Nickel-iron alloy supported over iron-alumina catalysts for steam reforming of biomass tar model compound

Article abstract of DOI:10.1021/cs400621p

Iron-alumina-supported nickel-iron alloy catalysts were tested in a fixed-bed reactor for steam reforming of toluene as a biomass tar model compound. The influence of the calcination temperature of the iron-alumina support was also explored for the steam reforming reaction. Ni supported on an Fe₂O₃-Al₂O₃ support calcined at 500 C [NFA(500)] gave superior catalytic performance in terms of activity and stability over other catalysts. NFA(500) gave a toluene conversion of more than 90% for a period of 26 h with a H₂/CO value of 4.5. According to XRD analysis, the Ni-Fe alloys were formed and stable throughout the reforming reaction. It was observed from XPS results that the surface of the reduced NFA(500) catalyst was enriched with Fe species, where the other catalysts were enriched with Ni species. These surface Fe species play the role of cocatalysts by increasing the coverage of oxygen species during the reforming reaction to enhance the reaction of toluene and suppresses coke formation. The presence of oxygen species in the reduced catalysts was confirmed by temperature-programmed surface reaction (TPSR) with toluene and water over NFA catalysts. A temperature-programmed oxidation (TPO) study on spent catalysts revealed that the NFA(500) and NFA(700) catalysts have an additional low-temperature oxidation peak at around 525 and 535 C, respectively, suggesting the presence of a higher amount of amorphous carbon compared with the NFA(900) catalyst. The presence of a low-temperature oxidation peak at 525 C for the NFA(500) catalyst is one of the reasons for its stable catalytic performance compared with other catalysts.

Full text of DOI:10.1021/cs400621p

Research Article
pubs.acs.org/acscatalysis
Nickel−Iron Alloy Supported over Iron−Alumina Catalysts for Steam
Reforming of Biomass Tar Model Compound
J. Ashok and S. Kawi*
Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore 119260, Republic of Singapore
ABSTRACT: Iron−alumina-supported nickel−iron alloy catalysts were tested in a
fixed-bed reactor for steam reforming of toluene as a biomass tar model
compound. The influence of the calcination temperature of the iron−alumina
support was also explored for the steam reforming reaction. Ni supported on an
Fe2O3−Al2O3 support calcined at 500 °C [NFA(500)] gave superior catalytic
performance in terms of activity and stability over other catalysts. NFA(500) gave
a toluene conversion of more than 90% for a period of 26 h with a H2/CO value of
4.5. According to XRD analysis, the Ni−Fe alloys were formed and stable throughout the reforming reaction. It was observed
from XPS results that the surface of the reduced NFA(500) catalyst was enriched with Fe species, where the other catalysts were
enriched with Ni species. These surface Fe species play the role of cocatalysts by increasing the coverage of oxygen species during
the reforming reaction to enhance the reaction of toluene and suppresses coke formation. The presence of oxygen species in the
reduced catalysts was confirmed by temperature-programmed surface reaction (TPSR) with toluene and water over NFA
catalysts. A temperature-programmed oxidation (TPO) study on spent catalysts revealed that the NFA(500) and NFA(700)
catalysts have an additional low-temperature oxidation peak at around 525 and 535 °C, respectively, suggesting the presence of a
higher amount of amorphous carbon compared with the NFA(900) catalyst. The presence of a low-temperature oxidation peak
at 525 °C for the NFA(500) catalyst is one of the reasons for its stable catalytic performance compared with other catalysts.
KEYWORDS: hydrogen production, nickel−iron alloy, biomass, toluene steam reforming, iron−alumina
oxides, and other metals on Ni-based systems have been
investigated.17,18 However, the development of a cheap, stable,
and coke-resistant catalyst for steam reforming of tar remains as
a challenge for many researchers. In a bid to reduce the cost of
the catalyst, alumina (Al2O3) has been used as the catalyst
support for Ni-based systems in many instances. However, the
main drawback of Ni/Al2O3 catalysts lies in the rapid catalyst
deactivation that occurs because of the encapsulation of carbon
and sintering or loss of the active Ni component at high
temperature.19,20 It is also reported that the addition of CeO2
to Ni catalysts enhances the catalytic performance by forming a
Ni−CeO2 nanocomposite. Good catalytic activity and high
resistance toward formation of coke were attributed to the
existence of the large interface between the Ni metal and the
Ce oxide surface.21 Similar to this work, we recently explored
Ni supported on CeO2-doped CaO−Al2O3 as a catalyst for
steam reforming of toluene and observed that the addition of
CeO2 to Ni/CaO−Al2O3 in an optimum quantity can
significantly enhance the steam reforming performance. As
reported earlier,10 the promotional effect of CeO2 is mainly due
to its redox properties as well as the fact that Ce oxide species
can interact with Ni species to form Ni−CeO2 nanocomposites.
Furthermore, alloying Ni22−24 with other metals can improve
the catalytic performance in terms of activity and stability, and
one such suitable modifier is Fe. In addition, like Ce species, Fe
1. INTRODUCTION
Catalytic steam reforming of biomass tar to produce H2-rich
synthesis gas is becoming more attractive because of the rising
interest in utilizing biomassderived from forest, agricultural,
and municipal solid waste materialsas a renewable energy
source.1−5 However, one of the most critical problems in
utilization of biomass via the gasification process is the
formation of tars, an undesirable reaction that is an impediment
to the commercialization of the biomass gasification process. In
addition, these tars are largely aromatic hydrocarbons, such as
benzene, toluene, and naphthalene, which harness large
amounts of energy, thereby lowering the energy content of
the product gas. These tars also cause operational problems
when cooled and condensed in heat exchangers, lowering heat
transfer efficiencies, plugging pipes, and so on. Therefore,
conversion of biomass tars is crucial in the biomass gasification
process. Among many processes, catalytic steam reforming of
biomass tar is a considerably attractive approach for tar removal
during biomass gasification since it produces syngas, which is a
high-value gas product.6−8
In steam reforming of tar, the catalysts should exhibit high
catalytic performance in terms of catalyst stability and high
resistance toward coke deposition.9,10 It has been known from
the literature that supported metal (e.g., Ni, Co, Pt, Pd)
catalysts are effective for steam reforming of biomass tar.11−16
Moreover, Ni-based catalysts have been widely investigated for
the steam reforming of biomass tar because of its low cost and
high activity.7 In the development process of highly active and
stable Ni-based catalysts, the effects of supports, additives,
Received: July 30, 2013
Revised: October 17, 2013
Published: November 7, 2013
© 2013 American Chemical Society
289
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
species have good redox properties. It has also been reported
that the improved performance in the steam reforming activity
of Ni−Fe/Al2O3 catalysts is due to the presence of an intimate
interaction between Fe and Ni, which results in the formation
of a Ni−Fe alloy. As has been reported,22 in the Ni−Fe/Al2O3
catalytic system, Fe plays the role of a cocatalyst. Since Fe has a
higher oxygen affinity than Ni,25 the addition of Fe to Ni
catalysts can increase the coverage of oxygen atoms during the
steam reforming reactions. This increment in oxygen coverage
suggests that Ni−Fe alloy species are vital in enhancing the
performance in steam reforming of tar. In addition, was also
recently reported that the presence of uniform Ni−Fe alloy
particles in Ni−Fe/Mg/Al catalysts enables a high catalytic
performance in the steam reforming of toluene and phenol.24
Thus, obtaining an intimate interaction between Ni and Fe
species to form Ni−Fe alloys is a crucial entity in preparing
Ni−Fe-based catalysts. Hence, in the present study, in order to
obtain an intimate interaction between most of the Ni species
and the Fe species in the Ni/Fe2O3−Al2O3 catalyst system, we
first synthesized Fe2O3−Al2O3 support materials by a
coprecipitation method and subsequently calcined them at
different temperatures in order to establish different kinds of
interactions between Fe and Al species present in the supports.
We chose to synthesize the Fe2O3−Al2O3 support materials
first because according to previous reports,26 the redox
properties and thermal stability of Fe oxides can increase
significantly in the presence of Al oxides. In all of the support
materials, the Fe to Al molar ratio was kept constant. This
Fe2O3−Al2O3 support was used to prepare Ni/Fe2O3−Al2O3
catalysts by the impregnation method. The thus-prepared Ni/
Fe2O3−Al2O3 catalysts were used to obtain FeOx−Al2O3-
supported Ni−Fe alloy catalysts upon reductive pretreatment
with H2 gas at 700 °C for 1 h to investigate their catalytic
performance for the toluene steam reforming process at a
reaction temperature of 650 °C and ambient pressure.
steam dealkylation:
C7H8 + 2H2O → C6H6 + CO2 + 3H2
(7)
steam reforming of methane:
CH4 + H2O → CO + 3H2
(8)
Other reactions such as carbon formation by toluene
decomposition and cracking are also included in the process.
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation. Fe2O3−Al2O3 mixed oxides
with the nominal Fe:Al molar ratio of 1:2 were prepared using a
coprecipitation method. In a typical preparation method, nearly
saturated solutions of nitrates (ferric nitrate and aluminum
nitrate) were homogeneously mixed in the required molar
ratios. The pH value of the solution was then slowly adjusted to
∼10 with 2 M aqueous ammonia solution, and the solution was
then stirred continuously for another 1 h to complete the
precipitation process. The precipitated solution was dried in an
oven at 100 °C for 24 h and then calcined at different
temperatures (500, 700, and 900 °C) over 4 h in an air
atmosphere. The as-prepared catalyst supports were labeled as
FA(x), where x represents the calcination temperature of the
support.
The 10 wt % Ni supported catalysts were prepared by a
simple wetness impregnation method wherein an aqueous
solution containing the required amount of Ni(NO3)2·6H2O
was mixed with the requisite amount of support material to
yield the respective Ni weight percent needed. The mixtures
were then dried under constant stirring at 80 °C and further
dried at 100 °C for another 24 h before they were calcined in
air at 700 °C for 4 h (under a ramping rate of 3 °C/min). The
thus-prepared catalysts were labeled as NFA(x), where x
represents the calcination temperature of the Fe2O3−Al2O3
support material.
In steam reforming of toluene (SRT), the final products can
be regarded as the result of competition of many parallel and
consecutive reactions.27 However, different operation con-
ditions such as the steam-to-carbon (S/C) ratio, temperature,
and catalysts will result in different products. These evolutions
of gases yields under different conditions can be explained by
the following reactions involved in toluene steam reforming.
2.2. Catalyst Characterization. The X-ray diffraction
(XRD) pattern of each catalyst was measured on a Shimadzu
XRD-6000 diffractometer using Cu Kα radiation. The catalyst
was placed on an aluminum slide and scanned from 2θ of 20°
to 80° at a ramp rate of 2°/min. The beam voltage and current
used were 40 kV and 30 mA, respectively. H2/CO Temper-
ature-programmed reduction (TPR) measurements for fresh
catalysts were performed on a Quantachrome ChemBET-3000
instrument. Prior to the TPR measurement, 0.04 g of catalyst
was outgassed in He for 1 h at 300 °C to remove any moisture
and then cooled to room temperature. Then 5% H2/N2 or 5%
CO/He gas was introduced to the catalyst while the
temperature of the furnace was increased at a heating rate of
10 °C/min to 1000 °C. The CO chemisorption was done in the
Quantachrome ChemBET-3000 instrument. Prior to the CO
pulse chemisorption measurement, the sample was reduced at
700 °C for 1 h under H2 gas before it was cooled to 0 °C in a
helium flow of 80 mL/min. The CO chemisorption was
measured at 0 °C by introducing pulses of CO using 5% CO/
He with a loop volume of 250 μL. The pulses were continued
until no further uptake of CO was measured.
steam reforming:
C7H8 + 7H2O → 7CO + 11H2
(1)
steam reforming:
C7H8 + 14H2O → 7CO2 + 18H2
(2)
water-gas shift:
CO + H2O → CO2 + H2
(3)
(4)
(5)
(6)
dry reforming:
C7H8 + 7CO2 → 14CO + 4H2
In order to understand the actual reduced species involved in
the SRT reaction, the following characterization experiments
were conducted in sequence: (1) H2 temperature-programmed
reduction (H2-TPR1), (2) temperature-programmed oxidation
(TPO), and (3) temperature-programmed reduction (H2-
TPR2). In these experiments, 5% H2/N2 and 5% O2/N2 were
used for the TPR and TPO steps, respectively, while the
hydrodealkylation:
C7H8 + H2 → C6H6 + CH4
steam dealkylation:
C7H8 + H2O → C6H6 + CO + 2H2
290
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
temperature of the system was increased from room temper-
ature to 700 °C at a ramp rate of 10 °C/min and dwelled for 1
h. These experiments were performed on the Quantachrome
ChemBET-3000 instrument. For temperature-programmed
surface reaction (TPSR) experiments with toluene and water,
about 30 mg of catalyst sample was loaded in an isothermal
zone of the reactor. Prior to the TPSR measurement, the
catalyst was reduced in H2 gas at 700 °C for 1 h. After the
sample was cooled to room temperature, the H2 gas was
switched to 80 mL/min of He gas saturated with toluene/water
at room temperature, and the temperature was increased to 800
°C at a rate of 10 °C/min. The reactor effluent gas was
analyzed by mass spectroscopy (Thermostat GSD300).
calcined at different temperatures. Table 1 presents the product
gas distributions, H2/CO molar ratios, and catalytic perform-
Table 1. Catalytic Performance of NFA(x) Catalysts after 1 h
a
Reaction Time
product distribution
(μmol min−1
)
toluene conversion
(%)
catalyst
blank
H2
CO
CO2
H2/CO
−
−
−
−
−
NFA(500)
NFA(700)
NFA(900)
80.9
68.2
48.7
1641.7
1426.4
743.2
521.8
423.7
419.7
521.4
487.8
153.1
3.1
3.3
1.8
The catalyst surface analysis was performed using X-ray
photoelectron spectroscopy (XPS) from a KRATOS AXIS Hsi
165 spectrometer equipped with a Mg Kα source (1253.6 eV).
The total amount of deposited carbon on the spent catalysts
was measured using thermogravimetric analysis (TGA) on a
Shimadzu DTG-60 thermogravimetric analyzer. About 10 mg
of spent catalyst was used in each TGA experiment and was
heated in air to 1000 °C at a heating rate of 10 °C/min. The
type of carbon deposited on spent catalysts was determined by
TPO, in which about 10 mg of spent sample was heated in an
80 mL/min flow of 5% O2/N2 at a rate of 10 °C/min up to 900
°C and held at this temperature for 15 min. The amount of
CO2 emission as a function of temperature was recorded using
mass spectroscopy (Thermostat GSD300).
2.3. Catalyst Evaluation. The catalytic reaction was carried
out in a fixed-bed quartz reactor with an inner diameter of 4
mm and a length of 400 mm as reported previously.28,29 The
catalyst in each test was held by quartz wool placed in the
middle of the reactor. The operating conditions were as
follows: catalysts amount = 0.03 or 0.10 g; toluene amount =
0.19 mmol (diluted by He gas); S/C ratio = 1.0 to 4.0; reaction
temperature = 650 °C. Prior to the catalytic reaction, the
catalyst was reduced in 30 mL/min H2 at 700 °C for 1 h and
then purged in 120 mL/min helium while the temperature
reached the desired reaction temperature (650 °C). Water and
toluene were vaporized at 300 °C in a preheater and mixed with
120 mL/min He gas before entering the main reactor in the
vapor state. The reaction products were then passed through a
cold trap at a temperature of 5 °C to condense the unreacted
water from the gasification reaction. The noncondensable gas
product was analyzed using a gas chromatograph (HP 6890)
equipped with a Carboxen column and a thermal conductivity
detector (TCD). The chromatogram showed peak areas for all
of the reacted gases, which were then converted to volume
percent using a calibration curve. The total flow rate of the
product gases was measured using a bubble flow meter. The
conversion of toluene was expressed in terms of carbon
conversion, which was calculated using the following formula:
a
Reaction conditions: toluene = 188 μmol min−1; steam = 4444 μmol
min−1; He = 5357 μmol min−1; W = 30 mg; reaction temperature =
650 °C; reaction time = 1 h.
ances for NFA catalysts after 1 h reaction time. The reaction
was performed at 650 °C with an S/C ratio of 3.4 as reported
earlier.28,29 In the blank test there was no conversion of toluene
at all, showing that presence of catalyst is very important for
this reaction. In this SRT process, the main product stream
consists of H2, CO, CO2, and trace amounts of CH4 (<1%).
The production of these gases can be explained by eqs 1 to 4.
According to Table 1, the toluene conversions for the
NFA(500), NFA(700), and NFA(900) catalysts were 80.9,
68.2, and 48.7%, respectively. It was also observed that the H2,
CO, and CO2 yields increased correspondingly with increasing
toluene conversion. In comparison, NFA(500) and NFA(700)
produced significantly a higher H2/CO value of ∼3.2 compared
with the NFA(900) catalyst’s H2/CO value of 1.8. This is a
result of the fact that the NFA(500) and NFA(700) catalysts
have higher amounts of available lattice oxygen species (to be
shown later from TPSR of toluene results), which apparently
promotes the water-gas shift (WGS) reaction (eq 3). Next, the
time-dependent catalytic performance of NFA catalysts for 6 h
reaction time is depicted in Figure 1, which shows that the
overall catalytic performance of the NFA catalysts was
influenced by the calcination temperature of the catalyst
support, as the catalyst support calcined at lower temperature of
500 °C, NFA(500), showed superior catalytic performance
compared with the others. It can also be observed from Figure
1 that all of the catalysts showed a decreasing trend of catalytic
performance with time on stream. This decrease in catalytic
performance with time is primarily due to the deposition of
carbon species over the catalysts during the SRT process at 650
°C. This deposited carbon can encapsulate the active metal
species, resulting in a decrease in the amount of catalytically
active species. The nature and amount of carbon species
deposited over spent NFA catalysts were characterized, and the
same will be discussed in later parts of this article. In addition,
structural changes within the NFA catalysts during the SRT
process can also play an important role in their catalytic
performance with time on stream. These structural changes
were identified by characterizing reduced and spent NFA
catalysts, and the obtained results will also be discussed later in
this article.
nCO + n
CO2
Xtoluene
=
× 100 %
7nr in
(9)
where nj is the molar flow rate of gas j.
3. RESULTS AND DISCUSSION
3.1. Catalyst Activity and Stability Study. 3.1.1. Steam
Reforming of Toluene over NFA Catalysts. The catalytic
performance for steam reforming of toluene as a biomass tar
model compound was carried out in a continuous fixed-bed
reactor over 10 wt % Ni supported on Fe2O3−Al2O3 supports
3.1.2. Effect of Steam Feed Concentration. The influence
of the steam-to-carbon ratio on the catalytic activity of NFA
catalysts was investigated by varying the S/C ratio from 1.0 to
4.0 at a reaction temperature of 650 °C after 1 h reaction time
on stream. The obtained results are depicted in Figure 2, which
291
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
Figure 2. Effect of steam-to-carbon ratio on the catalytic performance
of NFA catalysts: (A) toluene conversion (%); (B) H2/CO ratio.
Reaction conditions: toluene = 188 μmol min−1; He = 5357 μmol
min−1; W = 30 mg; reaction temperature = 650 °C; reaction time = 1
h.
that even at an S/C ratio of 2 or more, the toluene reforming
reaction did not proceed according to eq 2. Instead, it is more
appropriate to postulate that it proceeds according to SRT with
an S/C ratio of 1 (eq 1) followed by the WGS reaction (eq 3).
This suggests that increasing the steam amount in the feed can
enhance the SRT reaction by enhancing the WGS reaction to
produce more H2 gas. High amounts of H2 and CO2 were
produced at high S/C ratios via the WGS reaction since a high
steam partial pressure pushes the WGS equilibrium toward
formation of H2 and CO2. Furthermore, an excess of steam is
known to be beneficial to the suppression of coking because it
leads to partial gasification of the carbon formed.30 In Figure 2,
when the S/C ratio was raised to 3.4, the conversion of toluene
reached a maximum for all of the catalysts and is almost similar
to that with an S/C ratio of 4. Likewise, the product yield with
respect to the H2/CO value also reached a maximum at an S/C
ratio of 3.4 (Figure 2 B). This indicates that increasing H2O
beyond an optimum content not only promotes steam
reforming of toluene but also substantially favors the water-
gas shift reaction. Domine et al.31 investigated the effect of the
S/C ratio on hydrogen production over Pt and Rh noble-metal-
Figure 1. SRT performance of (A) NFA(500), (B) NFA(700), and
(C) NFA(900) catalysts. Reaction conditions: toluene = 188 μmol
min−1; steam = 4444 μmol min−1; He = 5357 μmol min−1; W = 30
mg; reaction temperature = 650 °C.
shows that the S/C ratio has a significant influence on the
catalytic activity and the distribution of the products. From
Figure 2 it can also be observed that all of the catalysts display
similar behaviors with varying S/C ratioincreasing toluene
conversion and H2/CO ratio with increasing S/C ratio
despite achieving different conversions and product yields. Ideal
SRT reactions at S/C ratios of 1 and 2 are shown in eqs 1 and
2, respectively. Equation 3 represents the WGS reaction to
produce CO2 and more H2. The CO2 produced by eqs 2 and 3
may react with toluene according to dry reforming of toluene
(eq 4). A comparison of the obtained results (Table 1, Figure 1,
and Figure 2) and the above possible reactions clearly shows
292
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
based catalysts. They concluded that the effect of the S/C ratio
depends on the nature of the metal sites on the catalyst. It is
evident from Figure 2 that an S/C ratio of 3.4 is optimum for
all of the NFA catalysts to attain their best toluene reforming
performances at a reaction temperature of 650 °C.
was able to maintain toluene conversion at a level of more than
90% for a period of 26 h reaction time. In contrast, the
NFA(700) catalyst attained a toluene conversion of more than
90% for ca. 20 h, with a subsequent gradual decrease in toluene
conversion as time proceeds. Eventually, a toluene conversion
of 85% after 26 h reaction time was obtained for the NFA(700)
catalyst. In stark contrast to the other catalysts, the NFA(900)
catalyst showed toluene conversion ranging from 75 to 40%
over a 20 h run time. The H2/CO values for the NFA catalysts
(Figure 3B) decrease in the order NFA(500) > NFA(700) >
NFA(900). This suggests that the NFA(500) catalyst showed
superior SRT catalytic performance in terms of conversion,
stability, and H2/CO value for ca. 26 h reaction time compared
with the NFA(700) and NFA(900) catalysts.
3.2. Characterizations of Fresh and Spent Catalysts.
3.2.1. DT/TG Analyses of Oven-Dried Fe−Al Support. To
understand the potential transformation of the oven-dried
Fe2O3−Al2O3 support precursor in the muffle furnace, TGA
and differential thermal analysis (DTA) were performed up to
1000 °C at a heating rate of 10 °C/min under an air
atmosphere, as shown in Figure 4. The DTA trace in Figure 4
3.1.3. Catalyst Stability Study. In order to integrate this
system with the industrial reforming process downstream, the
catalysts should be capable of completely reforming toluene to
H2, CO, and CO2 by preventing possible coke deposition in the
SRT reformer at ambient temperatures. Therefore, the
following severe experimental conditions were imposed to
examine the stability of the NFA catalysts for the SRT reaction:
toluene = 0.19 mmol (diluted by He gas); S/C ratio = 3.4;
reaction temperature = 650 °C; space velocity = 132 L g−1 h−1
at 25 °C. Figure 3A,B displays the conversion of toluene and
the product selectivities (H2/CO ratio), respectively, as
functions of reaction time. It is observed that the NFA(500)
and NFA(700) catalysts exhibited good activity and stability
under the present operating conditions. The NFA(500) catalyst
Figure 4. DTA and TGA of the oven-dried Fe2O3−Al2O3 catalyst
support precursor.
shows four endothermic peaks. The initial endothermic peak
(∼90 °C) indicates loss of physisorbed water present in the
precursor, while the other three endothermic peaks (130, 170,
and 272 °C) represent both weight loss and formation of
gaseous species that are caused by decomposition reactions
occurring in various distinguishable steps.32 The sample
undergoes three decomposition steps in the temperature
range of 120−320 °C, as evidenced by the huge sample weight
loss, sharp endothermic effect, and vigorous gas emissions
(NH3, NOx, and H2O). The TGA curve presented in Figure 4
shows that as the temperature was increased from ambient
temperature to around 300 °C, the weight of the oven-dried
Fe−Al support precursor decreased from 8.79 to 2.3 mg. This
major weight loss of 73% denotes that most of the precursor
decomposed below 300 °C to form various oxides.
Furthermore, a significant weight loss of >6% was recorded
as the temperature was increased from 300 °C to around 500
°C, and there was no substantial weight loss (<2%) observed
above 500 °C. The weight loss up to 1000 °C agrees well with
the theoretical value corresponding to complete decomposition
to Fe2O3 and Al2O3. The aforementioned results suggest that a
Figure 3. Stability study for steam reforming of toluene over Fe2O3−
Al2O3-supported Ni catalysts: (A) toluene conversion and (B) H2/CO
values against time on stream. Reaction conditions: toluene = 188
μmol min−1; steam = 4444 μmol min−1; He = 5357 μmol min−1; W =
100 mg; reaction temperature = 650 °C.
293
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
minimum temperature of 500 °C is required to prepare a stable
Fe2O3−Al2O3 support material.
3.2.2. Surface Area Analysis. Table 2 shows the Brunauer−
Emmett−Teller (BET) surface areas of bare supports and
Table 2. BET Surface Areas and H2 Consumption and CO
Uptake Values for FA and NFA Catalysts
H2
ABET
H2 consumption
consumption
CO uptake
c
a
b
sample
(m2/g)
(mmol/g)
(mmol/g)
(μmol/gcat)
FA(500)
169.8
93.4
92.4
81.1
20.3
15.1
−
8.98
−
10.06
−
19.15
NFA(500)
FA(700)
−
9.31
−
−
10.24
−
−
18.23
−
NFA(700)
FA(900)
NFA(900)
9.95
10.52
15.97
a
b
c
H2 consumption below 700 °C in TPR1 profiles shown in Figure 7.
H2 consumption below 700 °C in TPR2 profiles shown in Figure 7.
Obtained from CO pulse chemisorption at 0 °C.
supported Ni catalysts. The surface area of the bare FA support
was observed to be greatly reduced with increasing support
calcination temperature because of sintering of the support.33 It
was also observed that the supported Ni catalysts have lower
surface areas than the bare FA supports. This is primarily due to
the penetration of the active component into the pores of the
support during the preparation, which in turn results in the
dispersion of the active component on the support. This
decrease in surface area is greater for NFA(500) than for the
other catalysts because its corresponding FA support was
calcined at a lower temperature (i.e, 500 °C). The highest
surface area of 93 m2/g for the NFA(500) catalyst is possibly
one of the reasons for its superior catalytic performance
compared with the other catalysts.
3.2.3. XRD Analysis of Reduced and Spent Catalysts.
Figure 5A shows the XRD patterns of freshly reduced NFA
catalysts for 2θ = 42−48°. All of the patterns in Figure 5A show
a diffraction peak corresponding to the Ni metal phase that is
centered at around 2θ = 44.5°. However, a gradual shift to
smaller angle was observed for the patterns of NFA(500) and
NFA(700) catalysts. According to previous reports,34 the shift
is interpreted to result from the creation of defect sites in the
metallic Ni phase by Fe species, thus resulting in the formation
of a Ni-rich Ni−Fe alloy with a face-centered-cubic (fcc)
structure. The average particle size of metallic Ni was
determined by broadening of the (111) line of nickel at
around 44.5°. The Ni particle sizes were estimated to be 18.6,
21.7, and 22.7 nm for NFA(500), NFA(700), and NFA(900),
respectively (Table 3). The average Ni particle sizes were found
to increase with increasing calcination temperature, and this can
be primarily attributed to the decrease in the surface area of the
corresponding Fe−Al support. In addition to the peaks
belonging to metallic Ni, all of the profiles show a peak at
around 2θ = 43.5°, which is assigned to an Fe-rich Ni−Fe alloy
with fcc structure.35 The average particle size of the Ni−Fe
alloy phase was estimated at 2θ = 43.5° and was observed to
decrease with decreasing calcination temperature of the NFA
catalyst supports. Furthermore, the broadness of the Ni−Fe
alloy peak suggests that the Fe and Ni composition during alloy
phase formation is not uniform. Together with the TPR results
(which will be presented later in Figure 6), the results reveal
that reduced Fe species can interact with reduced Ni species to
form Ni−Fe bimetallic compounds. On the other hand, the
Figure 5. XRD patterns of (A) freshly reduced and (B) spent Fe2O3−
Al2O3-supported Ni catalysts.
XRD pattern of NFA(900) in Figure 5A has peak at around 2θ
= 43.1° that is assigned to the Al2O3-rich Fe2O3 corundum
phase.26 This phase is relatively inert, thereby inferring that it is
hard to reduce and thermally stable. It is noteworthy to
mention here that a peak at 2θ = 33.2° corresponding to an
Fe2O3-rich Fe2O3−Al2O3 hematite phase36 was observed in the
XRD patterns of freshly calcined NFA catalysts. However, this
peak disappeared in XRD patterns of the reduced NFA
catalysts. The bulk compositions of Ni−Fe alloy in the reduced
NFA catalysts presented in Table 3 were determined using
Vegard’s rule.24 The Fe/Ni molar ratios are in the order of 1.28,
1.22, and 0.98 for NFA(500), NFA(700), and NFA(900)
catalysts, respectively. This suggests that the reduced
NFA(500) catalyst had a slightly higher Fe-rich Ni−Fe alloy
phase than other NFA catalysts. The presence of Ni−Fe
bimetallic compounds on the NFA catalysts was also confirmed
after a life test for 6 h, as shown in Figure 5B. It can be
observed from Figure 5B that the peak intensities appeared to
be considerably lower than the XRD patterns of corresponding
reduced NFA catalysts. This is possibly due to the deposition of
amorphous carbon species over the NFA catalysts during the
SRT process. In comparison with the XRD patterns of reduced
NFA catalysts (Figure 5A), the peak assigned to the Fe-rich
294
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
Table 3. Ni−Fe Alloy Crystallite Sizes for Reduced and Spent NFA Catalysts
reduced
spent
metallic Ni phase
crystallite size (nm)
Fe-rich Ni−Fe alloy phase
Ni-rich Ni−Fe alloy phase
a
catalyst
2θ (deg)
2θ (deg)
crystallite size (nm)
2θ (deg)
crystallite size (nm)
Fe/Ni atomic ratio
NFA(500)
NFA(700)
NFA(900)
44.45
44.48
44.53
18.6
21.7
22.7
43.56
43.58
43.59
07.7
11.3
16.8
44.34
44.37
44.39
10.6
11.8
19.3
1.28
1.22
0.80
a
Calculated from the d(200) spacing from XRD patterns of reduced catalysts using Vegard’s law: dNi−Fe = dNi × Ni/(Ni + Fe) + dFe × Fe/(Ni + Fe),
where dNi and dFe are 0.1762 nm (JCPDS no. 01-070-1849) and 0.1823 nm (JCPDS no. 01-089-4185), respectively.
Ni−Fe alloy phase was weakened and almost vanished.
Correspondingly, the peak assigned to the metallic Ni phase
in Figure 5A was also weaker and shifted to lower angles. This
shifted peak observed in Figure 5B can be assigned to a Ni-rich
Ni−Fe alloy phase (at 2θ = 44.3°). In addition, this shift to
lower angles suggests that the bulk composition of Fe to Ni can
increase.22 The average particle sizes of this Ni-rich Ni−Fe alloy
phase were determined at around 2θ = 44.3° and were observed
to be 10.6, 11.8, and 19.3 nm for the NFA(500), NFA(700),
and NFA(900) catalysts, respectively. According to the particle
size values in Table 3, the average Ni metal size for spent
catalysts generally decreased during SRT reaction. The decrease
in the particle size was comparatively higher for the Ni/Fe−
Al(500) catalyst than others. This behavior is mainly due to the
presence of the higher amount of Fe species in the catalysts,
which plays an important role in forming the Ni−Fe alloy
phase. The decrease in Ni particle size for the spent catalysts is
possibly due to the formation of a reasonably low crystalline Ni-
rich Ni−Fe alloy phase during the SRT process at a reaction
temperature of 650 °C.
3.2.4. H2-TPR Analysis. Generally, the reduction behavior of
any bimetallic catalyst is a complex phenomenon. Likewise, for
Ni−Fe catalysts, its complexity is derived from the multiple
interactions that can arise during the catalyst preparation
process as well as the multiple reduction centers that Fe species
possess. However, in order to better understand the reduction
behavior of NFA catalysts, TPR profiles for an Al2O3-supported
Ni catalyst [NA(500)] up to 1000 °C were obtained, and the
profiles are presented in Figure 6. The TPR profile of NA(500)
shows a two-stage reduction behavior. A lower-intensity, lower-
temperature reduction peak centered at ∼450 °C is due to the
reduction (Ni2+ → Ni0) of NiO species not interacting with the
Al2O3 support. On the other hand, the major reduction peak
starting from 650 to 1000 °C is due to the reduction of Ni
species present in NiAl2O4 spinel.36 In contrast, the TPR
profiles of the NFA(x) catalysts have three reduction centers.
First, a low-temperature reduction peak ranging from 250 to
500 °C and centered at 460 °C can be attributed to the
reduction of Fe species (Fe2O3 to Fe3O4).36 According to both
the literature37 and the TPR profile of NA(500), uninteracted
Ni species were also reduced in this reduction temperature
range. However, because of the presence of alumina in the γ
phase, these Ni species are expected to have strong interactions
with support elements under the present catalyst preparation
conditions. This suggests that the low-temperature reduction
peak is mainly due to the reduction of Fe species. Second, a
moderate-temperature reduction peak ranging from 500 to 800
°C can be assigned to the conversion (to Ni0) of NiO species
that have interacted with Fe species. According to previous
reports,38 in Ni−Fe bimetallic catalysts, the Ni reduction center
shifts to higher temperature with an increase in the interaction
with Fe species in the catalyst. The TPR profile of the
NFA(900) catalyst suggests that it has weaker Ni−Fe
interactions than the other catalysts. The weaker Ni−Fe
interactions in the NFA(900) catalyst exist because the support
material of the NFA(900) catalyst was calcined at 900 °C prior
to preparing the supported Ni catalyst. This calcination
temperature of 900 °C is sufficient to establish strong
interactions between Fe and Al species in the support to
form an inert corundum Fe2O3−Al2O3 phase. The formation of
this inert phase was previously shown in Figure 5A,B for the
NFA(900) catalyst. Lastly, all of the NFA samples have a high-
temperature reduction peak starting from 750 to 1000 °C,
which can be attributed to the reduction of Ni species in
NiAl2O3 spinel. The same reduction center is also observed in
the TPR profile of the NA(500) catalyst. This high-temperature
reduction peak is more significant for NFA(500) than for the
other catalysts. However, in the TPR profile of the NFA(900)
catalyst, this peak appears to be a shoulder of the moderate-
temperature reduction peak. Finally, TPR analysis suggested
that Ni species in the NFA(500) catalyst have stronger
interactions with support species (Fe and Al species) compared
with the other catalysts.
3.2.5. H2-TPR1 and H2-TPR2 Study. The reduction behavior
of the NFA catalysts up to 1000 °C is presented in Figure 6.
However, according to the SRT performance conditions
reported in section 2.3, all of the catalysts underwent reductive
treatment at 700 °C for 1 h prior to use in the SRT process.
Thus, in order to understand the actual reduced species
involved in the SRT reaction, several consecutive TPSR studies
were conducted on fresh NFA catalysts using 5% H2/N2 and
5% O2/N2 gases. The thus-obtained TPR1 and TPR2 profiles
for the NFA(x) catalysts are depicted in Figure 7. The TPR1
Figure 6. H2-TPR profiles of the NA(500) and NFA(x) catalysts.
295
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
SRT reaction were mostly Ni−Fe alloy species in all of the
catalysts.
3.2.6. Temperature-Programmed Surface Reaction with
Toluene and Water. TPSR studies with water and toluene
were measured separately over reduced NFA catalysts in order
to understand the nature of the active centers and their
capability to activate reactant molecules, namely, water and
toluene. The TPSR study was also carried out over the
NA(500) catalyst for comparison. TPSR of toluene was carried
out over reduced NA(500) and NFA catalysts, and the results
are depicted in Figure 8A,B as H2 and CO evolution as
Figure 7. H2-TPR1 and H2-TPR2 profiles of Ni supported on Fe2O3−
Al2O3 supports calcined at different temperatures.
profiles for all of the NFA catalysts in Figure 7 show two
reduction peaks. According to conclusions made from Figure 6,
the low-temperature reduction peak at ∼470 °C is due to
reduction of Fe species (Fe2O3 → Fe3O4) as well as
uninteracted Ni species (NiO → Ni0), whereas the high-
temperature peak centered at ∼700 °C corresponds to
conversion of Ni species that have strongly interacted with
Fe species. Interestingly, this low-temperature reduction peak
was also observed in the TPR2 profiles for all of the NFA
catalysts. However, this low-temperature peak shifted by ∼30
°C to an even lower temperature region, and the H2
consumption for this peak increased significantly compared
with its corresponding TPR1 profile. On the other hand, the
total H2 consumption values for NFA catalysts during the TPR1
and TPR2 processes (Table 2) show that H2 consumption value
during TPR2 was generally higher than that during TPR1. This
increase in the H2 consumption was greater for the NFA(500)
catalyst than for the other catalysts. This increase is possibly
due to redispersion of reduced species upon exposure to the
oxygen environment. According to previous reports,22 during
the TPR1 process Fe species were reduced first, and these
reduced Fe species can interact intimately with Ni species to
form Ni−Fe alloy species. The thus-formed Ni−Fe alloy
species were oxidized during the TPO process and then
reduced at lower temperature during the TPR2 step. Since some
or most of the Ni species were involved in forming Ni−Fe alloy
species, the involved Ni species were also reduced in this low-
temperature region. Furthermore, a significant high-temper-
ature reduction peak (∼600 °C) due to the reduction of Ni
species that had interacted with Fe species was observed in the
TPR2 profile of the NFA(900) catalyst. This high-temperature
reduction peak was broader and appeared like a shoulder on the
low-temperature reduction peak in the TPR2 profiles of the
NFA(500) and NFA(700) catalysts. It is noteworthy to
mention here that the H2 consumption for the high-
temperature reduction peak for all of the TPR2 profiles is
comparatively lower than that for the TPR1 profiles. Finally,
these consecutive surface reaction experiments suggest that the
amount of reduced Ni species from TPR1 that were involved in
forming Ni−Fe alloy species was higher for the NFA(500) and
NFA(700) catalysts than for the NFA(900) catalyst. They also
confirm that the active components available to perform the
Figure 8. TPSR with toluene over NA(500) and NFA catalysts: (A)
H2 produced and (B) CO produced as functions of temperature.
functions of temperature, respectively. Two gaseous products,
H2 and CO, were observed during the TPSR of toluene, along
with solid carbon that was deposited on the catalyst. This
product distribution can be explained by eqs 10 and 11:39
M* + C7H8 → M*C + H2
(10)
(11)
M*C + S−O → M* + S + CO
where M* is an active metal center (Ni0 or Ni−Fe alloy) and
S−O is a lattice oxygen species present in the NFA catalyst.
From Figure 8A, H2 was produced from toluene decomposition
296
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
according to eq 6. This decomposition started at around 470
°C in all of the profiles. However, the rate of H2 production was
higher for NFA catalysts than for the NA(500) catalyst.
Similarly, CO was also produced during TPSR of toluene
according to eq 11. The CO production, as shown in Figure 8B,
also began at around 470 °C in all of the profiles. However, the
amount of CO produced seemed to be less for the NFA(900)
and NA(500) catalysts compared with the NFA(500) and
NFA(700) catalysts. According to eq 11, CO production
depends on the amount of available lattice oxygen species,
possibly from the support in the catalyst during TPSR of
toluene. This suggests that NFA(500) and NFA(700) catalysts
have higher amounts of lattice oxygen species available to
achieve better performance in the SRT reaction than the
NFA(900) catalyst.
peak is centered at 600 °C in the TPSR profiles of the
NFA(700) and NFA(900) catalysts and at 650 °C in the TPSR
profile of the NFA(500) catalyst. However, the intensities of
these peaks vary. From the TPSR profile of the NFA(500)
catalyst, it is evident that maximum H2 evolution occurs at 650
°C, coinciding with the actual temperature at which the SRT
reaction was conducted in the present study. This ability to
evolve the highest quantity of H2 at 650 °C is one of the
reasons why the NFA(500) catalyst is superior over other
catalysts. Combining the results obtained from Figures 8 and 9
shows that the superiority of the NFA(500) catalyst over the
others is due to the existence of higher amounts of available
oxygen species during the SRT reaction as well as its ability to
maximize water splitting at a temperature at which the SRT
reaction is conducted (i.e., 650 °C). On the contrary, despite
the ability of the NFA(900) catalyst to split water at a lower
temperature than the others, it has a lower catalytic
performance because of the lower amount of available oxygen
species in the catalyst. As will be shown subsequently in TPO
of spent catalysts, the lower oxygen availability results in a
higher amount of graphitic carbon deposition on the NFA(900)
catalyst.
Subsequently, TPSR of water was also carried out over
reduced NA(500) and NFA catalysts. Figure 9 shows the H2
3.2.7. CO-Pulse Chemisorption. The amount of active metal
content in the freshly reduced NFA catalysts was measured
using CO pulse titration at 0 °C as reported in the literature.40
The CO uptake results presented in Table 2 were found to be
19.15, 18.23, and 15.97 μmol/g for the NFA(500), NFA(700),
and NFA(900) catalysts, respectively. Since Ni is the main
active component in the present steam reforming reaction, an
intrinsic rate per metal site (turnover frequency, TOF) for the
NFA catalysts during the SRT reaction was evaluated by
considering the toluene conversions shown in Table 1 as well as
the CO uptake results presented in Table 2. The TOF values of
the NFA(500), NFA(700) and NFA(900) catalysts are 4.4, 3.9,
and 3.1 s−1, respectively. From these TOF values, it is clearly
observed that the NFA(500) catalyst has a higher TOF than
the other NFA catalysts. On the other hand, dispersion values
of 1.12, 1.07, and 0.93% were obtained for the NFA(500),
NFA(700), and NFA(900) catalysts, respectively. Thus, the
TOF value was highest for NFA(500) catalyst even though it
has the highest CO consumption value among all the catalysts.
This result affirms that the better performance of NFA(500)
can be attributed to the presence of a higher amount of surface
active metal species.
3.2.8. XPS Analysis of Reduced and Spent Catalysts. A
comparative XPS study was done in order to understand the
surface species present in the freshly reduced and spent NFA
catalysts. Figure 10 displays the binding energies (BEs) of Ni
2p3/2 species present in the NFA catalysts. Generally, the BE of
metallic Ni 2p is 852.4 0.4 eV, the BE of Ni 2p in NiO is 854
0.4 eV, the BE of Ni 2p in NiFe2O4 is 855 0.4 eV, and the
BE of Ni 2p in NiAl2O4 is around 857 0.4 eV.41 Figure 10A
shows Ni 2p BEs present in reduced NFA catalysts due to the
presence of Ni0 (∼852.2 eV) and Ni2+ (∼855 eV) species,
where the Ni2+ species interacts with Fe species present in the
NFA(x) catalysts. According to Figure 10A, moving from the
profile of NFA(500) to that of NFA(900), both BE values of Ni
2p shifted slightly to lower values. This indicates that Ni species
present in NFA(900) catalyst have weaker interactions with Fe
species when compared with those in the NFA(700) and
NFA(500) catalysts, affirming the observations in the TPR
profiles presented in Figure 6. Next, Ni 2p BEs of spent
NFA(x) catalysts are displayed in Figure 10B, which shows the
Figure 9. H2 produced during TPSR with water over NA(500) and
NFA catalysts.
evolution as a function of temperature during TPSR with water
over the NA(500) and NFA(x) catalysts. In the TPSR with
water experiments, only H2 evolved, and this can be explained
according to eqs 12 and 13:
M* + H2O → H2 + M*O
(12)
(13)
M*O + S → M* + S−O
where M* is an active metal center (Ni0 or Ni−Fe alloy) and S
is an oxygen-vacant center in the support.
Generally, in the NA(500) catalyst, only metallic Ni species
possess available active sites to activate water molecules.
However, in the NFA catalysts, in addition to metallic Ni, Ni−
Fe alloy species also exist to activate water molecules in order
to generate H2 gas. In Figure 9, NA(500) shows a single H2
evolution peak starting from 310 °C and centered at 370 °C.
This H2 evolution peak is solely due to active sites found on
metallic Ni. A similar H2 evolution peak can also be observed in
all of the other profiles shown in Figure 9. However, the
intensities of those H2 evolution peaks differ from each other,
and it is also observed that this H2 evolution peak shifts to
higher temperature in going from NFA(900) to NFA(500). In
addition, a high-temperature H2 peak is also observed in the
TPSR profiles of all NFA catalysts starting from 450 °C; this
297
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
Figure 11. Fe 2p binding energies of (A) freshly reduced and (B)
spent NFA catalysts.
Figure 10. Ni 2p binding energies of (A) freshly reduced and (B)
spent NFA catalysts.
Al2O3) phase observed in the XRD pattern of the NFA(900)
catalyst (Figure 5). Next, Fe 2p BEs of spent NFA catalysts are
displayed in Figure 11B. The BE values at around 710 0.2 eV
correspond to Fe species in FeO form, while BE values at 712
0.2 eV can be attributed to Fe species present in Fe3O4 form
as well as Fe species that are interacted with Ni species. It is
also evident that this BE value is shifted to a slightly higher
value for the NFA(900) catalyst compared with the other
catalysts. This is due to the stronger interaction of Fe species
with Al species in the catalytic support, resulting in a shift
similar to that previously observed in Figure 11A. On
comparison with Figure 11A, the absence of Fe 2p BE values
around 707 eV in Figure 11B, corresponding to Fe0 species,
suggests the exhibition of redox behavior. During the SRT
process, the Fe0 species present in reduced NFA catalysts are
partially oxidized to form intermediary FeO species, which have
a better redox nature than both Fe0 and Fe3O4 species. This
redox behavior enables Fe species to easily shuffle between
their oxidation states to enhance the reforming process by
providing necessary oxygen species to active Ni species.
presence of two kinds of Ni species, namely, Ni0 (BE at 852.8
eV) and Ni2+ (BE at 856.3 eV). All of the Ni 2p BEs of spent
catalysts were shifted to higher values compared with the Ni 2p
BE values observed in Figure 10A. This shift to higher BE
values in all of the spent NFA catalysts indicates that the
interactions between Ni species and support species were
significantly enhanced during the reforming process. This
hypothesis was further supported by the XRD patterns of spent
NFA catalysts as presented in Figure 5B.
Figure 11 displays BEs of Fe 2p3/2 species present in reduced
and spent NFA catalysts. Unlike Ni species, Fe species in NFA
catalysts can exist in multiple oxidation states ranging from +3
to 0, and thus, it can be a challenge to identify the exact
oxidation states of Fe species present in NFA catalysts. The
following interpretation of the BEs of Fe 2p in Figure 11 was
derived after consideration of numerous literature reports.
Figure 11A, which depicts Fe 2p BEs of reduced NFA catalysts,
shows Fe 2p BE values of 707 0.2 and 711 0.2 eV, which
correspond to Fe0 species and Fe species in Fe3O4 form,
respectively. However, according to the literature,41 the latter
Fe 2p BE also corresponds to Fe species that have interacted
with Ni species in NFA(x) catalysts. In addition, compared
with other catalysts, the shift to higher BE values of around 711
eV in the NFA(900) catalyst is possibly due to stronger
interactions of Fe species with Al species. This was further
confirmed through the presence of the corundum (Fe2O3−
The molar composition of various surface species present in
freshly reduced and spent NFA catalysts are presented in Table
4. It is noteworthy to mention here that the bulk compositions
of all of the NFA catalysts were similar. However, it can be
observed from Table 4 that the Fe/Ni molar ratios on the
surfaces of the reduced NFA(500), NFA(700), and NFA(900)
catalysts were 1.56, 0.74, and 0.38 respectively, suggesting that
298
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
Table 4. Surface Compositions of Elements Present in Reduced and Spent NFA Catalysts
surface composition (%)
catalyst
form
Ni0
Ni2+
Fe0
Fex+
Al
O
Fe/Ni surface atomic ratio
NFA(500)
NFA(500)
NFA(700)
NFA(700)
NFA(900)
NFA(900)
reduced
spent
0.27
0.20
0.44
0.40
3.96
0.48
1.02
1.07
2.75
1.65
9.84
2.08
0.77
1.25
6.85
1.33
6.69
3.47
5.32
35.46
59.59
30.29
60.49
23.62
65.00
65.21
31.92
64.14
30.73
57.27
27.11
1.56
5.39
0.74
3.20
0.38
2.07
−
1.04
reduced
spent
−
1.84
reduced
spent
−
the reduced NFA(500) catalyst is rich with surface Fe species.
On the other hand, the Fe/Ni molar ratio of spent NFA
catalysts suggests that during the course of the SRT reaction all
of the catalysts were enriched with Fe species. Furthermore, it
can be observed that there was a significant increase in the
surface Al species on the spent NFA catalysts compared with
the reduced NFA catalysts. However, this increment is
considerably lower for the NFA(500) catalyst than for the
others. In addition, the surface O species were decreased in the
spent NFA catalysts compared with their corresponding
reduced ones, suggesting that this decrease in surface oxygen
species is possibly due to their involvement in the steam
reforming process. It is well-known that metallic Ni catalysts are
widely explored for steam reforming of hydrocarbons because
metallic Ni has a strong ability to activate C−H and C−C
bonds in the hydrocarbon molecules on the Ni metal surface.42
According to Table 4, the amount of surface metallic Ni species
is lower in the spent NFA catalysts than in the reduced
catalysts. This decrease in metallic Ni species is possibly due to
the increase in surface Fe species that can encapsulate metallic
Ni species during SRT process with time on stream.
Additionally, the deposition of carbon can also be responsible
for the decrease of metallic Ni species during the SRT process,
suggesting that the deactivation behavior of all NFA catalysts
with time on stream during the SRT process (Figure 1) is
primarily due to the decrease in the amount of metallic Ni
species. Finally, the better catalytic performance of NFA(500)
catalysts is attributed to the presence of strong interactions
between Ni and Fe species to form Ni−Fe alloy particles that
are rich in surface Fe species. This extra Fe species are mostly
partially reduced Fe2+ species, which are highly redox-active in
nature.
deposition rate could be one of the reasons accounting for its
lower SRT performance. According to the literature,43 the
deposited carbon can encapsulate the active metal components
during the SRT reaction, thereby leading to deactivation of the
catalyst by decreasing the number of active metal sites
presented by the catalyst. Despite having a higher carbon
formation rate than the NFA(700) catalyst, the NFA(500)
catalyst achieved more superior catalytic performance than the
NFA(700) catalyst. The reason for this can be explained on the
basis of the nature of the carbon species deposited during the
SRT reaction, as presented in the next section.
3.2.10. TPO of Spent Catalysts. Figure 12 displays the TPO
profiles (CO2 emission plotted against the oxidation temper-
3.2.9. TGA of Spent Catalysts. The amounts of carbon
deposited over spent NFA catalysts after SRT at 650 °C for 6 h
were measured using TGA up to 1000 °C in air, and the results
are presented in Table 5. According to Table 5, the carbon
Figure 12. TPO profiles of spent NFA catalysts. Reaction conditions:
toluene = 188 μmol min−1; steam = 4444 μmol min−1; He = 5357
μmol min−1; W = 30 mg; reaction temperature = 650 °C; reaction
time = 6 h.
deposition rates are 21.2, 18.3, and 48.9 mgC gcat h−1 for the
−1
NFA(500), NFA(700), and NFA(900) catalysts, respectively.
Overall, the NFA(900) catalyst has the highest carbon
deposition rate among all of the catalysts. The high carbon
ature) of carbon deposited on spent NFA catalysts after SRT at
650 °C for 6 h. The oxidation of carbonaceous species starts at
around 400 °C for almost all of the catalysts. The asymmetric
shape of the CO2 peaks for all of the catalysts implies the
presence of several different carbonaceous species. Generally,
unsaturated hydrocarbons such as olefins and graphitic and
filamentous carbon tend to oxidize at higher temperatures
compared with hydrocarbons with saturated C−C bonds such
as paraffinic and amorphous carbon.44 From the profiles, both
amorphous carbons (Toxidation < 650 °C) and graphitic/
filamentous carbon (Toxidation > 650 °C) can be found in all
of the spent catalysts. According to Figure 12, the profile of the
NFA(900) catalyst has a single oxidation peak centered at Tmax
≈ 710 °C with most of the carbons being oxidized after 650 °C,
Table 5. Carbon Deposition Rates for Spent NFA(x)
a
Catalysts As Determined by TGA Analysis
−1
catalyst
carbon deposition rate (mgC gcat h−1
)
NFA(500)
NFA(700)
NFA(900)
21.2
18.3
48.9
a
Reaction conditions: reaction temperature = 650 °C; reaction time =
6 h; toluene = 188 μmol min−1; steam = 4444 μmol min−1; He = 5357
μmol min−1; W = 30 mg.
299
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
(6) De Lasa, H.; Salaices, E.; Mazumder, J.; Lucky, R. Chem. Rev.
2011, 111, 5404−5433.
whereas for the NFA(500) and NFA(700) catalysts, the profiles
show two stages of oxidation behavior with a low-temperature
peak centered at Tmax ≈ 525 and 535 °C, respectively, as well as
a high-temperature peak centered at Tmax ≈ 690 and 680 °C,
respectively. Furthermore, most of the deposited carbon was
oxidized below 650 °C for both the NFA(500) and NFA(700)
catalysts. This could be one of the reasons for the lower amount
of carbon deposited on NFA(500) and NFA(700) catalysts
compared with the NFA(900) catalyst during the SRT reaction.
According to other studies,43 this amorphous carbon can be
easily reformed compared with graphitic carbon during the
SRT process at 650 °C. On the contrary, the superior SRT
performance of NFA(500) over NFA(700) catalyst despite its
higher carbon formation rate can possibly be attributed to the
fact that its low-temperature oxidation peak is centered at a
lower temperature than that of NFA(700) catalyst.
(7) Swierczynski, D.; Libs, S.; Courson, C.; Kiennemann, A. Appl.
Catal., B 2007, 74, 211−222.
(8) Polychronopoulou, K.; Giannakopoulos, K.; Efstathiou, A. M.
Appl. Catal., B 2012, 111, 360−375.
(9) Guell, B. M.; Babich, I.; Nichols, K. P.; Gardeniers, J. G. E.;
Lefferts, L.; Seshan, K. Appl. Catal., B 2009, 90, 38−44.
(10) Kimura, T.; Miyazawa, T.; Nishikawa, J.; Kado, S.; Okumura, K.;
Miyao, T.; Naito, S.; Kunimori, K.; Tomishige, K. Appl. Catal., B 2006,
68, 160−170.
(11) Wang, L.; Hisada, Y.; Koike, M.; Li, D.; Watanabe, H.;
Nakagawa, Y.; Tomishige, K. Appl. Catal., B 2012, 121−122, 95−104.
(12) Guan, G.; Chen, G.; Kasai, Y.; Lim, E. W. C.; Hao, X.;
Kaewpanha, M.; Abuliti, A.; Fushimi, C.; Tsutsumi, A. Appl. Catal., B
2012, 115−116, 159−168.
(13) Ammendola, P.; Cammisa, E.; Chirone, R.; Lisi, L.; Ruoppolo,
G. Appl. Catal., B 2012, 113−114, 11−18.
(14) Virginie, M.; Courson, C.; Niznansky, D.; Chaoui, N.;
Kiennemann, A. Appl. Catal., B 2010, 101, 90−100.
(15) Kuhn, J. N.; Zhao, Z.; Felix, L. G.; Slimane, R. B.; Choi, C. W.;
Ozkan, U. S. Appl. Catal., B 2008, 81, 14−26.
4. CONCLUSIONS
Iron−alumina-supported nickel−iron alloy catalysts were
explored for steam reforming of toluene as a biomass tar
model compound. The NFA(500) catalyst showed the best
steam reforming performance in terms of higher catalytic
activity and stability for 26 h reaction time with a H2/CO value
of 4.5. The superior catalytic performance of NFA(500) is
mainly due to the presence of a higher amount of surface active
metal species, Fe-rich Ni−Fe alloy particles, strong metal−
support interactions, and a relatively low carbon deposition
rate. The high catalyst surface area and higher amount of
available lattice oxygen species also play important roles in
promoting the reforming activity of the NFA(500) catalyst over
the others. The synergy between Ni and Fe atoms is achieved
by forming Fe-rich Ni−Fe alloy particles, which are crucial for
the high activity of the NFA(500) catalyst. In addition, the
strong interaction between metal and support on the
NFA(500) catalyst can prevent metal sintering, thus achieving
high catalytic stability. Finally, this NFA(500) catalyst has great
potential for application in the steam reforming of biomass tar.
(16) Mei, D.; Lebarbier, V. M.; Rousseau, R.; Glezakou, V. A.;
Albrecht, K. O.; Kovarik, L.; Flake, M.; Dagle, R. A. ACS Catal. 2013,
3, 1133−1143.
(17) Nakamura, K.; Miyazawa, T.; Sakurai, T.; Miyao, T.; Naito, S.;
Begum, N.; Kunimori, K.; Tomishige, K. Appl. Catal., B 2009, 86, 36−
44.
(18) Li, D.; Wang, L.; Koike, M.; Nakagawa, Y.; Tomishige, K. Appl.
Catal., B 2011, 102, 528−538.
(19) Maluf, S. S.; Assaf, E. M. Fuel 2009, 88, 1547−1553.
(20) Wu, P.; Li, X.; Ji, S.; Lang, B.; Habimana, F.; Li, C. Catal. Today
2009, 146, 82−86.
(21) Miyazawa, T.; Kimura, T.; Nishikawa, J.; Kado, S.; Kunimori, K.;
Tomishige, K. Catal. Today 2006, 115, 254−262.
(22) Wang, L.; Li, D.; Koike, M.; Koso, S.; Nakagawa, Y.; Xu, Y.;
Tomishige, K. Appl. Catal., A 2011, 392, 248−255.
(23) Wang, L.; Li, D.; Koike, M.; Watanabe, H.; Xu, Y.; Nakagawa,
Y.; Tomishige, K. Fuel 2013, 112, 654−661.
(24) Koike, M.; Li, D.; Nakagawa, Y.; Tomishige, K. ChemSusChem
2012, 5, 2312−2314.
(25) Reed, T. B. Free Energy of Formation of Binary Compounds; MIT
Press: Cambridge, MA, 1971; pp 66−70.
AUTHOR INFORMATION
Corresponding Author
■
(26) Ishida, M.; Takeshita, K.; Suzuki, K.; Ohba, T. Energy Fuels
2005, 19, 2514−2518.
*Address: Department of Chemical and Biomolecular Engi-
neering, National University of Singapore, Singapore 119260,
Republic of Singapore. Tel.: +65 6516 6312. Fax: +65 6779
1936. E-mail: chekawis@nus.edu.sg.
(27) Duprez, D. Appl. Catal., A 1992, 82, 111−157.
(28) Oemar, U.; Ang, P. S.; Hidajat, K.; Kawi, S. Int. J. Hydrogen
Energy 2013, 38, 5525−5534.
(29) Ashok, J.; Kawi, S. Int. J. Hydrogen Energy 2013, 38, 13938−
13949.
Notes
The authors declare no competing financial interest.
(30) Kechagiopoulos, P. N.; Voutetakis, S. S.; Lemonidou, A. A.;
Vasalos, I. A. Energy Fuels 2006, 20, 2155−2163.
(31) Domine, M. E.; Iojoiu, E. E.; Davidian, T.; Guilhaume, N.;
Mirodatos, C. Catal. Today 2008, 133−135, 565−573.
(32) Shi, Z.; Zhang, Z.; Fan, R.; Gao, M.; Guo, J. J. Inorg. Organomet.
Polym. 2011, 21, 836−840.
ACKNOWLEDGMENTS
■
The authors gratefully thank National University of Singapore
and the National Environmental Agency (NEA-ETRP Grant
1002114 and RP 279-000-333-490) for generously supporting
this work.
(33) Chen, W.; Ding, Y.; Song, X.; Wang, T.; Luo, H. Appl. Catal., A
2011, 407, 231−237.
(34) Gheisari, K.; Javadpour, S.; Oh, J. T.; Ghaffari, M. J. Alloys
Compd. 2009, 472, 416−420.
REFERENCES
■
(35) Kuhrt, C.; Schultz, L. J. Appl. Phys. 1993, 73, 1975−1980.
(36) Kidambi, P. R.; Cleeton, J. P. E.; Scott, S. A.; Dennis, J. S.; Bohn,
C. D. Energy Fuels 2012, 26, 603−617.
(1) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044−
4098.
(2) Hu, X.; Lu, G. Appl. Catal., B 2009, 88, 376−385.
(3) Dauenhauer, P. J.; Dreyer, B. J.; Degenstein, N. J.; Schmidt, L. D.
Angew. Chem. 2007, 119, 5968−5971.
(4) Serrano, D. P.; Aguado, J.; Escola, J. M. ACS Catal. 2012, 2,
1924−1941.
(5) Dapsens, P. Y.; Mondelli, C.; Ramirez, J. P. ACS Catal. 2012, 2,
1487−1499.
(37) Richardson, J. T.; Turk, B.; Twigg, M. V. Appl. Catal., A 1996,
148, 97−112.
(38) Tian, D.; Liu, Z.; Li, D.; Shi, H.; Pan, W.; Cheng, Y. Fuel 2013,
104, 224−229.
(39) Lamacz, A.; Krzton, A.; Mariadassou, G. D. Catal. Today 2011,
176, 347−351.
300
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301

ACS Catalysis
Research Article
(40) Oyama, S. T.; Zhao, H.; Freund, H. J.; Asakura, K.; Wlodarczyk,
R.; Sierka, M. J. Catal. 2012, 285, 1−5.
(41) Handbook of X-ray Photoelectron Spectroscopy; Wagner, C. D.,
Riggs, W. M., Davis, L. E., Moulder, J. F., Muilenberg, G. E., Eds.;
Perkin-Elmer Corporation: Eden Prairie, MN, 1979; pp 76−80.
(42) Ahmed, S.; Aitani, A.; Rahman, F.; Al-Dawood, A.; Al-Muhaish,
F. Appl. Catal., A 2009, 359, 1−24.
(43) Djinovic, P.; Crnivec, I. G. O.; Erjavec, B.; Pintar, A. Appl. Catal.,
B 2012, 125, 259−270.
(44) Gould, B. D.; Chen, X.; Schwank, J. W. Appl. Catal., A 2008,
334, 277−290.
301
dx.doi.org/10.1021/cs400621p | ACS Catal. 2014, 4, 289−301