Facile preparation of Ni/Al2O3 catalytic formulations with the aid of cyclodextrin complexes: Towards highly active and robust catalysts for the direct amination of alcohols

Article abstract of DOI:10.1016/j.jcat.2017.10.006

A series of Ni/Al₂O₃ catalysts with variable Ni loading (2–20 wt%) were prepared by aqueous wet impregnation of a nitrate precursor using native cyclodextrins as metal complex hosts. The impact of β-CD was carefully characterized at different stages of the preparation by a set of complementary techniques including TG-MS analysis, mass spectrometry, X-ray diffraction, temperature-programmed reduction, CO pulse chemisorption, X-ray photoelectron spectroscopy and electron microscopy. It was found that the use of cyclodextrins afforded a much higher Ni dispersion and narrower distribution of Ni particle sizes, as well as a much higher availability of reduced surface Ni species. As a result, the cyclodextrin-assisted catalysts exhibited enhanced catalytic properties in the direct amination of benzyl alcohol with aniline and 1-octanol with ammonia, both operated via the hydrogen borrowing mechanism. Furthermore, the use of cyclodextrins allowed a significant improvement of the robustness of the catalysts by mitigating the nickel leaching during reaction.

Full text of DOI:10.1016/j.jcat.2017.10.006

Journal of Catalysis 356 (2017) 111–124
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Facile preparation of Ni/Al2O3 catalytic formulations with the aid of
cyclodextrin complexes: Towards highly active and robust catalysts for
the direct amination of alcohols
Ajay Tomer a,b, Frédéric Wyrwalski a, Cédric Przybylski c, Jean-François Paul a, Eric Monflier a,
Marc Pera-Titus b, , Anne Ponchel a,
⇑
⇑
a Univ. Artois, CNRS, Centrale Lille, ENSCL, Univ. Lille, UMR 8181, Unité de Catalyse et de Chimie du Solide (UCCS), F-62300 Lens, France
b Eco-Efficient Products and Processes Laboratory (E2P2L), UMR 3464 CNRS-Solvay, Shanghai 201108, PR China
c Sorbonne Universités, Université Pierre et Marie Curie, IPCM-CNRS UMR 8232, Paris, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of Ni/Al2O3 catalysts with variable Ni loading (2–20 wt%) were prepared by aqueous wet impreg-
nation of a nitrate precursor using native cyclodextrins as metal complex hosts. The impact of b-CD was
carefully characterized at different stages of the preparation by a set of complementary techniques
including TG-MS analysis, mass spectrometry, X-ray diffraction, temperature-programmed reduction,
CO pulse chemisorption, X-ray photoelectron spectroscopy and electron microscopy. It was found that
the use of cyclodextrins afforded a much higher Ni dispersion and narrower distribution of Ni particle
sizes, as well as a much higher availability of reduced surface Ni species. As a result, the cyclodextrin-
assisted catalysts exhibited enhanced catalytic properties in the direct amination of benzyl alcohol with
aniline and 1-octanol with ammonia, both operated via the hydrogen borrowing mechanism.
Furthermore, the use of cyclodextrins allowed a significant improvement of the robustness of the cata-
lysts by mitigating the nickel leaching during reaction.
Received 6 July 2017
Revised 4 October 2017
Accepted 5 October 2017
Keywords:
Amination
Cyclodextrin
Octanol
Borrowing hydrogen
Nickel-alumina
Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction
which has been well described for organometallic cata-lysts based
on Ru and Ir, [6–12] comprises three consecutive steps: (1) dehy-
Amines are N-containing intermediates with a broad variety of
applications, serving as building blocks in the manufacture of agro-
chemicals, pharmaceuticals, detergents, polymers, solvents and
dyes [1]. The world demand of amines is expected to increase by
8% in the following years mainly driven by China and other Asian
markets to reach a value of $ 20 billion US by 2020. Although the
amine industry has attained a position of maturity, it still faces
numerous challenges, especially when regarding the synthesis of
primary alkylamines. The existing processes often lack of selectiv-
ity, producing a mixture of amines and/or encompassing haz-
ardous reagents and waste (e.g., salts) [2]. As an alternative, the
direct amination with alcohols could become an eco-efficient strat-
egy, since water is formed as sole byproduct and the process is
compatible with biorefineries [3–5].
drogenation of the alcohol to an aldehyde/ketone, (2) formation
of an imine or enamine intermediate by fast condensation of the
carbonyl with ammonia or an amine, and (3) hydrogenation of
the imine to give the final amine. In this mechanism, H2 is tempo-
rally borrowed by the action of the catalyst from the alcohol to
the imine.
The most studied heterogeneous catalysts for the direct synthe-
sis of amines from long-chain (fatty) alcohols via the H2 borrowing
mechanism rely on Ni and Cu comprising Raney Ni [13–16] as well
as Ni [17–22], Cu [23–28], NiCu [29–32], NiCuFeOx [33] or NiCuZn
[34] supported over alkaline or amphoteric oxides. In particular,
Shimizu and co-workers recently showed that 10 wt% Ni/h-Al2O3
could afford high yields (70–96%) in the reaction of primary and
secondary alcohols with excess ammonia at 160 °C for 13–72 h
[20]. The major shortcoming of Ni and Cu catalysts is ascribed to
their large metal contents (most often >15 wt%), as well to their
heterogeneous particle distribution and poor dispersion, which
impact negatively their activity for amination. Furthermore, high
Ni/Cu loaded catalysts are prone to leaching upon exposure to
ammonia and polar solvents.
The direct amination of alcohols proceeds over metals via a H2
borrowing or H2 autotransfer mechanism. This tandem mechanism,
⇑
Corresponding authors.
E-mail addresses: marc.pera-titus-ext@solvay.com (M. Pera-Titus), anne.
ponchel@univ-artois.fr (A. Ponchel).
https://doi.org/10.1016/j.jcat.2017.10.006
0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

112
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
Since the size of metal clusters, their dispersion and degree of
ahydrate Ni(NO3)2ꢁ6H2O (>99 wt%) was procured from Sigma
Aldrich, China. Native b-cyclodextrin was supplied by Roquette
Frères (Lestrem, France) whereas a- and c-cyclodextrins were pur-
reduction largely determine the catalytic activity for direct alcohol
amination, the optimization of catalyst preparation through simple
and reliable protocols remains of fundamental importance. Conven-
tional metal-supported catalysts are often prepared by aqueous
impregnation of a water-soluble metal salt precursor (generally
nitrate) followed by calcination and reduction under H2. The direct
introduction of organic additives in the impregnation solution has
been reported as a simple method for improving the dispersion
and tuning the morphology of metal nanoparticles over supports
while reducing the metal-support interaction. This is the case for
instance of cyclodextrins (CDs) [35–37]. Native CDs are cyclic
chased from Wacker Chemie. The reactants aniline, benzyl alcohol,
1-octanol and ammonia, all supplied by J&K Scientific, China (99.5%
purity), were used in the catalytic tests. O-xylene (J&K, 99.5% pur-
ity) was used as solvent in the 1-octanol/ammonia catalytic tests.
N-benzylideneaniline, N-benzylaniline, N,N-dibenzylaniline, ben-
zonitrile, toluene, benzene, 1-octylamine, dioctylamine, triocty-
lamine and octanenitrile standards for GC calibration were all
purchased from J&K (purity 99.5%). All the reactants were used
without further purification.
oligosaccharides with 6 (a-CD), 7 (b-CD) or 8 (c-CD) glucopyranose
units connected through glycosidic
a-1,4 bonds generating a charac-
2.2. Catalyst preparation
teristic doughnut-shaped structure [38,39]. This particular structure
allows not only the formation of inclusion complexes with a wide
variety of organic compounds, but also the formation of adducts with
inorganic metal salts, which can find applications as capping agents
or templates in the synthesis of nanoparticles and materials [40–44].
In 2011, Khodakov and Monflier reported the first application of
CDs in the preparation of alumina-supported Co catalysts for
Fischer-Tropsch synthesis. The addition of b-CD to the cobalt
nitrate impregnation solution afforded higher metal dispersions
(after calcination at 400 °C) and a higher metal reducibility, result-
ing in a higher hydrocarbon productivity [35]. The use of CDs was
further extended to the preparation of zirconia-supported cobalt
oxide catalysts (5 wt% Co) for the total oxidation of formaldehyde.
Both the nature of the cobalt precursor and the b-CD/Co molar ratio
played a crucial role on the catalytic performance of Co3O4/ZrO2
catalysts by enhancing the dispersion and reducibility of the cobalt
active species [36,37]. The most striking effects were observed
with the optimized catalyst prepared from a cobalt nitrate precur-
sor with a b-CD/Co molar ratio of 0.1, resulting in a drop by 20 °C in
the light-off temperature. This optimal ratio was related to the
ability of b-CD to interact with Co(NO3)2 by forming ion-
molecule complexes, where Co(II) cations are coordinated between
two O atoms of the hydroxyl groups at the b-CD rim (note that b-
CD has 21 OH groups in total). The use of b-CD as additive for metal
impregnation over supports was also applied to other catalytic
transformations [45–47]. For example, He and coworkers reported
the beneficial effect of native CDs on the Ni dispersion over silica
supports (SBA-15, MCM-41, 5 wt% Ni), resulting in a higher activity
and resistance against coke deposition for the dry reforming of
methane [46,47].
Herein, we concentrate our attention on the preparation of Ni/
Al2O3 catalysts assisted by CDs (CD/Ni = 0.1) for the synthesis of pri-
mary amines from aliphatic alcohols via the H2 borrowing mecha-
nism. In particular, we studied the effect of b-CD addition with
respect to the metal content in the supported catalysts (range 2–20
wt% Ni) to assess the versatility of the preparation method. Detailed
characterization of the different CD-assisted Ni/Al2O3 catalysts was
conducted, underlying the main benefits of b-CD for achieving
homogeneous particle size distributions and preferential surface Ni
species while discouraging the formation of inactive Ni aluminates.
Finally, the catalytic performance was assessed in the direct amina-
tion of benzyl alcohol (chosen as a model reaction), and more impor-
tantly in the direct amination of 1-octanol with NH3, aiming at
promoting the selectivity to the value-added primary 1-octylamine.
A series of Ni oxide catalysts with Ni loading in the range 2–20
wt% were synthesized by wet impregnation using an aqueous solu-
tions of Ni nitrate and b-CD as follows: Ni(II) nitrate hexahydrate
solutions based on different concentrations was added to 250 mL
of an aqueous solution containing 0.1 M equivalent of native b-
CD (C42H47O35, M = 1134 g molꢀ1). This solution was kept under
stirring for 2 h at room temperature. The alumina support (5 g)
was then added to the solution and the as-generated solid suspen-
sion was further stirred for 2 h. After this period, water was slowly
removed at 60 °C until dryness using a rotary evaporator. The
recovered solid was dried overnight in an oven at 100 °C and cal-
cined at 400 °C for 4 h using a heating ramp of 2 °C minꢀ1 under
air flow (2 L(STP)/h). The final catalysts were denoted as xNi/Al-
CDy where x corresponds to the wt% Ni loading and y corresponds
to the b-CD/Ni molar ratio (y = 0.1). Some catalysts were also syn-
thesized using a-CD and c-CD at a CD/Ni molar ratio of 0.1. Like-
wise, a series of Ni-alumina catalysts were also synthesized by
the wet impregnation method without CD, which are hereinafter
designated as xNi/Al, where x stands for the wt% Ni loading.
2.3. Catalyst characterization
Thermogravimetry – Mass Spectrometry (TG-MS). TG measure-
ments were performed using a TA SDT 2960 instrument equipped
with a gas flow system. The solid was treated from room temper-
ature to 800 °C (5 °C min–1) under a gas mixture composed of He
(80 vol%) and O2 (20 vol%) with a flow rate of 75 mL(STP) min–1
.
Approximately, 10 mg of sample was heated in an open Pt crucible.
The temperature-programmed decomposition products were in-
situ analyzed using a Pfeiffer vacuum Omnistar GSD 320 mass
spectrometer.
Electrospray
ionization-Mass
Spectrometry
(ESI-MS).
Electrospray-mass spectrometry (ESI-MS) experiments were per-
formed using a LTQ-Orbitrap XL from Thermo Scientific (San Jose,
CA, USA) and operated in positive ionization mode, with the spray
voltage at +3.85 kV and a sheath and auxiliary gas flow at 45 and
15 a.u., respectively.
1/1 (v:v) with or without a nickel nitrate were continuously
infused at 5 ml minꢀ1 using a 250-
L syringe. The applied voltages
a-, b- and c-CD at 10 mM in water/methanol
l
were +40 and +100 V (positive mode) for the ion transfer capillary
and the tube lens, respectively. The ion transfer capillary was held
at 275 °C. The resolution level was set to 30,000 (m/z = 400) for all
the studies, while the m/z range was set to 300–2000 m/z in the
profile mode and in the normal mass ranges during full scan exper-
iments. The spectra were analyzed using the acquisition software
XCalibur 2.0.7 (Thermo Scientific, San Jose, CA, USA) without
smoothing and background substraction. Higher energy collision
dissociation (HCD) experiments were performed with an activation
time of 100 ms according to a previous study [48] and occurred in
an octopole collision cell aligned to the C-trap and detection in the
Orbitrap in centroid mode. This dedicated cell was supplied with
2. Experimental section
2.1. Materials
c
-Al2O3 (Puralox SCCA 5/170, 154 m2 gꢀ1) was purchased from
SASOL and used without further pre-treatment. Nickel nitrate hex-

A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
113
an rf voltage (2.6 MHz, 500 Vp-p), where the DC offset can be var-
ied by 250 V and crossed with nitrogen gas at a pressure of 5.103
mbar. The endcap voltage was controlled using normalized colli-
sion energy (NCE) technology, where the collision energy was
modified for each experiment and was expressed in% of NCE. The
peak-to-peak voltage was scaled as Vp-p = NCE/30 ꢂ (a(m/z) + b),
where a and b are the instrument variable parameters set to
0.002 and 0.4, respectively. The standard range was 0–100% corre-
sponding to 0–500 V. During the survival yield experiments, the
NCE was varied from 0 to 50%. The precursor selection window
was set to m/z = 2.5 during MS/MS experiments. The automatic
gain control (AGC) allowed accumulation up to 106 ions for FTMS
scans, 3.105 ions for FTMSn scans and 3.104 ions for ITMSn scans.
The maximum injection time was set to 500 ms for both FTMS
and FTMSn scans and 100 ms for ITMSn scans. For all scan modes,
1 mscan was acquired.
monochromated X-ray source (10 kV, 15 mA) with a pass energy
of 40 eV (0.1 eV/step) for high resolution spectra, and a pass energy
of 160 eV (1 eV/step) for survey spectrum in hybrid mode and slot
lens mode, respectively. The adventitious C1s binding energy
(285.0 eV) was used as an internal reference. Prior to the measure-
ments, the samples were reduced in a pre-treatment chamber at
580 °C for 30 min under a H2 flow (50 mL (STP)/min).
TEM & SEM-EDX. TEM/STEM was performed using a Tecnai
microscope operating at an accelerating voltage of 100 kV with a
resolution point of 4.5 Å. Before analysis, the solid powder was
directly dispersed over the carbon coated copper grid. Several
micrographs from different areas of the sample were taken. The
size distribution of the nanoparticles was estimated after the mea-
surement of 50 particles through a manual analysis of micrographs
with Image J software using Origin 8 Pro to generate the particle
size distributions.
BET analysis. N2 adsorption/desorption isotherms were mea-
sured at -196 °C using an ASAP2020 Micromeritics AutoChem
apparatus after degassing the samples overnight at 100 °C. The
specific area was computed using the Brunauer-Emmett-Teller
(BET) equation in the P/P0 range 0.05–0.25, while the pore size dis-
tribution was measured from the desorption branch using the
Barrett-Joyner-Halenda (BJH) method.
2.4. DFT calculations
The DFT calculations were performed using the VASP package
[49–51] relying on the GGA PBE [52] functional to compute ener-
gies and including an implicit solvation module [53,54]. The
Kohn-Sham equations were solved until the energy was converged
within 10ꢀ6 eV and the electron ion interactions were described
using PAW pseudo potential [55]. The cut off energy was set to
450 eV and due to the size of the unit cell (25 ꢂ 25 ꢂ 25 Å3) the cal-
culations were performed at the gamma-point. The molecules
placed on the center of the unit cell were fully optimized without
Room temperature and in-situ X-Ray diffraction (XRD). At normal
configuration (room temperature), the XRD experiments were con-
ducted on a Bruker D8 Advance diffractometer in Bragg Brentano
geometry equipped with a copper anode (k = 1.5418 Å) and a 1D
PSD Lynxeye detector. The scattering intensities were measured in
the range 5° < 2h < 80° with a step size of 0.02°. The diffraction pat-
terns were indexed by comparison with JCPDS files. In-situ temper-
ature controlled XRD patterns were measured to assess the
structural changes that may occur during the thermal reduction of
the catalyst under H2. The measurements were conducted on a Bru-
ker D8 diffractometer equipped with a Vantec 1 PSD detector and an
Anton Paar HTK 1200N chamber connected to a gas line. The powder
sample was placed in its calcined form on a Pt holder and the XRD
measurements (30°< 2h <70°) were performed under H2 flow (5%
any constraint until the force was lower than 0.03 eV Åꢀ1
.
2.5. Catalytic tests in amination reactions
The performance of the different Ni catalysts was evaluated
first in the model amination reaction of benzyl alcohol with ani-
line (Scheme 1). In typical conditions, the reaction was per-
formed using 2 mmol of aniline, 6 mmol of benzyl alcohol and
35 mg of catalyst at 140 °C for 1 h using Biotage microwave vials
under atmospheric N2 pressure and a stirring speed of 600 rpm.
Likewise, the reaction of 1-octanol with ammonia was performed
in a 30-mL stainless steel autoclave equipped with a pressure
gauge and a safety valve (Scheme 2). In a given experiment,
the reactor was charged with a known amount of 1-octanol
(1.3 mmol) and a known amount of pre-reduced catalyst (55
mg). The reactor was sealed and evacuated by applying vacuum
followed by ammonia charging (7 bar). The reactor was then
placed on a hot plate equipped with a magnetic stirrer and
the reaction was conducted at different temperatures (160 °C,
180 °C) for 4 h.
vol., 2 L(STP) h–1) from room temperature to 800 °C (5 °Cꢁminꢀ1
)
with a scan every 25 °C, hold at 800 °C for 1 h with two scans, and
then decrease in temperature(50 °Cꢁminꢀ1) with a scan every 100 °C.
H2-Temperature
programmed
reduction
(H2-TPR).
H2-
temperature programmed reduction was used to assess the
reducibility of the different catalysts. The reduction profiles were
measured using
a Micromeritics AutoChem 2920 instrument
equipped with a thermal conductivity detector (TCD) and a cold
trap before the detector. The profiles were recorded by reducing
ca. 75 mg of sample under 5% H2-Ar flow (40 mL(STP) minꢀ1
)
and heating the sample from 30 °C to 1000 °C using a heating rate
of 10 °C minꢀ1. The H2 consumption was used to measure the
reduction level of the different nickel oxide species present in
the calcined samples.
The reactants (aniline, benzyl alcohol and 1-octanol) and the
expected products for both reactions, i.e. N-benzylideneaniline
(BZIA), N-benzylaniline (BZA), N,N-dibenzylaniline (DBZA), octy-
lamine (OA), dioctylamine (DOA), trioctylamine (TOA), octaneni-
trile (ON) and 1-octanol were analyzed and quantified using an
Agilent 7890 GC equipped with a J&W Scientific HP-5 capillary col-
CO pulse chemisorption. Pulse CO chemisorption measurements
were carried out on a Micrometrics AutoChem 2920 instrument.
Briefly, a known amount of sample (ca. 100 mg) was pre-reduced
under H2 at 580 °C for 30 min using a heating rate of 10 °C minꢀ1
.
umn (30 m ꢂ 0.32 mm ꢂ 0.25
lm) with 5 wt% phenyl groups. In
Later on, the sample was cooled down to 50 °C and then CO
chemisorption analysis was performed by introducing successive
pulse doses of 5% CO-He. Most often, the chemisorption process
occurred in the first few pulses and the pulse area was smaller than
the sequential pulse area. The chemisorption was repeated until
the pulse area peaks exhibited no longer change.
X-ray photoelectron spectroscopy (XPS). The surface composition
of the synthesized catalysts was analyzed by XPS using a Kratos
Axis Ultra DLD apparatus equipped with a hemi-spherical analyzer
and a delay line detector. The spectra were recorded using an Al
addition to the above stated main products, toluene (Tol) was
formed in the model reaction as side product by dehydroxylation
of benzyl alcohol. Biphenyl was used as the internal standard. Mass
balances were accurate to within 5% for all the catalytic tests. The
conversion of the limiting reactant (LR: aniline and 1-octanol),
selectivity, normalized selectivity and yield to the N-containing
products were defined as follows:
nLR
noLR
Conversion ð%Þ ¼ 1 ꢀ
ꢃ 100

114
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
Scheme 1. Potential amination products in the reaction between benzyl alcohol with aniline.
Scheme 2. Potential amination products in the reaction between 1-octanol with ammonia.
ni
at a much broader temperature range 250–400 °C suggests the pres-
Selectivityi ð%Þ ¼
ꢃ 100
noLR ꢀ nLR
ence of Ni nitrate species with a high affinity for the support.
5Ni/Al-b-CD0.1 displays a markedly different thermal pattern
with three distinct weight loss regions (Fig. 1C). The total weight
loss of 20.5% is in good agreement with the value of 19.5% corre-
sponding to the theoretical weight loss for complete decomposi-
tion of Ni nitrate hexahydrate into NiO. The first weight loss
below 120 °C is attributed to physically adsorbed H2O, the second
and most distinct weight loss between 150 °C and 250 °C is due to
the decomposition of nitrate ions, and the third weight loss above
275 °C refers to the thermal decomposition of free b-CD, as previ-
ously evidenced in the thermal pattern of Al-b-CD0.1 (Fig. 1A). This
interpretation is consistent with the mass fragmentograms. The
decomposition of b-CD in 5Ni/Al-b-CD0.1 exhibits two CO2 bands
(m/z = 44) centered at 212 °C and 275 °C, respectively. The first
band, matching perfectly the NO release band, can therefore be
attributed to the interaction between b-CD and the nickel nitrate
salt. The second band can be at first sight ascribed to the decompo-
sition of b-CD free from interaction with the precursor. This
hypothesis can be further confirmed by comparing the bands at
275 °C for 5Ni/Al-b-CD0.1 and Al-b-CD0.1. Note also that the pres-
ence of a band at 210 °C, which can be attributed to nitrate decom-
position into NO (Fig. 1C, m/z = 30), is consistent with the
decomposition of part of the b-CD into CO2 and H2O, as evidenced
by the bands at 212 °C and 213 °C, respectively (Fig. 1C, m/z = 44
and m/z = 18). These similar temperatures strongly suggest that
both b-CD and the Ni nitrate precursor are intimately linked
together within the dried oxide catalyst and that their interaction
significantly accelerates the decomposition of the oligosaccharide
with a temperature difference of ꢄ70 °C. Furthermore, it is worth
mentioning that the signal accounting for Ni nitrate decomposition
in the MS spectrum of 5Ni/Al-b-CD0.1 is also much more intense
than that observed for 5Ni/Al (compare the profiles at m/z = 30,
Fig. 1C and B). The latter observation suggests that an important
portion of the nitrate species coming from the Ni salt precursor
is still present over 5Ni/Al-b-CD0.1 after the drying step (100 °C
overnight). This behavior can be connected again to the presence
of specific interactions occurring between b-CD and the Ni nitrate
precursor via the immobilization onto the support of non-covalent
complexes with enhanced thermal stability against transformation
into nickel oxide. Very similar decomposition patterns were also
observed for the dried catalysts synthesized with the two other
ni
ni
P
Normalized selectivity ð%Þ ¼
ꢃ 100
ni
noLR
Yieldi ð%Þ ¼
ꢃ 100
ð1Þ
Moles of formed product þ reactants
Carbon balance ¼
Moles of reactants
where noLR and nLR refer to the initial and final mole number of the
limiting reactant, respectively, whereas ni corresponds to the mole
number of the N-containing products formed. Each catalytic run
was performed in duplicate and the reported results are the average
between two runs with an uncertainty equal or less than 3%.
3. Results and discussion
3.1. Characterization of the dried catalysts by thermogravimetric
analysis coupled to mass spectrometry (TG-MS)
The ability of b-CD to interact with the Ni nitrate precursor was
first investigated in the dried state by TG-MS at a Ni/b-CD molar
ratio of 0.1. Fig. 1 plots the thermal profiles of the alumina support
wet-impregnated with Ni nitrate alone (5Ni/Al), and Ni nitrate with
b-CD (5Ni/Al-b-CD0.1) after drying at 100 °C (see Section 2 for fur-
ther details). For comparison, the TG-MS results obtained with a
Ni-free control sample prepared by b-CD impregnation onto the
parent alumina (Al-b-CD0.1) are also provided. The thermal profile
of Al-b-CD0.1 exhibits a two-step decomposition process with a total
weight loss of ca. 13% at 800 °C. The first weight loss (2%) in the
temperature range 50–200 °C corresponds to the removal of phys-
ically adsorbed water, whereas the second weight loss (11%) in
the range 260–370 °C with a major weight loss at ca. 277 °C is attri-
butable to the thermal decomposition of b-CD (Fig. 1A) [56]. The
thermal pattern of 5Ni/Al also exhibits a two-step decomposition
process with a first weight loss at a temperature below 120 °C cor-
responding to the desorption of physically adsorbed water, and a
second weight loss in the temperature range 200–500 °C corre-
sponding to nitrate decomposition (Fig. 1B). The total weight loss
for this sample is ca. 14%. This value is close to the weight loss
(16%) that can be expected by assuming total conversion of fully
hydrated Ni(NO3)2ꢁ6H2O into NiO. The low intensity of the mass
fragmentogram corresponding to the signal of NO release at m/z =
30 suggests that the nitrate precursor decomposed very early dur-
ing Ni impregnation, most likely along the drying step. The presence
of weak NO release bands decomposing at higher temperature and
native CDs (i.e.
a-CD and c-CD), underlining that whatever the
type, CDs behave as complexing molecules for Ni nitrate (Fig. S1).
3.2. Study of the metal-CD complex formation by Electrospray-Mass
Spectrometry (ESI-MS)
To gain further insight into the occurrence of an interaction
between CDs (a, b and c-CD) and the Ni nitrate precursor, we con-

A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
115
A
γ-Al2O3-β-CD0.1
B
5Ni/Al
C
5Ni/Al-β-CD0.1
Fig. 1. TG profiles (left side) and corresponding MS spectra of CO2 (m/z = 44), H2O (m/z = 18) and NO (m/z = 30) (right side) for the following dried materials: (A) Al-b-CD0.1, (B)
5Ni/Al and (C) 5Ni/Al-b-CD0.1
.
ducted a series of electrospray-mass spectrometry (ESI-MS) analy-
ses in positive mode. The experiments were carried out on Ni
(NO3)2 and CD (a-CD, b-CD and c-CD) mixtures in aqueous metha-
required to decompose 50% of the initial ion (CE50). This value
can be measured from the sigmoid survival yield curve using the
higher collision-induced dissociation (HCD) mode [48,57]. The sur-
vival yield curves issued from the decomposition of some com-
nol by setting the CD/Ni molar ratio to 0.1 (see Section 2). The most
intense ion in the positive MS spectrum of the b-CD/Ni nitrate sys-
tem appears at m/z = 596.15, which can be unambiguously
assigned to [b-CD + Ni]2+, supporting the direct complexation of
b-CD with Ni. Further evidence of complexation can be deduced
from the presence of additional ions at m/z = 624.11 and m/z =
1163.33, which can be ascribed to [b-CD + 2Ni-2H]2+ and [2b-CD
+ Ni]2+, respectively. Even if less prominent, numerous other com-
plexes encompassing the coordination of the nitrate counterion
with b-CD are also observed at m/z = 655.61, m/z = 687.10, m/z =
747.56 and m/z = 855.83, being attributed to [b-CD + 2Ni + NO3-
H]2+, [b-CD + 2Ni + 2NO3]2+, [b-CD + 3Ni + 3NO3-H]2+ and [2b-CD
+ 3Ni + 2NO3-H]3+, respectively. Noteworthy, similar complexes
plexes formed between Ni nitrate with either
are gathered in Fig. S4. For most of the complexes, the stability
was found to follow the trend b-CD > -CD > -CD. Interestingly,
a-CD, b-CD or c-CD
a
c
such complex formation can be explained on the basis of the inter-
action between the OH groups present on the CD rims and the Ni
nitrate precursor by assuming 2OH groups per 1Ni for molar ratio
of CD/Ni = 0.1.
3.3. DFT study of the metal-CD complex formation
Periodic DFT calculations were performed on b-CD-Ni com-
plexes to rationalize the interaction between the organic and the
metallic part, but also to define the most favorable interaction sites
for the formation of complexes. To take into account the presence
of an aqueous environment, an implicit solvation model was used
for b-CD, the aqua nickel cation [Ni(H2O)6]2+ and the resulting
were observed in the ESI-MS spectra of
Ni(NO3)2 (Fig. S3).
a-CD or c-CD mixed with
The strength of the interaction in non-covalent complexes can
be assessed from the center-of-mass energy, which is the energy

116
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
inclusion complex [53,54]. If the nickel cation is placed inside the
b-CD, the inclusion reaction is endothermic (DrE = 0.26 eV), which
confirms that the inclusion of the hexahydrated nickel ion is not
favorable in the hydrophobic cavity of b-CD. Alternative reaction
sites are formed by OH groups on the rims of the b-CD molecule,
which can substitute one water molecule on the aqua-nickel com-
plex. Due to the geometry of the secondary face, two bonds are
formed between the cation and b-CD, and the formation of the
compound is exothermic. The reaction energies are ꢀ0.59 eV or
ꢀ0.38 eV when the two OH groups belong to the same or different
glucose molecules, respectively. The cation can also interact with
one primary OH group on the other side of b-CD. This reaction
leads to the most stable b-CD-Ni complex with a reaction energy
of ꢀ0.75 eV. The final geometry after optimization is given by
way of illustration in Fig. 2. This type of complex formed has also
been detected in the ESI-MS experiments. Taken together, our cal-
culations provide further evidence on the fact that the hydroxyl
sites (either on the primary or secondary rim) can effectively con-
tribute to the formation of CD-Ni complexes and that various
geometries are possible due to the numerous OH groups present
on b-CD.
3.4.2. X-Ray diffraction
Fig. 3 plots the XRD patterns of the xNi/Al and xNi/Al-b-CD0.1
catalysts after calcination at 400 °C under air as a function of the
Ni loading (range 2–20 wt%).
In the absence of b-CD (Fig. 3A), a neat Ni(2 0 0) reflection line
can be visualized at 2h = 43.5°, becoming more intense with the Ni
loading, especially above 8 wt%. In contrast, the XRD patterns of
the catalysts synthesized using b-CD exhibit weaker and broader
NiO (2 0 0) lines (Fig. 3B). These observations point out a decrease
in the average size of the NiO crystallites mediated by b-CD, sug-
gesting a higher NiO dispersion over alumina after calcination at
400 °C irrespective of the Ni loading. For example, for the highest
Ni-loaded catalyst (20 wt%), the average NiO crystallite size can
be estimated at 9.4 nm for 20Ni/Al-b-CD0.1 against 15.8 nm for
20Ni/Al. These results unambiguously demonstrate that the addi-
tion of b-CD during Ni impregnation can induce the formation of
smaller NiO nanoparticles. We attribute this observation to the sta-
bilization of b-CD-Ni(II) adducts after water removal, thus preserv-
ing a good distribution of Ni(II) species in the vicinity of b-CD and
delaying the formation, growth and aggregation of NiO during cal-
cination. This behavior is perfectly in line with the results pre-
sented above from TG-MS analyses (Fig. 1).
To visualize the phase transformations occurring during reduc-
tion under H2, the in-situ temperature-controlled XRD patterns
were further measured at temperatures up to 800 °C. To this aim,
we selected two highly Ni-loaded calcined catalysts synthesized
without and with b-CD (i.e. 15Ni/Al and 15Ni/Al-b-CD0.1), for
which the NiO reflection lines were clearly visible
(Fig. 4A and B). The reflections at 2h = 43.5° (2 0 0) and 63° (2 2
0) are attributed to NiO, while the characteristic reflections belong-
ing to reduced Ni (Ni0) can be visualized at 2h = 44.3° (1 1 1) and
51.8° (2 0 0). The reflection lines of NiO at 2h = 43.5° and 2h =
63° are relatively intense between room temperature and 400 °C
for 15Ni/Al. These lines roughly disappear after 350 °C, whereas
the intensity of the (1 1 1) and (2 0 0) reflections belonging to Ni0
increase from 350 °C to ꢄ700 °C and then stabilize, suggesting a
complete reduction of NiO after 350 °C. Opposing this behavior,
15Ni/Al-b-CD0.1 exhibits less intense and broader NiO reflections
at lower temperature (T <250 °C), confirming the genesis of smaller
NiO particles. Note also that the only visible NiO (2 0 0) reflection
disappears at 300 °C as compared to 350 °C for 15Ni/Al, revealing a
faster NiO reduction into Ni0. Unlike 15Ni/Al, 15Ni/Al-b-CD0.1 does
not show any reflection belonging to NiO or Ni0 at 2h = 43.5° (2 0 0)
and 44.3° (1 1 1), respectively, in the mid temperature region
(300–450 °C, orange patterns). Overall, these results confirm the
3.4. Characterization of calcined catalysts
3.4.1. Textural properties
Table S1 lists the textural properties of the different catalysts
prepared in this study, with and without b-CD addition (i.e. xNi/
Al and xNi/Al-CD0.1, respectively). For comparison, the textural
properties of 5Ni/Al catalysts prepared with other native CDs (
a-
CD and -CD), as well of the parent -Al2O3, are also given. The
c
c
support shows a mesoporous structure with a BET specific surface
area of 154 m2 gꢀ1, a pore volume of 0.52 cm3 gꢀ1 and an average
pore size of 11.8 nm. Very similar textural properties are observed
for
c-Al2O3 impregnated with b-CD and calcined at 400 °C. It is
clear from this observation that b-CD does not alter the support
properties, and that, after thermal treatment under air at 400 °C,
b-CD is completely removed without any trace of coke deposition.
The surface area, pore volume and pore size decrease to 142 m2
gꢀ1, 0.46 cm3 gꢀ1 and 10.7 nm, respectively, after Ni impregnation
and further calcination (5Ni/Al), reflecting a partial blockage of the
alumina pores [58]. The specific surface area generally decreases
with the Ni loading, but this trend seems to be more moderate
for the series of catalysts prepared by the assistance of CD.
genesis of finer and better dispersed Ni nanoparticles over
c-
Al2O3 for the b-CD-assisted preparations.
3.4.3. Catalyst reducibility
The effect of b-CD on the reducibility of the two series of cata-
lysts prepared with and without b-CD was assessed by tempera-
ture programmed reduction (H2-TPR) as a function of the Ni
loading (range 2–20 wt%) (Fig. 5 and Table 1). In all cases, the total
H2 consumption increases with the Ni content, matching the
expected values calculated by assuming the complete reduction
of NiO to Ni.
In the absence of b-CD, the band corresponding to Ni2O3 [59–
62] shifts to lower temperatures, i.e. from 298 °C to 249 °C when
increasing the Ni loading in the xNi/Al catalysts (with x = 5 to x
= 20 wt% Ni) along with a slight decrease in its percent reduction
from 10% for 8Ni/Al to 5% for 20Ni/Al (Fig. 5A). Ni2O3 species are
generally formed due to NiO disproportionation, generating two
Ni species with different oxidation states (3NiO ? Ni + Ni2O3)
[61,62]. It is worth noting that the H2 consumption corresponding
to the second band increases dramatically with the Ni loading,
Fig. 2. Geometry of the most stable complex: (Ni in blue, C in grey, O in red and H in
white).
reflecting the presence of a larger density of bulk Ni species (a-

A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
117
NiO
A
NiAl2O4
γ-Al2O3
15.8 nm
10.3 nm
B
9.4 nm
Fig. 3. XRD patterns of the calcined xNi/Al catalysts prepared in the absence (A) and presence (B) of b-CD (b-CD/Ni = 0.1) for: (a)
c-Al2O3, (b) 2 wt%, (c) 5 wt%, (d) 8 wt%, (e) 10
wt%, (f) 12 wt%, (g) 15 wt% and (h) 20 wt%. The numbers in the right indicate the average crystallite size of NiO (2 0 0) estimated from the Scherrer equation.
Al2O3
Al2O3
Ni
Ni
NiO
NiO
NiO
(T= 350 oC)
Al2O3
(T= 300 oC)
(T= 350 oC)
Al2O3
Ni
Ni
50 o
800 o
600 o
400 o
200 oC
30 o
C
50 o
800 o
600 o
400 o
200 oC
30 o
C
A
B
C
C
C
C
C
C
C
C
Fig. 4. In-situ XRD patterns of calcined 15Ni/Al prepared without (A) and with b-CD (B) heated under 5% H2 atmosphere from 30 °C to 800 °C at a rate of 5 °C minꢀ1 and then
cooled down to 50 °C at a rate of 50 °C minꢀ1. The violet lines indicate the patterns recorded every 25 °C from room temperature to 775 °C, the red lines indicate the patterns
recorded at 800 °C after holding for 2 h, while the aqua blue lines refer to the patterns recorded every 100 °C during cooling (800 °C to 50 °C). The orange lines indicate the
temperature range where reflection lines of 2h = 43.5 (NiO) and 44.3 (Ni) are absent.
type [63–65]) generated by an uncontrolled growth of NiO species
during the early stage of the impregnation process. Moreover, the
band appearing in the highest temperature region ascribed to the
reduction of hard-to-reduce Ni aluminate species [63,64] appears
to decrease and to be shifted towards lower temperatures with
the Ni loading (compare profiles a and g).
This observation indicates that higher Ni loadings favor the for-
mation of more readily reducible Ni species [64]. Also note that for
the catalysts prepared by conventional Ni impregnation, the reduc-
tion level of surface Ni species (b1 + b2 type [66–69]) is very sen-
sitive to the Ni content. As evidenced from Fig. 6, the highest
percentage of reducible surface metal Ni is achieved for 8Ni/Al

118
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
CD, respectively against 502 °C for b-CD (Table 1, entries 15 and
16).
3.4.4. CO pulse chemisorption
CO pulse chemisorption was used for measuring the Ni disper-
sion, the average Ni particle size and the Ni surface area per gram
of catalyst or per gram of Ni metal on the different catalysts
(Table 1). Fig. 7 compares the CO uptake measured as a function
of the Ni loading for the catalysts prepared without and with b-CD.
In both cases, an increase of the CO uptake is observed with the
Ni loading due to a larger amount of metal Ni sites interacting with
CO molecules. Whatever the Ni loading, the CO uptake is signifi-
cantly higher on xNi/Al-b-CD0.1 than on xNi/Al catalysts (compare
for instance the CO up-take of 252
l
mol gꢀ1 for 15Ni/Al-b-CD0.1
A
and 165
l
mol gꢀ1 for 15Ni/Al). In line with this, a relatively high
Ni dispersion can be maintained with values ranging from 20% to
22% in the range 5–15 wt% Ni, while a further increase to 20 wt%
Ni induces a considerable reduction of the metal dispersion to
13% (Table 1). The resulting Ni particle size estimated from the
CO uptake shows lower values for xNi/Al-b-CD0.1 catalysts as com-
pared to control xNi/Al catalysts (4.5–7.8 nm vs. 6.7–10.8 nm,
respectively).
3.4.5. X-ray photoelectron spectroscopy
Fig. S5 plots the Ni2p XPS spectra measured on 5Ni/Al and 5Ni/
Al-b-CD0.1 after calcination and in-situ reduction at 600 °C. Note-
worthy, no band belonging to the C1s core level is observed in
the survey XPS spectrum of 5Ni/Al-b-CD0.1, confirming the com-
plete removal of b-CD upon calcination. The deconvoluted spectra
of the Ni2p core level for the calcined catalyst (Fig. S5-A,C) exhibit
two main bands centered at 856.5 and ꢄ861 eV that are indicative
of the Ni2+ state and shake-up satellite bands of NiO or Ni2O3,
respectively [70]. An additional band at 852 eV is observed for
the in-situ reduced catalysts (Fig. S5-B,D), which can be attributed
to reduced Ni (Ni0) [71]. The surface concentrations measured from
the XPS spectra indicate that a larger density of surface Ni species
are present in the calcined catalyst synthesized with b-CD (Ni/
Al = 0.10 for 5Ni/Al vs. Ni/Al = 0.14 for 5Ni/Al-b-CD0.1). A similar
trend is observed for the reduced catalysts, showing Ni/Al ratios
of 0.05 and 0.07, respectively. Interestingly, a gain of 30% surface
Ni species has been estimated for reduced 5Ni/Al-b-CD0.1 as com-
pared to reduced 5Ni/Al. This body of results unambiguously con-
firms that the controlled addition of b-CD during Ni impregnation
over alumina allows a marked enhancement of the density of sur-
face Ni species on the support.
B
Fig. 5. H2-TPR profiles of calcined xNi/Al catalysts prepared without (A) and with b-
CD (B); (a) 2Ni/Al, (b) 5Ni/Al, (c) 8Ni/Al, (d) 10Ni/Al, (e) 12Ni/Al, (f) 15Ni/Al and (g)
20Ni/Al.
(73%) and decreases dramatically with the Ni loading down to 37%
for 20Ni/Al due to the formation of bulk Ni oxide species with lar-
ger crystallite sizes.
A significant effect of b-CD addition on the reduction behavior
of NiO species can be observed on the H2-TPR profiles (Fig. 5B).
Except for 2Ni/Al-b-CD0.1, for which the main species are Ni alumi-
nates (84%), the use of b-CD affects the reduction behavior of the Ni
species. On the one hand, the area of the reduction band assigned
to Ni2O3 (Fig. 5B, profiles b-g) is much lower for the b-CD-assisted
samples (xNi/Al-b-CD0.1 with x = 5–20 wt% Ni) than for the control
catalysts prepared without b-CD (Fig. 5A, profiles b-g). Likewise,
3.4.6. TEM and STEM imaging
To confirm the positive effect of b-CD on reducing the Ni parti-
cle size, a dedicated TEM study was carried out on two selected
reduced catalysts, i.e. 10Ni/Al and 10Ni/Al-b-CD0.1, after reduction
under H2 flow at 580 °C. Figs. 8 and 9 show the corresponding TEM
images. The Ni particles in 10Ni/Al are somewhat spherical in
the reduction percentage of bulk NiO species (a type) in the cata-
lysts prepared with b-CD (Table 1, entries 4, 6, 8, 10, 12, 14) is
lower than for those prepared without b-CD (Table 1, entries 3, 5,
7, 9, 11, 13). On the other hand, the decline in the relative propor-
tion of bulk NiO or Ni2O3 species compares well with the higher
formation of surface Ni species irrespective of the Ni loading
(Fig. 6). Indeed, for the xNi/Al-b-CD0.1 series (Fig. 5B, profiles d-
shape
being
heterogeneously
dispersed
over
alumina
(Fig. 8A and B). The Ni particles show a size in the range 3–10
nm with an average particle size centered on ꢄ6–7 nm (Fig. 8D).
Note that additional STEM imaging of 10Ni/Al clearly indicates that
the smallest particles are Ni rich (Fig. S6).
In the case of 10Ni/Al-b-CD0.1, the Ni particles appear to be
more homogeneously dispersed over the support with a majority
of particles having a diameter of around 3–4 nm, which is signifi-
cantly smaller than that estimated for the 10Ni/Al sample (Fig. 9D).
Note that no particles with a size larger than 6 nm are observed.
Although the number of counted particles (about 50 in each case)
is relatively small, requiring careful consideration, the average par-
ticle sizes measured by TEM are in good agreement with the values
g), the area of
a-type species increases fast with the Ni loading
in the range 10–20 wt%, encompassing an increase of the reduction
level ranging from 3 to 21% (Table 1). Besides the effect of b-CD
addition on the reducibility of NiO species, it can be noticed that
a
-CD and
ꢄ80% of NiO surface species can be reduced at slightly higher
reduction temperatures, i.e. 522 °C and 511 °C for -CD and c-
c-CD also show a similar reduction behavior. Overall,
a

A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
119
Table 1
Reducibility and physicochemical properties of xNi/Al catalysts synthesized without and with b-CD.
Entry Sample
Total H2
T (°C) Surf
Ni (b1 + b2) Ni2O3
%Redn
%Redn
%Redn Surf
Nid (b1
+ b2)
%Redn
NiAl2O4
CO uptake
Dispersion
(%D)g
Metal surf.
Particle
size, Sm
(nm)i
Bulk Nic
(l )
mol gꢀ1
area, Am (m2Ni
b
e
f
uptake
l
mol
gꢀ1a
(
a)
(
c
)
g
)
Ni
ꢀ1
h
1
2
3
4
5
6
7
8
2Ni/Al
2Ni/Al-b-CD0.1
5Ni/Al
5Ni/Al-b-CD0.1
8Ni/Al
8Ni/Al-b-CD0.1
10Ni/Al
10Ni/Al-b-CD0.1 1373
12Ni/Al 1570
12Ni/Al-b-CD0.1 1588
15Ni/Al 1934
15Ni/Al-b-CD0.1 1995
20Ni/Al 2583
20Ni/Al-b-CD0.1 2446
330 (106)
266 (89)
793(105)
674 (92)
1098
–
–
38
16
13
6
11
4
18
3
25
11
31
6
–
–
61
84
21
23
6
8
6
6
3
4
2
4
2
2
12
11
23.0
27.5
60.5
84.7
13.5
16.2
14.2
19.8
15.0
20.3
13.5
21.9
14.1
21.4
12.9
19.7
9.3
90.2
107.9
94.7
132.6
100.1
135.0
90.2
146.0
94.5
142.5
86.3
131.6
62.3
86.9
81.9
7.5
6.2
7.1
5.1
6.7
5.0
7.6
4.5
7.1
4.7
7.8
5.1
10.8
7.8
8.2
5.0
0
0
0
10
2
9
2
7
3
7
3
5
3
0
0
558
502
568
538
563
529
560
545
565
531
558
540
522
511
49
70
73
86
67
89
65
81
60
87
37
74
79
81
102.3
137.9
115.3
186.5
144.9
218.5
165.3
252.1
159.0
221.8
52.3
1002 (93)
1364 (94)
9
10
11
12
13
14
15
16
56
21
9
13.0
12.3
20.4
5Ni/Al-
5Ni/Al-
a
-CD0.1
726
703
c
-CD0.1
8
87.0
136.2
a
Estimated by integrating the H2-TPR profile from RT to 1000 °C. The numbers in brackets represent the % of reduction of Ni oxide with respect to the theoretical H2
consumption calculated by considering the Ni loading measured by ICP.
b
Calculated by dividing the integrated area in the temperature region 100–300 °C by the total H2 consumption (range 30–1000 °C).
Calculated by dividing the integrated area in the temperature range 300–500 °C for non-CD assisted catalysts and 300–400 °C for CD assisted catalysts by the total H2
c
consumption (range 30–1000 °C).
d
Calculated by dividing the integrated area in the temperature range 450–750 °C for non-CD assisted catalysts and 400–750 °C for CD assisted catalysts by the total H2
consumption (range 30–1000 °C).
e
Calculated by dividing the integrated area in the temperature range 750–950 °C by the total H2 consumption (range 30–1000 °C).
f
Estimated from the area under the CO pulse signals.
g
Calculated from the CO uptake using equation: (Vchem * S.F. * M.W.)/(c/100) * 100; where Vchem is the amount of CO uptake (mol gꢀ1), S.F. is the stoichiometric factor
(Ni/CO = 2), M.W. is the atomic weight of supported Ni and c is the Ni loading (wt%).
h
Calculated from the equation: (Vchem * NA. * S.F. *
Mean particle size calculated from the following equation: 60c/(Am * q) where Am is the Ni surface area per gram catalyst and q is the Ni density.
r
m. * 10ꢀ18)/c * 100; where NA is the Avagadro number and rm is the Ni cross section area (0.0649).
i
100
80
60
40
20
0
Without CD
With CD
15
2
5
8
10
12
20
Nominal Ni loading (wt. % )
Fig. 6. Percentage of reduction of surface Ni species (b1 + b2) as a function of the Ni
loading for the catalysts synthesized without and with b-CD as measured by H2-TPR
Fig. 7. CO uptake as a function of the Ni loading for the catalysts synthesized
without and with b-CD.
in the temperature range 300–750 °C without considering Ni aluminate (
c type)
species.
sizes (2–5 nm) and a higher reducibility of surface-active Ni spe-
cies (b1 + b2). Notably, this trend is readily preserved over a wide
range of Ni loadings.
measured from CO pulse chemisorption and reflect fairly well the
overall beneficial impact of b-CD on the Ni dispersion observed.
To assess the effect of CD on the properties of the different cat-
alysts for conducting amination processes, the reaction of benzyl
alcohol with aniline at 140 °C was initially used as a model reaction
using 35 mg of pre-reduced xNi/Al and xNi/Al-b-CD0.1 catalysts (x =
2–20 wt% Ni) (Fig. 10 and Table S2). These conditions were set to
facilitate the comparison between the different catalysts.
Interestingly, irrespective of the Ni loading (except for 2 wt%
Ni), the xNi/Al-b-CD0.1 catalysts exhibit much higher activities
compared to xNi/Al catalysts prepared without b-CD. The highest
activity is observed for 10Ni/Al-b-CD0.1 with 62% aniline conver-
sion and 61% BZA selectivity. Nevertheless, the fact that 12Ni/Al-
3.5. Catalytic performance
3.5.1. Amination of benzyl alcohol with aniline (model reaction)
Taken together, the above characterization results for the cata-
lysts both in the dried and calcined forms demonstrate the ability
of native CDs to interact with Ni nitrate in the first steps of the
impregnation process by forming complexes, which in turn signif-
icantly affect the final structure of the catalysts. Ni complexation
helps to achieve both a better dispersion with smaller Ni particle

120
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
A
B
C
D
Fig. 8. TEM micrographs of reduced 10Ni/Al: (A) 100 nm, (B) 20 nm, (C) electron diffraction pattern and (D) normalized particle size histogram.
A
B
C
D
Fig. 9. TEM micrographs of reduced 10Ni/Al-b-CD0.1: (A) 50 nm, (B) 50 nm, (C) 20 nm and (D) normalized particle size histogram.
b-CD0.1 contains an abnormally high proportion of bulk nickel
oxide (11%) at the expense of active surface nickel species (74%)
may explain the observed deviation of the catalytic activity from
the overall trend that is emerging in this series of catalysts pre-
pared with the aid of b-CD. Note also that xNi/Al catalysts with x
= 5 wt% and 8 wt% Ni do not show any activity when Ni impregna-
tion is not assisted by b-CD. For these catalysts, the reaction tem-
perature needs to be increased to 160 °C for detecting a
conversion (Table S2, entry 3).
Fig. 11 plots the turnover number (TON, calculated by dividing
the number of reacted moles of aniline at 1 h by the number of sur-
face Ni atoms) against the mean particle size of the Ni particles
measured from pulse CO chemisorption for the different catalysts
prepared in this study. Overall, the catalysts with smaller Ni parti-

A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
121
Fig. 10. Performance of the different Ni-alumina catalysts synthesized without and with b-CD (b-CD/Ni = 0.1) in the amination reaction of benzyl alcohol with aniline.
Reaction conditions: Aniline – 2 mmol, Benzyl alcohol – 6 mmol, Cat – 35 mg, T – 140 °C, Time – 1 h, rpm – 600, no solvent, Ar atm. The number in the histograms indicates the
aniline conversion and selectivity to the different N-products.
cles exhibit higher activity. Furthermore, irrespective of the Ni
loading, the catalysts prepared in the presence of b-CD displaying
smaller Ni particles sizes afford comparatively higher activities.
These results suggest that the most active sites for amination are
located on smaller Ni particles (i.e. edges, corners or sites) at the
Ni-alumina interface. The highest TON is attained for 8Ni/Al-b-
CD0.1 exhibiting the highest Ni dispersion (i.e. lower average Ni
particle size, ꢄ5 nm) and a large fraction of reducible surface Ni
species (86%).
ing the activity between b-CD and
ing the activity between b-CD and
c
a
-CD and by 34% when compar-
-CD clearly indicates that b-CD
is the most active CD. The contrasting activity displayed by the dif-
ferent native CDs can be explained on the basis of a different inter-
action strength between the CD and the Ni precursor. This can be
attributed in turn to the different number of OH groups present
on the CD and to the stability of the as-formed adduct during the
thermal treatment (i.e. drying and calcination). As confirmed from
the ESI-MS experiments, the most stable complex tends to be
formed between b-CD and nickel nitrate (Fig. S4-B) and therefore
the decomposition pattern of the complex and the subsequent
growth of NiO particles should differ with respect to other native
We also surveyed the effect of other native CDs (a-CD and c-
CD) on the performance of xNi/Al catalysts in the model amination
reaction of benzyl alcohol with aniline at 160 °C for 1 h (Fig. S7).
The following activity trend is observed: b-CD >
no CD. The observed activity, which dropped by 12% when compar-
c
-CD >>
a
-CD >
CDs (
0.1, the OH/Ni ratio in the complex is slightly different (i.e. 2.1
for b-CD vs 1.8 and 2.4 for -CD and -CD respectively), likely
a and c-CD) during calcination. Thus, for a CD/Ni ratio of
a
c
affecting the final surface properties of the catalysts, both in terms
of surface density and reducibility. Finally, our data show that the
use of the b-CD-assisted catalyst with a CD/Ni ratio of 0.1 presents
a higher amount of reducible surface Ni species, thus leading to
most significant effects in the model amination reaction.
Without CD
with CD
g
100
80
60
40
20
0
c
d
e
3.5.2. Amination of 1-octanol with ammonia (target reaction)
f
In a subsequent step of our study, we assessed the performance
of the xNi/Al and xNi/Al-b-CD0.1 catalysts (x = 2, 5, 8, 10 and 15 wt%
Ni) in the amination reaction of 1-octanol with ammonia. The cat-
alytic tests were carried out at 160 °C and 180 °C for 4 h using 55
mg of catalyst pre-reduced at 580 °C for 30 min. Fig. 12 (160 °C),
Table S3 (entries 11–16, 180 °C) and Fig. S8 (180 °C) compile the
results obtained. The 2Ni/Al catalysts prepared with or without
b-CD do not display any activity. The catalytic activity is enhanced
with the Ni loading beyond 5 wt%, especially when b-CD is used as
additive during Ni impregnation. As a matter of fact, the 1-octanol
conversion raises by 20% at 160 °C when the Ni loading increases
from 5 wt% to 10 wt% for the Ni/Al catalysts prepared without b-
CD, whereas the OA yield reaches almost 60% (Table S3, entries 3
and 7). Meanwhile, an enhancement of the catalytic activity is
observed when b-CD is used as additive. For example, when com-
g
b
f
e
c
d
a
b
a
0
2
4
6
8
10
12
Nickel mean parꢀcle diameter (nm)
Fig. 11. Turnover number of the different Ni-alumina catalysts at 1 h in the
amination of benzyl alcohol with aniline as a function of the average Ni particle size
for the catalysts synthesized without and with b-CD at variable Ni loading.
Indexation of alphabets represents the Ni loading: (a) 2 wt%, (b) 5 wt%, (c) 8 wt%, (d)
10 wt%, (e) 12 wt%, (f) 15 wt%, and (g) 20 wt%. Reaction conditions as in Fig. 10.

122
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
paring the catalytic performance of 5Ni/Al and 5Ni/Al-b-CD0.1, the
1-octanol conversion increases from 57% to 68%, while the OA yield
is enhanced from 45% to 56%. A similar type of effect is observed
for highly Ni-loaded catalysts, the OA yield showing a maximum
value of 69% for 15Ni/Al-b-CD0.1. These results point out the posi-
tive effect of b-CD on enhancing the catalytic properties of xNi/Al
catalysts for 1-octanol amination. Indeed, the catalysts synthesized
with b-CD can save up to 50% of Ni compared to the catalysts pre-
pared without b-CD and also reported in the literature at compara-
ble 1-octanol conversion (Fig. 12) [20]. For example, at short
reaction time (4 h), the 5Ni/Al-b-CD0.1 catalyst prepared in this
study exhibits a similar TON (87) as that reported by Shimizu
and coworkers on 10Ni/h-Al2O3 (TON = 83), but obtained at longer
reaction time (13 h) (compare entries 5 and 17 in Table S3) [20].
3.5.3. Recycle study and characterization of the spent catalysts
To further assess the potentials of CDs in stabilizing Ni nanopar-
ticles against leaching, a recycling study was conducted on 5Ni/Al-
b-CD0.1 and 5Ni/Al for the direct amination reaction of 1-octanol
with ammonia in the same conditions (160 °C for 4 h). In both
cases, after the 1st run, the reaction mixture was centrifuged and
the catalyst was filtered, dried overnight at 100 °C for 10 h and
finally calcined at 400 °C for 2 h. The resulting catalyst was further
reduced at 580 °C for 30 min and the reaction was conducted in a
2nd run. Fig. 13 and Table S4 collect the results for the fresh and
spent catalysts.
It is apparent that the catalyst synthesized without b-CD (i.e.
5Ni/Al) deactivates after the 1st run as the 1-octanol conversion
dramatically drops from 57% to 5%. In contrast, a higher stability
is noticed for the catalyst synthesized with b-CD (i.e. 5Ni/Al-b-
CD0.1), which shows only a slight drop in activity (68% to 55%).
Chemical analysis of the spent 5Ni/Al reveals a Ni loss due to leach-
ing after the 1st run (from 4.4 wt% Ni to 3.9 wt% Ni), which further
increases after the 2nd run (from 3.9 wt% Ni to 3.2 wt% Ni). Oppos-
ing this behavior, no significant Ni loss is apparent for 5Ni/Al-b-
CD0.1 after the 1st run, and a slight Ni loss is observed after the
2nd run (from 4.6 wt% Ni to 4.1 wt% Ni) (Table S4, entries 3 and
4). These results point out the greater ability of b-CD to retain
Fig. 13. Recycling study of 5Ni/Al-b-CD0.1 and 5Ni/Al catalysts synthesized with
and without b-CD. Reaction conditions: 1-octanol – 1.3 mmol, NH3 – 7 bar Cat – 55 mg,
T – 160 °C, Time – 4 h, rpm – 600, Solvent – 3 mL o-xylene. The numbers in brackets
indicate the Ni loadings measured on the solid catalysts by ICP analysis after each
reaction run.
the catalyst composition and its structure during 1-octanol amina-
tion with ammonia for at least two consecutive runs. The higher
stability of CD-assisted catalysts against leaching can be explained
on the basis of a much more homogeneous distribution of smaller
Ni particles (2–5 nm) in stronger interaction with the alumina sup-
port. These trends are in line with the study reported by Claus and
co-workers pointing out a higher resistance to leaching for cata-
lysts with low Ni loadings and for small particle size [72].
The dried spent samples (synthesized with and without b-CD)
were further subjected to TGA analysis to assess any coke forma-
tion during the amination reaction. The thermal profiles of the
spent catalysts (Fig. S9, profiles b and d) show in both cases two
weight loss regions occurring at low and high temperatures. The
weight loss of 2% at about 100 °C can be attributed to physically
adsorbed water, whereas the larger weight loss (9%) in the region
Fig. 12. Performance of the different Ni-alumina catalysts synthesized without and with b-CD (b-CD/Ni = 0.1) in the amination reaction of 1-octanol with ammonia. Reaction
conditions: 1-octanol – 1.3 mmol, NH3 – 7 bar, Cat – 55 mg, T – 160 °C, Time – 4 h, rpm – 600, Solvent – 3 mL o-xylene. The number in the histograms indicates the 1-octanol
conversion and selectivity to the different N-products.

A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
123
[9] S. Bähn, S. Imm, L. Neubert, M. Zhang, H. Neumann, M. Beller, ChemCatChem 3
(2011) 1853–1864.
300–400 °C can be ascribed to the combustion of amorphous car-
bonaceous deposits [73]. No appreciable differences are observed
between the thermal profiles of the spent 5Ni/Al and 5Ni/Al-b-
CD0.1, indicating the same nature of carbon deposits. However,
since the spent catalysts are re-calcined at 400 °C before the 2nd
run, this temperature ensures complete coke removal and regener-
ation of the catalyst, as supported by the corresponding TG profiles.
[10] J.L. Klinkenberg, J.F. Hartwig, Angew. Chem. Int. Ed. 50 (2011) 86–95.
[11] C. Gunanathan, D. Milstein, Science 341 (2013) 1229712.
[12] Q. Yang, Q. Wang, Z. Yu, Chem. Soc. Rev. 44 (2015) 2305–2329.
[13] Fluzard, Bo, Huzard, Internal Rhodia Report No. 2174, 1954.
[14] Bo, Perras, Internal Rhodia Report No. 2932, 1958.
[15] G. Rice, E.J. Koh, J. Org. Chem. 77 (1955) 4052–4054.
[16] A. Mehta, A. Thaker, V. Londhe, S.R. Nandan, Appl. Catal. A: General 478 (2014)
241–251.
[17] G.A. Vedage, L.A. Emig, H.-X. Li, J.N. Armor, US Patent 5 917 092 (1998), to Air
Products and Chemicals, Inc.
[18] G.A. Vedage, K.S. Hayes, M. Leeaphon, J.N. Armor, US Patent 5 932769 (1998),
to Air Products and Chemicals, Inc.
4. Conclusions
In summary, we have described the preparation of a series of
alumina-supported Ni catalysts using native cyclodextrins as a
water-soluble additive via wet impregnation. The catalysts were
synthesized with a controlled amount of cyclodextrin to the metal
precursor and characterized in dried and calcined forms. The use of
cyclodextrin during Ni impregnation afforded enhanced catalytic
properties in the direct amination of alcohols. Among the native
CDs, the catalysts synthesized with b-CD outstand in activity. The
enhanced activity of b-CD-assisted catalysts was attributed to the
ability of generating stable complexes with the Ni nitrate precur-
sor, delaying its decomposition upon calcination. As a result, the
Ni dispersion, reducibility and speciation over alumina could be
finely controlled and tuned. In particular, a higher density of sur-
face Ni species could be achieved in the CD-assisted catalysts.
Interestingly, this effect could be preserved over a wide range of
Ni loading. Last but not least, b-CD boosted the stability of the cat-
alysts by discouraging Ni leaching during the amination reaction,
thereby maintaining the catalytic activity after at least two consec-
utive catalytic runs.
[19] K.-I. Shimizu, K. Kon, W. Onodera, H. Yamazaki, J.N. Kondo, ACS Catal. 3 (2013)
112–117.
[20] K.-I. Shimizu, K. Kon, W. Onodera, H. Yamazaki, J.N. Kondo, ACS Catal. 3 (2013)
998–1005.
[21] K.-I. Shimizu, S. Kanno, K. Kon, S.M.A.H. Siddiki, H. Tanaka, Y. Sakata, Catal.
Today 232 (2014) 134–138.
[22] K-I. Shimizu, Catal. Sci. Technol. 5 (2015) 1412–1427.
[23] A. Baiker, W. Richarz, Ind. Eng. Chem. Prod. Res. Dev. 16 (1977) 261–266.
[24] R.C. Lemon, S. Depot, R.C. Myerly, US Patent 3 022 349 (1962), to Union
Carbide Corporation.
[25] J. He, K. Yamaguchi, N. Mizuno, Chem. Lett. 39 (2010) 1182–1183.
[26] M. Dixit, M. Mishra, P.A. Joshi, D. Shah, Catal. Commun. 33 (2013) 80–83.
[27] F. Santoro, R. Psaro, N. Ravasio, F. Zaccheria, ChemCatChem 4 (2012) 1249–
1254.
[28] F. Santoro, R. Psaro, N. Ravasio, F. Zaccheria, RSC Adv. 4 (2014) 2596–2600.
[29] A. Baiker, J. Kijenski, Catal. Rev. Sci. Eng. 27 (1985) 653–697.
[30] H. Kimura, H. Taniguchi, Appl. Catal. A: General 287 (2005) 191–196.
[31] Y. Li, Q. Li, L. Zhi, M. Zhang, Catal. Lett. 141 (2011) 1635–1642.
[32] J. Sun, X. Jin, F.W. Zhang, W. Hu, J. Liu, R. Li, Catal. Commun. 24 (2012) 30–33.
[33] X. Cui, Y. Dai, Y.Q. Deng, F. Shi, Chem. Eur. J. 19 (2013) 3665–3675.
[34] Q. Li, G. Zhang, S. Peng, J. Surf. Detergents 5 (2002) 229–233.
[35] A. Jean-Marie, A. Griboval-Constant, A.Y. Khodakov, E. Monflier, F. Diehl, Chem.
Commun. 47 (2011) 10767–10769.
[36] L. Bai, F. Wyrwalski, J.-F. Lamonier, A.Y. Khodakov, E. Monflier, A. Ponchel,
Appl. Catal. B: Environ. 138–139 (2013) 381–390.
[37] L. Bai, F. Wyrwalski, M. Safariamin, R. Bleta, J.-F. Lamonier, C. Przybylski, E.
Monflier, A. Ponchel, J. Catal. 341 (2016) 191–204.
[38] J. Szejtli, Chem. Rev. 98 (1998) 1743–1753.
Acknowledgements
[39] M.Y. Rekharsky, Y. Inoue, Chem. Rev. 98 (1998) 1875–1918.
[40] S. Noël, B. Léger, A. Ponchel, K. Philippot, A. Denicourt-Nowicki, A. Roucoux, E.
Monflier, Catal. Today 235 (2014) 20–32.
[41] J.D. Senra, G.M. Viana, L.B. Malta, A.B.C. Simas, L.C.S. Aguiar, ChemCatChem 8
(2016) 192–199.
[42] S. Polarz, B. Smarsly, L. Bronstein, M. Antonietti, Angew. Chem. Int. Ed. 40
(2001) 4417–4421.
[43] L. Bai, F. Wyrwalski, C. Machut, P. Roussel, E. Monflier, A. Ponchel,
CrystEngComm 15 (2013) 2076–2083.
[44] R. Bleta, C. Machut, B. Léger, E. Monflier, A. Ponchel, Macromolecules 46 (2013)
5672–5683.
[45] S. Scirè, S. Giuffrida, C. Crisafulli, P.M. Riccobene, J. Mol. Catal. A: Chem. 353
(2012) 87–94.
[46] H. Liu, Y. Li, H. Wu, H. Takayama, T. Miyake, D. He, Catal. Commun. 28 (2012)
168–173.
[47] H. Liu, Y. Li, H. Wu, W. Yang, D. He, Fuel 136 (2014) 19–24.
[48] C. Przybylski, V. Bonnet, Rapid Commun. Mass Spectrom. 27 (2013) 75–78.
[49] G. Kresse, J. Hafner, Phys. Rev. B 48 (1993) 13115–13118.
[50] G. Kresse, J. Hafner, Phys. Rev. B 49 (1994) 14251–14269.
[51] G. Kresse, J. Furthmuller, Comp. Mater. Sci. 6 (1996) 15–50.
[52] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865–3868.
[53] K. Mathew, R. Sundararaman, K. Letchworth-Weaver, T.A. Arias, R.G. Hennig, J.
Chem. Phys. 140 (2014) 084106.
The authors would like to express their gratitude to Solvay,
CNRS and the French Agency of Research (ANR) for funding
through SHAPes program (grant No. ANR-13-CDII-0004). The
authors also acknowledge the Chevreul Institute (FR 2638), Min-
istère de L’Enseignement Supérieur et de la Recherche, Région
Nord-Pas de Calais and FEDER program for supporting and partially
funding this work. The authors thank Olivier Gardoll (UCCS, Univ.
Lille1) and Ahmed Addad (UCCS, Univ. Lille1) for their assistance
in TG-MS and TEM analyses. The authors are also grateful to Dr.
Nicolas Kania (UCCS, Univ. Artois) and Laurence Burylo (UCCS,
Univ. Lille1) for the XRD measurements, as well as to Martine
Trentesaux (Univ. Lille1) and Dr. Joelle Thuriot (REALCAT, Ecole
Centrale Lille) for their assistance in XPS and ICP analysis, respec-
tively. Finally, the authors would like to express their gratitude
to Roquette Frères (Lestrem, France) for the generous gift of b-
cyclodextrin.
[54] K. Mathew, R.G. Hennig, arXiv (2016) [arXiv:601.03346].
[55] G. Kresse, D. Joubert, Phys. Rev. B 59 (1999) 1758–1775.
[56] S. Kohata, K. Jyodoi, A. Ohyoshi, Thermochim. Acta 217 (1993) 187–198.
[57] C. Przybylski, V. Bonnet, C. Cezard, Phys. Chem. Chem. Phys. 17 (2015) 19288–
19305.
[58] Y. Seo, Y.S. Jung, W.L. Yoon, I.G. Jang, T.W. Lee, Int. J. Hydrogen Energy 36
(2011) 94–102.
[59] J. Loosdrecht, A.M. Kraan, A.J. Dillen, J.W. Geus, J. Catal. 170 (1997) 217–226.
[60] A. Zhao, W. Ying, H. Zhang, H. Ma, D. Fang, Catal. Commun. 17 (2012) 34–38.
[61] F.S.-S. Chien, Y.T. Wu, G.L. Lai, Y.H. Lai, Appl. Phys. Lett. 98 (2011) 153513–
153514.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.jcat.2017.10.006.
References
[1] H.A. Wittcoff, B.G. Reuben, J.S. Plotkin, Industrial Organic Chemicals, second
ed., Wiley, New York, 2004.
[62] C.-W. Hu, J. Yao, H.-Q. Yang, Y. Chen, A.-M. Tian, J. Catal. 166 (1997) 1–7.
[63] X.J. Zou, X.G. Wang, L. Li, Int. J. Hydrogen Energy 35 (2010) 12191–12200.
[64] J. Yang, X.G. Wang, L. Li, K. Shen, X.G. Lu, W.Z. Ding, Appl. Catal. B 96 (96)
(2010) 232–237.
[2] T.W.G. Solomons, C. Fryhle, S.C. Snyder, Organic Chemistry, 12th ed., Wiley,
New York, 2000.
[3] A. Lawrence, Amines: Synthesis, Properties and Applications, Cambridge
University Press, 2004.
[65] K.Y. Koo, H.S. Roh, Y.T. Seo, D.J. Seo, W.L. Yoon, S.B. Park, Int. J. Hydrogen
Energy 32 (2008) 2036–2043.
[4] Z. Rappoport, The Chemistry of Anilines, Wiley, New York, 2007.
[5] M. Pera-Titus, F. Shi, ChemSusChem 7 (2014) 1–4.
[66] B. Scheffer, P. Molhoek, J.A. Mounlijn, Appl. Catal. 46 (1989) 11–30.
[67] J.M. Rynkowski, T. Paryjczak, M. Lenik, Appl. Catal. A 106 (1993) 73–82.
[68] B. Vos, E. Poels, A. Bliek, J. Catal. 198 (2001) 77–88.
[6] G. Guillena, D.J. Ramón, M. Yus, Chem. Rev. 110 (2009) 1611–1641.
[7] T.D. Nixon, M.K. Whittlesey, J.M.J. Williams, Dalton Trans. 5 (2009) 753–762.
[8] G.E. Dobereiner, R.H. Crabtree, Chem. Rev. 110 (2010) 681–703.

124
A. Tomer et al. / Journal of Catalysis 356 (2017) 111–124
[69] C. Guimona, A. Auroux, E. Romeroc, A. Monzon, Appl. Catal. A 251 (2003) 199–
[71] D. Srinivas, C.V.V. Satyanarayana, H.S. Potdar, P. Ratnasamy, Appl. Catal. A 246
(2003) 323–334.
214.
[70] S.A. Krasnikov, A.B. Preobrajenski, T. Chasse, R. Szargan, Thin Solid Films 246
(2003) 201–205.
[72] S. Schimpf, C. Louis, P. Claus, Appl. Catal. A 318 (2007) 45–53.
[73] H. Guo, H. Lou, X. Zheng, Carbon 45 (2007) 1314–1321.