Hydrogen production by steam reforming of ethanol over Ni-Sr-Al2O3-ZrO2 aerogel catalyst

Article abstract of DOI:10.1016/j.molcata.2016.09.013

Ni-Sr-Al₂O₃-ZrO₂ aerogel (NSAZ-Ae) catalyst was prepared by an one-step epoxide-driven sol-gel method and a subsequent supercritical CO₂ drying method. For comparison, Ni-Sr-Al₂O₃-ZrO₂ xerogel (NSAZ-Xe) catalyst was also prepared by a sol-gel method and a subsequent conventional evaporative drying method. NSAZ-Ae catalyst retained higher surface area and larger pore volume than NSAZ-Xe catalyst. NSAZ-Ae catalyst also showed strong interaction between active metal and support, which led to small nickel particles after the reduction. H₂-TPD (temperature-programmed desorption) results revealed that NSAZ-Ae catalyst with smaller nickel particle size had higher nickel dispersion and higher nickel surface area. In addition, NSAZ-Ae catalyst showed higher ethanol adsorption capacity than NSAZ-Xe catalyst. Performance and stability of the catalyst in the ethanol steam reforming reaction was enhanced with the employment of supercritical drying as observed for NSAZ-Ae catalyst. Enhanced physicochemical property and high nickel surface area (high nickel dispersion) of NSAZ-Ae catalyst led to excellent catalytic activity and stability in the hydrogen production by steam reforming of ethanol.

Full text of DOI:10.1016/j.molcata.2016.09.013

Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Hydrogen production by steam reforming of ethanol over
Ni-Sr-Al O -ZrO aerogel catalyst
2
3
2
∗
Ji Hwan Song, Seung Ju Han, Jaekyeong Yoo, Seungwon Park, Do Heui Kim, In Kyu Song
School of Chemical and Biological Engineering, Institute of Chemical Processes, Seoul National University, Shinlim-dong, Kwanak-ku, Seoul 151-744, South
Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 16 August 2016
Received in revised form
Ni-Sr-Al O -ZrO aerogel (NSAZ-Ae) catalyst was prepared by an one-step epoxide-driven sol-gel method
2
3
2
and a subsequent supercritical CO2 drying method. For comparison, Ni-Sr-Al2O3-ZrO2 xerogel (NSAZ-
Xe) catalyst was also prepared by a sol-gel method and a subsequent conventional evaporative drying
method. NSAZ-Ae catalyst retained higher surface area and larger pore volume than NSAZ-Xe catalyst.
NSAZ-Ae catalyst also showed strong interaction between active metal and support, which led to small
nickel particles after the reduction. H2-TPD (temperature-programmed desorption) results revealed that
NSAZ-Ae catalyst with smaller nickel particle size had higher nickel dispersion and higher nickel surface
area. In addition, NSAZ-Ae catalyst showed higher ethanol adsorption capacity than NSAZ-Xe catalyst.
Performance and stability of the catalyst in the ethanol steam reforming reaction was enhanced with
the employment of supercritical drying as observed for NSAZ-Ae catalyst. Enhanced physicochemical
property and high nickel surface area (high nickel dispersion) of NSAZ-Ae catalyst led to excellent catalytic
activity and stability in the hydrogen production by steam reforming of ethanol.
1
2 September 2016
Accepted 12 September 2016
Available online 13 September 2016
Keywords:
Hydrogen production
Steam reforming of ethanol
Ni-Sr-Al2O3-ZrO2 aerogel catalyst
Supercritical drying
Nickel surface area
©
2016 Elsevier B.V. All rights reserved.
1
. Introduction
Various metals have been investigated as active components in
the ethanol steam reforming. For example, noble metal-based cat-
alysts such as Rh, Pt, and Pd show high hydrogen yield [10,11].
Application of other transition metals such as Ni and Co to steam
reforming of ethanol have also been studied [12,13]. As for sup-
porting materials, extensive researches concerning the usage of
mixed-metal oxides, hydrotalcites, and perovskites have been
made [14–16]. Among these numerous catalytic systems, nickel
catalyst supported on alumina is known to be the most feasible
and widely studied catalytic system for ethanol steam reforming,
because nickel shows high activity for ethanol steam reforming
reaction in spite of its low price [17,18]. Reaction pathways of
ethanol steam reforming on the supported nickel catalyst leads
to the formation of various carbon-containing intermediates due
to ethanol dehydrogenation (Eq. (2)), ethanol decomposition (Eq.
(3)), and ethanol dehydration (Eq. (4)) [19].
Concerns about environmental problems and shortage of con-
ventional fossil fuels have accelerated researches on alternative
energy sources [1,2]. Hydrogen energy is one of the promising can-
didates of alternative energy sources due to its high specific energy
density, ampleness, and eco-friendly nature [3,4]. Apart from the
advantages listed above, hydrogen can also be utilized in various
industrial fields such as internal engines and fuel cells [5]. Extensive
researches have been made on various methods to produce hydro-
gen in massive scale with efficiency [5,6]. Although industrial-scale
hydrogen production is mainly based on methane steam reforming
process, ethanol is also gaining attention as a reactant candidate of
steam reforming for hydrogen production [7,8]. Ethanol is easy to
handle and can be produced from fermentation process of biomass
materials [8]. Ideal steam reforming of ethanol occurs as presented
in Eq. (1), where ethanol and water react with a stoichiometric fac-
tor of H O/C H5OH = 3 to produce 6 molecules of hydrogen per one
◦
C H5OH → CH CHO + H
(ꢀH
K = + 68 kJ/mol)
= + 49 kJ/mol)
(2)
(3)
(4)
2
3
2
29
2
2
◦
ethanol molecule [9].
C H OH → CH + CO + H
(ꢀH
29 K
2
5
4
2
◦
◦
C H5OH → C H + H O
(ꢀH
K = + 45 kJ/mol)
C H5OH + 3H O → 2CO + 6H
(ꢀH
29
K = + 174 kJ/mol)(1)
2
2
4
2
29
2
2
2
2
Ethylene by-product formed by dehydration of ethanol poly-
merizes to form carbon species which suppresses activity and
stability of the catalyst [20]. Ethanol dehydration has a tendency to
be catalyzed on the acid sites of alumina [21]. To solve this problem,
∗ Corresponding author.
E-mail address: inksong@snu.ac.kr (I.K. Song).
http://dx.doi.org/10.1016/j.molcata.2016.09.013
1
381-1169/© 2016 Elsevier B.V. All rights reserved.

J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
343
◦
diverse modifications were made to Ni/Al O3 catalyst. For exam-
resultant gel was dried in a convection oven at 80 C for 5 days
instead of in a supercritical CO2 stream. The obtained powder
2
ple, alkaline-earth metal elements have been added to neutralize
acid sites of alumina, and thus, to suppress ethanol dehydration
and to improve catalytic activity and stability [17]. It has been
◦
was calcined at 550 C for 5 h to yield a Ni-Sr-Al O -ZrO xero-
2
3
2
gel (denoted as NSAZ-Xe) catalyst. In both catalysts, nickel and
strontium contents with respect to alumina-zirconia were fixed at
15 wt% and 6 wt%, respectively.
reported that promotion of Ni/Al O3 catalyst with Mg results in
2
diminished acidity and enhanced nickel dispersion [22]. Addition
of other alkaline-earth metal species such as Ca and Sr into Ni/Al O3
2
catalyst also shows a positive effect in a sense that the reduced acid-
ity leads to lower ethylene selectivity and higher hydrogen yield
2.2. Characterization
[
23,24]. Apart from alkaline-earth metal, another element widely
Textural properties of calcined and reduced catalysts were
analyzed with nitrogen adsorption-desorption experiments using
a BELSORP-mini II (BEL Japan) device. Surface impurities were
studied for modification of steam reforming catalyst is zirconium,
which forms a mixed metal oxide with alumina support. Introduc-
tion of Zr species on Ni/Al O3 is reported to induce enhanced H O
◦
removed by degassing the samples at 150 C for 4 h with a
2
2
adsorption/dissociation capacity on the catalyst surface [22,25].
Modification of steam reforming catalyst can be attained by
altering chemical compositions and/or by improving textural prop-
erties of the catalyst. Development of mesoporous structure within
the catalytic system can enhance dispersion of active metal and
can increase transfer rate of reactant species [26,27]. Several works
concerning the employment of various surfactant materials, which
act as a template during the formation of mesoporous structure,
have been reported previously [28,29]. For instance, an aerogel
catalyst prepared by supercritical drying retains excellent textural
properties compared to a xerogel catalyst with the same chemi-
cal composition, leading to enhanced catalytic activity of aerogel
catalyst [30]. In this respect, direct comparison between aero-
gel and xerogel forms of Ni-Sr-Al O -ZrO catalysts for ethanol
vacuum pump prior to nitrogen adsorption-desorption measure-
ments. Calculation of surface area was made by applying the BET
(Brunauer-Emmett-Teller) method to the isotherm plots. Chem-
ical compositions of the catalysts were determined by ICP-AES
(inductively coupled plasma atomic emission spectroscopy) analy-
ses with an ICPS-100IV (Shimadzu) device. XRD (X-ray diffraction)
analyses were carried out with a D-Max2500-PC (Rigaku) device,
while the operating conditions were kept at 100 mA and 50 kV
with Cu-K␣ radiation (␭ = 1.541 Å). For the TPR (temperature-
programmed reduction) measurements, 0.03 g of each calcined
sample was charged into a quartz reactor. Furnace temperature
◦
was then elevated from room temperature to 1000 C at a rate
◦
of 10 C/min under a flow of hydrogen (2 ml/min) diluted with
nitrogen (20 ml/min). TPR profiles were collected using a TCD
(thermal conductivity detector) attached to a gas chromatograph
(ACME 6000, Younglin). TEM (transmission electron microscopy)
images of reduced and used samples were obtained with a JEM-
2000EXII device (Jeol). Nickel surface areas of reduced catalysts
2
3
2
steam reforming prepared by different drying procedures would
be worthwhile.
In this work, Ni-Sr-Al O -ZrO aerogel (NSAZ-Ae) catalyst was
2
3
2
prepared by an epoxide-driven sol-gel method and a subse-
quent supercritical drying in a stream of CO . For comparison,
were calculated from H -TPD (temperature-programmed desorp-
2
2
Ni-Sr-Al O -ZrO xerogel (NSAZ-Xe) catalyst with the same chem-
tion) profiles by assuming that adsorption stoichiometry of H/Ni is
2
3
2
−
20
2
ical composition was also prepared by an epoxide-driven sol-gel
method and a subsequent conventional drying method. The effect
of drying method on the physicochemical properties and cat-
1 and cross-sectional area of single nickel atom is 6.49 × 10
m .
H -TPD analyses were conducted with a BELCAT-B instrument
2
(BEL Japan), where 0.02 g of each catalyst was reduced under 5%
◦
alytic activities of Ni-Sr-Al O -ZrO catalysts in the ethanol steam
hydrogen/argon (50 ml/min) at 700 C for 2 h. Hydrogen adsorp-
2
3
2
reforming was investigated.
tion was done by flowing 5% hydrogen/argon flow (50 ml/min) at
◦
5
0 C for 30 min, followed by purging the reactor under an argon
◦
flow (50 ml/min) at 100 C for 1 h to remove physisorbed hydro-
gen molecules. Furnace temperature was increased from 50 C to
◦
2
. Experimental
◦
◦
1
000 C at a rate of 5 C/min under an argon flow (30 ml/min),
2
.1. Preparation of catalysts
and the amount of desorbed hydrogen was detected with a TCD.
In order to conduct EtOH-TPD experiments, 0.02 g of each cata-
lyst was reduced under a flow of hydrogen (3 ml/min) diluted with
A mesoporous nickel-strontium-alumina-zirconia aerogel cata-
◦
lyst was prepared by a single-step epoxide-driven sol-gel method
followed by a supercritical CO2 drying [31]. 0.74 g of zirconium
precursor (zirconium oxynitrate hydrate, Sigma-Aldrich, 99%) and
0
9
helium (30 ml/min) at 700 C for 2 h. After cooling down the reactor
to room temperature, 10 ml of vaporized ethanol was pulsed into
the reactor for several times along with 5 ml/min of helium flow to
saturate the catalyst surface with ethanol. Physisorbed ethanol was
.18 g of strontium precursor (strontium nitrate, Sigma-Aldrich,
9%) were dissolved in 30 ml of ethanol (Solution A). 6 g of alu-
◦
removed by evacuating the sample at 100 C for 1 h under a flow of
minum precursor (aluminum nitrate nonahydrate, Sigma-Aldrich,
8%) and 0.89 g of nickel precursor (nickel nitrate hexahydrate,
Sigma-Aldrich, 97%) were separately dissolved in 30 ml of ethanol
Solution B). 5 ml and 3 ml of hydrochloric acid (hydrochloric acid,
helium (15 ml/min). Furnace temperature was elevated from room
◦
◦
9
temperature to 700 C at a rate of 10 C/min under a helium flow
(10 ml/min). Desorbed species were monitored using a GC–MSD
(6890N GC-5975MSD, Agilent) device. Different m/z values were
assigned to various desorbed species; 16, 28, 29, 31, and 44 for
(
Samchun, 36%) were then added into Solution A and Solution B,
respectively. Solution A and Solution B were then mixed under vig-
orous stirring to obtain a transparent solution (Solution C). After
Solution C was further stirred for a few minutes, 25 ml of propy-
lene oxide (Samchun, 99%) was introduced into the solution as a
gelation agent. The obtained green opaque gel was aged for 2 days.
The resultant gel was then transferred to an autoclave and it went
CH , CO, CH CHO, C H5OH, and CO , respectively. The amount of
4
3
2
2
carbon deposition on the used catalysts was determined by CHNS
elemental analyses with a CHNS 932 device (Leco).
2.3. Steam reforming of ethanol
◦
◦
through supercritical CO2 drying at 50 C and 120 atm for 12 h. The
Catalytic activity of NSAZ-Ae and NSAZ-Xe catalysts in the
ethanol steam reforming was evaluated in a fixed-bed reactor at
resulting powder was calcined at 550 C for 5 h to yield a Ni-Sr-
◦
Al O -ZrO aerogel (denoted as NSAZ-Ae) catalyst. For comparison,
450 C for 15 h under a continuous flow and atmospheric pressure.
2
3
2
a nickel-strontium-alumina-zirconia xerogel catalyst was also pre-
pared with the same procedures mentioned above, except that the
Prior to the reaction, 0.05 g of each catalyst sample was reduced
at 700 C for 2 h under a mixed flow of hydrogen (3 ml/min) and
◦

3
44
J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
Table 1
nitrogen (30 ml/min). Nitrogen (30 ml/min) was used as a carrier
gas during the reforming reaction. A liquid mixture of ethanol
and deionized water was continuously introduced into the reac-
tor using a syringe pump (US/KDS-101, KDScientific) at a rate of
◦
Physicochemical properties of NSAZ-Ae and NSAZ-Xe catalysts calcined at 550 C for
5
h.
Catalyst
Ni content Sr content Surface area
Pore volume
(cm /g)
Average pore
diameter
a
a
2
b
3
c
(
wt%)
(wt%)
(m /g)
0.5 ml/h, where the molar ratio of steam to ethanol in the liquid
(
nm)
mixture was fixed at 6. Inlet and the outlet lines of the reactor
◦
NSAZ-Ae 15.0
NSAZ-Xe 14.7
6.1
6.3
385
279
1.35
0.56
13.7
8.0
were heated at 220 C to maintain a vapor phase. Total feed rate
with respect to catalyst weight was set as 46,280 ml/(h gcat). Com-
positions of reaction products were analyzed using an on-line gas
chromatograph (ACME 6000, Younglin) equipped with a TCD. Pora-
pak N and Molecular Sieve 5A columns were used for product
separation. Ethanol conversion, hydrogen yield, and selectivity for
carbon-containing products were calculated by the following equa-
a
Determined by ICP-AES analyses.
Calculated by the BET equation at relative pressure (P/P0) of 0.05-0.30.
Total pore volume at P/P0 ∼ 0.990.
b
c
Detailed physicochemical properties of calcined catalysts are
tions, where n in Eq. (7) represented the stoichiometric factor of
summarized in Table 1. Although both samples maintained a meso-
porous structure after the calcination process, NSAZ-Ae catalyst
retained higher surface area, larger pore volume, and larger aver-
age pore diameter than NSAZ-Xe catalyst. This result implies that
supercritical drying during the catalyst preparation process was
helpful to maintain well-structured mesopores, where the col-
lapse of gel structure was suppressed by effective removal of
solvent molecules confined within the gel structure [30]. It was
also revealed that actual contents of nickel and strontium in the
NSAZ-Ae and NSAZ-Xe catalysts were in good agreement with the
designed values.
i
the carbon-containing species.
FEtOH,in − FEtOH,out
Ethanol conversion (%) =
Hydrogen yield (%) = 3
× 100
(5)
F
EtOH,in
F
H ,out
2
× 100
(6)
× (FEtOH,in
− FEtOH,out
)
n × F
i
i,carbon containing product
Si,carbon containing product (%) =
× 100 (7)
2
× (FEtOH,in − FEtOH,out
)
3
. Results and discussion
.1. Characterization of calcined catalysts
Calcined NSAZ-Ae and NSAZ-Xe catalysts were characterized
XRD patterns of calcined NSAZ-Ae and NSAZ-Xe catalysts are
presented in Fig. 2. No noticeable diffraction peaks related to alu-
mina, zirconia, and strontium species were detected, indicating that
these species were in an amorphous state. The presence of broad
NiO diffraction peaks were detected in two catalysts, implying that
nickel mainly existed as nickel oxide in both catalysts after the cal-
3
by nitrogen adsorption-desorption measurements. As shown in
Fig. 1(a), both NSAZ-Ae and NSAZ-Xe catalysts exhibited type IV
isotherms, which were found in samples where capillary conden-
sation occurred in pore structures with mesoporous range [31,32].
However, two samples showed different hysteresis loops. NSAZ-
Ae showed H1-type hysteresis loop indicating the presence of
pore network formed from spherical agglomerates, while NSAZ-
Xe exhibited H2-type hysteresis loop resulting from ‘ink-bottle’
shaped pore structure [32,33]. This disparity in pore structure
between two samples was due to different drying methods
employed during the preparation process. In case of NSAZ-Xe,
conventional drying of ethanol solvent resulted in shrinkage and
rearrangement of pore structure due to vapor-liquid interfacial ten-
sion which formed mesopores with smaller pore opening than body
cination. Existence of a weak peak corresponding to NiAl O4 was
2
also observed in the NSAZ-Ae catalyst, indicating that a small por-
tion of nickel was incorporated into Al O3 to form an aluminate
2
phase. Diffraction peaks related to nickel showed a broad shape
rather than a sharp form, implying that the mesoporous structure
developed in two catalysts prevented the crystallization of nickel
or alumina phases during the calcination process.
TPR profiles of NSAZ-Ae and NSAZ-Xe catalysts are presented in
Fig. 3. Both catalysts showed a reduction band within the range of
◦
600–700 C. Reduction of NiO (the major nickel species in Fig. 2)
into metallic nickel was responsible for this reduction band [35].
◦
Peak temperatures of reduction bands were 641 C for NSAZ-Ae
◦
and 615 C for NSAZ-Xe, respectively. TPR profile of NSAZ-Ae also
[
34]. On the other hand, NSAZ-Ae experienced less pore shrink-
showed a shoulder peak at high temperature region, which was
age, because the solvent was exchanged with liquid CO during the
attributed to the reduction of NiAl O4 [36]. The above result indi-
2
2
supercritical CO2 drying process. Fig. 1(b) shows the pore size dis-
tributions of the samples. The result showed that NSAZ-Ae retained
larger pore size than NSAZ-Xe.
cates that the metal-support interaction between alumina and
nickel species was stronger in the NSAZ-Ae than in the NSAZ-Xe.
NSAZ-Ae catalyst with excellent textural properties is anticipated
◦
Fig. 1. (a) Nitrogen adsorption-desorption isotherms of NSAZ-Ae and NSAZ-Xe catalysts calcined at 550 C for 5 h, and (b) BJH pore size distribution plots of the catalysts.

J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
345
Table 2
◦
Physicochemical properties of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for
2
h.
Catalyst
Surface area
Pore volume
(cm /g)
Average pore
diameter (nm)
2
a
3
b
(
m /g)
NSAZ-Ae
NSAZ-Xe
315
255
1.02
0.50
13.4
7.8
a
Calculated by the BET equation at relative pressure (P/P0) of 0.05-0.30.
Total pore volume at P/P0 ∼ 0.990.
b
Table 3
Nickel particle size of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for 2 h.
◦
Catalyst
Particle size (nm)
NSAZ-Ae
NSAZ-Xe
6.9
14.1
a
Calculated from Ni (2 0 0) diffraction peak in Fig. 5.
However, these physicochemical properties of reduced catalysts
showed decreased values when compared to those of calcined cat-
alysts. It is believed that pore blockage due to the release of reduced
nickel species from alumina lattice to support surface and sinter-
ing due to high temperature might be responsible for diminished
physicochemical properties of reduced catalysts [38]. Pore size dis-
tributions presented in Fig. 4(b) showed that NSAZ-Ae catalyst still
retained larger pore size than NSAZ-Xe catalyst.
◦
Fig. 2. XRD patterns of NSAZ-Ae and NSAZ-Xe catalysts calcined at 550 C for 5 h.
Crystalline structures of reduced NSAZ-Ae and NSAZ-Xe cata-
lysts were analyzed with XRD measurements as shown in Fig. 5.
Both samples showed sharp peaks indicative of metallic nickel,
while diffraction peaks corresponding to other nickel species which
were observed before the reduction (Fig. 2) disappeared. This
indicates that nickel species in the NSAZ-Ae and NSAZ-Xe were
completely reduced into metallic nickel after the reduction. In addi-
tion, diffraction peak intensity of metallic nickel of NSAZ-Xe was
much stronger than that of NSAZ-Ae. This indicates that NSAZ-Ae
retained smaller nickel crystallite particles than NSAZ-Xe. Nickel
particle sizes in the reduced catalysts were calculated by applying
the Scherrer equation on Ni (2 0 0) peaks of Fig. 5 as presented in
Table 3. Higher surface area (Table 1) and stronger metal-support
interaction (Fig. 3) of NSAZ-Ae caused better dispersion of nickel
species and stronger resistance against sintering, which led to
smaller nickel particle size. On the contrary, nickel species in the
NSAZ-Xe catalyst were more susceptible to sintering due to weak
metal-support interaction, leading to large nickel particle size after
the reduction [39]. To confirm the analyses concerning the disper-
sion and particle size of nickel in the reduced catalysts, TEM images
of reduced NSAZ-Ae and NSAZ-Xe were obtained as shown in Fig. 6.
It is evident that nickel particles were more finely dispersed in the
NSAZ-Ae (Fig. 6(a)) than in the NSAZ-Xe (Fig. 6(b)). Distributions of
nickel particle sizes in both samples are shown in Fig. 6. NSAZ-Xe
exhibited broader distribution of particle sizes than NSAZ-Ae. No
nickel particles larger than 10 nm were observed in the NSAZ-Ae,
while a considerable portion of nickel particles in the NSAZ-Xe was
larger than 10 nm.
◦
Fig. 3. TPR profiles of NSAZ-Ae and NSAZ-Xe catalysts calcined at 550 C for 5 h.
to provide the condition for fine dispersion of nickel species, lead-
ing to intimate metal-support interaction [37]. Specific reduction
bands related to strontium were not detected in both samples.
3.2. Characterization of reduced catalysts
3.3. Nickel surface area of reduced catalysts
Physicochemical properties of reduced NSAZ-Ae and NSAZ-Xe
catalysts were characterized with nitrogen adsorption-desorption
measurements as shown in Fig. 4. Both catalysts retained type
IV isotherms along with H1-type (NSAZ-Ae) and H2-type (NSAZ-
Xe) hysteresis loops (Fig. 4(a)). This indicates that the mesoporous
structure was still maintained after the reduction process at high
Fig. 7 shows the H -TPD results of reduced NSAZ-Ae and
2
NSAZ-Xe catalysts. NSAZ-Ae and NSAZ-Xe catalysts showed sev-
eral desorption peaks reaching over a wide temperature range.
To calculate the amount of hydrogen desorbed from metallic
nickel sites, H -TPD profiles were deconvoluted. Three different
2
◦
temperature (700 C). Table 2 summarizes the detailed physico-
domains resulting from deconvolution were attributed to desorbed
hydrogen molecules of various identities. First domain with the
chemical properties of reduced catalysts. Reduced NSAZ-Ae and
2
◦
NSAZ-Xe still retained high surface area (>255 m /g), large pore
lowest desorption peak temperature (<300 C) originates from des-
3
volume (>0.50 cm /g), and large average pore diameter (>7.8 nm).
orbed hydrogen weakly interacted with nickel sites. Next domain

3
46
J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
◦
Fig. 4. (a) Nitrogen adsorption-desorption isotherms of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for 2 h, and (b) BJH pore size distribution plots of the catalysts.
Table 4
◦
H2-TPD results of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for 2 h.
Catalyst
Amount of desorbed hydrogen
Nickel surface area
(m /g-Ni)
a
2
b
(
␮mol-H2/g)
Weak site
Strong site
(300–600 C)
Total
◦
◦
(
<300 C)
NSAZ-Ae
NSAZ-Ae
19.5
8.3
96.6
42.9
116.1
51.2
60.4
26.6
a
Calculated from deconvoluted peak area of H2-TPD profiles in Fig. 7.
Calculated by assuming H/Niatom = 1.
b
3.4. Ethanol adsorption capacity of reduced catalysts
The behavior of reduced NSAZ-Ae and NSAZ-Xe catalysts in
the adsorption of ethanol was studied with EtOH-TPD analyses
as presented in Fig. 8. Adsorbed ethanol on nickel-based catalyst
was desorbed in diverse forms due to various surface reaction
pathways [23]. The desorbed species were analyzed with a mass-
spectrometer, and thus, desorption profiles of reduced samples
were separated into individual signals of C H5OH, CH , CH CHO,
2
4
3
CO, and CO2 according to m/z value. In both NSAZ-Ae (Fig. 8(a))
and NSAZ-Xe (Fig. 8(b)) catalysts, ethanol (m/z = 31) and acetalde-
hyde (m/z = 29) were the first to be desorbed at around 200 C.
◦
Fig. 5. XRD patterns of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for 2 h.
◦
When ethanol goes through steam reforming on the surface of
nickel catalyst, ethanol is initially adsorbed as an ethoxide form
*
due to dehydrogenation of hydroxyl (CH CH O ). Then a subse-
3
2
◦
with intermediate desorption peak temperature (300–600 C) is
quent dehydrogenation from C␣ of the ethoxide species occurs to
*
ascribed to desorbed hydrogen which is strongly bonded to metal-
lic nickel. However, the last domain with the highest desorption
peak temperature (>600 C) is due to hydrogen desorbed from
form acetaldehyde species (CH CHO ) [42]. Desorption profiles of
3
ethanol and acetaldehyde arise as ethanol adsorbed on the cata-
lyst surface is desorbed in its original state or is dehydrogenated
to be desorbed as acetaldehyde. Adsorbed acetaldehyde species
then goes through C C bond cleavage to yield adsorbed methyl
◦
the support sublayer as a result of hydrogen spillover [40,41].
Therefore, the last domain at the highest temperature was not con-
sidered when calculating the nickel surface area. In other words,
the amount of desorbed hydrogen and nickel surface area were
calculated from the peak areas of the other two domains attributed
to active nickel sites by assuming the adsorption of one hydrogen
atom per single active nickel site [41].
The amount of hydrogen desorbed from two domains of the
samples and nickel surface area calculated from the amount of
desorbed hydrogen are summarized in Table 4. Total amount of
hydrogen desorbed from NSAZ-Ae was much larger than that
of NSAZ-Xe. Although almost identical amount of nickel species
existed in both samples, NSAZ-Ae catalyst exhibited higher nickel
surface area. Excellent textural properties of aerogel-based catalyst
*
*
*
*
group, carbon oxides, and hydrogen (CH CHO → CH + CO + H )
3
3
[43]. This is related to the fact that other carbon-containing species
such as CH , CO and CO2 are then desorbed at higher tempera-
4
ture. In particular, carbon dioxide exhibits desorption peak with the
broadest temperature range in both samples, reaching almost up to
◦
550 C. Sequential dehydrogenation of methyl groups adsorbed on
catalyst surface results in carbon species, and these carbon species
react with oxygen to produce carbon oxides that is desorbed at suf-
ficiently high temperature [44]. Similarity of NSAZ-Ae and NSAZ-Xe
in the sequence of desorbed intermediates with regard to temper-
ature suggests that adsorbed ethanol behaves in an analogous way
on the surface of both catalysts.
(
NSAZ-Ae) is expected to enable better dispersion of active nickel
To calculate the ethanol adsorption capacity of the catalysts, it
was assumed that the sum of various desorbed carbon-containing
species was equal to the total quantity of adsorbed ethanol ini-
phase, leading to smaller nickel particle size (Table 3) and higher
nickel surface area.

J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
347
◦
Fig. 6. TEM images and particle size distributions of (a) NSAZ-Ae and (b) NSAZ-Xe catalysts reduced at 700 C for 2 h.
Table 5
◦
EtOH-TPD results of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for 2 h.
Catalyst
NSAZ-Ae
0.691
NSAZ-Xe
0.626
Amount of ethanol adsorbed(mmol-EtOH/g-catalyst)
tially existing on the surface of the reduced catalysts. Although each
carbon-containing intermediate was assigned according to m/z
value, desorbed ethanol (m/z = 31) and carbon dioxide (m/z = 44)
were also known to express weak signals at m/z = 29 and m/z = 28,
respectively. Thus, a portion of acetaldehyde (m/z = 29) and car-
bon monoxide (m/z = 28) desorption peaks was regarded as ethanol
and carbon dioxide when calculating the amount of desorbed
species. Table 5 summarizes the ethanol adsorption capacity of
the catalysts. NSAZ-Ae catalyst showed higher capacity for ethanol
adsorption than NSAZ-Xe catalyst. Therefore, it is expected that
NSAZ-Ae with higher nickel surface area (Table 4) and higher spe-
cific surface area (Table 2) provides more active sites where ethanol
reactant can adsorb on.
◦
Fig. 7. H2-TPD profiles of NSAZ-Ae and NSAZ-Xe catalysts reduced at 700 C for 2 h.
3.5. Hydrogen production by steam reforming of ethanol
Catalytic performance of NSAZ-Ae and NSAZ-Xe catalysts in the
NSAZ-Ae catalyst. Excellent textural properties and high nickel
dispersion of NSAZ-Ae catalyst were responsible for such improve-
ment. Large pore volume promotes the reactant molecules to be
transferred more efficiently during the reforming reaction, while
high nickel dispersion provides more accessible active sites where
ethanol can preferably adsorb on.
Hydrogen yield, ethanol conversion, and selectivity for by-
products after a 15 h-reaction are summarized in Table 6. Ethanol
supplied onto the catalysts resulted in complete conversion.
◦
ethanol steam reforming at 450 C is shown in Fig. 9. It was revealed
that NSAZ-Ae catalyst was more effective in the steam reforming
of ethanol, leading to higher hydrogen yield. In addition, NSAZ-Ae
catalyst maintained a stable catalytic activity without deactivation
for 15 h, while NSAZ-Xe catalyst experienced a slight catalyst deac-
tivation. Thus, introduction of supercritical drying method during
the catalyst preparation process had a positive effect in the ethanol
steam reforming reaction by increasing activity and stability of

3
48
J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
◦
Fig. 8. EtOH-TPD profiles of (a) NSAZ-Ae and (b) NSAZ-Xe catalysts reduced at 700 C for 2 h.
Table 7
Amount of carbon deposition on NSAZ-Ae and NSAZ-Xe catalysts after a 15 h-
reaction.
Catalyst
NSAZ-Ae
56.2
NSAZ-Xe
35.7
Amount of carbon deposition (wt%)a
a
Determined by CHNS elemental analyses.
Fig. 9. Hydrogen yield with time on stream in the steam reforming of ethanol over
◦
◦
NSAZ-Ae and NSAZ-Xe catalysts at 450 C. All the catalysts were reduced at 700 C
for 2 h prior to the reaction.
Table 6
Detailed catalytic performance of NSAZ-Ae and NSAZ-Xe catalysts in the steam
◦
reforming of ethanol at 450 C after a 15 h-reaction.
Catalyst
NSAZ-Ae
NSAZ-Xe
Ethanol conversion (%)
Hydrogen yield (%)
Selectivity for CH4 (%)
Selectivity for C2H4 (%)
Selectivity for CO (%)
Selectivity for CO2 (%)
100
119.4
24.5
0
15.9
59.6
100
98.1
26.3
0
24.1
49.6
Fig. 10. XRD patterns of NSAZ-Ae and NSAZ-Xe catalysts after a 15 h-reaction.
3.6. Characterization of used catalysts
NSAZ-Ae and NSAZ-Xe catalysts obtained after a 15 h-reaction
Methane, carbon monoxide, and carbon dioxide were formed
as intermediate products during the steam reforming reaction.
Ethylene formed by ethanol dehydration (Equation (4)) was not
detected in the outlet gas over both catalysts, suggesting that
ethylene species were reformed and/or acted as a coke precur-
sor on the catalyst surface. There was no great difference in
methane selectivity between NSAZ-Ae and NSAZ-Xe. Selectivity
for CO over NSAZ-Xe was higher than that over NSAZ-Ae, while
CO2 selectivity was lower. This implies that water-gas shift reac-
tion (CO + H O → CO + H ) preferentially occurred over NSAZ-Ae
were characterized with CHNS, XRD, and TEM analyses. The amount
of carbon deposition on the used catalysts is summarized in Table 7.
NSAZ-Ae catalyst, which exhibited higher hydrogen yield and more
stable performance, showed a higher degree of carbon deposition.
It is believed that NSAZ-Ae with higher area of exposed catalyst
surface also retains more surface acid sites on Al O [21], providing
2
3
the conditions favorable for carbon deposition.
XRD patterns of used NSAZ-Ae and NSAZ-Xe samples are shown
in Fig. 10. A strong peak corresponding to carbon was detected in
both catalysts due to severe carbon deposition during the ethanol
steam reforming reaction. Peak intensity of metallic nickel was still
2
2
2
catalyst, contributing to high hydrogen yield [45].

J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
349
Fig. 11. TEM images of (a) NSAZ-Ae and (b) NSAZ-Xe catalysts after a 15 h-reaction.
stronger in the NSAZ-Xe than in the NSAZ-Ae. However, peak inten-
sity of metallic nickel was attenuated in both catalysts after the
reaction when compared to that in Fig. 5. This decrement of peak
intensity was due to dilution effect caused by considerable quantity
of carbon deposited on NSAZ-Ae and NSAZ-Xe catalysts [46]. TEM
images of used NSAZ-Ae (Fig. 11(a)) and NSAZ-Xe (Fig. 11(b)) indi-
cate the formation of filamentous carbon species on both samples
after a 15 h-reaction. It is known that filamentous carbon does not
pose a severe threat to catalytic activity by encapsulating the active
[2] A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy Fuels 19 (2005)
098–2106.
3] K. Mazloomi, C. Gomes, Renew. Sustain. Energy Rev. 16 (2012) 3024–3033.
4] M. Momirlan, T.N. Veziroglu, Int. J. Hydrogen Energy 30 (2005) 795–802.
2
[
[
[5] W. Lubitz, W. Tumas, Chem. Rev. 107 (2007) 3900–3903.
[
[
6] K. Takeishi, H. Suzuki, Appl. Catal. A: Gen. 260 (2004) 111–117.
7] K. Vasudeva, N. Mitra, P. Umasankar, S.C. Dhingra, Int. J. Hydrogen Energy 21
1996) 13–18.
(
[8] M. Benito, J.L. Sanz, R. Isabel, R. Padilla, R. Arjona, L. Daza, J. Power Sources 151
(2005) 11–17.
[
9] C. Montero, A. Ochoa, P. Casta n˜ o, J. Bilbao, A.G. Gayubo, J. Catal. 331 (2015)
181–192.
sites in conditions of high H /CO or steam/hydrocarbon ratio [47].
[10] A.C. Basagiannis, P. Panagiotopoulou, X.E. Verykios, Top. Catal. 51 (2008) 2–12.
[11] S. Cavallaro, V. Chiodo, S. Freni, N. Mondello, F. Frusteri, Appl. Catal. A: Gen.
2
Thus, it is expected that carbon deposition on the used NSAZ-Ae
and NSAZ-Xe catalysts was not related to encapsulation of nickel,
and both catalysts did not suffer from severe deactivation despite
the considerable amount of deposited carbon.
2
49 (2003) 119–128.
12] C.C. Hung, S.L. Chen, Y.K. Liao, C.H. Chen, J.H. Wang, Int. J. Hydrogen Energy 37
2012) 4955–4966.
[13] R. Padilla, M. Benito, L. Rodrigues, G. Mu n˜ oz, L. Daza, Int. J. Hydrogen Energy
5 (2010) 8921–8928.
[
(
3
[
14] S.M. de Lima, A.M. Silva, I.O. da Cruz, G. Jacobs, B.H. Davis, L.V. Mattos, F.B.
Noronha, Catal. Today 138 (2008) 162–168.
4
. Conclusions
Ni-Sr-Al O -ZrO catalysts were prepared by an epoxide-driven
[
15] S.M. de Lima, A.M. da Silva, L.O.O. da Costa, J.M. Assaf, L.V. Mattos, R. Sarkari,
A. Venugopal, F.B. Noronha, Appl. Catal. B: Environ. 121 (2012) 1–9.
2
3
2
[16] L. Barattini, G. Ramis, C. Resini, G. Busca, M. Sisani, U. Constantino, J. Chem.
Eng. 153 (2009) 43–49.
[
sol-gel method. Aerogel-based catalyst (NSAZ-Ae) was obtained via
supercritical CO2 drying of the composite gel. Xerogel-based cat-
alyst (NSAZ-Xe) was obtained via conventional drying. Although
both NSAZ-Ae and NSAZ-Xe catalysts maintained well-structured
mesopores, NSAZ-Ae retained excellent textural properties com-
pared to NSAZ-Xe. NSAZ-Ae showed stronger interaction between
metal and support, maintaining smaller nickel particles after the
17] R. Trane-Restrup, S. Dahl, A.D. Jensen, Int. J. Hydrogen Energy 38 (2013)
5105–15118.
18] J.R. Rostrup-Nielsen, J. Sehested, J.K. Nørskov, Adv. Catal. 47 (2002) 65–139.
1
[
[19] M.H. Youn, J.G. Seo, H. Lee, Y. Bang, J.S. Chung, I.K. Song, Appl. Catal. B:
Environ. 98 (2010) 57–64.
[
[
20] D.L. Trimm, Catal. Today 49 (1999) 3–10.
21] H.S. Bengaard, J.K. Nørskov, J. Sehested, B.S. Clausen, L.P. Nielsen, A.M.
Molenbroek, J.R. Rostrup-Nielsen, J. Catal. 209 (2002) 365–384.
reduction process. From H -TPD analyses, it was revealed that
[22] M.C. Sánchez-Sánchez, R.M. Navarro, J.L.G. Fierro, Catal. Today 129 (2007)
36–345.
23] C.K.S. Choong, L. Huang, Z. Zhong, J. Lin, L. Hong, L. Chen, Appl. Catal. A: Gen.
07 (2011) 155–162.
[24] J.H. Song, S.J. Han, J. Yoo, S. Park, D.H. Kim, I.K. Song, J. Mol. Catal. A: Chem.
18–419 (2016) 68–77.
2
3
NSAZ-Ae showed higher nickel surface area due to better disper-
sion state of nickel. Higher nickel surface area was also related to
higher ethanol adsorption capacity of NSAZ-Ae catalyst. Catalytic
activity and stability of NSAZ-Ae was found to be higher than those
of NSAZ-Xe, due to larger amount of active nickel sites and greater
affinity toward ethanol of NSAZ-Ae catalyst. It was concluded that
aerogel-based Ni-Sr-Al O -ZrO catalyst (NSAZ-Ae) prepared by a
[
4
4
[
25] S. Therdthianwong, A. Therdthianwong, C. Siangchin, S. Yongprapat, Int. J.
Hydrogen Energy 33 (2008) 991–999.
[26] Z.-X. Li, F.-B. Shi, L.-L. Li, T. Zhang, C.-H. Yan, Phys. Chem. Chem. Phys. 13
2011) 2488–2491.
27] J.G. Seo, M.H. Youn, J.C. Jung, I.K. Song, Int. J. Hydrogen Energy 35 (2010)
738–6746.
[28] W. Li, Q. Yue, Y. Deng, D. Zhao, Adv. Mater. 25 (2013) 5129–5152.
(
2
3
2
[
supercritical drying method served as an efficient ethanol steam
reforming catalyst.
6
[
29] S.J. Han, Y. Bang, J. Yoo, S. Park, K.H. Kang, J.H. Choi, J.H. Song, I.K. Song, Int. J.
Hydrogen Energy 39 (2014) 10445–10453.
Acknowledgement
[30] Y. Bang, S.J. Han, J. Yoo, S. Park, J.H. Choi, Y.J. Lee, J.H. Song, I.K. Song, Int. J.
Hydrogen Energy 39 (2014) 4909–4916.
[
31] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.
Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619.
This work was supported by the Global Frontier R&D Program
on Center for Multiscale Energy System funded by the National
Research Foundation under the Ministry of Science, ICT & Future
Planning, Korea (20110031575).
[32] G. Leofanti, M. Padovan, G. Tozzola, B. Venturelli, Catal. Today 41 (1998)
07–219.
[
2
33] T.F. Baumann, A.E. Gash, S.C. Chinn, A.M. Sawvel, R.S. Maxwell, J.H. Satcher,
Chem. Mater. 17 (2005) 395–401.
[
[
[
34] C.L. Glaves, C.J. Brinker, D.M. Smith, P.J. Davis, Chem. Mater. 1 (1989) 34–40.
35] J.M. Rynkowski, T. Paryjczak, M. Lenik, Appl. Catal. A: Gen. 106 (1993) 73–82.
36] J. Zieli n´ ski, J. Catal. 76 (1982) 157–163.
References
[1] M.S. Dresselhaus, I.L. Thomas, Nature 414 (2001) 332–337.

3
50
J.H. Song et al. / Journal of Molecular Catalysis A: Chemical 424 (2016) 342–350
[
[
[
[
[
37] B. Mile, D. Stirling, M.A. Zammitt, A. Lovell, M. Webb, J. Catal. 114 (1988)
17–229.
38] J. Yoo, Y. Bang, S.J. Han, T.H. Kang, J. Lee, I.K. Song, J. Mol. Catal. A: Chem. 380
2013) 28–33.
39] C.K.S. Choong, Z. Zhong, L. Huang, Z. Wang, T.P. Ang, A. Borgna, J. Lin, L. Hong,
L. Chen, Appl. Catal. A: Gen. 407 (2011) 145–154.
[42] D. Zanchet, J.B.O. Santos, S. Damyanova, J.M.R. Gallo, J.M.C. Bueno, ACS Catal. 5
(2015) 3841–3863.
[43] A.N. Fatsikostas, X.E. Verykios, J. Catal. 225 (2004) 439–452.
[44] A.L. Alberton, M.M.V.M. Souza, M. Schmal, Catal. Today 123 (2007) 257–264.
[45] S.J. Han, Y. Bang, J. Yoo, K.H. Kang, J.H. Song, J.G. Seo, I.K. Song, Int. J. Hydrogen
Energy 38 (2013) 15119–15127.
2
(
40] Y. Cesteros, P. Salagre, F. Medina, J.E. Sueiras, Appl. Catal. B: Environ. 25 (2000)
[46] A. Carrero, J.A. Calles, A.J. Vizcaíno, J. Chem. Eng. 163 (2010) 395–402.
[47] C.H. Bartholomew, Appl. Catal. A: Gen. 212 (2001) 17–60.
2
13–227.
41] S. Velu, S.K. Gangwal, Solid State Ionics 177 (2006) 803–811.