Effect of lanthanide promotion on catalytic performance of sol-gel Ni/Al2O3 catalysts in steam reforming of propane

Article abstract of DOI:10.1016/j.molcata.2005.07.017

The effect of lanthanide elements (La, Ce, and Yb) on the catalytic behavior of sol-gel Ni/Al₂O₃ catalysts in propane steam reforming was investigated. Steady-state reaction experiments show that the addition of small amounts (2 wt.%

Full text of DOI:10.1016/j.molcata.2005.07.017

Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
Effect of lanthanide promotion on catalytic performance of sol–gel
Ni/Al O catalysts in steam reforming of propane
2
3
∗
Sittichai Natesakhawat, Okan Oktar, Umit S. Ozkan
The Ohio State University, Department of Chemical and Biomolecular Engineering, 140 W, 19th Avenue, Columbus, OH 43210, USA
Received 26 April 2005; received in revised form 7 July 2005; accepted 7 July 2005
Available online 16 August 2005
Abstract
The effect of lanthanide elements (La, Ce, and Yb) on the catalytic behavior of sol–gel Ni/Al
2
3
O catalysts in propane steam reforming was
investigated. Steady-state reaction experiments show that the addition of small amounts (2 wt.%) of lanthanide elements improves the catalytic
activity and stability significantly. The changes in reaction performance are related to catalyst reducibility, nickel surface area and resistance
to deactivation. Temperature-programmed reduction (TPR) and X-ray photoelectron spectroscopy (XPS) characterization results reveal that
the presence of lanthanide elements enhances the catalyst reducibility; the positive effect is most evident with 20% Ni–2% Ce/Al catalysts.
2
O
3
H
2
chemisorption data indicate that 20% Ni–2% Ce/Al catalysts exhibit larger nickel surface area compared to 20% Ni/Al O catalysts.
2
O
3
2 3
Nickel surface area of sol–gel catalysts is also found to be strongly dependent on calcination and reduction temperatures. In situ DRIFTS
experiments coupled with TPD and TPReaction suggest that formate and carbonate species are formed upon the reaction of adsorbed propane
and water on reduced catalysts.
©
2005 Elsevier B.V. All rights reserved.
Keywords: Steam reforming; Propane; Ni/Al2O3 catalysts; Sol–gel; Lanthanides; H2 chemisorption; DRIFTS
1
. Introduction
steam reforming processes, suffer from catalyst deactivation
by coke formation more severely when higher hydrocarbons
are reformed at low steam/carbon ratios.
Hydrogen is the ultimate clean energy carrier. When it is
combusted, heat and water are the only products. When it
is used as a fuel for fuel cells, then it has the potential of
providing much higher efficiencies compared to combustion.
Thus, hydrogen offers a potentially non-polluting and effi-
cient fuel for today’s rising energy demands [1]. However, in
the transition period before renewable sources of hydrogen
can be developed, production of hydrogen from fossil fuels
remains an important technology. Currently, steam reform-
ing of natural gas, which is composed mainly of methane,
is used to produce most of hydrogen in US and about half
of the world’s hydrogen supply. Naphtha fractions with a
Although supported precious metals such as Pd, Pt, and Rh
have been reported to be active and stable for steam reforming
of hydrocarbons [3–8], the cost of the precious metals is still a
major drawback. The low-cost and long-proven performance
of Ni-based catalysts, therefore, warrant the effort to optimize
these catalysts for steam reforming applications, particularly
fuel processing technology from existing liquid fuels such as
gasoline and diesel.
There is evidence in the literature pointing to an improve-
ment in catalytic activity of Ni-based catalysts in hydrocar-
bon steam reforming with the addition of lanthanide ele-
ments [9–11]. Zhuang et al. [9] investigated the effect of
cerium oxide as a promoter in supported Ni catalysts for
methane steam reforming. It showed a beneficial effect by
not only decreasing the carbon deposition rate, but also
increasing the steam reforming activity. The authors believed
that cerium oxide accelerated the reaction of steam with
◦
final boiling point of less than 220 C are also considered as
suitable feedstocks for hydrogen production [2]. However,
current Ni-based catalysts, which are used for natural gas
∗
Corresponding author. Tel.: +1 614 292 6623; fax: +1 614 292 3769.
E-mail address: ozkan.1@osu.edu (U.S. Ozkan).
1
381-1169/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2005.07.017

1
34
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
adsorbed species on the nickel surface, thus decreasing the
carbon deposition as well as increasing or maintaining the
catalytic activity. Su and Guo [10] also reported an improve-
ment in catalytic activity and resistance to Ni sintering of
Ni/Al2O3 catalysts doped with rare earth oxides in methane
steam reforming. The growth of Ni particles and the for-
mation of inactive NiO and NiAl2O4 phases were sup-
pressed by the addition of rare earth oxides. Moreover, the
oxides of heavy rare earth elements (Gd, Er, Dy) exhib-
ited a more pronounced effect than those of the light ones
parameters were varied during the synthesis: nickel con-
tent = 10–25 wt.%, lanthanide content = 2–5 wt.%, H2O/ATB
molar ratio = 3.6–5.0, and pH of the resulting gels = 4.8–8.5.
Initially, ATB was placed in ethanol and stirred. The aqueous
solutions with the desired amounts of nickel and lanthanide
were added dropwise into the suspension using a syringe
3
pump at the flow rate of 0.5 cm /min. It is noted that the aque-
ous lanthanide nitrate solution was added first followed by
that of nickel. The pH of the resulting green gel was measured
and adjusted by adding HNO3 (Fisher) or NH4OH (Fisher).
The gels were stirred for an additional 15 min and were kept
at room temperature for 30 min. The samples were then dried
(La, Pr, Nd).
The use of lanthanide-promoted Ni/Al2O3 catalysts in
◦
dry reforming of methane [12–15] and benzene hydrogena-
tion [16] has also been reported. The effects of the addition
of lanthanide oxides (La2O3 and CeO2) and preparation
procedures were investigated in the reforming of methane
with carbon dioxide over impregnated Ni/Al2O3 catalysts
overnight in the oven at 110 C to remove the remaining water
and ethanol. The dry samples were ground to a fine powder
◦
◦
andwerecalcinedunderO2 at450 C(ramprate = 10 C/min)
for 4 h.
[12]. When lanthanides were impregnated prior to nickel,
2.2. Catalyst characterization
an enhanced reducibility of nickel and a decrease in nickel
particle size were observed. However, the reforming activ-
ity was not affected to a large extent by the impregna-
tion order of nickel and lanthanides. The presence of CeO2
BET surface areas of calcined catalysts were measured
by N2 adsorption–desorption at 77 K using a Micromerit-
ics ASAP 2010 instrument. Before measurements, samples
◦
(1–5 wt.%) in impregnated Ni/␥-Al2O3 catalysts was found
were degassed under vacuum at 130 C overnight. XRD pat-
to reduce the chemical interaction between nickel and the
alumina support, resulting in an increase in reducibility and
higher nickel dispersion [13,14]. Moreover, carbon depo-
sition was greatly suppressed due to the oxygen storage
capacity of CeO2. The catalytic activity, sulfur tolerance,
and thermal stability were remarkably improved after adding
La, Sm, and Yb into Ni–B/Al2O3 catalysts for benzene
hydrogenation [16]. The positive effect of lanthanide promo-
tion was associated with easier reduction of nickel oxides,
smaller nickel particles, and larger surface area of active
nickel.
In this study, Ni/Al2O3 catalysts promoted with lan-
thanide elements (La, Ce, and Yb) were prepared by a
modified sol–gel technique and were examined in propane
steam reforming. Characterization was performed using
the BET N2 adsorption method, X-ray diffraction (XRD),
temperature-programmed reduction (TPR), X-ray photoelec-
tron spectroscopy (XPS), and H2 chemisorption. In situ
DRIFTS studies were also conducted to identify adsorbed
species under C3H8 and C3H8 + H2O flows as well as
to elucidate the reaction pathways for propane steam
reforming.
terns of calcined and reduced catalysts were obtained with
a Bruker D8 Advance X-ray diffractometer equipped with
atmosphere and temperature control stage and using Cu K␣
˚
radiation (λ = 1.542 A) operated at 40 kV and 50 mA. In this
experiment, 5% H2/N2 was used as a reducing agent. TPR
of catalysts was conducted using a laboratory-made gas flow
system described in detail elsewhere [17]. Catalyst samples
(50 mg) were placed in a 1/4 in. o.d. quartz U-tube reac-
tor. The catalysts were then re-calcined under 10% O2/He
◦
at 450 C for 1 h followed by cooling to room temperature
under Ar. The reduction was performed with 10% H2/Ar
3
(40 cm (STP)/min). The temperature of the catalysts was
◦
◦
raised using a ramp rate of 10 C/min to 900 C and held at
that level for 10 min. The effluent from the reactor was passed
through a silica gel water trap to remove moisture formed dur-
ing reduction. H2 consumption was measured using a thermal
conductivity detector (TCD) connected to a data-acquisition
computer.
XPS of reduced catalysts was performed with an AXIS
Ultra XPS spectrometer with an Al anode operated at 14 kV
and 10 mA. Reduction was performed with 20% H2/N2 at
◦
600 C for 2 h. Prior to the measurements, all reduced cat-
alysts were kept in tightly sealed reactors and were subse-
quently mounted on a sample holder with conductive tape in
a dry glovebox filled with Ar. The sample holder was then
carefully transferred to the analysis chamber of the spectrom-
eter with a controlled-atmosphere transporter. Spectra were
corrected using the C 1s signal located at 284.5 eV.
2
. Experimental
2
.1. Catalyst preparation
Catalysts were prepared using a modified sol–gel tech-
nique. Nickel and lanthanide precursors (Aldrich) used were
in nitrate form. For the aluminum precursor, aluminum tri-
H2 chemisorption of catalysts was conducted using a
Micromeritics ASAP 2010 Chemisorption System. Prior
to adsorption measurements, calcined samples (0.2 g) were
reduced under H2 at the desired reduction temperature for
sec-butoxide (ATB; Aldrich) was used. Ethanol (Alfa Aesar;
3
◦
1
30 cm ) was used as the solvent. The following sol–gel
2 h. Subsequently, the samples were cooled down to 35 C

S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
135
followed by evacuation to 10 5 mmHg. The adsorption
−
%H-product selectivity for B
ꢀ
ꢁ
isotherms were measured at equilibrium pressures between
0 and 500 mmHg. H2 uptakes at monolayer coverage of
moles of H in product B
moles of H in all H − containing products
5
=
× 100
the Ni crystallites were used to estimate Ni dispersion and
metallic surface area, assuming that each surface Ni atom
chemisorbs one hydrogen atom (H/Nisurface = 1).
DRIFTS experiments were performed using a Bruker
IFS66 instrument equipped with an MCT detector and a KBr
C-containing products are CO, CO2, and CH4. H-
containing products are H2 and CH4. Reproducibility of the
results were tested using different catalyst batches run under
identical conditions.
beamsplitter. Catalysts were reduced under 20% H2/N2 at
00 C for 2 h and then were placed in a sample cup inside
◦
6
a Spectratech diffuse reflectance cell equipped with KBr
windows and a thermocouple that allowed direct measure-
ments of the surface temperature. Spectra for each experi-
ment were averaged over 1000 scans in the mid-IR range
3
. Results and discussion
3.1. Physical properties of sol–gel catalysts
−1 −1
(400–4000 cm ) to a nominal 3 cm resolution. Prior to
The effect of synthesis parameters on the physical prop-
the collection of spectra, the catalysts were re-reduced in situ
◦
erties of sol–gel Ni/Al2O3 catalysts is reported in Table 1.
Parameters that were varied during the synthesis included Ni
content, pH of the resulting gels, and H2O/ATB molar ratio.
BET surface area of sol–gel Ni/Al2O3 catalysts is seen to
decrease with increasing Ni contents possibly due to nickel
aggregation to form larger particles on the Al2O3 support sur-
face during thermal treatment. The decrease in surface area
is even more pronounced at higher Ni contents. Furthermore,
an increase in pH and H2O/ATB ratio causes an increase in
surface area of 20 wt.% Ni/Al2O3 catalysts. It is noted that
a decrease in the pore diameter from 10.5 to 7.4 nm occurs
when H2O/ATB ratio is increased from 3.6 to 5.0. This phe-
nomenon is related to an increase in cross-linking as a result
of an increase in the extent of hydrolysis at high H2O/ATB
ratios [18].
under 20% H2/He at 450 C for 1 h. The background spectra
were taken at different temperature intervals under He. The
adsorption of C3H8 was performed at room temperature for
1
h followed by flushing under He for 30 min. The spectra
were subsequently collected at different temperatures up to
◦
4
50 C.
2
.3. Reaction studies
The steady-state reaction experiments were performed
using a stainless steel fixed-bed flow reactor (1/4 in. o.d.).
Catalysts were compared in propane steam reforming using
2
◦
equal surface area loading (9 m ) in the reactor at 400–550 C
and H2O/C = 1.3. The feed percentages for these experiments
were C3H8/H2O/N2 = 1/4/95 (feed flow rate = 300–400 cm3
BET surface areas of lanthanide-promoted sol–gel
Ni/Al2O3 catalysts prepared at different lanthanide contents
are listed in Table 2. In general, catalysts have a lower sur-
face area than the pure support. At lower loading levels, the
lanthanide-promoted catalysts exhibit a higher surface area
than the unpromoted catalyst, but a loss in surface area with
higher lanthanide content is observed with Ni–Ce/Al2O3 and
Ni–Yb/Al2O3 catalysts possibly due to pore blockage by
excess metal. In addition, an increase in calcination temper-
(
STP)/min). Unless otherwise noted, catalysts were reduced
3
◦
in situ under 20% H2/N2 (50 cm (STP)/min) at 600 C for
h prior to the reaction. The effluent from the reactor was
2
analyzed using an automated Shimadzu GC-14A equipped
with FID and TCD detectors. Separations were performed
under Ar using two columns: Porapak Q (12 ft × 1/8 in. SS,
8
6
0/100 mesh) and Molecular Sieve 13X (5 ft × 1/8 in. SS,
0/80 mesh). A GOW-MAC 069-50 ruthenium methanizer
◦
operated at 350 C was used with the FID for the sensitive
determination of CO as low as 10 ppm. Reaction data were
taken after steady state was reached. Conversion, yield, and
selectivities were defined as follows:
◦
ature from 450 to 600 C causes a decrease in surface area
Table 1
ꢀ
ꢁ
moles of carbon converted
BET surface area of various sol–gel Ni/Al2O3 catalysts
%
C3H8 conversion =
2
total moles of carbon in the feed
Synthesis parameters
Ni content (wt%)
Surface area (m /g)
×
100
pH
H2O/ATB ratio
1
15
0
4.8
4.8
4.8
4.8
4.8
7.0
8.8
4.8
4.8
4.8
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
4.0
5.0
248
244
209
180
209
215
259
209
203
248
ꢀ
2
ꢁ
× moles of H2 produced
%
H2 yield =
× 100
2
2
2
2
2
2
0
5
0
0
0
0
total moles of hydrogen in the feed
%
C-product selectivity for A
ꢀ
ꢁ
moles of C in product A
=
moles of C in all carbon − containing products
20
20
×
100

1
36
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
Table 2
BET surface area of lanthanide-promoted sol–gel Ni/Al2O3 catalysts cal-
◦
cined at 450 C
Catalyst
Surface area
2
(m /g)
2
2
2
2
2
2
2
0% Ni/Al2O3
209
286
300
304
228
278
234
342
0% Ni–2% La/Al2O3
0% Ni–5% La/Al2O3
0% Ni–2% Ce/Al2O3
0% Ni–5% Ce/Al2O3
0% Ni–2% Yb/Al2O3
0% Ni–5% Yb/Al2O3
Al2O3
of Ni–Ce/Al2O3 catalysts. The surface area of the catalyst
◦
2
2
calcined at 600 C is 261 m /g compared to 304 m /g for the
◦
catalyst calcined at 450 C.
3
.2. X-ray diffraction
The XRD patterns of calcined and reduced 20% Ni–2%
Ce/Al2O3 catalysts are compared in Fig. 1. The support cal-
◦
cined at 450 C is relatively amorphous as evidenced by
broad peaks that correspond to ␥-Al2O3 (Fig. 1a). The pos-
◦
sible crystalline phases in the catalyst calcined at 450 C are
NiO, NiAl2O4, and ␥-Al2O3 (Fig. 1b). However, XRD results
are not conclusive about the presence of any Ni-containing
crystalline phases, suggesting a high-level of dispersion of
Ni. Furthermore, no cerium-related crystalline phases are
detected, suggesting that cerium species could be well dis-
persed over the alumina support or could exist within the NiO
Fig. 2. TPR profiles of various sol–gel 20% Ni/Al2O3 catalysts promoted
with 2 wt.% lanthanide elements (calcined at 450 C). The effect of calcina-
tion temperature on reducibility of 20% Ni–2% Ce/Al2O3 catalysts is shown
as an inset.
◦
3.3. Promotion effect on catalyst reducibility and nickel
surface area
matrix undetectable by XRD. After in situ reduction under
◦
5
% H2/N2 at 600 C, significant amounts of metallic nickel
were formed as confirmed by the diffraction lines of a metal-
3.3.1. Temperature-programmed reduction
◦
◦
lic nickel phase [(1 1 1), (2 0 0), (2 2 0)] at 2θ = 44.5 , 51.8 ,
TPR experiments on lanthanide-promoted sol–gel
Ni/Al2O3 catalysts were performed using 10% H2/Ar as
a reducing agent. As shown in Fig. 2, the TPR profiles
exhibit two peaks, representing two different reduction sites
on all catalysts. A small low-temperature (LT) reduction
◦
and 76.4 , respectively (Fig. 1c).
◦
peak around 400 C could be assigned to the reduction of
trace amounts of NiO clusters. A broad high-temperature
◦
(
HT) reduction peak above 500 C could be due to the
reduction of a well-dispersed and possibly amorphous NiO
phase that interacts strongly with the alumina support.
◦
For lanthanide-promoted catalysts calcined at 450 C, the
position of the HT peak maximum shifts toward lower
temperatures compared to that of the unpromoted catalyst.
This suggests that the presence of lanthanide elements results
in a greater reduction of nickel species; the promotion effect
is most evident on the Ni–Ce/Al2O3 catalyst. This could be
due to a decreased interaction of Ni with the alumina support
when a lanthanide element is present as a promoter. It is
noted that the introduction of lanthanides to nickel catalysts
showed no additional reduction peak under these conditions.
In general, calcination temperature plays a major role
on the strength of metal oxide and support interaction for
◦
Fig. 1. XRD patterns of (a) sol–gel Al2O3 calcined at 450 C (b) 20% Ni–2%
◦
Ce/Al2O3 catalyst calcined at 450 C (c) 20% Ni–2% Ce/Al2O3 catalyst
◦
reduced at 600 C.

S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
137
supported metal catalysts, which in turn strongly affects the
Table 3
Comparison of the ANi /(ANi + ANi2+) XPS ratio of the catalysts after
0
0
degree of reduction. The effect of calcination temperature
on the catalyst reducibility of Ni–Ce/Al2O3 was also inves-
tigated. An increase in calcination temperature from 450
◦
reduction at 600 C
ANi /(ANi + ANi2+
0
0
)
Catalyst
◦
to 600 C results in a shift of the HT peak maximum to a
20% Ni/Al2O3
20% Ni–2% La/Al O
0.51
0.51
0.55
0.54
◦
2
3
higher temperature around 680 C as displayed in an inset
Fig. 2). Moreover, the reduction peak at 400 C is more
2
2
0% Ni–2% Ce/Al2O3
0% Ni–2% Yb/Al2O3
◦
(
◦
prominent with the catalyst calcined at 600 C implying that
the formation of crystalline NiO could be enhanced as the cal-
cination temperature is increased. TPR results show that the
interaction of Ni species with the alumina support becomes
stronger at higher calcination temperatures, thus suppressing
the reduction of Ni species to metallic Ni [19,20]. This could
alsobeindicativeoftheformationofanickelaluminatespinel
at higher calcination temperatures.
870.2 eV, respectively [21–24]. The Ni 2p3/2 signal corre-
sponding to Ni2+ is accompanied by a broader peak located
at 860.8 eV due to a strong shake-up process [25]. Ni 2p spec-
tra indicate that the reduction of nickel species to the metallic
state in the sol–gel catalysts is not complete due to a strong
interaction of nickel oxide with the alumina support. It is
noted that the satellite peak at 860.8 eV still remains in XPS
spectra following TPR experiments (not shown). This sug-
gests that the signal of Ni2+ species could be due to NiAl2O4,
which is more difficult to reduce. However, the possibility of
a NiO phase cannot be excluded since the Ni 2p binding ener-
gies for these two species are very close together. In addition,
all catalysts show similar Ni 2p spectra, suggesting that nickel
species in these catalysts exist in a similar coordination envi-
ronment.
3
.3.2. X-ray photoelectron spectroscopy
XPS was further conducted to study the nature of surface
nickel species over the catalysts after reduction under 20%
◦
H2/N2 at 600 C for 2 h. The catalysts were transferred to the
XPS chamber without exposing them to the atmosphere. The
Ni 2p spectra of reduced lanthanide-promoted catalysts are
shown in Fig. 3. The Ni 2p spectra of the catalysts exhibit
features that are assigned to Ni2+ with Ni 2p3/2 and Ni 2p1/2
binding energies at 855.8 and 873.3 eV, respectively and Ni0
Deconvolution of the Ni 2p spectra was performed by
Gaussian–Lorentzian curve-fitting method to determine the
with Ni 2p3/2 and Ni 2p1/2 binding energies at 852.7 and
peak areas for obtaining relative surface concentrations of
nickel species. The ratio of peak areas ANi0/(ANi0 + A
2+
)
Ni
is used to illustrate the degree of reducibility. As displayed
in Table 3, an increase in the ANi0/(ANi0 + A
2+
) ratio for
Ni
lanthanide-promoted Ni/Al2O3 catalysts is seen after reduc-
tion compared to that of the unpromoted catalyst, indicative
of an enhancement in the catalyst reducibility with lanthanide
promotion. This is in agreement with TPR results obtained
with the lanthanide-promoted catalysts, which have shown
a shift in the maximum of the high-temperature peak corre-
sponding to the reduction of NiO toward lower temperatures.
3.3.3. H2 chemisorption
H2 chemisorption was performed to determine nickel dis-
persion and nickel surface area of the catalysts. Unless oth-
erwise noted, the reduction was conducted under H2 (UHP,
◦
9
9.999%) at 600 C for 2 h. The catalysts were then evacu-
◦
ated at 610 C to remove adsorbed hydrogen from the sur-
face and cooled to 35 C. The first analysis was performed
◦
to measure both strong and weak sorption in combination
whereas a repeat analysis was conducted to measure only
weak (reversible) hydrogen uptake. The difference in the
volume of hydrogen adsorbed from the first and repeat anal-
yses was then plotted versus the hydrogen pressure and was
extrapolated to zero pressure to obtain hydrogen uptake due
to chemisorption on nickel. The nickel dispersion and nickel
surface area were calculated assuming that one hydrogen
atom is adsorbed on one surface nickel atom [26–28].
As listed in Table 4, the nickel dispersion is observed to
be quite poor (2.3–6.4%) for all catalysts. The apparent low
Fig. 3. Ni 2p X-ray photoelectron spectra of lanthanide-promoted sol–gel
◦
2
0% Ni/Al2O3 catalysts reduced at 600 C.

1
38
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
Table 4
Table 5
H2 chemisorption data on various lanthanide-promoted sol–gel Ni/Al2O3
catalysts reduced under H2 at 600 C for 2 h
Effect of calcination (Tcal) and reduction (Tred) temperatures on nickel
dispersion and surface area of sol–gel 20% Ni/Al2O3 and 20% Ni–2%
Ce/Al2O3 catalysts
◦
2
b
Catalyst
Percentage Da
MSA (m /g)
Catalyst
Tcal
Tred
( C)
Percentage
D
MSA
(m /g)
2
2
2
2
2
2
2
0% Ni/Al2O3
4.9
3.7
3.3
6.4
2.3
5.0
2.8
6.6
4.9
4.5
8.5
3.1
6.6
3.8
◦
◦
a
2
b
(
C)
0% Ni–2% La/Al2O3
0% Ni–5% La/Al2O3
0% Ni–2% Ce/Al2O3
0% Ni–5% Ce/Al2O3
0% Ni–2% Yb/Al2O3
0% Ni–5% Yb/Al2O3
20% Ni/Al2O3
450
500
600
700
3.7
4.9
4.8
5.0
6.6
6.4
600
500
3.3
5.2
5.4
4.4
7.0
7.2
600
a
Percentage nickel dispersion.
Nickel surface area, m /g.
700
b
2
20% Ni–2% Ce/Al2O3
450
500
5.2
6.4
6.2
6.9
8.5
8.3
6
7
00
00
nickel dispersion on the high surface area of sol–gel alumina
could be due to the formation of NiAl2O4, which is rather
difficult to reduce [29,30]. The amount of H2 chemisorbed on
pure sol–gel alumina is negligibly small, thus no correction
on H2 uptake values was made. The nickel dispersion and
nickel surface area increase in the following order:
600
500
00
700
3.8
6.2
6.7
5.0
8.2
8.9
6
a
Percentage nickel dispersion.
Nickel surface area, m /g.
b
2
2
0% Ni–2% La/Al2O3 < 20% Ni/Al2O3
20%Ni − 2%Yb/Al2O3 < 20%Ni–2% Ce/Al2O3
Additionally, the nickel dispersion and nickel surface area
catalysts exhibit higher nickel dispersion and larger nickel
◦
surface area at the reduction temperature of 700 C compared
≈
◦
to those calcined at 450 C; and (iii) regardless of calcination
and reduction temperatures, Ni–Ce/Al2O3 catalysts possess
higher nickel dispersion and larger nickel surface area than
Ni/Al2O3 catalysts.
are seen to decrease at higher lanthanide content; a decrease
is more pronounced with Ni–Ce/Al2O3 and Ni–Yb/Al2O3
catalysts. This implies that higher amounts of lanthanides
could cover or block nickel sites for hydrogen adsorption.
The effect of calcination and reduction temperatures on
the nickel dispersion and nickel surface area of Ni/Al2O3
and Ni–Ce/Al2O3 catalysts is further examined. As reported
in Table 5, there are three major trends exhibited by the H2
3.4. Catalytic activity
3.4.1. Effect of synthesis parameters
The effect of catalyst synthesis parameters on the cat-
◦
alytic activity of sol–gel Ni/Al2O3 catalysts at 500 C and
◦
chemisorption data: (i) for the catalysts calcined at 450 C,
the nickel dispersion and nickel surface area increase with
increasing reduction temperature and reach a maximum at
H2O/C = 1.3 is shown in Table 6. The parameters examined
are Ni content, pH, and H2O/ATB ratio.
As reported in Table 6, C3H8 conversion and H2 yield
are seen to increase with increasing Ni content and reach a
maximum at 20 wt.%, with C3H8 conversion and H2 yield
at 53 and 37%, respectively. Further increase in Ni content
to 25 wt.% leads to a decline in catalytic activity. More-
over, the variations of pH and H2O/ATB ratio also have
a pronounced effect on catalytic activity of 20% Ni/Al2O3
◦
◦
6
00 C. Further increase in reduction temperature to 700 C
leads to a decrease in both parameters. It is conceivable that
growth of nickel crystallites becomes more prominent with
higher reduction temperatures; (ii) the nickel dispersion and
nickel surface area increase with increasing reduction tem-
◦
perature for the catalysts calcined at 600 C. In addition, the
Table 6
Propane steam reforming reaction data obtained with sol–gel Ni/Al2O3 catalysts—effect of synthesis parameters
Synthesis parameters
Ni content (wt.%)
C3H8 conversion (%)
H2 yield (%)
Selectivity (%)
CO2
pH
H2O/ATB ratio
CH4
CO
1
1
2
2
2
2
2
2
2
2
0
5
0
5
0
0
0
0
0
0
4.8
4.8
4.8
4.8
4.8
7.0
8.8
4.8
4.8
4.8
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
4.0
5.0
16
27
53
43
53
31
27
53
43
23
21
33
37
29
37
28
26
37
30
23
85
73
54
52
54
61
68
54
63
77
0
19
36
31
36
20
20
36
30
15
15
8
10
17
10
19
12
10
7
8
◦
3
2
Reaction conditions: 500 C, C3H8/H2O/N2 = 1/4/95, feed flow rate = 300 cm /min, equal surface area reactions (9 m ).

S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
Table 7
139
catalysts. The selectivities to C-containing products are also
Reaction comparison for propane steam reforming (dilution experiments) at
shown in Table 6. The C-containing products from propane
steam reforming are CO, CO2, and CH4 with CO2 being the
main one. CO2 selectivity is seen to increase with a decrease
in CH4 selectivity at lower C3H8 conversions. No CH4 for-
mation is observed at 10 wt.%.
◦
500 C
Catalyst
C3H8
H2 yield Selectivity (%)
CO2 CH4 CO
conversion (%) (%)
2
2
2
2
2
0% Ni/Al2O3
53
37
40
29
46
38
50
44
54
59
55
50
56
46
50
36
28
29
38
33
38
35
10
13
16
12
11
16
15
0% Ni–2%La/Al2O3 51
0% Ni–5%La/Al2O3 29
0% Ni–2%Ce/Al2O3 72
0% Ni–5%Ce/Al2O3 53
3
.4.2. Effect of lanthanide promotion
Several sol–gel 20% Ni/Al2O3 catalysts promoted with
lanthanide elements (La, Ce, Yb) were subsequently prepared
at pH = 4.8 and H2O/ATB = 3.6. The influence of lanthanide
promotion (2 wt.%) on catalytic performance at different
reaction temperatures is shown in Fig. 4. C3H8 conversion
20% Ni–2%Yb/Al2O3 84
20% Ni–5%Yb/Al2O3 65
◦
Reaction conditions: 500 C, C3H8/H2O/N2 = 1/4/95, feed flow rate = 300
3
2
cm /min, equal surface area reactions (9 m ).
(
Fig. 4a) and H2 yield (Fig. 4b) are seen to increase with
est C3H8 conversion (84%) and H2 yield (50%) are achieved
with 20% Ni–2% Yb/Al2O3 catalyst. The effect of lanthanide
promotion is more evident at higher reaction temperatures.
When the performances in propane steam reforming at dif-
increasing reaction temperature on all catalysts. Hydrogen
yield that would be obtained at equilibrium is also included
for comparison. 20% Ni–2% Ce/Al2O3 and 20% Ni–2%
Yb/Al2O3 catalysts show an improvement in C3H8 con-
version and H2 yield compared to 20% Ni/Al2O3 catalyst.
Although 20% Ni–2% La/Al2O3 catalyst exhibits similar
C3H8 conversion to 20% Ni/Al2O3 catalyst, it still shows
◦
ferent lanthanide contents are compared at 500 C (Table 7),
reaction data show that C3H8 conversion and H2 yield
decrease with increasing lanthanide content from 2 to 5 wt.%.
H2 chemisorption data have also shown a decreasing trend
of nickel surface area at higher lanthanide content. This
indicates that the catalytic activity of lanthanide-promoted
sol–gel Ni/Al2O3 catalysts is strongly related to nickel sur-
face area. Thus, loss of activity with higher lanthanide content
is due to the blockage of active sites by excess amounts of
lanthanides introduced in the catalysts. It is noted that the
distribution of C-containing products is not much affected
by an increase in lanthanide content.
◦
◦
an increased H2 yield above 450 C. At 500 C, the high-
To further investigate catalytic performance at higher reac-
tiontemperature, steady-statereactionexperimentswerecon-
◦
ducted at 550 C. However, the feed flow rate was raised
3
to 400 cm (STP)/min to avoid complete C3H8 conversion.
Interestingly, the highest C3H8 conversion (80%) and H2
yield (52%) are obtained with 20% Ni–2% Ce/Al2O3 catalyst
◦
at550 C(Table8). Furthermore, 20%Ni–2%La/Al2O3 cata-
lyst also shows a significant improvement in steam reforming
activity compared to 20% Ni/Al2O3 catalyst under these
reaction conditions. Moreover, an increase in CO selectiv-
ity is observed at higher reaction temperature. This sug-
gests an increased contribution from the reverse water-gas
shift reaction (CO2 + H2 → CO + H2O) to the overall reac-
tion network. Another possible explanation is that at higher
Table 8
Reaction comparison for propane steam reforming (dilution experiments) at
◦
5
50 C
Catalyst
C3H8
H2 yield Selectivity (%)
CO2 CH4 CO
conversion (%) (%)
2
2
2
2
0% Ni/Al2O3
48
36
52
52
50
48
38
31
33
21
25
32
30
31
37
37
37
0% Ni–2%La/Al2O3 66
0% Ni–2%Ce/Al2O3 80
0% Ni–2%Yb/Al2O3 70
Fig. 4. Effect of reaction temperature on catalytic performance of
lanthanide-promoted sol–gel 20% Ni/Al2O3 catalysts in propane steam
reforming (a) C3H8 conversion (b) H2 yield [reaction conditions:
◦
3
◦
4
00–500 C, C3H8/H2O/N2 = 1/4/95, feed flow rate = 300 cm /min, equal
Reaction conditions: 550 C, C3H8/H2O/N2 = 1/4/95, feed flow rate = 400
cm /min, equal surface area reactions (9 m ).
2
3
2
surface area reactions (9 m )].

1
40
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
conversion, water becomes the limiting reactant, and leading
to an increased formation of CO as opposed to CO2 from the
steam reforming reaction. Increased coking on the surface at
higher temperatures could also lead to higher CO formation
through the gasification of surface carbon.
Table 9
Reaction comparison for propane steam reforming (neat experiments)
◦
Catalyst
T ( C) C3H8
H2 yield Selectivity (%)
conversion (%)
%)
(
CO2 CH4 CO
5
5
00
50
35
43
14
18
68
69
25
20
7
“
Neat” experiments were performed with 20% Ni/Al2O3,
2
2
2
0% Ni/Al2O3
11
2
0% Ni–2% Ce/Al2O3, and 20% Ni–2% Yb/Al2O3 cata-
5
00
550
43
95
15
31
65
58
28
28
7
lysts in propane steam reforming without N2 dilution at
0% Ni–2%Ce/Al2O3
0% Ni–2%Yb/Al2O3
◦
14
the following conditions: T = 500–550 C, H2O/C = 4, total
3
feed rate = 270 cm (STP)/min, equal surface area reaction
18 m ). Demineralized distilled water was delivered by an
500
550
59
83
19
27
60
58
32
30
8
2
12
(
◦
3
Eldex metering pump (model: A-10-S) connected with a
Reaction conditions: 500–550 C, H2O/C = 4, total feed rate = 270 cm /min,
equal surface area reaction (18 m ).
2
4
0 psi back pressure regulator and subsequently vaporized
◦
in a preheater which was operated above 150 C. A column
packed with ceramic beads was installed after the preheater to
help C3H8 and steam mix effectively. The steady-state reac-
tion experiments were performed using 1/2 in. o.d. stainless
steel flow reactor to avoid high pressure drop due to higher
amounts of catalysts loaded in the reactor. Consistent with
reaction data obtained from dilution experiments, lanthanide-
promoted catalysts give higher C3H8 conversion and H2 yield
compared to the unpromoted catalyst (Table 9).
◦
through a maximum at the reduction temperature of 600 C
and decreases at higher reduction temperature. By comparing
the performance of 20% Ni–2% Ce/Al2O3 catalysts calcined
at 450 and 600 C, we see that higher calcination temper-
ature causes a decline in catalytic activity (Fig. 5b). The
optimum reduction temperature is 600 C regardless of cal-
cination temperature. This observation is in good agreement
with TPR results. When Ni–Ce/Al2O3 catalysts are calcined
◦
◦
◦
3
.4.3. Effect of calcination and reduction temperatures
at higher temperature (600 C), NiO interacts with Al O to
2 3
When the effect of reduction temperature was examined
2 4
form a very stable NiAl O spinel which is more difficult
◦
by keeping the calcination temperature constant at 450 C, a
to reduce [31–35]. However, we cannot explain the lower
◦
◦
maximum is observed at a reduction temperature of 600 C
activity of the catalyst calcined at 600 C when it is reduced
◦
(
Fig. 5). Both 20% Ni/Al2O3 and 20% Ni–2% Ce/Al2O3
at 700 C. As previously discussed, the catalysts calcined
◦
catalysts give the highest H2 yield when reduced at 600 C.
However, 20% Ni–2% Ce/Al2O3 catalyst gives superior per-
formance at every reduction temperature. An improvement
in H2 yield with cerium promotion is much more evident
at higher reduction temperatures. Furthermore, the catalytic
at higher temperature exhibit higher nickel dispersion and
◦
nickel surface area when reduced at 700 C (Table 5). Our in
situ X-ray diffraction studies performed to observe the phase
changesduringreductionhaveshownincompletereductionat
lower temperatures while Ni sintering was observed at higher
temperatures [36].
◦
activity of both catalysts calcined at 450 C is seen to go
Fig. 5. Effect of calcination and reduction temperatures on catalytic performance in propane steam reforming (a) comparison between 20% Ni/Al2O3 and 20%
◦
◦
Ni–2% Ce/Al2O3 catalysts calcined at 450 C; (b) effect of calcination temperature on 20% Ni–2% Ce/Al2O3 catalysts [reaction conditions: 550 C, H2O/C = 2,
GHSV = 248,000 h ].
−1

S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
141
Table 11
Comparison of CO-methanation activity of 20% Ni/Al2O3 and 20% Ni–2%
Ce/Al2O3 catalysts
Catalyst
CH4 yield (%)
CO2 yield (%)
◦
◦
◦
◦
5
00 C
550 C
500 C
550 C
2
2
0% Ni/Al2O3
0% Ni–2%Ce/Al2O3
59
71
46
61
5
5
5
5
Reaction conditions: FCO = 10 cm3 (STP)/min, CO/H2/N2 = 1/6/13,
−
1
GHSV = 372,000 h
.
◦
Ce/Al2O3 catalysts at 500 and 550 C. TOF is seen to increase
with increasing reaction temperature. The increase is more
pronounced with 20% Ni–2% Ce/Al2O3 catalyst. Further-
more, the higher activity of promoted catalysts is clearly
seen when TOF values are compared for the 20% Ni–2%
Ce/Al2O3 and 20% Ni/Al2O3 catalysts.
The variation of product selectivities with C3H8 conver-
sion over 20% Ni/Al2O3 and 20% Ni–2% Ce/Al2O3 catalysts
is shown in Fig. 7. In this study, CO2 selectivity is defined
as the ratio of moles of CO2 produced to the total moles of
C-containing products (i.e., CO2, CO, and CH4) whereas H2
selectivity is defined as 2 × moles of H2 produced/total moles
of hydrogen in the H-containing products, which are H2 and
CH4. Interestingly, CH4 selectivity is seen to increase at the
expense of both H2 and CO2 selectivities at lower GHSVs,
i.e., higher conversions. This suggests that CO2 methana-
tion (CO2 + 4H2 → CH4 + 2H2O) becomes more appreciable
whenresidencetimeisincreased, leadingtoanincreasedCH4
selectivity. CO selectivity, on the other hand, is not much
affected by the variation of C3H8 conversion over either cat-
alyst.
Fig. 6. Effect of GHSV on H2 yield in propane steam reforming over 20%
Ni/Al2O3 and 20% Ni–2% Ce/Al2O3 catalysts: (a) 500 C and (b) 550 C.
◦
◦
3
.4.4. Effect of gas hourly space velocity (GHSV)
The catalytic performance of 20% Ni/Al2O3 and 20%
The CO- and CO2-methanation reactions over 20%
Ni/Al2O3 and 20% Ni–2% Ce/Al2O3 catalysts were further
Ni–2% Ce/Al2O3 catalysts in propane steam reforming
was further investigated at various space velocities. The
reaction conditions for these experiments were as fol-
◦
investigated at 500–550 C. The H2/CO and H2/CO2 ratios
for these experiments were kept constant at 6. CH4 yield
is defined as moles of carbon in CH4 produced/total moles
of carbon in the feed. As shown in Tables 11 and 12, CH4
yield is seen to decrease with increasing reaction tempera-
ture over both catalysts. Ce-promoted catalyst shows higher
CH4 yields from both CO- and CO2-methanation reactions
compared to unpromoted catalyst. It is worth noting that
a lower level of CH4 produced from the CO2-methanation
◦
lows: 500–550 C, H2O/C = 2. The variation of GHSV was
performed by adjusting both the amount of the catalyst
3
(
10–20 mg) and the feed flow rate (200–400 cm (STP)/min).
As expected, H2 yield increases with decreasing GHSV
−
1
from 496,000 to 124,000 h over both catalysts (Fig. 6). At
◦
5
00 C, the beneficial effect of cerium promotion becomes
◦
much more pronounced at lower GHSVs. At 550 C, an
improvement in H2 yield with the addition of cerium is evi-
dent for the whole range of GHSVs examined. C3H8 conver-
sion rates at steady state were further normalized to surface
◦
reaction at 550 C implies an increased contribution from the
reverse water-gas shift reaction at higher temperature. Ross
and co-workers [37,38] studied the CO-methanation reaction
◦
nickel atoms, measured by H2 chemisorption at 35 C, in
the form of turnover frequencies (TOFs). Table 10 compares
average TOFs obtained with 20% Ni/Al2O3 and 20% Ni–2%
Table 12
Comparison of CO2-methanation activity of 20% Ni/Al2O3 and 20% Ni–2%
Ce/Al2O3 catalysts
Table 10
Catalyst
CH4 yield (%)
CO yield (%)
Effect of reaction temperature on turnover frequency
◦
◦
◦
◦
TOF (s−1)
500 C
550 C
500 C
550 C
Catalyst
2
2
0% Ni/Al2O3
0% Ni–2%Ce/Al2O3
14
30
13
22
28
24
36
34
◦
◦
5
00 C
550 C
2
2
0% Ni/Al2O3
0% Ni–2%Ce/Al2O3
0.35
0.50
0.65
1.20
3
Reaction conditions: FCO = 15 cm (STP)/min, CO2/H2/N2 = 1/6/20,
2
−
1
GHSV = 496,000 h
.

1
42
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
Fig. 7. Variation of product selectivities with C3H8 conversion.
on coprecipitated Ni/Al2O3 catalysts promoted with La2O3
and CeO2. An enhancement in methanation activity was
observed with an increase in apparent activation energy upon
lanthanide promotion. The role of lanthanum and cerium
in the methanation reaction was reported to be associated
with higher amounts of CO and H2 adsorbed on nickel sur-
face due to an interaction between lanthanides and nickel in
the catalysts. Our H2 chemisorption studies have shown that
3.5. DRIFTS studies
3.5.1. C3H8-TPD/TPReaction
TPD/TPReaction experiments with DRIFTS were per-
formed over reduced 20% Ni–2% Ce/Al2O3 catalyst, where
strongly adsorbed propane was allowed to react with water.
Adsorption of propane was carried out at room tempera-
ture for 1 h. The system was subsequently purged and heated
under He and spectra were collected in steps under He flow up
2
0% Ni–2% Ce/Al2O3 catalyst shows an increased H2 uptake
◦
◦
compared to 20% Ni/Al2O3 catalyst, thus resulting in higher
nickel dispersion and larger nickel surface area (Table 4). Fur-
thermore, the CO uptakes measured by CO chemisorption at
to 450 C. The introduction of H2O was performed at 450 C
for different times on stream (i.e., 5, 10, and 15 min) followed
by He flushing before spectra were taken.
◦
3
3
2
5 C are 3.39 and 10.16 cm /gcat over 20% Ni/Al2O3 and
Fig. 8 shows DRIFT spectra taken at 25, 100, 200, 250,
◦
◦
0% Ni–2% Ce/Al2O3 catalysts reduced at 600 C, respec-
300, and 450 C under He after C3H8 adsorption over reduced
tively. Consequently, an improvement in methanation activity
when cerium is added into sol–gel Ni/Al2O3 catalysts could
be due to an increased CO and H2 adsorption on nickel crys-
tallites. A higher amount of CO2 adsorbed on nickel upon
cerium promotion could also be another possibility.
20% Ni–2% Ce/Al2O3 catalyst. After propane adsorption
and He flushing, bands remained at 2940, 2932 (shoul-
der), 2824, 1630, 1476, 1435 (shoulder), 1373, 1189, and
1103 cm at room temperature. The band at 1630 cm
could be attributed to an olefinic C C vibration [39,40]
−
1
−1

S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
143
negative peak at 3750 cm 1, which is characteristics of sur-
−
face hydroxyl groups on the alumina support, was observed
[43,44]. This indicates that the surface hydroxyl groups either
interacted with propane or were consumed upon the adsorp-
tion of propane. Although the intensities of the bands at
−
1
2
940 and 2824 cm decreased at higher temperatures, they
◦
were thermally stable up to 450 C. The features at 2940
and 2932 cm are assigned to asymmetric CH3 stretch-
−
1
ing [νas(CH3)] and asymmetric CH2 stretching [νas(CH2)],
−1
respectively [39,42, 45–47]. The feature at 2824 cm could
be attributed to either symmetric CH2 stretching [νs(CH2)] or
symmetric CH3 stretching [νs(CH3)] [42,48]. The features at
−
1
1
476, 1437, 1373, 1189, and 1103 cm decreased in inten-
sity with temperature and subsequently almost disappeared
◦
at 300 C. These features could be associated with adsorbed
ethylidyne/propylidyne species that were depleted upon the
◦
reaction with H2O. At 450 C and before the introduction
−1
of water, new bands at 1600 and 1374 cm were present
accompanied by two small C-H stretching bands at 2954 and
904 cm . Although it is conceivable that new CH3 groups
−
1
2
could be forming due to desorption/re-adsorption processes,
a more likely explanation is that the features at 2954, 2904,
Fig. 8. DRIFT spectra taken over reduced 20% Ni–2% Ce/Al2O3 catalysts
after C3H8 adsorption; *H2O was introduced for 5, 10, and 15 min and the
system was flushed under He for 5 min.
−1
1
600, and 1374 cm are due to formate species [49,50].
◦
With the introduction of water at 450 C, the bands asso-
ciated with the formate species grew much stronger and were
seen in higher intensity with time on stream. In the course
of propane steam reforming, formate species are probably
produced by the reaction of CO with the water introduced.
In the absence of water, surface hydroxyl groups can serve
as a source of oxygen to form formate species. Our formate
band assignments are consistent with those which arise from
adsorption of CO on 1% Au/ceria and 1% Pt/ceria catalysts
and has been observed in a separate propylene adsorp-
tion/desorption DRIFTS experiment performed in our lab-
2
oratory (Fig. 9). However, bands typical of olefinic sp C H
−
1
stretching present above 3000 cm were not observed after
propane adsorption which leads us to assume that the band at
−1
1
630 cm could be assigned as the O-H bending frequency
−1
of adsorbed water as observed by Gates and co-workers [41].
The remaining bands are characteristic of adsorbed propyli-
dyne [39,42] and are similar to resultant spectra of propane
adsorbed on a reduced Ir/Al2O3 catalyst [41]. In addition, a
[51]. Two additional features at 1455 and 1255 cm , which
could be attributed to carbonate species, were also observed
◦
after H2O exposure at 450 C [49,50].
In order to help in the band assignments for the C3H8-
TPD/TPReaction experiment, DRIFT spectra were taken
after propylene and CO2 adsorption in separate experiments.
◦
Fig. 9 shows DRIFT spectra taken at 25, 250, and 450 C
under He flow over reduced 20% Ni–2% Ce/Al2O3 cata-
lyst during the C3H -TPD experiment. The strong presence
6
−
1
of the band at 1630 cm on the catalyst surface following
propylene adsorption signals the C C stretching vibration.
Presence of olefinic C H vibrations that would be expected
−
1
above 3000 cm cannot be discerned from this spectrum
due to a very broad feature in that region. It is possible that
adsorbed propylene could be dehydrogenated on the surface,
giving rise to water (surface hydroxyl groups), which would
−
1
show an IR band around 1630 cm as well as a broad fea-
ture above 3000 cm 1. The feature at 1373 cm , which was
quite prominent following propane adsorption, could not be
observed after propylene adsorption. This observation sup-
−
−1
−1
ports the assignment of the band at 1373 cm to symmetric
CH3 bending [δs(CH3)] of propylidyne. This assignment is
in agreement with what is reported by Shahid and Shep-
Fig. 9. DRIFT spectra taken over reduced 20% Ni–2% Ce/Al2O3 catalysts
after C3H6 adsorption at room temperature.
−1
pard [42]. In addition, the feature at 2824 cm was absent,

1
44
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
Fig. 10. DRIFT spectra taken over reduced 20% Ni–2% Ce/Al2O3 catalysts
after CO2 adsorption at room temperature.
and the 2940 cm 1 band was much weaker and disappeared
−
Fig. 11. DRIFT spectra taken over reduced 20% Ni–2% Ce/Al2O3 catalysts
after consecutive adsorption of C H and H O.
−
1
3
8
2
very quickly at higher temperatures. The band at 1476 cm
,
which was quite strong following propane adsorption was
−
1
◦
barely noticeable (∼1456 cm ) after propylene adsorption.
This observation makes us believe that this feature could be
due to a methyl group, which may have a different orientation
in species resulting from propylene adsorption. This expla-
nation would also be consistent with our earlier assertion that
propylidyne species are the predominant hydrocarbon deriva-
tives on the surface following propane adsorption, but this is
not necessarily the case following propylene adsorption.
From DRIFT spectra of CO2-TPD over the reduced
tra were collected in steps under He flow up to 450 C. It
is noted that H O was re-introduced at each temperature for
2
5 min followed by He flushing before taking spectra.
After C H8 and H O introduction at room tem-
perature, bands remained at 3750–3500 (very broad),
2940, 2824, 2055, 1630, 1435 (weak), 1373 (weak), and
(Fig. 11). The broad band around
is characteristic of the OH stretching
region while the feature at 2055 cm is assigned to lin-
early adsorbed CO on the catalyst surface and disappeared
above 200 C [55–57]. Compared to the DRIFT spectrum
after C H adsorption at room temperature, the intensity
of the band located at 1630 cm increased significantly
when H O was introduced at room temperature. This band
3
2
−
1
1
1255 (weak) cm
3750–3500 cm
−
−
1
2
1
1
0% Ni–2% Ce/Al2O3 catalyst, bands at 3619, 2355, 2325,
◦
643, 1543 (shoulder), 1438, 1387 (shoulder), 1229, and
−
1
084 cm are seen at room temperature (Fig. 10). The bands
3
8
−
1
−1
at 3619, 1643, 1438, and 1229 cm are assigned to bicarbon-
ate species and decrease in intensity significantly at 250 C
◦
2
[
49,50,52]. Conversely, carbonate species are characterized
decreased in intensity at higher temperatures and eventu-
ally shifted to 1600 cm at 450 C. The band at 1600 cm
−
1
−1
◦
−1
by the shoulder bands at 1543, and 1387 cm as well as the
1
−
1
084 cm band [49,50,52]. In addition, the only other fea-
is attributed to a formate species as previously observed
from the C H -TPD/TPReaction experiment. The features
−
1
ture observed is a doublet at 2355 and 2325 cm , which
corresponds to adsorbed CO2 [53,54]. The disappearance
of carbonate and bicarbonate species at higher tempera-
tures leads us to believe that formate and carbonate species
observedfromtheC3H8-TPD/TPReactionexperimentarethe
result of the surface transformation of propylidyne species
rather than from the re-adsorption of CO2.
3
8
−
1
at 2940 and 2824 cm were seen in much lower intensity
and disappeared above 250 C whereas the bands at 2954
and 1455 cm started to appear at 200 C and grew stronger
at higher temperatures. Furthermore, two additional bands
at 2904 and 1374 were seen at 450 C. It is evident that
◦
−
1
◦
◦
most surface species such as formate and carbonate ions are
formed upon the reaction of adsorbed C3H8 and H2O above
◦
3
.5.2. C3H8 + H2O-TPReaction
C3H8 + H2O-TPReaction with DRIFTS was also per-
250 C. In a separate experiment with C3H8 + H2O adsorp-
−1
tion in a reverse order (not shown), a large band at 1645 cm
formed over reduced 20% Ni–2% Ce/Al2O3 catalyst. C3H8
was adsorbed at room temperature for 1 h. The system was
then purged under He for 30 min and H2O was subsequently
introduced for 5 min at room temperature followed by He
flushing for 5 min. The system was heated and DRIFT spec-
was also observed after water adsorption at room tempera-
ture. This observation provides additional support for our
earlier assertion that the 1630 cm feature seen in the C3H8-
TPD/TPReaction and C3H8 + H2O-TPReaction experiments
is due to adsorbed water. Interestingly, no additional features
−
1

S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
145
hydroxyl groups [50]. Moreover, the fraction of the formate
remaining adsorbed could react with hydroxyl groups aris-
ing from adsorbed water to form carbonates and hydrogen at
higher temperatures. While the DRIFTS experiments do not
specifically indicate the participation of other surface inter-
mediates, it could be possible that adsorbed species such as
oxygen atoms, hydrogen atoms, and hydroxyl groups could
also be spilling over to the nickel surface from the alumina
support or ceria, thus enhancing the overall adsorption and
participation of water.
4. Conclusions
Various sol–gel Ni/Al2O3 catalysts promoted with lan-
thanide elements (La, Ce, and Yb) were studied in the steam
reforming of propane. The promotion effect is most pro-
nounced on the Ni–Ce/Al2O3 catalysts. An improvement in
catalytic activity with respect to C3H8 conversion and H2
yield appears to be due to easier reduction of nickel species
to a metallic state and larger nickel surface area. Additionally,
cerium promotion enhances the CO- and CO2-methanation
activity, leading to an increased CH4 yield. DRIFT experi-
ments suggest adsorbed propylidyne to be the predominant
surface species following C H adsorption at room temper-
ature and to subsequently react with hydroxyl groups arising
from adsorbed water to form formate and carbonate species
at higher temperatures. The second paper of this series [36]
examines the effect of lanthanide promotion in terms of the
deactivation characteristics of these catalysts under steam
reforming conditions.
Fig. 12. Possible mechanistic steps during propane steam reforming over
reduced Ni–Ce/Al2O3 catalyst as observed using in situ DRIFTS.
wereseenfollowingC3H8 adsorption, suggestingwatercould
be primarily adsorbed on metallic nickel sites and therefore
inhibits the adsorption of propane.
3
8
Based on the observed surface specie transformations
from in situ DRIFTS studies, the following major reaction
pathways for adsorbed propane in the presence of steam over
reduced sol–gel Ni–Ce/Al2O3 catalysts are proposed as illus-
trated in Fig. 12. The first step is believed to be adsorption
of propane on metallic nickel sites to form propylidyne. To a
lesser degree, adsorbed propane can further decompose into
hydrocarbon fragments (i.e., CHx) and even carbon oxides in
the presence of adsorbed water. Carbon deposition, especially
at higher temperatures, also is a possibility as seen through
our coking studies [36]. During this step, it is likely that the
formation of carbon monoxide could be occurring on nickel
sites. Adsorbed water from the catalyst can combine with
adsorbed hydrocarbon fragments or adsorbed carbon monox-
ide to form formate species. It is also possible that adsorbed
hydrocarbon fragments could react with other sources of oxy-
gen (hydroxyls, ceria lattice) to form formates in the absence
Acknowledgments
The financial support provided by Honda Research Insti-
tute, USA Inc. and the National Science Foundation (NSF-
DMR grant # 0114098) is gratefully acknowledged. The
authors thank Dr. Rick B. Watson for his helpful discussion
of the DRIFTS studies.
References
◦
of water, as observed at 450 C (Fig. 8). Additionally, it is
possible that water could dissociatively adsorb to produce
gaseous hydrogen and surface oxygen species. These surface
oxygen species have been found to react with methane to
form CO and hydrogen in methane steam reforming over var-
ious oxide supports, including ceria [58–60]. Unfortunately,
we cannot obtain information pointing to the dissociative
adsorption of water to form hydrogen from the DRIFTS
experiments. However, the contribution from hydroxyls cre-
ated from adsorbed water is expected to be much higher
than that from surface oxygen. The thermal decomposition of
adsorbed formates can result in the formation of carbon diox-
ide and hydrogen, or the formation of carbon monoxide and
[
1] J.N. Armor, Appl. Catal. A 176 (1999) 159.
[2] D.E. Ridler, M.V. Twigg, in: M.V. Twigg (Ed.), Catalyst Handbook,
Wolfe, London, 1989, p. 225.
[3] J.T. Richardson, S.A. Paripatyadar, Appl. Catal. A 61 (1990) 293.
[4] A. Igarashi, T. Ohtaka, S. Motoki, Catal. Lett. 13 (1991) 189.
[5] J.R. Rostrup-Nielsen, J.H. Hansen, J. Catal. 144 (1993) 38.
[6] D. Qin, J. Lapszewicz, Catal. Today 21 (1994) 551.
[7] R. Craciun, B. Shereck, R.J. Gorte, Catal. Lett. 51 (1998).
[8] X. Wang, R.J. Gorte, Catal. Lett. 73 (2001) 15.
[9] Q. Zhuang, Y. Qin, L. Chang, Appl. Catal. 70 (1991) 1.
[
10] B.L. Su, S.D. Guo, in: B. Delmon, G.F. Froment (Eds.), Catalyst
Deactivation, Elsevier, Amsterdam, 1999, p. 325.
11] R.M. Sambrook, J.R.H. Ross, US Patent 4 469 815, 1984.
[
[12] Z. Cheng, Q. Wu, J. Li, Q. Zhu, Catal. Today 30 (1996) 147.

1
46
S. Natesakhawat et al. / Journal of Molecular Catalysis A: Chemical 241 (2005) 133–146
[
[
[
13] S. Wang, G.Q. Lu, Appl. Catal. B 19 (1998) 267.
[39] F. Solymosi, R. Nemeth, L. Ovari, L. Egri, J. Catal. 195 (2000) 316.
[40] M.A. Chesters, C.D.L. Cruz, P. Gardner, E.M. McCash, P. Pudney,
G. Shahid, N. Sheppard, J. Chem. Soc. Faraday Trans. 86 (1990)
2757.
14] S. Wang, G.Q. Lu, J. Chem. Technol. Biotechnol. 75 (2000) 589.
15] J.A. Montoya, E.R. Pascual, C. Gimon, P.D. Angel, A. Monzon,
Catal. Today 63 (2000) 71.
[
[
16] R. Zhang, F. Li, Q. Shi, L. Luo, Appl. Catal. A 205 (2001) 279.
17] U.S. Ozkan, Y. Cai, M.W. Kumthekar, L. Zhang, J. Catal. 142 (1993)
[41] A.M. Argo, J.F. Goellner, B.L. Philips, G.A. Panjabi, B.C. Gates, J.
Am. Chem. Soc. 123 (2001) 2275.
1
82.
[42] G. Shahid, N. Sheppard, Spectrochimica Acta 46A (1990) 999.
[43] D.H. Lee, R.A. Condrate, Mater. Lett. 23 (1995) 241.
[44] A. Hinz, M. Skoglundh, E. Fridell, A. Andersson, J. Catal. 201
(2001) 247.
[45] N. Sheppard, C.D.L. Cruz, Adv. Catal. 41 (1996) 1.
[46] P.D. Holmes, G.S. McDougall, I.C. Wilcock, Catal. Today 9 (1991)
15.
[
[
[
18] R.D. Gonzalez, T. Lopez, R. Gomez, Catal. Today 35 (1997) 293.
19] Y.G. Chen, J. Ren, Catal. Lett. 29 (1994) 39.
20] C.E. Quincoces, S. Dicundo, A.M. Alvarez, M.G. Gonzalez, Mater.
Lett. 50 (2001) 21.
21] Y.J. Huang, J.A. Schwarz, J.R. Diehl, J.P. Baltrus, Appl. Catal. 36
[
(1988) 163.
[
22] K.T. Ng, D.M. Hercules, J. Phys. Chem. A 80 (1976) 2094.
[47] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coor-
dination Compounds, third ed., John Wiley & Sons, 1978.
[48] F.C. Tompkins, Chemisorption of Gases on Metals, Acedamic Press,
1978.
[23] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain
(Ed.), Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer,
Minnesota, 1992.
[
[
[
[
[
24] A. Velon, I. Olefjord, Oxid. Met. 56 (2001) 415.
25] X. Wang, U.S. Ozkan, J. Phys. Chem. B 109 (2005) 1882.
26] C.H. Bartholomew, R.J. Farrauto, J. Catal. 45 (1976) 41.
27] C.H. Bartholomew, R.B. Pannell, J. Catal. 65 (1980) 390.
28] R.J. Farrauto, C.H. Bartholomew, Fundamentals of Industrial Cat-
alytic Processes, Blackie Academic & Professional, London, 1997.
29] Z. Zhang, X.E. Verykios, Appl. Catal. A 138 (1996) 109.
30] R.G. Ding, Z.F. Yan, Catal. Today 68 (2001) 135.
31] J.R.H. Ross, M.C.F. Steel, A. Zeini-Isfahani, J. Catal. 52 (1978) 280.
32] J. Zielinski, J. Mol. Catal. 83 (1993) 197.
[49] V.F. Kiselev, O.V. Krylov, Adsorption and Catalysis on Transition
Metals and Their Oxides, Berlin, vol. 9, 1989.
[50] D. Bianchi, T. Chafik, M. Khalfallah, S. Jean, Appl. Catal. A 105
(1993) 223.
[51] G. Jacobs, P.M. Patterson, L. Williams, E. Chenu, D. Sparks, G.
Thomas, B.H. Davis, Appl. Catal. A 262 (2004) 177.
[52] R. Philip, K. Fujimoto, J. Phys. Chem. A 96 (1992) 9035.
[53] M.A. Bollinger, M.A. Vannice, Appl. Catal. B 8 (1996) 417.
[54] V. Ermini, E. Finocchio, S. Sechi, G. Busca, S. Rossini, Appl. Catal.
A 190 (2000) 157.
[55] S.I. Fujita, M. Nakamura, T. Doi, N. Takezawa, Appl. Catal. A 104
(1993) 87.
[56] J. Ryczkowski, Catal. Today 68 (2001) 263.
[57] D. Liu, G.H. Que, Z.X. Wang, Z.F. Yan, Catal. Today 68 (2001)
155.
[
[
[
[
[
[
[
[
33] J. Zielinski, Appl. Catal. A 94 (1993) 107.
34] C. Li, Y.W. Chen, Thermo. Acta 256 (1995) 457.
35] B. Vos, E. Poels, A. Bliek, J. Catal. 198 (2001) 77.
36] S. Natesakhawat, R.B. Watson, X. Wang, U.S. Ozkan, J. Catal., in
press.
[
37] H.G.J.L. Rotgerink, R.P.A.M. Paalman, J.G. Ommen, J.R.H. Ross,
Appl. Catal. 45 (1988) 257.
[58] K. Hou, R. Hughes, Chem. Eng. J. 82 (2001) 311.
[59] Y. Matsumura, T. Nakamori, Appl. Catal. A 258 (2004) 107.
[60] E. Ramirez-Cabrera, A. Atkinson, D. Chadwick, Appl. Catal. B 47
(2004) 127.
[38] H.G.J.L. Rotgerink, R.P.A.M. Paalman, J.G. Ommen, J.R.H. Ross,
Appl. Catal. 45 (1988) 281.