
6
7
P. Azadi, K. M. Syed and R. Farnood, Appl. Catal., A, 2009, 358, 65–72.
P. Azadi, A. A. Khodadadi, Y. Mortazavi and R. Farnood, Fuel Process.
Technol., 2009, 90, 145–151.
8
9
P. Azadi, R. Farnood and E. Meier, J. Phys. Chem. A, 2010, 114, 3962–
3968.
D. C. Elliott, T. R. Hart and G. G. Neuenschwander, Ind. Eng. Chem.
Res., 2006, 45, 3776–3781.
Conclusions
1
0 D. C. Elliott, L. J. Sealock and E. G. Baker, Ind. Eng. Chem. Res., 1993,
2, 1542–1548.
11 T. Minowa and S. Inoue, Renewable Energy, 1999, 16, 1114–1117.
The performance of supported nickel catalysts for the gasifica-
tion of a biomass model compound in supercritical water was
studied. The support materials covered a wide range of physical
and chemical characteristics, allowing for the determination of
the effects of different parameters (such as chemistry, surface
area, and particle size) on the catalytic performance. Among the
tested catalysts, α-Al O , carbon nanotube (CNT), and MgO
3
1
2 P. Azadi, S. Khan, F. Strobel, F. Azadi and R. Farnood, Appl. Catal., B,
012, 117, 330–338.
3 T. Minowa and Z. Fang, Catal. Today, 1998, 45, 411–416.
2
1
14 L. Zhang, P. Champagne and C. Xu, Int. J. Hydrogen Energy, 2011, 36,
591–9601.
9
1
1
5 P. Azadi and R. Farnood, Int. J. Hydrogen Energy, 2011, 36, 9529–9541.
6 D. C. Elliott, Biofuels, Bioprod. Biorefin., 2008, 2, 254–265.
17 R. D. Cortright, R. R. Davda and J. A. Dumesic, Nature, 2002, 418, 964–
2
3
supports resulted in the highest carbon conversions, while SiO2,
Y O , hydrotalcite, yttria-stabilized zirconia (YSZ), and TiO
2
showed modest activities and other catalysts including zeolites
967.
2
3
1
8 G. W. Huber, J. W. Shabaker and J. A. Dumesic, Science, 2003, 300,
075–2077.
2
showed negligible activities. γ-Al O supports resulted in a wide
2
3
1
9 J. W. Shabaker, D. A. Simonetti, R. D. Cortright and J. A. Dumesic,
J. Catal., 2005, 231, 67–76.
range of catalytic activities from almost inactive to highly active.
No clear relationship was found between the catalyst surface area
20 J. W. Shabaker and J. A. Dumesic, Ind. Eng. Chem. Res., 2004, 43,
105–3112.
1 J. W. Shabaker, G. W. Huber and J. A. Dumesic, J. Catal., 2004, 222,
80–191.
22 R. Molina and G. Poncelet, J. Catal., 1998, 173, 257–267.
3
(and particle size) and the catalytic activity. However, it is poss-
2
ible that other factors; such as acidity, nickel–support interaction
and stability, played a more dominant role under the conditions
used in this study. The hydrogen selectivity significantly
increased by increasing the nickel loading on the α-Al O cata-
1
2
3 J. G. Seo, M. H. Youn, S. Park and I. K. Song, Int. J. Hydrogen Energy,
008, 7427–7434.
4 I. Chen, S. Y. Lin and D. W. Shiue, Ind. Eng. Chem. Res., 1988, 27, 926–
29.
25 L. Zhang, J. Lin and Y. Chen, J. Chem. Soc., Faraday Trans., 1992, 88,
97–502.
2
2
3
2
lyst. The maximum hydrogen selectivity was obtained using a
0% Ni/α-Al O catalyst vat 380 °C. Addition of alkali promo-
9
2
2
3
4
ters enhanced the carbon conversion whereas addition of tin
decreased the catalyst activity.
2
2
2
6 C. H. Bartholomew and R. J. Farrauto, Fundamentals of Industrial Cata-
lytic Processes, John Wiley and Sons, NJ, 2nd edn, 2006.
7 K. H. Becker, L. Cemic and K. Langer, Geochim. Cosmochim. Acta,
1
983, 47, 1573–1578.
8 H. Weingartner and E. U. Franck, Angew. Chem., Int. Ed., 2005, 44,
672–2692.
References
2
2
3
9 A. Kruse and E. Dinjus, J. Supercrit. Fluids, 2007, 39, 362–380.
0 H. X. Hao, L. J. Guo, Z. Mao, Z. M. Zhang and X. J. Chen,
Int. J. Hydrogen Energy, 2003, 28, 55–64.
1
T. Yoshida, Y. Oshima and Y. Matsumura, Biomass Bioenergy, 2004, 26,
1–78.
T. Yoshida and Y. Matsumura, Ind. Eng. Chem. Res., 2001, 40, 5469–
474.
E. Afif, P. Azadi and R. Farnood, Appl. Catal., B, 2011, 105, 136–143.
P. Azadi, R. Farnood and C. Vuillardot, J. Supercrit. Fluids, 2011, 55,
7
2
3
1 J. A. Onwudili and P. T. Williams, Int. J. Hydrogen Energy, 2009, 34,
5
5
645–5656.
2 A. Kruse, D. Meier, P. Rimbrecht and M. Schacht, Ind. Eng. Chem. Res.,
000, 39, 4842–4848.
3
4
3
2
1
038–1045.
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1766–1777 | 1777