Highly stable Ni-Al2O3 catalyst prepared from a Ni-Al layered double hydroxide for ethanol decomposition toward hydrogen

Article abstract of DOI:10.1016/j.apcata.2015.10.007

The preparation, characterization, activity, and stability of a Ni-Al₂O₃ catalyst derived from reduction of a Ni-Al layered double hydroxide precursor (LDH, Ni₆Al₂(OH)₁₆(CO₃)_0.75(OH)_0.25·4H₂O) are reported in this paper. In-situ X-ray adsorption spectroscopy shows that reduction of Ni from the LDH precursor to form a highly loaded 80% Ni-Al₂O₃ catalyst (Ni-Al₂O₃-LDH) is faster than reduction of a 10% impregnated Ni-Al₂O₃ alumina (Ni-Al₂O₃-I) catalyst. The reduced Ni-Al₂O₃-LDH catalyst exhibits highly dispersed Ni nanoparticles (3-5 nm) distributed on top, partially embedded nanoparticles, and some encapsulated in the Al₂O₃ matrix. The nanoparticles impregnated on alumina (Ni-Al₂O₃-I) are larger (～7-15 nm) and appear on top of the alumina support. Conversion vs time on stream (TOS) results during ethanol decomposition at 250°C on Ni-Al₂O₃-LDH exhibits only a slight deactivation during 100 h TOS, while the Ni-Al₂O₃-I catalyst shows rapid deactivation with no conversion after 2 h TOS. X-ray photoelectron spectroscopy shows that the carbon content increases up to 48% after 100 h TOS on the Ni-Al₂O₃-LDH catalyst, while a similar increase occurs after 2 h TOS on the Ni-Al₂O₃-I catalyst. TEM shows that after 100 h TOS either a thin layer of amorphous carbon or carbon nanotubes forms on Ni on top of the alumina matrix and on partially embedded Ni nanoparticles on the Ni-Al₂O₃-LDH catalyst. Total surface area of the Ni-Al₂O₃-LDH catalyst increased during TOS, which may be suplying fresh surface Ni from the encapsulated Ni nanoparticles that sustain the high activity.

Full text of DOI:10.1016/j.apcata.2015.10.007

Accepted Manuscript
Title: Highly Stable Ni-Al O Catalyst Prepared from a Ni-Al
2
3
Layered Double Hydroxide for Ethanol Decomposition
toward Hydrogen
Author: Khachatur V. Manukyan Allison J. Cross
ArmenuhiV. Yeghishyan Sergei Rouvimov Jeffrey J. Miller
Alexander S. Mukasyan E.E. Wolf
PII:
S0926-860X(15)30184-8
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.10.007
APCATA 15583
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
23-7-2015
3-10-2015
6-10-2015
Please cite this article as: Khachatur V.Manukyan, Allison J.Cross,
ArmenuhiV.Yeghishyan, Sergei Rouvimov, Jeffrey J.Miller, Alexander S.Mukasyan,
E.E.Wolf, Highly Stable Ni-Al2O3 Catalyst Prepared from a Ni-Al Layered Double
Hydroxide for Ethanol Decomposition toward Hydrogen, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2015.10.007
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Highly Stable Ni-Al
2
O Catalyst Prepared from a Ni-Al Layered Double
3
Hydroxide for Ethanol Decomposition toward Hydrogen
1
2
2
Khachatur V. Manukyan , Allison J. Cross , ArmenuhiV. Yeghishyan
3
4
2
2
Sergei Rouvimov , Jeffrey J. Miller , Alexander S. Mukasyan , and E. E. Wolf *
1Department of Physics, University of Notre Dame, Notre Dame, IN, 46556
2Department of Chemical & Biomolecular Engineering, University of Notre Dame,
Notre Dame, IN, 46556
3
Department of Electrical Engineering, University of Notre Dame,
Notre Dame, IN, 46556
4
School of Chemical Engineering, Purdue University, West Lafayette, IN 47907
*
To whom correspondence should be addressed. Phone: +1-574.631.5897;
E-mail: ewolf@nd.edu
1

Graphical abstract
2

Highlights




A Ni-Al layered double hydroxide is an efficient precursor to produce a 80% Ni-Al
catalysts with highly dispersed and embedded Ni nanoparticles on an Al matrix.
LDH Ni-Al catalyst exhibits excellent stability towards deactivation during ethanol
decomposition to produce H as compared to catalysts prepared by impregnation .
2
O
3
2
O
3
2
O
3
2
We hypothesize that carbon formation occurs mainly on top Ni whereas on partially
embedded Ni particles is slow and none occurs in encapsulated Ni particles
Encapsulated Ni particles, when exposed by spalling, can provide fresh surface area to
sustain the reaction for longer TOS at high conversion.
3

ABSTRACT
The preparation, characterization, activity, and stability of a Ni-Al
2
O
3
catalyst derived from
precursor (LDH,
O) is reported in this paper. In-situ X-ray adsorption
spectroscopy show that reduction of Ni from the LDH precursor to form a highly loaded 80%Ni-
Al catalyst (Ni-Al -LDH) is faster than reduction of a 10% impregnated Ni-Al alumina
Ni-Al -I) catalyst. The reduced Ni-Al -LDH catalyst exhibits highly dispersed Ni
nanoparticles (3-5 nm) distributed on top, partially embedded nanoparticles, and some
encapsulated in the Al matrix. The nanoparticles impregnated on alumina (Ni-Al -I) are
larger (~7-15 nm) and appear on top of the alumina support. Conversion vs time on stream
TOS) results during ethanol decomposition at 250°C on Ni-Al -LDH exhibits only a slight
deactivation during 100 h TOS, while the Ni-Al -I catalyst shows rapid deactivation with no
conversion after 2 h TOS. X-ray photoelectron spectroscopy shows that the carbon content
increases up to 48% after 100 h TOS on the Ni-Al -LDH catalyst, while a similar increase
occurs after 2 h TOS on the Ni-Al -I catalyst. TEM shows that after 100 h TOS either a thin
layer of amorphous carbon or carbon nanotubes forms on Ni on top of the alumina matrix and on
partially embedded Ni nanoparticles on the Ni-Al -LDH catalyst. Total surface area of the Ni-
Al -LDH catalyst increased during TOS, which may be suplying fresh surface Ni from the
encapsulated Ni nanoparticles that sustain the high activity.
reduction
of
a
Ni-Al
layered
double
hydroxide
Ni Al (OH)16(CO
6
2
3
)
0.75(OH)0.25•4H
2
2
O
3
2
O
3
2
O
3
(
2
O
3
2
O
3
2
O
3
2
O
3
(
2
O
3
O
2 3
2
O
3
2
O
3
2
O
3
2
O
3
Key words: ethanol decomposition, Ni-Al
2
3
O supported catalysts, Ni-Al layered double
hydroxide, activity, deactivation stability.
4

INTRODUCTION
Ethanol (C
2
5
H OH) is a potential candidate as a hydrogen chemical carrier for fuel cell
applications [1-3]. Ethanol is non-hazardous compared to methanol and can be produced readily
by the fermentation of sugar cane, corn grains, and other starch-rich materials [4-7]. In addition,
several novel technologies are being developed to produce bioethanol from agricultural wastes
and forestry residues [1,6,8].
The following reactions are possible routes for hydrogen production from ethanol: steam
reforming (Eq. 1), partial oxidation (Eq. 2), and ethanol decomposition (Eq. 3,4),
C
2
H
5
OH + 3H
2
O = 2CO
= 4CO
CHO + H
2
+ 6H
2
(1)
(2)
(3)
(4)
2
C
2
H
5
OH + 3O
2
2
+ 5H
2
C
C
2
2
H
5
5
OH = CH
3
4
2
H
OH = CH
+ CO + H
2
Although the steam reforming of ethanol provides a high hydrogen yield, it is a highly
endothermic reaction. Therefore, high operating temperatures are necessary in order to achieve
high conversion [2,7-15]. The partial oxidation of ethanol is an exothermic reaction that exhibits
fast start-up and response times [16-23], both of which are critically important for vehicle fuel
cell applications. Drawbacks for the partial oxidation of ethanol include lower selectivity and
yields due to hydrogen combustion and the difficulty in controlling the reaction due to hot-spot
formation. Ethanol decomposition is another reaction pathway for hydrogen generation from
ethanol [16, 24-28], and relatively less literature is available for this reaction as compared to
steam reforming and partial oxidation pathways. A main drawback of ethanol decomposition is
the formation of carbon that deactivates the catalyst, and to a lesser extent, the formation of
undesirable byproducts such as acetaldehyde and carbon monoxide. Regardless of economic
5

considerations, this reaction was selected in this study as a probe reaction to examine catalyst
deactivation using a novel catalyst preparation method. Detailed mechanistic discussion of the
various reaction pathways and the role of Ni in the activity and selectivity of the reaction have
been presented in detail in the literature [1,16,24,29] so they will be not be discussed hereafter
except in connection to specific issues related to the results obtained.
Catalytic stability towards deactivation, however, is one of the most important challenges
for hydrogen production from ethanol. Regardless of the metal and support used, significant
deactivation has been reported in the literature for applications involving Pt [30,31], Pd [32,33],
Rh [33], Ru [10,31], Co [34-37] and Ni-based [38-41] catalysts. Deactivation aspects of catalysts
during ethanol conversion were thoroughly reviewed in the recent article by Mattos et al [1].
Catalyst deactivation is attributed to the deposition of carbonaceous species as well as sintering
and/or oxidation of metal crystallites.
Filamentous and amorphous carbons are two general types of carbonaceous species that
form on metal catalysts. The nature of the carbon formed depends both on the operating
conditions (primarily the reaction temperature), and the nature of the metal. Polymerization of
ethylene intermediates and the disproportionation of carbon monoxide (i.e., the Boudouard
reaction) are the main carbon formation (coking) reactions at low temperatures. At higher
temperatures, decomposition of intermediate hydrocarbons (e.g., methane, ethylene) and
byproducts (e.g. acetaldehyde) are the main pathways to form carbon.
On Ni- and Co-supported catalysts, at high temperatures, the deposited carbon species
diffuse into the metal particles to nucleate the growth of filaments or nanotubes at the metal-
support interface. Depending on the microstructure of the crystallite, in some cases, the
nanotubes can lift the metal particles off the support [42]. However, the surface at the tip of the
6

metal particle may still remain accessible to reactants, without deactivation albeit of significant
carbon formation. At lower temperatures, the catalyst may deactivate through deposition of
amorphous coke. The amorphous carbon may cover both the metal particles and the support
[
1,34,43].
Another deactivation route [1,44] is the oxidation of ultra-small metallic crystallites,
leading to a lower activity and selectivity of the oxidized metal surface. This problem may be
controlled by careful balancing of the oxygen-to-ethanol ratio above the value in which oxidation
of the metal occurs [45]. Yet, another pathway of catalyst’s deactivation during ethanol
reforming is the sintering of metal particles. High reaction temperatures or the exothermic nature
of partial oxidation reaction can significantly contribute to the sintering of catalysts nanoparticles
[
1]. To prevent sintering, supports, such as ceria, or zirconia that exhibit a strong interaction with
the small metal nanoparticles have been studied [19,46].
Previously, we prepared Ni catalysts for the ethanol reforming reaction using a simple
combustion synthesis method [28]. Here, we demonstrate a different strategy of catalyst
preparation for ethanol reforming using a Ni-Al layered double hydroxide (LDH) as precursor.
Recent review articles suggested that reduction of LDHs permits formation of highly dispersed
metal particle-oxide nanomaterials that could be used as catalysis [47-49]. Our results show that
short thermal treatment of LDH under a reductive atmosphere produces in-situ a highly porous
8
0% Ni-Al
2
3
O catalysts in which small size (3-5nm) metal crystallites on the surrounding
alumina matrix. Here we investigated the time on stream (TOS) stability towards deactivation of
LDH-derived catalyst during ethanol decomposition and compared it with a Ni catalyst prepared
by impregnation. Textural and spectroscopic characterization of the Ni-Al
2
O
3
-LDH catalyst are
7

presented to explain the high activity and stability of this catalyst for the ethanol decomposition
reaction.
2
. EXPERIMENTAL METHODS
The Ni-Al LDH precursor was prepared by the co-precipitation method. Briefly, an aqueous
solution (50 mL) containing 10.89 g of Ni(NO ·6H O and 4.68 g of Al(NO ·9H O (the total
concentration of cations was 1 mol/L with a Ni/Al molar ratio of 3:1) was added dropwise into a
0 mL continually stirred solution containing 4.24 g of Na CO (0.8 mol/L) and 2.40 g of NaOH
1.2 mol/L), while the pH value of the suspension was adjusted to 10 using 3 mol/L NaOH
3
)
2
2
3
)
3
2
5
2
3
(
aqueous solution. The suspension was then kept at 65 °C for 24 h to obtain a precipitate with Ni-
Al LDH structure. The precipitate was recovered by filtration and repeatedly washed with
deionized water. After drying at 80°C, the resulting material was reduced at 500 °C for one hour
in pure hydrogen to get the Ni-Al
2
3
O -LDH catalyst with a Ni loading of 80 wt.%. It should be
noted that no LDH structure was detected at lower Ni concentration, so to compare results with a
typical impregnated catalysts two alternative formulations were prepared.
A catalyst with 10 wt% of Ni loading was prepared by incipient wetness impregnation of
Al
2
O
3
support. The Al
2
O
3
support was prepared by the same precipitation method described
o
above, but without addition of nickel nitrate. The washed precipitate was heated to 500 C for one
hour in hydrogen to get a transition -Al
of Ni(NO ·6H O followed by drying at 80°C for 8 h and reduction at 500°C for 1 h in pure
hydrogen. This catalyst is referred as Ni-Al -I and to compare its activity and stability with the
LDH-derived catalysts, we prepared a physical mixture of the Ni-Al -LDH and Al with a
2
O3. Then, it was impregnated with an aqueous solution
3
)
2
2
2
O
3
2
O
3
2
O
3
1
0% Ni content, referred as 10% Ni-Al
2
O
3
-LDH/Al
2
O
3
. The physical mixture avoids introducing
8

chemical changes during preparation, but it allows to scale down of Ni content so that
deactivation could be observed in a shorter TOS.
FTIR spectra of catalyst precursors were measured on a Bruker Tensor 27 spectrometer
−
1
−1
after 64 scans within 4000–400 cm at a resolution of 4 cm by measuring the infrared
absorbance of sample using diffuse reflection sampling technique.
Catalytic activity and selectivity during ethanol decomposition were measured in a
continuous flow fixed bed quartz reactor at atmospheric pressure as described elsewhere [25,28].
The Ni-Al LDH precipitates were ground into a fine powder, pressed into a wafer, ground again,
and then sieved to select particles from 0.6 to 1 mm in size to eliminate mass transfer diffusion.
0
.4 g of the precursor was placed into the reactor and reduced with hydrogen at 500 °C for 1 h.
After purging with nitrogen for 2 hours, the nitrogen flow was diverted into a bubbler filled with
ethanol and sent to the reactor, operating at 250 °C. The products in the reactor effluent were
analyzed using two gas chromatographs connected in parallel to measure the resulting gas
concentrations. Conversion and product selectivity were calculated as described elsewhere [28]
in Equations 5-6, where 푛 is the molar flow rate:
푖
푛퐸푡푂퐻,푖ꢀ − 푛퐸푡푂퐻,표푢푡
(5)
푋퐸푡푂퐻
=
× 100
푛퐸푡푂퐻,푖ꢀ
푛푖
(6)
휒푖 =
× 100
∑
푛푖
X-ray diffraction (XRD) patterns were obtained in a D8 Advance powder diffractometer
(Bruker), using CuKα monochromatic radiation (=0.1541 nm) operated at 40 kV and 40 mA at
o
-1
a scan rate of 1 min in angular range (2) from 20 to 80 degrees. X-ray photoelectron
spectroscopy (XPS) measurements were carried out in a PHI VersaProbe II spectrometer with an
9

Al Kα X-ray source operating at 1486.6 eV and a 90° take-off angle for near surface analysis of
C 1s, Ni 2p3, O 1s, Al 2p electronic transitions. Catalyst powders were adhered to stainless steel
mounts and loaded into the analysis chamber. Samples were outgassed in the vacuum system
−
7
maintained at a pressure less than ∼10 Pa. Binding energy values were referenced to the C 1s
peak (284.8 eV) that resulted from the adventitious contamination layer. The spectra were
analyzed using the Multipak XPS software package.
XANES measurements of in-situ reduction by hydrogen were made using X-ray
absorption spectroscopy (XAS) at the Advanced Photon Source (APS) at Argonne National
Laboratory as previously reported [24]. The measurements were obtained in transmission mode
with ionization chambers optimized for the maximum current with linear response (∼1010
photons detected/s). A cryogenically cooled double-crystal Si (111) monochromator with
resolution (ΔE) better than 2.5 eV at 8.979 keV (Cu K edge) was used in conjunction with a Rh-
coated mirror to minimize the presence of harmonics. 10 mg of the samples were loaded in a six
chamber sample holder by pressing a pellet in a chamber. The sample holder was then placed in
a quartz tube fitted with polyimide film on either end to allow X-rays to bombard the sample and
to be run in transmission mode. The tube featured a gas inlet and outlet fittings to allow
hydrogen flow through the samples. The tube was placed in a heater operated at temperatures up
to 500 °C and adjusted to be in line with the X-ray beam. A Ni foil spectrum was simultaneously
obtained with each sample for energy calibration. WinXAS and Athena software programs are
used to analyze XAS spectra data. A linear combination between the height of the absorption
edge of an elemental oxide state and its metallic state was used to determine the percentage of
reduction of a given sample.
1
0

BET specific surface area determinations were made using nitrogen
o
adsorption/desorption (-196 C) by an ASAP 2020 apparatus (Micromeritics). Before analysis,
the samples were dried at room temperature and then vacuum degassed at 373 K for 12 hrs.
A Titan 80-300 (FEI, USA) transmission electron microscope (TEM) with resolution of
0
.136 nm and with 0.1 nm information limit in scanning TEM (STEM) and high resolution TEM
mode, respectively, was used for structural and compositional characterization of the Ni
crystallites at nm- and atomic levels. The Titan TEM was equipped with an energy dispersive X-
ray spectrometer (EDS, Oxford Inc.) with a spectral energy resolution of 130 eV. Droplets of
small catalyst particles dispersed by sonication in an ethanol solution were deposited on a holey
carbon grid for TEM analysis.
A Costech ECS 4010 element analyzer was used to determine carbon content of catalysts.
Prior to analysis, catalysts were wrapped in a tin foil capsules and deposited into an oxygen rich
combustion zone, where the contents of the capsule combust. A capillary separation column (2 to
3
m in length) and a thermal conductivity detector were used to separate gas products of
combustion and measure CO concentration.
2
3
3
. RESULTS AND DISCUSSION
. 1 Precursor and catalyst characterization.
The Ni-Al –LDH catalyst’s precursor prepared by co-precipitation method was studied
2
O
3
by XRD and FTIR spectroscopy. Figure 1 shows XRD and the transmission FTIR spectrum of
the air dried precipitate. The XRD pattern (Figure 1A) shows that the dried precipitate exhibits
major diffraction lines of the takovite-type structure (Ni
15-0087) shown as an inset in Figure 1A.
6
Al
2
(OH)16(CO
3
)
0.75(OH)0.25•4(H O), ( JSPDS
2
#
1
1

2
-
The FTIR spectrum of the precipitate in Figure 1B shows bands typical of CO
3
ion-
intercalated LDH material, with a broad band at 3467 cm (OH), a peak at 1616ꢀcm− 1
−
1
−
1
(
H O), bands at 740, 563, ~400 cm (due to Ni(Al)-O vibrations and Ni(Al)–O–H bending),
2
−
1
2-
and the sharp and strong band at 1362 cm , which clearly shows the presence of the CO
3
ion in
the sample.
The reduction of the both the catalysts prepared by co-precipitation (Ni-Al
2
3
O -LDH) and
by impregnation (Ni-Al -I) methods were studied by XANES using 5%H -Ar mixture at
2
O
3
2
various temperatures. Spectra were taken approximately every 4 minutes as the temperature was
o
increased to 500°C at a rate of 3 C/min. Both samples were reduced simultaneously in the
beam-line and data was collected for each sample every other point switching between the two
samples (a data point was taken every 8 min per each sample). The fraction of reduced Ni was
calculated by linear interpolation between the height of the edge signal of NiO and a Ni
reference. Figure 2 shows that the reduction of Ni(II) in either the Ni-Al
Al -I did not occur before 300°C. Nickel in LDH precursor is reduced to ~10% metallic Ni at
80°C and becomes ~80% reduced Ni at 500°C. The catalyst prepared by impregnation,
however, is reduced to ~10% metallic Ni at 400°C and is only reduced to 45% Ni at 500°C.
Thus, at this temperature, Ni-Al -I is reduced at a lower rate, resulting in only half of the
extent of reduction than in the LDH precursor. Such remarkable differences in the kinetics of
2
O
3
-LDH or the Ni-
2
O
3
3
2
O
3
reduction may be explained by the nature of the precursors. The Al
2
3
O support used in
2
preparation of Ni-Al
2
O
3
-I precursor has 160 m /g (Table 1), which is slightly higher than LDH
2
one (140 m /g). Continuous structural rearrangement of the LDH precursor during reduction due
to the removal of CO
2
and water molecules occurs at a higher rate than removal of oxygen from
NiO in the impregnated catalyst. Table 1 shows that the BET surface area of reduced catalysts are 110
1
2

2
2
m /g and 135 m /g for Ni-Al
O
2 3
-LDH and Ni-Al
2
3
O -I, respectively. Both catalysts exhibit similar
broad pore sizes distribution ranging from 3 to 30 nm with an average size ~18 nm. Such large
structural transformation affects the resulting location, crystallite size, and dispersion of the
o
resulting Ni crystallites. The XANES results also justify the choice of 500 C in pure hydrogen
for 1 h as the reduction conditions. Note that dilute hydrogen was used in the XANES
experiments as per safety requirements set by the APS. Our XANES results are consistent with
TPR experiments on the reducibility of LDH type structures reported in the literature [50].
Scanning TEM images (STEM) of both reduced catalyst are shown in Figure 3. These
results indicate that reduction of LDH precursor yields a Ni-Al O3-LDH catalyst with highly
2
dispersed Ni nanoparticles. Statistical analysis shows (see inset in Figure 3) that the size
distribution of Ni nanoparticles is narrow with mean size of 4 nm. A high magnification STEM
image inserted in Figure 3A shows that near cubic shape Ni nanoparticles are partially embedded
in the Al
particles and the dark ones are the alumina matrix. Figure 3B shows Ni nanoparticles on the
alumina support prepared by impregnation (Ni-Al -I). On average these particles are larger
~11 nm) compared to those prepared from the LDH precursor.
2
O
3
matrix. Note that in this mode of STEM imaging, the bright images are the metal
2
O
3
(
3
.2 Conversion, selectivity, and time on stream (TOS) results
Figure 4 shows ethanol conversion vs TOS results obtained at 250°C on Ni-Al
2
3
O -LDH,
Ni-Al
TOS, the Ni-Al
1% ethanol conversion (curve a), while the Ni-Al
no conversion after 2 h TOS (curve b). While we expect that Ni-Al
2
O
3
-I and Al
2
O
3
–diluted Ni-Al
-LDH catalyst only shows slight deactivation from 100% to approximately
-I catalyst shows rapid deactivation with
-LDH will eventually
2
O
3
-LDH (with 10% Ni content) catalysts. During 100 h of
2
O
3
9
2
O
3
2
O
3
1
3

deactivate, it will require too long time to observe deactivation on this high Ni content catalysts.
Because of the scaling down of the Ni concentration on the diluted catalyst the breakthrough of
the activity decrease on the 10% diluted Ni-Al
2
3
O -LDH catalysts takes 12 h TOS to completely
deactivate. These results indicate that depositing Ni on Al
2
O
3
by impregnation leads to higher
rate of carbon formation and hence a higher rate of deactivation than on the Ni-Al
catalyst.
2
3
O -LDH
The selectivity of the reaction over the stable Ni-Al
2
O
3
-LDH catalyst is shown in Figure
5
. There is an initial change of the effluent concentration due to the transient startup of the
reaction. These changes are attributed to structural changes caused by the reactants that led to the
formation of the special active sites exhibiting resistance to deactivation obtained from the LDH
precursor.
However, after several hours, hydrogen selectivity (curve a) increases throughout the
reaction from 35 to 52% while methane (curve b) and carbon monoxide selectivity (curve c)
decrease from 40 to 25% and 25 to 20%, respectively. This indicates a reaction pathway
occurring according to reaction 4. The change on the hydrogen production with TOS that can be
attributed to decomposition of methane and to structural changes as indicated by the increase in
reduction time.
Over Ni-Al
amounts (~10%) of CO and CH
concentration of H obtained over the Ni-Al
this catalyst. It should be noted that due to the synthesis requirement, Ni-Al
higher Ni content than the 10% on Ni-Al -I. In the later catalyst, it is not possible to load such
2
O
3
-I catalyst, ethanol reacts to yield acetaldehyde, hydrogen, and small
before a rapid decrease of conversion to zero. The higher
-LDH demonstrates the high activity/stability of
-LDH has a
4
2
2
O
3
2
O
3
2
O
3
1
4

a large amount of Ni, hence the advantage of Ni-Al
2
3
O -LDH catalysts. It should also be noted
that unsupported Ni deactivates even faster than the impregnated Ni catalyst [24], so adding
more Ni does not prevent deactivation by carbon unless it is embedded into the alumina matrix.
XRD analysis of the reduced and spent Ni-Al
2
3
O -LDH catalysts after various TOS are
shown in Figure 6. It can be seen that the XRD lines tend to broaden with increasing TOS.
Several NiO diffraction peaks become more visible on the spent catalyst, while others disappear.
The second peak in Figure 6a is the overlap of the NiO peak and the Ni(111) peak, so the former looks
stronger than the latter, but actually is the opposite. This result may be related to partial surface
oxidation of Ni particles upon exposure to air in the XRD diffractometer operating in a standard
ex-situ mode. The surface area of Ni-Al
3
3
O -LDH catalyst increases with TOS (Table 1),
indicating that structural transformations have taken place during reaction.
3
.3. Carbon Formation Pathway
The catalytic results of ethanol conversion (Figure 4) indicate that high activity and stability of
the Ni-Al -LDH catalyst is related to the high concentration (80%) and structure of dispersed
2
O
3
Ni nanoparticles and their interaction with and the location on the alumina matrix. We used
ultrahigh vacuum-XPS (UHV-XPS, requiring transfer to the spectrometer sample holder in air),
carbon analysis, and high resolution TEM to explore carbon formation on the catalyst’s surface
and its relation to decrease in activity.
XPS results. Ni-Al
2
O
3
-LDH and Ni-Al
2
O
3
-I catalysts after reduction and extended TOS were
studied by ex-situ XPS, and even though precautions were taken to minimize air exposure, some
3
/2
oxidation due to air can not be ruled out. All peaks (Ni 2p , O 1s, C 1s, Al 2p) were
deconvoluted and the calculated surface concentrations are reported in Table 2. The large
difference in Ni concentration, which is due to the higher loading in the Ni-Al
2
3
O -LDH (80% Ni)
1
5

compared to 10% in Ni-Al
and B). The intensity of these peaks in Ni-Al
larger than in Ni-Al -I (19 to 1% Ni on the surface as reported in Table 2, respectively).
Metallic Ni peaks are observed at 852.5, while NiO peaks appear at ~856 eV. In the Ni-Al O -
2
O
3
-I, is reflected in the Ni 2p3/2 peak of these catalysts (Figure 7A
2
O
3
-LDH (Figure 7A) is approximately 12 times
2
O
3
2
3
LDH catalyst (Figure 7A), no surface metallic Ni is observed on the reduced catalyst whereas the
surface concentration of metallic Ni increases slightly after reaction from 29 to 34%. XRD
analysis of these catalysts, however, shows that the intensity of some NiO-related peaks
increases (Figure 6). As XPS is near-surface sensitive, it detects surface oxidation of the reduced
catalyst due to its higher reactivity towards oxygen, and hence the disappearance of the reduced
Ni from the XPS spectra. SEM micrographs, show that the spent catalyst is partly covered with
layered structures that protects it from deactivation. This film protects the catalyst from further
reduction and thus, reduced Ni is observed in the XRD patterns. The spent catalyst is further
o
reduced during reaction, which explains the increase in the Ni signal in the XPS spectra with
TOS. Figure 7B shows (note the different scales with Fig. 7A) that the 1% surface concentration
of Ni (Table 2) of Ni-Al
catalyst.
2
3
O -I is mostly metallic Ni, which is related to larger particles of this
The main C peak (284.5 eV) of the Ni-Al
Table 2, increased from 24 to 48% total concentration after 100 h TOS. Figure 8B shows a
similar increase of the C signal on the Ni-Al -I catalyst but at only after 2 h TOS, increasing
2
3
O -LDH catalyst, shown in Figure 8A and
2
O
3
from 21 to 43% C (Table 2). Therefore, the LDH synthesis does not completely protect Ni from
carbon formation but the amount of carbon formed is much lower than on the impregnated Ni
catalysts, and it does not seem to affect the conversion. Obviously this is partly due to the higher
Ni loading of the Ni-Al
2
3
O -LDH catalyst, but loading is not the only factor involved as an
1
6

unsupported Ni catalyst with a similar loading deactivates very rapidly [24]. The secondary
carbon peak (at 288 eV) is due to C-O or C-N bond formation and is attributed to residual
precursors of carbonate left on the catalyst surface during synthesis as it remains constant.
Carbon formation seen on Ni-Al
seem to contribute to deactivating the catalyst as a similar increase (106%) in total elemental
surface carbon is detected on Ni-Al -I but only after 2 h of reaction, at which point this
catalyst is nearly completely deactivated. The diluted 10% Ni-Al -LDH catalyst has a higher
resistance to deactivation than the impregnated catalysts with similar loading as it does not
deactivate until 12h TOS. Therefore, the rate of carbon formation on the Ni-Al -LDH is
2
O
3
-LDH by UHV-XPS in Figure 8 after 100 h TOS does not
2
O
3
2
O
3
2
O
3
slower than on the impregnated catalyst wherein crystallites are deposited on top of the support
completely exposed to the reactants. The surface oxidation detected by XPS, results in the
formation of a surface layer that protects the surface from further oxidation. Exposure of
o
catalysts to air resulted in an increase of temperature less than 50 C rather than sudden ignition
affecting surface carbon. The location of the Ni crystallites on the alumina matrix clearly has an
effect on carbon formation and deactivation.
Elemental analysis. The percentage of carbon measured by elemental analyzer for the catalyst
after various TOS is shown in Table 3. These results clearly indicate that carbon quantity for Ni-
Al
2
O
3
-LDH catalysts increase from 0 to 9.45 wt.% after 100 h TOS, whereas on Ni-Al
2
3
O -I ~11
wt.% carbon builds-up onto the catalyst after only 2 h of reaction.
TEM analysis: The structure and composition of Ni-Al
their changes after TOS was studied by TEM. Figure 3 shows typical STEM images of reduced
Ni-Al -LDH (A) and Ni-Al -I (B) catalysts. Some Ni crystallites are observed located on
the surface while others appears embedded in the alumina matrix and just partially exposed to
2
O
3
-LDH and Ni-Al
2
3
O -I catalysts and
2
O
3
2
O
3
1
7

the atmosphere (Figure 3A). Other crystallites may be encapsulated by the alumina matrix. The
location of the particles may have substantial effect on their behavior during reaction. High
resolution TEM images of Ni-Al
crystallites exhibit a bimodal distribution with some nanoparticles distributed on top of the Al
Figure 9A), whereas others are embedded in Al . The electron diffraction pattern (inset)
shows weak diffraction rings that could be attributed both to Ni and NiO phases. A HRTEM
image of the Ni-Al -LDH catalysts after 100 h TOS is shown in Figure 9B and 9C. A carbon
layer covering Ni nanoparticles (Figure 9B) is observed on the exposed crystallites on top of
Al . The carbon was identified by EDS analysis. Figure 9C shows a carbon nanotube (ordered)
2
O
3
-LDH catalyst after 24 h TOS are shown in Figure 9. Ni
2
O
3
(
2
O
3
2
O
3
2
O
3
structure, such carbon filaments or nanotubes structures on Ni are well documented in the
literature, albeit at higher temperatures and from other more stable reactants such as methane
[
51]. We have examined many samples and observed similar features of carbon overlayers
further supporting the results presented in Figure 9. The growth of carbon filaments (or
nanotubes) around Ni particles at the tip has been known in catalysis for many years so these
observations are in agreement with the previous literature [42].
The TEM results of the used catalysts again show evidence that there are at least two
types of active Ni nanoparticles in the Ni-Al
and crystallites partially embedded into Al
2
O
3
-LDH catalysts: free-standing Ni nanoparticles
and. It is also possible that some Ni nanoparticles
matrix but these are not initially accessible to the
2
O
3
are completely encapsulated inside the Al
reactants.
2
O
3
Based on the characterization results, we propose the following model to explain the
stability of the DHL catalyst. The free standing Ni particles on the surface of the alumina matrix
interact more weakly with it and could be lifted from the surface by growing carbon material, as
1
8

shown in many TEM studies (41). Eventually, these particles will be completely covered by
carbon that would prevent transport of the reactants to the surface and therefore, will no longer
participate in the reaction. On the other hand, the Ni particles that were partially embedded in the
Al
carbon and carbon nanotubes around them and delay coverage of Ni by carbon. This model is
consistent with the characterization results and explains the long activity of the Ni-Al -LDH
2
O
3
matrix are strongly interacting with it, which would decrease the rate of formation of
2
O
3
catalyst. In addition, the Ni particles encapsulated by the matrix may be another source of
surface Ni, refreshing it when these particles are partially exposed to reactants after structural
transformation occurring during reaction. These complex structural transformations occurring on
Ni-Al
2
O
3
-LDH during TOS, are evidenced by the increase in the BET surface area of this
2
catalyst from 110 to 155 m /g (Table 1), and by the increase in hydrogen selectivity during
reaction. The phenomena of spalling, or the break up of the support by carbon formation which is
known to occur on Ni alumina catalysts at high temperatures, may contribute to these structural
tranformations and supply fresh Ni nanoparticles to the surface that sustain the high activity for
such a long period of time compared with the catalyst made by impregnation.
4
. CONCLUSIONS
In summary, we show that Ni-Al
2
3
O catalysts prepared from a LDH precursor is an effective way
to obtain a catalyst with high Ni loading in a similar volume occupied by the impregnated
catalysts. This results in a significant increase in its lifetime during hydrogen production from
ethanol decomposition. The LDH precursor reduces at lower temperature than a 10% Ni
impregnated on a similar alumina support with comparable surface area. Ni-Al
exhibit highly dispersed Ni nanoparticles (3-5 nm) located on the surface, partially embedded or
encapsulated by Al the alumina matrix. Ni nanoparticles impregnated on alumina are larger
2
3
O -LDH catalysts
2
O
3
1
9

(
~7-15 nm) and appear mainly on top of the support. Conversion vs TOS results at 250°C show
that Ni-Al -LDH exhibits only a slight deactivation during 100 h TOS, while catalysts
prepared by impregnation shows rapid deactivation with no conversion after 2 h TOS. The
hydrogen selectivity on the Ni-Al -LDH catalyst increases with TOS from 35 to 52%, while
2
O
3
2
O
3
methane and carbon monoxide selectivity decrease from 40 to 25% and 25 to 20%, respectively.
Such changes are attributed to further methane decomposition and structural transformation of
the catalyst. Catalyst prepared by impregnation exhibits conversion of ethanol to acetaldehyde,
hydrogen, and small amounts (~10%) of CO and CH before a rapid decrease to zero conversion.
4
XPS analysis show that the carbon content increases from 24 to 48% after 100 h TOS on Ni-Al-
LDH whereas a similar increase is seen on the impregnated Ni catalysts but at only 2 h TOS.
High resolution TEM analysis of Ni-Al
2
O
3
-LDH after 100 h TOS show a carbon layer covering
, and Ni nanoparticles on top of nanotubes.
some of the Ni nanoparticles on top of Al
2
O
3
The results suggest that encapsulated Ni crystallites, which initially might not participate
in the reaction, may refresh the surface when exposed by spalling. The high concentration of Ni
in the Ni-Al
2
O
3
-LDH catalyst assists in maintaining the high activity during long TOS as more
catalyst using
Ni surface is available. This model was tested by designing a 20wt.%Ni-SiO
2
combustion synthesis and obtained a 100 % conversion in ethanol which exhibited no
deactivation for 100 hrs [29], similar to the results resported here. Such results (manuscript in
preparation) give us confidence in the validity of the model proposed in this work.
ACKNOLEDGMENTS
This research was funded by the Notre Dame Sustainable Energy Initiative (SEI), a part of the
Center for Sustainable Energy at Notre Dame (cSEND). Also, we like to acknowledge the Notre
Dame Integrated Imaging Facility (NDIFF) for TEM results supporting this research. Use of the
2
0

Advanced Photon Source is supported by the U. S. Department of Energy, Office of Science, and
Office of Basic Energy Sciences, under Contract DE-AC02-06CH11357. MRCAT operations are
supported by the Department of Energy and the MRCAT member institutions.
REFERENCES
[
[
[
[
[
[
[
[
[
[
[
1] L.V. Mattos, G. Jacobs, B.H. Davis, F.B. Noronha, Chem. Rev. 112 (2012) 4094–4123.
2] A.Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy & Fuels 19 (2005) 2098–2106.
3] G.A. Deluga, J.R. Salge, L.D. Schmidt, X.E. Verykios, Science 303 (2004) 993–997.
4] A. Demirbas, Prog. Energy Combust. Sci. 33 (2007) 1–18.
5] M. Ni, D.Y.C. Leung, M.K.H. Leung, Int. J. Hydrogen Energy 32 (2007) 3238–3247.
6] D.B. Levin, R. Chahine, Int. J. Hydrogen Energy 35 (2010) 4962–4969.
7] P. Ramírez de la Piscina, N. Homs, Chem. Soc. Rev. 37 (2008) 2459–2467.
8] P.D. Vaidya, A.E. Rodrigues, Chem. Eng. J. 117 (2006) 39–49.
9] A.M. Karim, Y. Su, M.H. Engelhard, D.L. King, Y. Wang, ACS Catal. 1 (2011) 279–286.
10] D.K. Liguras, D.I. Kondarides, X.E. Verykios, Appl. Catal. B. 43 (2003) 345–354.
11] W. Xu, Z. Liu, A.C. Johnston-Peck, S.D. Senanayake, G. Zhou, D. Stacchiola, E.A.Stach,
J.A. Rodriguez, ACS Catal. 3 ( 2013) 975–984.
[
[
[
12] E. Martono, J.M. Vohs, ACS Catal. 1 (2011) 1414–1420.
13] I.I Soykal, H. Sohn, U.S. Ozkan, ACS Catal. 2 (2012) 2335–2348.
14] J. Zhang, Z. Zhong, X.M. Cao, P. Hu, M.B. Sullivan, L. Chen, ACS Catal. 4 (2014) 448–
4
56.
[
[
15] I.I Soykal, H. Sohn, D. Singh, J. Miller, U.S. Ozkan, ACS Catal. 4 (2014) 585-592.
16] A. Kumar, A.S. Mukasyan, E.E. Wolf, Appl. Catal. A 401 ( 2011) 20–28.
2
1

[
[
[
17] L.V. Mattos, F.B. Noronha, J. Catal. 233 (2005) 453–463.
18] J.R. Salge, G.A. Deluga, L.D. Schmidt, J. Catal. 235 (2005) 69–78.
19] W. Cai, F. Wang, E. Zhan, A.C. Van Veen, C. Mirodatos, W.J. Shen, Catal. 257 (2008)
9
6–107.
[
20] S.M. De Lima, I.O. da Cruz, G. Jacobs, B.H. Davis, L.V. Mattos, F.B. Noronha, J. Catal.
57 (2008) 356–368.
2
[
[
21] G. Rabenstein, V. Hacker, J. Power Sources 185 (2008) 1293–1304.
22] B. Beck, M. Harth, N.G. Hamilton, C. Carrero, J.J. Uhlrich, A. Trunschke, S.
Shaikhutdinov, H. Schubert, H.J. Freund, R. Schlögl, J. Sauer, R. Schomäcker, J. Catal.
2
96 (2012) 120–131.
[
23] J.J. Murcia, M.C. Hidalgo, J.A. Navío, V. Vaiano, P. Ciambelli, D. Sannino, Catal. Today
96 (2012) 101–109.
1
[
[
24] A. Kumar, J.T. Miller, A.S. Mukasyan, E.E. Wolf, Appl. Catal. A 467 ( 2013) 593–603.
25] A. Cross, A. Kumar, E.E. Wolf, A.S. Mukasyan, Ind. Eng. Chem. Res. 51 (2012) 12004–
1
2008.
[
[
[
[
26] H. Ren, D.A. Hansgen, A.L. Stottlemyer, T.G. Kelly, J.G. Chen, ACS Catal. 1 (2011)
90–398.
3
27] D.Z. Mezalira, L.D. Probst, S. Pronier, Y. Batonneau, C. Batiot-Dupeyrat, J. Mol. Catal.
A 340 ( 2011) 15–23.
28] A. Cross, S. Roslyakov, K.V. Manukyan, S. Rouvimov, A.S. Rogachev, D. Kovalev, E.E.
Wolf, A.S. Mukasyan, J. Phys. Chem. C 118 (2014) 26191–26198.
29] A.J. Cross, Catalyst design for hydrogen production from ethanol, PhD thesis, 2014,
University of Notre Dame.
[
[
30] M. Dömök, M. Tóth, J. Raskó, A. Erdohelyi, Appl. Catal. B 69 ( 2007) 262–272.
31] A. Erdohelyi, J. Raskó, T. Kecskés, M. Tóth, M. Dömök, K. Baán, Catal. Today 116
(2006) 367–376.
[
[
[
32] M.A. Goula, S.K. Kontou, P.E. Tsiakaras, Appl. Catal. B 49 ( 2004) 135–144.
33] A. Platon, H.S. Roh, D.L. King, Y. Wang, Top. Catal. 46 (2007) 374–379.
34] J.C. Vargas, S. Libs, A-C. Roger, A. Kiennemann, Catal. Today 107-108 (2005) 417–425.
2
2

[
[
35] G. Garbarino, P. Riani, M.A. Lucchini, F. Canepa, S. Kawale, G. Busca, Int. J. Hydrogen
Energy 38 (2013) 82–91.
36] A.S.P. Lovón, J.J. Lovón-Quintana, G.I. Almerindo, G.P. Valença, M.I.B. Bernardi, V.D
Araújo, T.S. Rodrigues, P.A. Robles-Dutenhefner, H.V. Fajardo, J. Power Sources 216
(2012) 281–289.
[
[
37] W. Cai, P.R. de la Piscina, N. Homs, Appl. Catal. B 145 ( 2014) 56–62.
38] K. Urasaki, K. Tokunaga, Y. Sekine, M. Matsukata, E. Kikuchi, Catal. Commun. 9 (2008)
6
00–604.
[
[
39] F. Mariño, M. Boveri, G. Baronetti, M. Laborde, Int. J. Hydrogen Energy 29 (2004) 67–
1.
7
40] A.J. Vizcaíno, M. Lindo, A. Carrero, J.A. Calles, Int. J. Hydrogen Energy 37 (2012)
985–1992.
1
[
[
41] C. Melchor-Hernández, A. Gómez-Cortés, G. Díaz, Fuel 107 (2013) 828–835.
42] B. Kitiyanan, W.E. Alvarez, J.H. Harwell, D.E. Resasco, Chem. Phys. Lett. 317 (2000)
4
97–503.
[
[
43] H. Wang, Y. Liu, L. Wang, Y.N. Qin, Chem. Eng. J. 145 (2008) 25–31.
44] F. Frusteri, S. Freni, V. Chiodo, S. Donato, G. Bonura, S. Cavallaro, Int. J. Hydrogen
Energy 31 (2006) 2193–2199.
[
[
45] N. Laosiripojana, S. Assabumrungrat, S. Charojrochkul, Appl. Catal. A 327 ( 2007) 180–
1
88.
46] W. Cai, F. Wang, A.C. Van Veen, H. Provendier, C. Mirodatos, W. Shen, Catal. Today
38 ( 2008) 152–156.
1
[
[
[
[
47] Q. Wang, D.O’Hare, Chem. Rev. 112 (2012) 4124−4155
48] S. He, Z. An, M. Wei, D.G. Evans, X. Duan, Chem. Commun., 49 (2013), 5912-5920
49] G. Fan, F. Li, D.G. Evans, X. Duan, Chem. Soc. Rev., 43 (2014) 7040-7066
50] L. He, Y. Huang, A. Wang, X. Wang, X. Chen, J.J. Delgado, T. Zhang, Angew. Chem,
1
24 (2012) 6295 –6298
[
51] P. Chen, H.B. Zhang, G.D. Lin, Q. Hong, K.R. Tsm, Carbon 35 (1997) 1495-1501.
2
3

Figure 1. (A) XRD pattern and (B) FTIR spectrum of Ni-Al
2
3
O -LDH
Figure 1
2
4

Figure 2. Ni % reduction from in-situ XANES vs. reduction temperature for: (a) Ni-Al
and (b) Ni-Al -I.
2
3
O -LDH
2
O
3
Figure 2
2
5

Figure 3. Scanning TEM images of reduced: (A) Ni-Al
2
O
3
-LDH and (B) Ni-Al
2
3
O -I catalysts
2
6

Figure 4. Ethanol conversion on: (a) Ni-Al
2
O
3
-LDH, (b) Ni-Al
2
O
3
-I , and (c) Al
2
3
O -diluted Ni-
Al -LDH catalysts over 100 h TOS
2
O
3
Figure 4
2
7

Figure 5. Selectivity for: (a) hydrogen, (b) methane and (c) carbon monoxide on Ni-Al
catalysts after 100 h TOS.
2
3
O -LDH
2
8

Figure 6. XRD patterns of reduced (a) Ni-Al
h TOS.
2
3
O -LDH, (b) spent catalyst after 24 h, and (c) 100
Figure 6
2
9