Importance of size and distribution of Ni nanoparticles for the hydrodeoxygenation of microalgae oil

Article abstract of DOI:10.1002/chem.201301005

Improved synthetic approaches for preparing small-sized Ni nanoparticles (d=3 nm) supported on HBEA zeolite have been explored and compared with the traditional impregnation method. The formation of surface nickel silicate/aluminate involved in the two pr

Full text of DOI:10.1002/chem.201301005

FULL PAPER
DOI: 10.1002/chem.201301005
Importance of Size and Distribution of Ni Nanoparticles for the
Hydrodeoxygenation of Microalgae Oil
Wenji Song,[a] Chen Zhao,*[a] and Johannes A. Lercher*[a, b]
Abstract: Improved synthetic ap-
port. The lower Brønsted acid concen-
trations of these two Ni/HBEA cata-
lysts compared with the parent zeolite
caused by the partial exchange of
Brønsted acid sites by Ni2+ cations do
not influence the hydrodeoxygenation
proaches for preparing small-sized Ni
nanoparticles (d=3 nm) supported on
HBEA zeolite have been explored and
compared with the traditional impreg-
nation method. The formation of sur-
face nickel silicate/aluminate involved
in the two precipitation processes are
inferred to lead to the stronger interac-
tion between the metal and the sup-
rates, but alter the product selectivity.
Higher initial rates and higher stability
have been achieved with these opti-
mized catalysts for the hydrodeoxyge-
nation of stearic acid and microalgae
oil. Small metal particles facilitate high
initial catalytic activity in the fresh
sample and size uniformity ensures
high catalyst stability.
Keywords: biomass · decarbonyla-
tion · fatty acids · hydrogenation ·
nanoparticles · nickel
Introduction
(rate-determining step), whereas the zeolite Brønsted acid
sites catalyze water elimination.[5] However, the Ni/zeolite
bifunctional catalyst prepared by conventional incipient wet-
ness impregnation method as reported[4] suffers from large
and non-uniform Ni particles, leading to relatively low Ni
dispersion and resultantly low hydrogenation activity.[6] In
addition to impregnation, supported Ni catalysts are usually
prepared by ion-exchange and sol–gel techniques, both of
which have limitations. For instance, the sol–gel method can
generate homogeneously dispersed Ni particles at higher
metal content, but the resulting particle size increases with
the increasing Ni content (typically larger than 10 nm, when
the Ni content exceeds 10 wt%).[7] However, the ion-ex-
change approach can achieve high metal dispersion, but the
attainable Ni loading is rather low and is limited by the ex-
change capacity of the zeolite.
Developing alternative energy sources has become a key
factor to ensure the sustainable growth through minimizing
the carbon footprint. In this context, microalgae have been
envisioned as one of the promising biomass resources be-
cause of their rapid growth rate[1] and high lipid content.[2]
Cultivation of microalgae is also considered to be socially
and environmentally benign, as it does not directly compete
with agricultural products or freshwater supply.[3] In a recent
contribution, Ni nanoparticles supported onto or/and into
HBEA zeolites or ZrO2 have been reported as sulfur-free
catalysts, which are able to realize quantitative transforma-
tion of crude microalgae oil into diesel-range alkanes as
high-grade second-generation biofuels under mild conditions
(2608C, 40 bar H2).[4]
Incorporation of nickel into zeolite framework enables
the preparation of dual-functional catalysts that combine
the hydrogenation–dehydrogenation abilities of a metal and
the carbocation rearrangement functions of solid Brønsted
acids. Nickel catalyzes the first hydrogenation reaction
The deposition–precipitation (DP) method developed by
Geus et al.[8] has been considered as a promising approach
to prepare the Ni/SiO2 catalyst with high metal dispersion at
relatively high loading, and effort was also made to rational-
ize the feasibility of extending the DP method to zeolite
support.[9] DP consists of precipitating a NiII phase by basifi-
cation of a suspension containing nickel salt solution and
[a] W. Song, Dr. C. Zhao, Prof. Dr. J. A. Lercher
Department of Chemistry and
Catalysis Research Center
the support. Urea (COACHTNUTRGNEUNG(NH2)2) is used as the precipitation/
basification agent. It begins to hydrolyze at elevated temper-
ature (908C), so that hydroxide (OHÀ) ions are generated
gradually and homogeneously to avoid local supersaturation
and the sole precipitation of Ni hydroxide in solution.
Attempts are also made to realize a two-step synthesis, in
which monodispersed Ni nanoparticles (NPs) are prepared
separately, followed by grafting of the as-synthesized Ni NPs
onto/into the support by mechanical mixing. Solution-phase
synthesis of Ni NPs normally involves reduction of NiII
metal precursors either in the presence of amine and/or car-
boxylic acid,[10] or by a strong reducing agent such as boro-
Technische Universitꢀt Mꢁnchen
Garching 85747 (Germany)
Fax : (+49)89-28913544
E-mail: chenzhao@mytum.de
johannes.lercher@ch.tum.de
[b] Prof. Dr. J. A. Lercher
Institute for Integrated Catalysis
Pacific Northwest National Laboratory
902 Battelle Boulevard
Richland WA 99352 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301005.
Chem. Eur. J. 2013, 19, 9833 – 9842
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9833

hydride derivatives.[11] The former synthesis involves com-
plex mechanisms including intermediate formation and
acid–base equilibria, which makes tailored synthesis hardly
achievable. In the latter case of chemical reduction, nuclea-
tion is usually too rapid to control.[12] The other challenge
with respect to solution-phase synthesis of Ni NPs lies in the
selection of appropriate protecting ligands (i.e., surfactants)
that stabilize the NPs,[13] because robust protecting agents
binding strongly to Ni0 usually show detrimental effects to
the catalytic activities. The efforts to remove surfactants
usually require high temperatures, which inevitably lead to
the aggregation of NPs. It is recently reported that monodis-
persed Ni NPs, synthesized by reducing NiII acetylacetonate
with a borane tributylamine complex in the presence of
oleylamine and oleic acid acting as solvent and surfactant,
showed high activities on hydrolytic dehydrogenation of the
ammonia borane complex even without the necessity to
remove the surfactants.[14]
Table 1. Physicochemical properties of parent HBEA and Ni/HBEA pre-
pared by different methods.
Method
Ni loading
[wt%]
BET surface
Pore volume
[cm3 gÀ1
N
area [m2 gÀ1
]
N
]
Micro
0.20
Meso
Total
0.37
Parent
HBEA
Impreg.
Ion ex/pre
DP
–
615
0.17
5.2
5.1
5.9
5.8
603
562
508
596
0.19
0.17
0.14
0.19
0.19
0.23
0.24
0.18
0.38
0.40
0.38
0.37
Ni NPs
5.2 wt% Ni content in Ni/HBEA (see Table 1). Ion ex-
change with NiII nitrate hexahydrate solution was performed
on HBEA zeolite in aqueous suspension three times succes-
sively, and the resulting Ni content, determined by atomic
absorption spectroscopy (AAS), was only 0.6 wt%, which
was tested to be insufficient to catalyze hydrodeoxygenation
(HDO) of C18 fatty acid. With the aim to increase the Ni
content to around 5 wt%, an ammonia solution was added
to adjust the pH value of the suspension. Consequently, pre-
cipitated NiII hydroxide phases as well as ion-exchanged
Ni2+ existed simultaneously in the resultant catalyst precur-
sor, annotated herein as “ion exchange/precipitation”. In the
DP synthesis, Ni contents in the final product increased
almost linearly from 5 to 15 wt% with DP durations from 1
to 3 h in a highly reproducible manner (Figure S1, the Sup-
porting Information). The evolution of the pH value during
the first hour of DP is shown in Figure S2 (the Supporting
Information). An increase in the pH (initial: 3.5) was firstly
observed due to the fast generation of OHÀ ions by urea hy-
drolysis at 908C. Then, the pH of mixture reached a maxi-
mum at 5.6 and stabilized at a slightly lower value of 5.5
due to the dynamic equilibrium between the formation of
OHÀ ions and their consumption for precipitation of NiII hy-
drooxoaqua species on the surface of HBEA (Figure S2, the
Supporting Information). The Ni loadings at different DP
duration and the shape of the pH variation curve are similar
to the DP cases with NiII on both silica and HBEA zeo-
lite,[6,9] demonstrating that DP is a controllable and reprodu-
cible method for synthesizing Ni/HBEA catalysts. In the
fourth case, the synthetic strategy with Ni NPs involved
firstly the preparation of soluble Ni NPs by reducing [Ni-
In general, the processes for synthesizing Ni/HBEA cata-
lysts by using four preparation routes including impregna-
tion, ion exchange/precipitation, deposition/precipitation,
and Ni NPs methods are summarized in Scheme 1. We ex-
Scheme 1. The processes for forming supported Ni nanoparticles onto/
into HBEA with four routines.
plore and compare the catalytic properties and performan-
ces of these four Ni/HBEA catalysts, with the goal to obtain
small and homogeneously dispersed Ni particles, in order to
realize high activity and stability with respect to the catalytic
hydrodeoxygenation of stearic acid and crude microalgae
oil.
AHCTUNGTERG(NUNN acac)2] (acac=acetylacetonate) with borane tributylamine
(BTB) in the presence of oleylamine (OAm) and oleic acid
at 908C, followed by grafting the as-synthesized NPs onto
the zeolite under stirring at ambient temperature in hexane.
Physicochemical properties: All Ni/HBEA samples had sim-
ilar Ni contents of approximately 5 wt% (see Table 1). The
nitrogen adsorption–desorption isotherms were nearly iden-
tical for parent HBEA and Ni/HBEA materials prepared by
impregnation and Ni NPs methods (see Table 1), indicating
minimal impact of Ni introduction on the surface property
(Brunauer–Emmett–Teller (BET) surface area: 500–
600 m2 gÀ1) and pore structure (pore volume: 0.37–
0.40 cm3 gÀ1) of the zeolite by these two methods. For sam-
Results and Discussion
Comparison of four synthesis methods for Ni/HBEA cata-
lysts: In the present work, four synthetic methods were se-
lected to prepare Ni/HBEA catalysts. The first traditional
incipient impregnation method incorporated NiII nitrate
hexahydrate aqueous solution with HBEA and reached
9834
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9833 – 9842

Size and Distribution of Ni Nanoparticles
FULL PAPER
ples prepared by ion exchange/precipitation and DP, the re-
markable decrease in micropore volumes by 0.03 and
0.06 cm3 gÀ1 infers that Ni nanoparticles are partially located
in the micropores of HBEA.
The nature of the different preparation methods has led
to the different morphologies of Ni nanoparticles presented
in the reduced Ni/HBEA samples (see Figure 1). Ni/HBEA
Figure 2. TPR profiles of calcined Ni/HBEA precursors prepared by the
first three methods, and the as-synthesized catalyst sample prepared by
using the Ni NP method. *=Layered structure at Ni-support interface
consisting of nickel silicate/nickel aluminate species.
ing that NiO clusters have a weak interaction with the zeo-
lite support after impregnation.[17]
The TPR profiles of catalyst precursors prepared by ion
exchange/precipitation and DP both showed broad peaks
with maxima at 555 and 5398C, respectively. These peaks
can be assigned to the reduction of surface Ni silicate in
agreement with literature,[9,15] or Ni aluminate species, de-
pending on the position of hydroxyl groups of the HBEA
surface that react with the NiII hydrooxoaqua complex
during the synthesis. Burratin et al.[9] confirmed the exis-
tence of Ni silicate in the calcined Ni/SiO2 sample prepared
by DP method, whereas in our case with the altered support
of HBEA, Ni-silicate and Ni-aluminate are both likely to be
Figure 1. TEM images of reduced Ni/HBEA samples prepared by differ-
ent methods; the respective insets are the histograms of Ni particle size
distribution (after counting 300 particles).
À
À
formed due to the coexistence of surface Si OH and Al
OH groups. The presence of either one or two Ni species
appears to strengthen the interaction of Ni with HBEA sup-
port.
prepared by impregnation showed much larger Ni particles
with average diameters of 25 nm compared with sizes of 2–
4 nm obtained from the other three methods. The DP
method resulted in the smallest Ni particles (2.5 nm) with
the most uniform size distribution (standard deviation:
0.7 nm), whereas ion exchange/precipitation led to similar
Ni particle size but in a wider range of size distribution with
1.4 nm deviation. This difference is attributed to a sudden
increase of local pH value upon the addition of ammonia in
the method of ion exchange/precipitation, which caused un-
evenly precipitated NiII species. In contrast, in the case of
DP, hydroxide ions are slowly and homogeneously generat-
ed through hydrolysis of urea at elevated temperature, lead-
ing to more evenly dispersed Ni particles. In the solution-
phase synthetic routine, grafting Ni NPs onto HBEA also
produced uniform Ni particles with a derivation of 1.0 nm at
an average size of 3.9 nm.
To further investigate the underlying differences on
metal-support interaction, temperature-programmed reduc-
tion (TPR) was performed on the calcined catalyst precur-
sors prepared by the three incorporation methods, and the
sample prepared by the Ni NPs method (see Figure 2). For
the impregnated sample, the most prominent reduction peak
occurred at 3628C, which is characteristic for the reduction
temperature of a supported “NiO”-like phase,[15,16] suggest-
À
The catalyst precursor prepared by the Ni NP method
produced two asymmetric peaks with maxima at 222 and
5778C, respectively (Figure 2). Although metallic Ni nano-
particles were readily synthesized, these neo-formed 4 nm
Ni0 clusters were inevitably oxidized upon exposure to the
atmospheric environment.[11b,18] The first peak in TPR pro-
files at 2228C is attributed to the reduction of oxidized Ni
nanoparticles by air in the atmospheric environment. This
“re-oxidized” NiO phase (reduced at 2228C) differs from
the bulk NiO (reduced at 3628C)[19] in nature, leading to the
lower reduction temperature. The other insignificant peak
appeared at 5778C, which is close to the reduction tempera-
ture of Ni silicate/aluminate. Such a peak indicates that an
interaction between metallic Ni nanoparticles and HBEA
exists, probably due to the oxidation of NPs by the oxygen
and/or water in the pores of zeolite. In general, the en-
hanced metal-support interaction is manifested by the sur-
face Ni-silicate/aluminate species presented in the latter
three catalyst precursors. It is speculated that the stronger
interaction is important in preventing sintering of Ni during
reduction.
To investigate the influence of reduction temperature on
the catalyst activity, Ni/HBEA catalysts prepared by the DP
Chem. Eur. J. 2013, 19, 9833 – 9842
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9835

C. Zhao, J. A. Lercher, and W. Song
method were reduced at 460, 560, and 6608C respectively,
and tested in the hydrodeoxygenation of stearic acid (see
Figure S3a, the Supporting Information). The results
showed that catalyst precursors reduced at 4608C had the
highest catalytic activity. Although higher temperatures
would be beneficial for the complete reduction of Ni spe-
cies, it also leads to Ni sintering, which was visible in the
TEM images (see Figure S3b, the Supporting Information).
Thus, 4608C was chosen as the reduction temperature for Ni
samples prepared by the four methods in this study.
tion (solid black lines in Figure 3). Moreover, disappearance
of the bands representing Ni+(CO)2 geminal dicarbonyls
was accompanied by the emergence of a new band at
around 2100 cmÀ1 in the spectra after outgassing. This band
is characteristic for the stable linear Ni+ monocarbon-
yl.[21–23,25,27] Such a species is resistant against evacuation and
more stable than Ni2+ CO. Its higher stability is attributed
À
to the p backbonding between Ni+ and CO as well as to the
synergistic effect between the p and s bonds.[25] The carbon-
yls formed with ionic Ni species were also observed on Ni/
HBEA prepared by impregnation, since a small proportion
of Ni silicate is also possibly formed.[31]
The other two pronounced bands in the IR spectra after
evacuation appeared at 2040–2045 cmÀ1 (linearly bonded
CO adsorption on metallic Ni0) and 1910–1970 cmÀ1 (two-
or three-fold bridged CO adsorption on Ni0).[19–21,24,27,32] Ni/
HBEA prepared by ion exchange/precipitation, DP, and Ni
NP methods showed much higher integral intensities under
these two bands representing linearly and bridged-bonded
CO, indicating higher concentrations of accessible metallic
Ni species present in these samples. This observation also in-
dicates that much smaller Ni nanoparticles with higher
metal dispersions are achieved by the optimized synthetic
methods (see Table 1).
The acidities of four Ni/HBEA samples were probed by
IR spectroscopy of adsorbed pyridine (pyridine-IR) and
temperature-programmed-desorption of NH3 (TPD-NH3).
Compared with parent HBEA (119 mmolgÀ1), Ni/HBEA
catalysts prepared by DP and ion-exchange methods pos-
sessed lower concentrations of Brønsted acid sites of
36 mmolgÀ1 (Figure 4A), possibly due to the heteroconden-
Figure 3. IR spectra of adsorbed CO at 408C. (b) recorded at 0.5 mbar
CO partial pressure, (c) recorded after outgassing for 15 min. All
spectra have been normalized by the weights of the catalyst pellets.
The state of Ni species as well as the dispersion of sup-
ported metallic Ni NPs was also explored by IR spectrosco-
py using CO as a probe molecule (Figure 3). In the spectra
with adsorbed CO under equilibrium pressure, bands above
2100 cmÀ1 were detected, which are typical for carbonyls
formed with Ni cations.[20–23] The reduction temperature at
4608C was lower than the maxima of reduction in the TPR
profiles for the latter three samples (Figure 2). This led to
the presence of cationic Ni. The bands at 2210–2212 cmÀ1
[24–27]
are assigned to Ni2+ CO,
whereas the two bands at
À
2136–2137 and 2092–2094 cmÀ1 are ascribed to the symmet-
ric and asymmetric CO stretching of geminal Ni+(CO)2 di-
carbonyls.[21,22,28,29] The Ni+ ion is speculated to originate
from the reduction of Ni2+ in the reducing environment of
CO.[30] The fact that two CO molecules are attached to one
Ni+ cation compared with Ni2+ CO can be rationalized by
À
the higher ionic radius of Ni+.[25] Both Ni2+ CO and Ni
+
À
Figure 4. A) IR spectra of adsorbed pyridine. B) Profiles of temperature-
programmed-desorption of NH3.
(CO)2 were unstable and readily decomposed upon evacua-
9836
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9833 – 9842

Size and Distribution of Ni Nanoparticles
FULL PAPER
sation between acidic hydroxyl groups and NiII hydrooxo-
ACHTUNGTRENNUNGaqua complexes during these two syntheses. In addition, in-
HBEA catalysts prepared by ion-exchange/precipitation and
Ni NP approaches (see Table 2). When normalized to the
amount of the surface Ni atoms, turnover frequencies
creased concentrations of Lewis acid sites were also ob-
served with Ni/HBEA compared with the parent HBEA
(see Table S2, the Supporting Information). These Ni-based
samples had unreduced ionic Ni species as evidenced by the
IR spectra of adsorbed CO (see Figure 3) and TPR-H2 (see
Figure 2), which are speculated to be potential Lewis acid
sites,[34] causing the increase in the Lewis acid concentration.
Evolution of total acidity for the four catalysts measured by
(TOFs) for the hydrogenation of stearic acid were 125, 243,
À1
178, and 140 mol mol
hÀ1 over the four Ni/HBEA
Surf Ni atom
samples, indicating that hydrogenation of stearic acid is in-
sensitive to the structure of Ni nanoparticles. The yields of
cracking products were below 5% (among products in the
liquid phase) after 2 h. Analysis of the product distribution
in the liquid phase (Table 2 and Figure 6) showed that a rel-
atively higher selectivity towards heptadecane (C17) and
lower selectivity to isooctadecane (C18) was observed for Ni/
HBEA catalysts synthesized by the later three methods as
compared with those catalyzed by the impregnated sample.
We have explored the mechanism of stearic acid hydro-
deoxygenation with Ni/HBEA in the previous work.[4a] It
showed that stearic acid is firstly hydrogenated to octadeca-
nal on Ni sites and then converted to 1-octadecanol through
a dehydrogenation–hydrogenation equilibrium (favored at
high hydrogen pressure). Parallel routes of decarbonylation
with Ni sites and dehydration on acidic sites can take place
with 1-octadecanol, producing C17 and C18 hydrocarbons, re-
spectively. In the current reaction network of C18 stearic
acid conversion, hydrodeoxygenation (HDO) proceeds pref-
erentially through a hydrogenation–dehydration–hydrogena-
tion pathway, with C18 octadecane yields exceeding 60%.
However, the C17 and C18 hydrocarbon product distributions
were quite different over the four Ni/HBEA catalysts.
both NH3-TPD and pyridine-IR techACHTUNTRGNEUNGniques showed a similar
trend (see Table S2, the Supporting Information).
Activity and stability for stearic acid hydrodeoxygenation:
The catalytic activities were tested for hydrodeoxygenation
of stearic acid at 2608C at 40 bar H2. The catalyst prepared
by DP with the smallest Ni particles had the highest hydro-
deoxygenation activity with
an initial rate of
18 mmolgÀ1 hÀ1, whereas the traditional impregnation
method resulted in the slowest initial rate of
3.5 mmolgÀ1 hÀ1 (Figure 5 and Table 2). The initial rates of
17 and 14 mmolgÀ1 hÀ1 were obtained over the other two Ni/
With Ni samples prepared by impregnation, the relatively
large Ni particles led to a slower metal-catalyzed decarbony-
lation step, and thus, the relatively more rapid dehydration–
hydrogenation and isomerization steps resulted in higher se-
lectivity to n- and iso-C18 hydrocarbons, maintaining a C18/
C17 ratio of 2.2 (see Table 2). For the other three Ni/HBEA
catalysts, the small and evenly dispersed Ni nanoparticles
enhanced the rate of decarbonylation, and the decreased
Brønsted acidity suppressed the dehydration and isomeriza-
tion steps, leading to a lower C18/C17 ratio of 1.9 among the
alkane products (see Table 2). The C18 hydrocarbon with its
selectivity higher than 60% was still the major product,
which suggests that with the improved catalysts prepared by
DP and ion exchange/precipitation, dehydration is still much
faster than decarbonylation on the C18-alcohol intermediate.
Therefore, the reduced Brønsted acidity caused by these
two precipitation methods would have a negligible impact
on the overall hydrodeoxygenation activity. In addition, dis-
tearyl ether was also observed as a reaction intermediate,
which was produced through an intermolecular dehydration
Figure 5. Hydrodeoxygenation (HDO) of stearic acid as a function of
time over Ni/HBEA prepared by four different methods. Reaction condi-
tions: stearic acid (1.0 g), catalyst (0.2 g), dodecane (100 mL), 40 bar H2
(at reaction temperature 2608C), at a stirring speed of 700 rpm.
Table 2. Hydrodeoxygenation of stearic acid over Ni/HBEA prepared by
four methods.[a]
Method
Conv. Initial TOF[c]
[%]
Yield of liquid products [C%]
rate[b]
n-
C18
iso-
C18
n-
C17
C18-
OH
Crack-
ing
À
of the in situ-formed 1-octadecanol. The C O aliphatic
Impreg.
Ion ex/
pre
DP
Ni NP
33
97
3.5
17
125
243
16
61
3.4
2.5
9.2
33
0.21
0.01
4.1
3.5
ether bond was in turn cleaved by Ni sites through hydroge-
nolysis as the reaction proceeded, followed by dehydration
of the C18 alcohol and subsequent hydrogenation of the re-
spective alkene to produce the C18 hydrocarbon.
100
100
18
14
178
140
61
61
2.1
2.3
34
32
–
0.1
3.3
3.8
[a] The reaction conditions are identical to that listed in Figure 5, reac-
Ni leaching was explored by heating hexane to reflux with
Ni catalysts in a Soxhlet extractor. The Ni content extracted
in the solvent after seven days was below the detection limit
of atomic absorption spectroscopy (1 ppm), indicating that
tion time: 2 h. [b] Unit: mmolgÀ1 hÀ1. [c] Unit: mol mol
hÀ1, Ni
À1
Surf Ni atom
dispersions are 3.3, 8.2, 9.9, 9.8% for Ni/HBEA catalysts synthesized by
impregnation, ion exchange/precipitation, DP, and Ni NP methods, re-
spectively, as determined by H2 chemisorption measurement.
Chem. Eur. J. 2013, 19, 9833 – 9842
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9837

C. Zhao, J. A. Lercher, and W. Song
Figure 6. Product distributions on stearic acid conversion over Ni/HBEA prepared by four different methods. Reaction conditions: stearic acid (1.0 g),
catalyst (0.2 g), dodecane (100 mL), 40 bar H2 (at reaction temperature 2608C), at a stirring speed of 700 rpm.
these four supported Ni catalysts are very stable against
leaching in nonpolar organic solvents.
showing comparable rates to the fresh catalysts. The higher
activities and stabilities of these two catalysts are attributed
to the smaller and more evenly distributed Ni particles
(TEM images in Figures 1 and 7B). In recycled samples of
DP and Ni NP, not only were the average diameters of Ni
particles much smaller (around 5.0 nm), more importantly,
these particles were more evenly distributed with deviations
of 3–4 nm. Notably, the NP method led to the most stable
Ni/HBEA catalyst against sintering, with only 1.1 nm
growth in average particle size as well as fewest aggregated
large Ni particles (>7 nm) after recycling. Impregnated cat-
alysts with much larger Ni particles (see Figure 1) were less
prone to further aggregation during the reaction compared
with nanometer-range Ni in the other three samples. But
still, it had the lowest activities throughout the recycling
tests. In contrast, the ion-exchange/precipitation method re-
sulted in a small Ni size, which facilitated high initial reac-
tion rates in its fresh sample. However, the large differences
in particle size reduced the stability of nanoparticles, and
the catalyst showed a dramatic particle size increase togeth-
er with activity decrease during recycling (Figure 7). There-
fore, we conclude that the average size and size uniformity
of Ni particles are both crucial for maintaining a highly
active and stable catalyst system, that is, concentration of ac-
cessible Ni facilitates high initial activity of the fresh sample,
whereas good uniformity ensures high catalyst stability over
prolonged usage.
In a following step, the stabilities of the four Ni/HBEA
catalysts were tested by evaluating their catalytic activities
during the recycling runs on stearic acid conversion at
2608C in the presence of 40 bar H2. After each reaction, the
catalysts were washed with acetone followed by drying in air
at 708C, and afterwards reused for the next catalytic run.
Significant catalyst deactivation was observed for Ni/HBEA
prepared by ion exchange/precipitation (see Figure 7A),
which lost 60% of its initial activity after four reaction runs.
As Ni leaching in organic solvent has been excluded, the ac-
tivity loss is attributed to the aggregation of Ni nanoparti-
cles during the successive reactions. This was confirmed by
TEM measurements on the recycled catalyst samples (see
Figure 7B). For Ni/HBEA synthesized by ion exchange/pre-
cipitation, the Ni particle size increased from 2.6 to 8.5 nm
after four runs of reaction, accompanied with a drastic in-
crease in size discrepancy (the standard deviation increased
from 1.4 to 7.5 nm). Moreover, a large portion (about 18%)
of the strongly aggregated Ni particles (17–30 nm) was ob-
served, which were absent in its fresh sample (see Figure 1).
Therefore, it could be concluded that the most severe aggre-
gation of Ni particles in sample prepared by ion exchange/
precipitation was responsible for the highest degree of cata-
lyst deactivation during the successive reactions.
On the other hand, Ni/HBEA prepared by DP and Ni NP
methods still retained high activities even after four runs,
9838
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9833 – 9842

Size and Distribution of Ni Nanoparticles
FULL PAPER
Figure 8. Mechanisms for the metal nanoparticle growth through A) coa-
lescence, and B) Ostwald ripening.
Ostwald ripening and faster catalyst deactivation as shown
in TEM images of Figure 7B.
Hydrodeoxygenation of crude microalgae oil: Crude micro-
algae oil was hydrodeoxygenated with the four Ni/HBEA
catalysts in the batch reactor at 2608C in the presence of
40 bar H2 (see Figure 9). Microalgae oil mainly consists of
Figure 7. A) Stability tests for the Ni/HBEA catalysts prepared by four
different methods. Reaction conditions: stearic acid (2.5 g), catalyst
(0.4 g), dodecane (100 mL), 40 bar H2 (at reaction temperature 2608C),
at stirring speed of 700 rpm. B) TEM images and the size distribution his-
tograms of Ni/HBEA samples after four runs.
Figure 9. Product distributions on conversion of crude microalgae oil. Re-
action conditions: microalgae oil (1.0 g), Ni/HBEA catalyst (0.2 g), do-
decane (100 mL), 40 bar H2 (at 2608C), at a stirring speed of 700 rpm.
neutral lipids such as mono-, di-, and triglyceride. The mi-
croalgae oil (provided by Verfahrenstechnik Schwedt
GmbH) used here is composed of unsaturated C18 fatty
acids (88.4 wt%), saturated C18 fatty acids (4.4 wt%), as
well as some other C14, C16, C20, C22, and C24 fatty acids
(7.1 wt% in total). The mechanism for hydrodeoxygenation
of microalgae oil[4a] has been proposed to proceed through
an initial metal-catalyzed hydrogenation of the double
bonds in the alkyl chain, followed by hydrogenolysis of the
formed saturated triglyceride leading to fatty acid and pro-
pane production. The subsequent hydrogenation of the car-
boxylic group of fatty acid leading to the corresponding al-
dehyde, for example, octadecanal, is the rate-determining
Generally speaking, two mechanisms are applied to ex-
plain the growth of the metal nanoparticle (shown in
Figure 8), namely coalescence and Ostwald ripening.[35] The
latter suggests a dynamic process of atom exchange between
nanoclusters, and is usually considered for the discussion of
catalyst growth and sintering. Ostwald ripening demon-
strates the behavior of catalyst deactivation reported in the
current work with four Ni/HBEA samples, in which larger
Ni clusters grow further at the expense of the smaller ones
which eventually shrink and disappear. Therefore, the wider
size-distribution, for instance with Ni nanoparticles prepared
by ion exchange/precipitation, will lead to more pronounced
Chem. Eur. J. 2013, 19, 9833 – 9842
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9839

C. Zhao, J. A. Lercher, and W. Song
step. Followed by either decarbonylation of octadecanal to
produce C17 n-heptadecane and carbon monoxide (minor
route), or hydrogenation of octadecanal to 1-octadecanol
(major route). Acid-catalyzed alcohol dehydration and
metal-catalyzed alkene hydrogenation produce the final n-
octadecane.
eral individual steps in the overall hydrodeoxygenation reac-
tion including hydrogenation of double bonds, hydrogenoly-
sis of triglyceride, decarbonylation of aldehyde, as well as
the hydrogenation of fatty acids catalyzed by Ni sites.
Among these steps, the remarkable enhancement on fatty
acid reduction (the rate-determining step) with the opti-
mized Ni particles is clearly visible. TOFs (normalized
by accessible surface Ni atoms) for hydrodeoxygenation
of crude microalgae oil attained 92, 244, 232,
185 moloil molÀSu1rf Ni atom hÀ1 over Ni/HBEA catalysts prepared
by impregnation, ion-exchange/precipitation, DP, and Ni
NPs methods (see Table 3), respectively. It has been demon-
strated that the hydrogenation of stearic acid to stearic alde-
hyde is not sensitive to the Ni particle size (Table 2). How-
ever, the TOF for the overall alkane formation from micro-
algae oil with the impregnated Ni catalyst (92 hÀ1) was 2–3
times lower than those with the other three Ni catalysts
(185–244 hÀ1). This is related to the fact that alkane produc-
tion from microalgae oil passes through a series of individu-
al reactions including hydrogenation, hydrogenolysis, decar-
bonylation, as well as dehydration in a cascade, among
The yields of major intermediates of stearic acid and stea-
ric acid propyl ester firstly increased to the maxima of 50–60
and 10 wt%, respectively, and subsequently decreased as a
function of time. The stearic acid propyl ester was formed
from the partial hydrogenolysis of glycerides at the begin-
ning of the reaction, which was further hydrogenolyzed to
stearic acid and propane at Ni sites as the reaction proceed-
ed. Stearic acid was the dominant intermediate (selectivity
>70 wt%), which verifies that fatty acid hydrogenation is
the rate-determining step. Only small concentrations of oc-
tadecanol were detected during the reaction (<2 wt% for
the impregnated sample, and <1 wt% for the other three),
demonstrating that the dehydration rate is much faster than
the rates of triglyceride hydrogenolysis and fatty acid hydro-
genation.
À
After a reaction time of 2 h, the impregnated Ni/HBEA
sample only achieved 26 wt% yield of total liquid hydrocar-
bons, including 13.6 wt% C18 and 4.6 wt% C17 hydrocarbons,
as well as other hydrocarbons of C14–C24, leaving 33 wt%
stearic acid still unconverted. Whereas by comparison, the
other three catalysts realized a high total liquid alkanes
yield of 79 wt% within 2 h, which is very close to the theo-
retical maximum hydrocarbon yield (84 wt%). The initial
rates of total liquid alkane formation reached 4.0, 4.6,
3.6 goil gcatÀ1 hÀ1 over Ni/HBEA catalysts prepared by ion ex-
change/precipitation, DP, and Ni NP methods, respectively
(see Table 3). Such a rate sequence is in line with the order
which the initial primary step of C O bond hydrogenolysis
of ester has been found to be structure sensitive.[36] There-
fore, the larger impregnated Ni particles (25 nm) would be
expected to result in a much lower TOF for the overall cas-
cade hydrodeoxygenation reaction compared with the other
three catalysts (3 nm particle diameter).
Conclusion
Ni nanoparticles supported onto and/or into HBEA zeolite
prepared by four techniques, namely impregnation, ion ex-
change/precipitation, DP, and Ni NP methods were investi-
gated for the hydrodeoxygenation of stearic acid and crude
microalgae oil. In the catalytic tests on stearic acid conver-
sion, Ni/HBEA catalysts prepared by DP and Ni NP meth-
ods maintained very high activities even after four runs of
reaction due to the presence of more accessible, uniformly
dispersed, and sintering-resistant Ni nanoparticles, con-
firmed by TEM, IR of adsorbed CO, and TPR-H2 measure-
ments. In a comparison, catalysts prepared by using ion ex-
change/precipitation methods achieved high initial reaction
rate in its fresh sample, but showed a dramatic particle-size
increase accompanied with a loss in activity during the recy-
cling test. The impregnated Ni/HBEA sample steadily
showed the lowest activity in the first and recycling runs.
This indicates that the small-sized metal particles control
the initial catalytic activity, and the good size uniformity en-
sures high catalyst stability for the longer utilization. It also
highlights the importance of the narrow size-distribution of
the particles, as it can minimize the growth of larger parti-
cles at the expense of smaller ones through the Ostwald rip-
ening mechanism.
Table 3. Data of hydrodeoxygenation of microalgae oil over Ni/HBEA
catalysts prepared by four methods.[a]
Method
Initial TOF[c]
rate[b]
Yield of liquid products [wt%]
n-C18 iso-C18 n-C17 Other Total
HC[d]
Impreg.
Ion ex/pre 4.0
DP
Ni NP
0.6
92
9.7
42
43
40
3.9
3.0
2.7
2.4
4.6
26
25
27
5.1
8.0
7.5
8.2
26
79
79
79
244
232
185
4.6
3.6
[a] The reaction conditions are identical to that listed in Figure 8, reac-
tion time: 2 h. [b] Initial rates for total liquid alkane formation base on
the weight of catalyst, unit: goil gcatÀ1 hÀ1. [c] Initial rates for total liquid
alkane formation base on the surface Ni atoms of catalysts, assuming the
average molecular weight of oil is 240 gmolÀ1, unit: moloil molÀSu1rf Ni atom hÀ1
[d] HC: hydrocarbons.
.
of Ni particle size shown from the TEM images (2.6, 2.5,
and 3.9 nm). The impregnated Ni/HBEA sample only at-
tained an alkane formation rate of 0.6 goil gcatÀ1 hÀ1, which is
6–8 times lower than those observed for the other three cat-
alysts. The higher rates are attributed to the much smaller
and highly-dispersed Ni nanoparticles, which promotes sev-
9840
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9833 – 9842

Size and Distribution of Ni Nanoparticles
FULL PAPER
the as-synthesized Ni NPs, the parent HBEA zeolite (2.5 g) was carefully
added into the obtained NP colloid, and the mixture was stirred over-
night. Before applying the catalyst to reaction, n-hexane was vacuum fil-
tered and Ni/HBEA was dried at 708C and reduced by H2 at 4608C for
Experimental Section
Chemicals: All chemicals were obtained from commercial suppliers and
used as received. Stearic acid (Fluka, analytical standard), dodecane
(Sigma–Aldrich, ꢀ99%, Reagent Plus), NiII nitrate hexahydrate (Sigma–
Aldrich, ꢀ 98.5%), NiII acetylacetonate (Aldrich, 95%), ammonium hy-
droxide solution (Sigma–Aldrich, 28.0–30.0% NH3 basis, ACS reagent),
urea (Sigma–Aldrich, BioReagent), oleylamine (Aldrich, 70%, technical
grade), oleic acid (Fluka, Ph Eur), borane triethylamine complex (Al-
drich, 97%), n-hexane (Sigma–Aldrich, ꢀ97%), ethanol (Roth,
ꢀ99.5%), HBEA (Sꢁd-Chemie AG Mꢁnchen, Si/Al=75).
5 h with
150 mLminÀ1
a a hydrogen flow of
heating rate of 18CminÀ1 and
.
Catalyst characterization: Elemental analysis was measured by atomic
absorption spectroscopy (AAS) on a UNICAM 939 AA-Spectrometer.
The BET specific surface area and pore volume were determined by ni-
trogen adsorption-desorption isotherms measured at À1968C by using a
PMI automatic BET-Sorptometer. The sample was activated in vacuum
at 2008C for 2 h before measurement.
Catalyst preparation: Preparation of Ni/HBEA by incipient wetness im-
Transmission electron microscopy (TEM) images were taken by a JEM-
2010 Jeol transmission microscope operating at 120 kV. The sample was
firstly made suspension in ethanol by ultrasonication, and a drop of such
suspension was deposited onto a carbon-coated Cu grid. More than 300
particles were counted for the histograms of particle size distribution.
pregnation: In a typical synthesis, NiACTHNUTRGNE(UNG NO3)2·6H2O (2.91 g) was dissolved
in doubly distilled water (10 g), and then the solution was slowly dropped
onto HBEA zeolite (10 g) with continuous stirring at ambient tempera-
ture for a total of 4 h. After completing this procedure, the material was
firstly dried overnight at ambient temperature and then further at 1108C
for 12 h. Afterwards, the catalyst precursor was calcined in air (flow rate:
100 mLminÀ1) at 4008C for 4 h and reduced with pure hydrogen (flow
rate: 150 mLminÀ1) at 4608C for 5 h with a temperature increment of
The temperature-programmed reduction (TPR) measurement was per-
formed on a self-constructed instrument, by using a stream of 3% H2/He
mixture and a heating rate of 58CminÀ1. The H2O formation during the
TPR experiments was monitored by mass spectrometry.
18CminÀ1
.
Preparation of Ni/HBEA by ion exchange: For a standard ion-exchange
procedure, the parent HBEA zeolite (5 g) was suspended in an aqueous
solution of NiACHTUNGTRENNUNG(NO3)2 (0.14m, 250 mL). The suspension was stirred for 6 h
Temperature-programmed desorption (TPD) of NH3 was performed in a
6-fold parallel reactor system. The catalysts were activated in He at
5008C with a heating rate of 58CminÀ1 for 1 h. NH3 was adsorbed with a
partial pressure of 1 mbar at 1008C. Afterwards, the samples were
purged with He (30 mLminÀ1) for 2 h to remove physisorbed molecules.
For the temperature-programmed desorption measurement, the samples
were heated in flowing He from 100 to 7608C with an increment of
108CminÀ1 to desorb ammonia. The desorbed species were monitored by
mass spectrometry (Balzers QME 200). For the quantification of acid
sites presented in the samples, a reference (HZSM-5, Si/Al=45 from
Sꢁd-Chemie AG) with known acidity was used to calibrate the signal.
at 708C. Later, the solid was separated by filtration, and the resulting ma-
terials were washed three times with doubly distilled water and oven-
dried at 1108C overnight. Subsequently, the catalysts precursors were cal-
cined at 4008C for 4 h with a heating rate of 18CminÀ1. The exchange
procedures were repeated three times. Finally, the calcined catalyst was
reduced at 4608C for 5 h with a heating rate of 18CminÀ1 and a hydrogen
flow of 150 mLminÀ1
.
Preparation of Ni/HBEA by ion exchange/precipitation with ammonia sol-
ution: Firstly, HBEA zeolite (5 g) was suspended in aqueous solution of
NiACHTUNGTRENNUNG(NO3)2 (0.14m, 250 mL), and a solution of ammonia was added drop-
The Infrared (IR) spectra of adsorbed pyridine were recorded with a
Perkin–Elmer 2000 spectrometer at a resolution of 4 cmÀ1. The catalyst
sample was prepared as self-supporting wafer and activated in vacuum
(P=10À7 mbar) at 4508C for 1 h (heating rate=108CminÀ1). After
cooing to 1508C, the sample was equilibrated with 0.1 mbar of pyridine
for 30 min followed by evacuating the system for 1 h, a spectrum was re-
corded thereafter. For quantification, molar integral extinction coeffi-
cients of 0.73and 0.96 cmmmolÀ1 were used for Brønsted and Lewis acid
sites, respectively. These were determined from a standard (HZSM-5, Si/
Al=45 from Sꢁd-Chemie AG).
wise until the pH value of the suspension reached 7.5, and the suspension
was then kept under constant stirring at room temperature for 90 min.
Afterwards, the solid was filtered out and washed three times with
doubly distilled water. The samples were dried at 1108C overnight and
calcined at 4008C for 4 h with a heating rate of 18CminÀ1. The calcined
catalyst was reduced at 4608C for 5 h with a heating rate of 18CminÀ1
and a hydrogen flow of 150 mLminÀ1
.
Preparation of Ni/HBEA by deposition–precipitation (DP) method: The
typical synthesis adopted procedures described in literature.[6] An aque-
The IR spectra of adsorbed CO were recorded on a Bruker VERTEX 70
spectrometer at a resolution of 4 cmÀ1 with 128 scans. The catalyst wafer
was firstly in-situ-reduced in hydrogen (1 bar) at 4508C for 1 h (heating
rate=108CminÀ1), and then allowed outgassing for 1 h (P=10À6 mbar).
CO was adsorbed at 408C, and a spectrum was taken at 0.5 mbar of CO
equilibrium pressure. Afterwards, the system was evacuated to remove
physically adsorbed CO, and the IR spectra were recorded every five mi-
nutes until no changes observed. To directly compare the surface cover-
age of adsorbed CO, all spectra were normalized with the weight of the
respective wafers.
ous solution of Ni
which was used to make a suspension with 2 g of the parent HBEA zeo-
lite. Urea was dissolved in the rest of Ni(NO3)2 solution (40 mL), and
ACHTUNGTRENNUNG(NO3)2 (0.14m, 250 mL) was prepared, and 210 mL of
AHCTUNGTRENNUNG
added drop-wise into the zeolite suspension when the latter was heated
to 708C. The mixture was then brought up to 908C at which DP started.
After 1 h, the suspension was cooled to room temperature, vacuum fil-
tered, and the solid was washed three times with doubly distilled water.
Subsequently, the samples were dried at 1108C overnight. The catalyst
precursors were calcined at 4008C for 4 h with
18CminÀ1, and then reduced at 4608C for 5 h with a heating rate of
18CminÀ1 and a hydrogen flow of 150 mLminÀ1
a heating rate of
Catalytic measurement: Catalytic reactions on stearic acid and microalgae
oil: A typical experiment was carried out as follows: Stearic acid (or mi-
croalgae oil) (1.0 g), dodecane (100 mL), and Ni/HBEA catalyst (0.2 g)
were charged into a batch autoclave (Parr Instrument, 300 mL). The re-
actor was firstly flushed with N2 at ambient temperature, and then heated
up to 2608C when 40 bar of H2 was purged, and the reaction started at a
stirring speed of 700 rpm. The liquid products were sampled in situ
during the catalytic runs, and analyzed by a Shimadzu GC-MS equipped
with a FID detector and a HP-5 capillary column. For the conversion of
microalgae oil, docosane was used as an internal standard for quantifica-
tion of liquid products. Conversion=(weight of converted reactant/
weight of the starting reactant)ꢃ100%. Yield of liquid products (C%)=
(C atoms in liquid products/C atoms in the starting reactant)ꢃ100%. Se-
lectivity (C%)=(C atoms of each product/C atoms in all the liquid prod-
ucts)ꢃ100%.
.
Preparation of Ni/HBEA by grafting as-synthesized Ni nanoparticles
(NPs): The Ni NPs were synthesized by using the following procedure
developed by Metin et al.[14] Firstly, [Ni
ACTHNUTRGENUN(G acac)2] (1.03 g, 4 mmol) was
mixed with oleylamine (OAm; 60 mL) and oleic acid (OA; 1.27 mL,
4 mmol). The mixture was heated up to 1108C under the protection of ni-
trogen and maintained for 1 h to remove air and moisture. Subsequently,
it was cooled down to 908C and then borane triethylamine (BTE,
1.2 mL, 8 mmol) was quickly injected. The formation of Ni0 NPs was ob-
served through a visible color change from green to brownish-black. The
resultant solution was kept at 908C for another 1 h before cooling down
to room temperature. Afterwards, ethanol (120 mL) was added into the
suspension, and the Ni NPs were separated by centrifugation and then
re-dispersed in n-hexane (80 mL) by means of ultrasonication. To graft
Chem. Eur. J. 2013, 19, 9833 – 9842
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9841

C. Zhao, J. A. Lercher, and W. Song
Dillen, J. W. Geus, J. Catal. 1997, 170, 217–226; c) M. Che, Z.
Cheng, C. Louis, J. Am. Chem. Soc. 1995, 117, 2008–2018.
[16] B. Mile, D. Stirling, M. A. Zammitt, A. Lovell, M. Webb, J. Catal.
1988, 114, 217–229.
Acknowledgements
W.S. gratefully acknowledges support from the graduate school (Faculty
Graduate Center of Chemistry) of the Technische Universitꢀt Mꢁnchen
and the Elitenetzwerk Bayern (graduate school NanoCat). We also ap-
preciate the financial support from the EADS Algenflugkraft project.
ˇ
[17] E. Marceau, M. Che, J. Cejka, A. Zukal, ChemCatChem 2010, 2,
413–422.
[18] a) A.-H. Lu, E. L. Salabas, F. Schꢁth, Angew. Chem. 2007, 119,
1242–1266; Angew. Chem. Int. Ed. 2007, 46, 1222–1244; b) D.
Chen, J. Li, C. Shi, X. Du, N. Zhao, J. Sheng, S. Liu, Chem. Mater.
2007, 19, 3399–3405.
[1] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–
4098.
[19] K. Hadjiivanov, M. Mihaylov, D. Klissurski, P. Stefanov, N. Abadjie-
va, E. Vassileva, L. Mintchev, J. Catal. 1999, 185, 314–323.
[20] a) A. G. Sault, C. H. F. Peden, E. P. Boespflug, J. Phys. Chem. 1994,
98, 1652–1662.
[21] L. Bonneviot, F. Cai, M. Che, M. Kermarec, O. Legendre, C. Lepe-
tit, D. Olivier, J. Phys. Chem. 1987, 91, 5912–5921.
[22] P. H. Kasai, R. J. Bishop Jr., D. McLeod Jr., J. Phys. Chem. 1978, 82,
279–285.
[23] K. G. Ione, V. N. Romannikov, A. A. Davydov, L. B. Orlova, J.
Catal. 1979, 57, 126–135.
[24] J. Leglise, A. Janin, J. C. Lavalley, D. Cornet, J. Catal. 1988, 114,
388–397.
[25] K. Hadjiivanov, H. Knçzinger, M. Mihaylov, J. Phys. Chem. B 2002,
106, 2618–2624.
[26] a) T. C. Sheng, B. Rebenstorf, Langmuir 1991, 7, 2659–2668;
b) C. L. Angell, P. C. Schaffer, J. Phys. Chem. 1966, 70, 1413–1418.
[27] J. B. Peri, J. Catal. 1984, 86, 84–94.
[2] a) P. M. Schenk, S. R. Thomas-Hall, E. Stephens, U. C. Marx, J. H.
Mussgnug, C. Posten, O. Kruse, B. Hankamer, Bioenerg. Res. 2008,
1, 20–43; b) T. M. Mata, A. A. Martins, N. S. Caetano, Renewable
Sustainable Energy Rev. 2010, 14, 217–232.
[3] Y. Chisti, Biofuels 2010, 3, 233–235.
[4] a) B. Peng, Y. Yao, C. Zhao, Angew. Chem. Int. Ed. 2012, 51, 2072–
2075; b) B. Peng, X. Yuan, C. Zhao, J. A. Lercher, J. Am. Chem.
Soc. 2012, 134, 9400–9405; c) B. Peng, C. Zhao, S. Kasakov, S. Forai-
ta, J. A. Lercher, Eur. J. Chem. 2013, 19, 4732–4741.
[5] L. M. Velichkina, A. N. Pestryakov, A. V. Vosmerikov, I. V. Tuzov-
skaya, N. E. Bogdanchikova, M. Avalos, M. Farias, H. Tiznado, Pe-
troleum Chemistry 2008, 48, 355–359.
[6] R. Nares, J. Ramꢄrez, A. Gutiꢅrrez-Alejandre, C. Louis, T. Klimova,
J. Phys. Chem. B 2002, 106, 13287–13293.
[7] M. A. Cauqui, J. M. Rodriguez-Izquierdo, J. Non-Cryst. Solids 1992,
147–148, 724–738.
[8] a) J. W. Geus, Dutch Pat. Appl. 6705,259, 1967; 6813,236, 1968;
b) L. A. M. Hermans, J. W. Geus in Preparation of Catalysts II (Eds.:
B. Delmon, P. Grange, P. A. Jacobs, G. Poncelet), Elsevier, Amster-
dam, 1979, p.113.
[9] a) P. Burattin, M. Che, C. Louis, J. Phys. Chem. B 1997, 101, 7060–
7074; b) P. Burattin, M. Che, C. Louis, J. Phys. Chem. B 1998, 102,
2722–2732.
[28] M. Kermarec, D. Olivier, M. Richard, M. Che, F. Bozon-Verduraz, J.
Phys. Chem. 1982, 86, 2818–2827.
[29] E. Garbowski, J. C. Vedrine, Chem. Phys. Lett. 1977, 48, 551–554.
[30] M. Mihaylov, K. Hadjiivanov, Langmuir 2002, 18, 4376–4383.
[31] C. Louis, Z. Cheng, M. Che, J. Phy. Chem. 1993, 97, 5703–5712.
[32] a) M. Mihaylov, K. Hadjiivanov, H. Knçzinger, Catal. Lett. 2001, 76,
59–63; b) S. Derrouiche, D. Bianchi, Appl. Catal. A 2006, 313, 208–
217.
[10] M. Han, Q. Liu, J. He, Y. Song, Z. Xu, J. Zhu, J. Adv. Mater. 2007,
19, 1096–1100.
[33] D. G. Blackmond, E. I. Ko, J. Catal. 1985, 96, 210–221.
[34] a) B. I. Mosqueda-Jimꢅnez, A. Jentys, K. Seshan, J. A. Lercher, J.
Catal. 2003, 218, 348–353; b) D. L. Hoang, H. Berndt, H. Miessner,
E. Schreier, J. Volter, H. Lieske, Appl. Catal. A 1994, 114, 295–311.
[35] a) S. T. Gentry, S. F. Kendra, M. W. Bezpalko, J. Phys. Chem. C
2011, 115, 12736–12741; b) P. Dagtepe, V. Chikan, J. Phys. Chem. C
2010, 114, 16263–16269.
[36] a) J. He, C. Zhao, J. A. Lercher, J. Am. Chem. Soc. 2012, 134,
20768–20775; b) B. Coq, F. Figueras, Coord. Chem. Rev. 1998, 178–
180, 1753–1783.
[11] a) H. Bçnnemann, W. Brijoux, R. Brinkmann, T. Joussen, Angew.
Chem. 1990, 102, 324–326; Angew. Chem. Int. Ed. Engl. 1990, 29,
273–275; b) Y. Hou, S. J. Gao, J. Mater. Chem. 2003, 13, 1510–1512.
[12] S. Carenco, C. Boissiꢆre, L. Nicole, C. Sanchez, P. Le Floch, N. Mꢅ-
zailles, Chem. Mater. 2010, 22, 1340–1349.
[13] a) J. S. Bradley, B. Tesche, W. Busser, M. Maase, M. T. Reetz, J. Am.
Chem. Soc. 2000, 122, 4631–4636; b) N. Cordente, M. Respaud, F.
Senocq, M.-J. Casanove, C. Amiens, B. Chaudret, Nano Lett. 2001,
1, 565–568; c) M. R. Knecht, J. C. Garcia-Martinez, R. M. Crooks,
Chem. Mater. 2006, 18, 5039–5044.
[14] ꢇ. Metin, V. Mazumder, S. ꢇzkar, S. Sun, J. Am. Chem. Soc. 2010,
132, 1468–1469.
[15] a) A. Gil, A. Dꢄas, L. M. Gandꢄa, M. Montes, Appl. Catal. A 1994,
109, 167–179; b) J. van de Loosdrecht, A. M. van de Kraan, A. J. van
Received: March 15, 2013
Published online: June 21, 2013
9842
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9833 – 9842