Highly selective direct amination of primary alcohols over a Pd/K-OMS-2 catalyst

Article abstract of DOI:10.1016/j.jcat.2013.10.003

A new Pd-substituted octahedral molecular sieve (Pd/K-OMS-2) catalyst has been prepared for the direct amination of alcohols with primary amines operating under the borrowing hydrogen mechanism. The catalyst offered full conversion and high selectivity toward N-benzylaniline in the model alkylation reaction of aniline with benzyl alcohol at mild temperature (160 C) for 3 h with neither production of the tertiary amine nor toluene. Pd/K-OMS-2 performed as a tandem tri-functional catalyst, first oxidizing benzyl alcohol to benzaldehyde, behaving as a Lewis acid for imine formation, and finally reducing the imine to the secondary amine. The catalyst was characterized in depth using BET, XRD, H₂-TPR, XPS, FTIR, TEM, TGA/DTG, and ICP-AES / EDX to elucidate the nature of the active sites. The unexpectedly high performance of the Pd/K-OMS-2 catalyst can be ascribed, at least partially, to the in situ generation of a very active, selective and partially recyclable Pd-substituted/supported hausmannite phase (i.e., Pd/Mn₃O₄) in the early stage of the reaction with a high density of surface oxygen moieties. We argue about a possible role of a Pd(IV)/Pd(II) redox pump for exchanging hydrogen during the amination reaction.

Full text of DOI:10.1016/j.jcat.2013.10.003

Journal of Catalysis 309 (2014) 439–452
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Highly selective direct amination of primary alcohols over a Pd/K-OMS-2
catalyst
⇑
M. Ousmane, G. Perrussel, Z. Yan, J.-M. Clacens, F. De Campo, M. Pera-Titus
Eco-Efficient Products and Processes Laboratory (E2P2L), UMI 3464 CNRS – Solvay, 3066 Jin Du Road, Xin Zhuang Ind. Zone, 201108 Shanghai, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A new Pd-substituted octahedral molecular sieve (Pd/K-OMS-2) catalyst has been prepared for the direct
amination of alcohols with primary amines operating under the borrowing hydrogen mechanism. The
catalyst offered full conversion and high selectivity toward N-benzylaniline in the model alkylation reac-
tion of aniline with benzyl alcohol at mild temperature (160 °C) for 3 h with neither production of the
tertiary amine nor toluene. Pd/K-OMS-2 performed as a tandem tri-functional catalyst, first oxidizing
benzyl alcohol to benzaldehyde, behaving as a Lewis acid for imine formation, and finally reducing the
imine to the secondary amine. The catalyst was characterized in depth using BET, XRD, H -TPR, XPS, FTIR,
2
TEM, TGA/DTG, and ICP-AES / EDX to elucidate the nature of the active sites. The unexpectedly high per-
formance of the Pd/K-OMS-2 catalyst can be ascribed, at least partially, to the in situ generation of a very
active, selective and partially recyclable Pd-substituted/supported hausmannite phase (i.e., Pd/Mn O ) in
3 4
Received 21 July 2013
Revised 16 September 2013
Accepted 7 October 2013
Keywords:
Direct amination
Borrowing hydrogen
Heterogenous catalyst
OMS-2
Doping
the early stage of the reaction with a high density of surface oxygen moieties. We argue about a possible
role of a Pd(IV)/Pd(II) redox pump for exchanging hydrogen during the amination reaction.
Ó 2013 Elsevier Inc. All rights reserved.
1
. Introduction
Amines are key chemical intermediates widely used as building
reaction proceeds via an intermediate imine or enamine affording
only water as by-product [3]. Nonetheless, unlike reductive amina-
tion, no additional reductive species (usually H ) is required for
2
blocks for the synthesis of polymers, dyes, pharmaceuticals, and
agrochemicals [1]. The industrial synthesis of amines usually in-
volves hazardous reagents (e.g., HCN) or generates hazardous by-
products (e.g., HCl). Moreover, the preparation of primary amines
by direct nucleophilic substitution of organic halides suffers most
often from a lack of selectivity, the reaction usually rendering a
mixture of primary, secondary, and tertiary amines. The use of acid
catalysts to activate the alcohol group can also be responsible for
the generation of alkenes through uni/bimolecular elimination
mechanisms [2]. An alternative process based on the addition of
ammonia or amines on alkenes has also been proposed for the syn-
thesis of tertiary amines over acid catalysts (e.g., t-butylamine
from isobutene), but with limited selectivities [3]. Finally, amines
can also be produced by the reduction of amides, but generally un-
der harsh conditions [4].
imine or enamine hydrogenation. The borrowing hydrogen mech-
anism overcomes the lack of reactivity of reductive amination cat-
alysts (most often based on Ni and NiCuMg formulations) by the
2
temporary removal of ‘‘H ’’ from the alcohol by the catalyst to give
a carbonyl, which is more reactive toward the amine or ammonia.
The catalyst returns the ‘‘borrowed H
the alkylated amine product.
2
’’ by reducing the imine into
The first studies on heterogeneous catalytic formulations for
alcohol amination via the borrowing hydrogen mechanism focused
on Ni-Raney [8,9] and Ni and Co nanoparticles supported on silica
and alumina [10–15]. However, as in reductive amination, the
2
reaction was carried out under H atmosphere to maintain the
activity of the catalyst. The introduction of small amounts of noble
metals on Ni and Co catalysts (typically 15–20% Ni or Co + 0.5–3%
Pd on alumina, silica, or titania) was shown to enhance the
catalytic activity and reduce the activation temperature [16].
Unsupported Co–Fe catalysts prepared by co-precipitation
[17,18], and Cu/Ni/Ca/Ba colloidal mixed-oxide catalysts prepared
from stearate precursors exhibited selectivity for the synthesis of
primary diamines from diols in (supercritical) ammonia [19]. Fur-
thermore, Baiker and co-workers [20] prepared Co–Fe catalysts
Recently, transition metal complex-catalyzed direct N-alkyl-
ation of amines and ammonia with alcohols has been reported as
an alternative route to amines through the so-called borrowing
hydrogen or hydrogen auto-transfer mechanism [5–7], encom-
2
passing a net transfer of H from the alcohol to the intermediate
imine. As in reductive amination of carbonyls, the amination
based on Fe/Co
of 1,3-propanediol with ammonia in the pressure range
0–150 bar. The absence of strong acidic and basic sites was
3 4
O with b-Co metal cluster sites for the amination
⇑
Corresponding author.
E-mail address: marc.pera-titus-ext@solvay.com (M. Pera-Titus).
5
0
021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.10.003

4
40
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
claimed as crucial for suppressing acid/base-catalyzed side reac-
tions (e.g., retro-aldol reaction, alkylation, disproportionation,
dimerization, and oligomerization). A small amount of Fe additive
could efficiently hinder the transformation of b-Co into a-Co phase
and prevent catalyst deactivation.
[53] has also been accomplished with a variable extent of substitu-
tion. Finally, in addition to cation substitution, oxide/metal nano-
particles grown over OMS-2 (e.g., CuO [54], Pt [55,56], Au
[57,58]) were also achieved via incipient wetness impregnation
using a metal salt precursor at short contact times.
More recent examples include Ni, Cu, and NiCu supported on
silica and alumina [21,22] and iron oxide [23] for amination reac-
tions of aromatic amines and alcohols. In the case of aliphatic sys-
tems, high yields have been recently reported on CuAl/hydrotalcite
In addition to total oxidation of VOCs, doped OMS catalysts
have been applied with success to partial oxidation [42,58–61],
epoxidation [62], and hydrogenation [63,64] reactions. To our
knowledge, the sole study on the application of the raw K-OMS-2
to the N-alkylation of aromatic and aliphatic amines was reported
by Suib and co-workers [65], showing the preferential generation
of imines, but with no amine formation. The reaction was carried
out in the presence of an apolar solvent (toluene or o-xylene) at
moderate temperature (110–140 °C) under air for in situ regenera-
tion of the catalyst. An excess of oxygen promoted the generation
of primary amides in the presence of ammonia by oxidation of the
imine to an intermediate nitrile species that could be further hy-
drated to the final amide [66]. In a step further, here we show that
a Pd-substituted K-OMS-2 formulation can be competitive for
amine synthesis from an imine intermediate via a hydrogen bor-
rowing mechanism under inert atmosphere.
(
1
+K
60 °C for 9–12 h under air or N
showing yields >70% toward the secondary amine. In the case of
supported noble metals, Au/MgO [26], Au/TiO [27], and Au/ZrO
28,29] catalysts showed a moderate capacity for imine formation
2
CO
3
additive) [24] and CuNi/
2 3
c-Al O [25] formulations at
2
atmosphere, the latter catalysts
2
2
[
by promoting alcohol deprotonation in the first step of the reac-
tion, but low capacity for amine formation. The combined dehydro-
genation capacity of metal nanoparticles (mean size < 3 nm) with
the aniline and benzyl alcohol adsorption capacity of the support
via Br/nsted basic and Lewis acid sites, respectively, was argued
as crucial for promoting secondary amine formation [28]. Partially
active and selective formulations based on Ag were reported by Shi
et al. [30] over Ag–MoO
31] over alumina-entrapped Ag nanoparticles doped with Cs
Heterogeneous bimetallic Pt–Sn/ -Al catalysts (Pt with SnO
x
oxides and by Jaenicke and co-workers
[
2
3
CO .
c
O
2 3
2
)
2. Experimental
showed activity in the direct synthesis of secondary and tertiary
diamines from the reaction of aliphatic diols with aniline [32].
Ru- and Pd-based catalysts have shown the most promising
potentials for the amination of alcohols. On the one hand, liquid-
2.1. Materials
Manganese sulfate hydrate (MnSO ꢁH O, 99.5%) and potassium
4
2
phase amination of cyclohexanol over a Ru/Al
batch reactor was reported to generate cyclohexylamine with high
selectivity [33,34], whereas Ru(OH) /Al [35] and Ru(OH) /TiO
36] catalysts demonstrated the N-alkylation of alcohols via the
2
O
3
catalyst in a
permanganate (KMnO , 99.5%), both supplied by Sigma–Aldrich,
4
were used as reactants for the synthesis of the OMS-2 phases
x
2
O
3
x
2
and Mn O and MnO2 supports. Aerosil 200 silica (99%, Evonik),
2
3
[
2 3 2
c-Al O (99.9%), TiO (99.9%) and CuO (99.9%), the latter three oxi-
preferential alcohol adsorption on weakly basic/acid Al-OH and
Ti-OH sites. On the other hand, Shi and co-workers prepared a
Pd/Fe O catalyst by co-precipitation and further calcination,
2 3
des supplied by Sinopharm, were used as supports. Aluminum,
iron, and copper nitrate salts (99.5% in all cases), all supplied by
Sinopharm, were used as precursors for catalyst preparation by
co-precipitation. Tetraamminepalladium (II) nitrate (Pd(NH3)4
(NO ) , 99.99%), provided by Sigma–Aldrich, was used for Pd
impregnation/substitution over the different supports or co-
precipitation with other metal nitrate precursors. The Pd/C catalyst
(5 wt.% Pd) was provided by Sigma–Aldrich. Aniline (99.5%), benzyl
alcohol (99.5%), N-phenyl benzylimine (99.5%), N-phenyl benzyl-
amine (99.5%), N,N-dibenzylaniline (99.5%), and biphenyl (>99%),
all purchased at J&K, were used in the catalytic tests and as
standards for GC calibration.
achieving full conversion and selectivity to the secondary amine
at 160 °C in the reaction of aniline with benzyl alcohol [37]. The
authors claimed that the active species consisted of a mixture of
Pd(IV) and Pd(II) moieties activated by the support. Corma and
co-workers [38] reported promising yields to N-phenyl benzyl
amine in the reaction of benzyl alcohol with aniline at 180 °C at
short times (<2 h) using Pd(0.8–5 wt.%)/MgO basic catalysts. The
initial reaction rate was found to increase inversely with the Pd
nanoparticle size (mean size < 5 nm) with a partial role of the sup-
port in stabilizing hydride species.
3
2
In this study, we concentrate our attention on metal-doped
cryptomelane-type manganese oxide Octahedral Molecular Sieves
2.2. Catalyst synthesis
(
OMS-2) as potential candidates for amination reactions. OMS-2
materials (also termed K-hollandites or -MnO ) are based on
edge-shared MnO octahedra hosting both Mn(III) and Mn(IV),
c
2
K-OMS-2 manganese octahedral molecular sieve was prepared
6
by the reflux method [67]. Briefly, an aqueous solution of KMnO
(40 g in 680 mL deionized water) was added to a solution of
MnSO O (54.446 g in 180 mL deionized water) and concen-
trated nitric acid (20 mL) in a 1-L round-bottom flask equipped
with a condenser. The final mixture was refluxed at 110 °C for
24 h. Finally, the dark brown solid was washed by filtration (/
4
and 2 ꢀ 2 1D microtunnels incorporating ex-framework compen-
+
sation cations (usually K ) [39]. The direct doping of OMS-2 by so-
4
ꢁH
2
lid-state conversion or aqueous-phase acid synthesis can only be
achieved in seldom cases, since other more stable manganese
+
phases are usually favored [40]. The direct exchange of K by other
alkaline cations is often discouraged due to the stability of the for-
mer in OMS-2, requiring most often a preliminary acid treatment
at long times [41,42]. Higher exchange capacities can be achieved
by the synthesis of Na–(Mg)–birnessite (OL-1) precursors with the
desired metal nitrate salts (e.g., alkaline cations [43], Ag [44,45], Co
= 0.02 lm) with deionized water (5 L) until neutral pH and dried
overnight at 120 °C. An acid OMS-2 support (labeled here as K,H-
OMS-2) was prepared following the protocol above stated, but
washed until pH 4.
The amorphous MnO
tion of KMnO and MnSO
preparation, KMnO (11.6 g in 200 mL deionized water) was added
to a solution of MnSO O (18.6 g in 60 mL deionized water) and
2
phase was prepared via the redox reac-
[
44], Cu [44], Pb [46], and Cr [47]) followed by solid-state transfor-
mation to the OMS-2 phase by hydrothermal synthesis at
20–200 °C from a few hours to a few days [43]. Framework sub-
4
4
ꢁH O at room temperature. In a typical
2
4
1
4
ꢁH
2
stitution of octahedral Mn by low- and high-valent metal cations
such as Ni(II) [48,49], Co(II) [48,49], Cu(II) [48–50], Zn(II) [48],
Cr(III) [47,49], Fe(III) [40,49], Ti(IV) [51], V(V) [52], and W(VI)
stirred under vigorous stirring for 6 h. Finally, the dark brown solid
was washed by washed and filtrated several times until neutral pH
(10 L of deionized water) and dried overnight at 120 °C. The

M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
441
crystalline Mn
heating at 700 °C for 2 h under air atmosphere.
2
O
3
phase obtained from the amorphous solid by
Thermogravimetric analysis (TGA) was employed to study the
thermal behavior of the samples. The experiments were performed
on a TA SDT Q600 Instrument. The samples were heated from room
Ru, Pd, Au, and Ag incorporation over the calcined K-OMS-2,
K,H-OMS-2, MnO2, and Mn
2
O
3
samples (2.0 wt.% theoretical load-
2
temperature to 1400 °C at a rate of 5 °C/min both under N and air
ing) was carried out by ion exchange using a diluted aqueous solu-
tion (10 M) of the corresponding metal precursor,
M(NH ) (NO ) , at 80 °C for 2 h under N atmosphere. Standard
3 4 3 x 2
atmosphere (100 mL(STP)/min). The heat flow data were dynami-
cally normalized using the instantaneous weight of the sample at
the respective temperature.
ꢂ4
Pd-supported samples were prepared by the incipient wetness
impregnation (IWI) and co-precipitation (CP) methods over 6 com-
mercial supports. On the one hand, in the former method, 48 mL of
2
H -TPR profiles were collected on a Micromeritics AutoChem
II2920 system equipped with a quartz U-type tubular reactor and
a TCD detector. A cold trap was used before the detector to avoid
any interference of water in the TPR plot. In each test, 50 mg of
the sample was loaded into the reactor and purged with 30 mL/
min He at 300 °C for 1 h to remove adsorbed moisture and vapors
and then cooled down to room temperature. The temperature was
then increased up to 750 °C at a heating rate of 10 °C/min under a
Pd(NH
sponding support and heated at 80 °C for 2 h under N
On the other hand, the co-precipitated samples were prepared by
mixing Pd(NH (NO and the corresponding metal precursor,
M(NO , the pH of the final solution was adjusted at 11.0. The mix-
ture was then heated at 80 °C for 2.0 h under N atmosphere. After
3
)
4
3
(NO )
2
(1.4 g/L) was added dropwise over 3 g of the corre-
2
atmosphere.
3
)
4
3 2
)
3 x
)
2
H
Ag
2
flow diluted in Ar (10 v/v%). The system was calibrated using an
O standard (>99% purity) to measure the amount of H con-
sumed in the TPR tests.
the synthesis, both the Pd-impregnated and co-precipitated sam-
ples were subjected to vacuum evaporation (150 mbar) at 70 °C,
dried overnight at 100 °C, and finally calcined at 300 °C for 2 h un-
der air. In all cases, the theoretical Pd loading of the samples was
about 2.3 wt.%.
2
2
Infrared spectra of the materials were recorded on calcined
powders dispersed in KBr (2 mg sample in 300 mg KBr) using a
ꢂ1
Perkin-Elmer One FTIR spectrometer with a resolution of 4 cm
ꢂ1
operating in the range 500–2000 cm with 4 scans per spectrum.
2.3. Catalyst characterization
2.4. Catalytic tests
The phases present in the different samples were analyzed by
powder X-ray diffraction (PXRD). PXRD analyses were carried out
using a Rigaku D/Max-2200/PC Diffractometer provided with Cu
Ka radiation (k = 1.5418 Å) and a beam voltage of 45 kV. The pat-
terns thus obtained were indexed using the Joint Committee on
Powder Diffraction (JCPDS) database.
The morphology and structure of the Pd-supported manganese
samples were inspected by transmission electron microscopy
The catalytic activity of the different Pd-impregnated catalysts
was investigated in the model amination reaction of benzyl alcohol
with aniline (Fig. 1). The reaction was performed using 2 mmol of
aniline (A) and 6 mmol of benzyl alcohol (BA) and 60 mg of catalyst
2
in a Radley’s multicell reactor under N or air atmosphere at 160 °C
for 3 h and stirring speed of 600 rpm. The reactants (aniline and
benzyl alcohol) and expected products, i.e. (a) N-phenylbenzyli-
mine (PBI), (b) N-phenylbenzylamine (PBA), and (c) N,N-
dibenzylaniline (DBA), were analyzed and quantified using an
Agilent 7890 GC equipped with a HP-5 capillary column with
(
HRTEM) on a JEOL JEM-2100F (200 kV) microscope equipped with
energy dispersive X-ray (EDX) analysis (semiquantitative) using a
Philips PV 9800 spectrometer operated under the SuperQuant
software.
In addition to EDX analysis, the bulk Pd composition in the me-
tal-supported/exchanged samples was analyzed by inductively
coupled plasma (ICP-AES) on a Perkin-Elmer optical emission spec-
trometer (Optima 8000) equipped with an auto-sampler. Further-
more, the surface Mn, K, and Pd composition of the samples was
analyzed by X-ray photoelectron analysis (XPS) on an Axis Ultra
5
wt.% phenyl groups using biphenyl as internal standard. Mass
balances were accurate to within 5% in all the catalytic tests.
3. Results
3.1. Catalyst screening
DLD spectrometer (Kratos Analytical) provided with Al K
tion (h = 1486.6 eV) and operating at 150 W (10 mA to 15 kV).
Carbon (C1s = 284.5 eV) was used as internal standard.
The specific surface area and pore volume of the different sam-
ples were measured from N adsorption–desorption isotherms at
7.4 K using a Micromeritics ASAP 2010 surface area analyzer.
The surface areas were calculated by the Brunauer–Emmett–Teller
BET) method in the relative pressure range 0.05 < P/P < 0.25,
while the pore volumes were measured at P/P = 0.99. The
Barret–Joyner–Halenda (BJH) method was used for measuring
intercrystalline mean pore sizes. The t-plot method was used to
measure the external surface of the catalysts and micropore vol-
ume. Prior to the measurements, the silica samples were degassed
at 200 °C under vacuum (0.5 mbar) during 15 h.
a
radia-
A first series of metal-doped OMS-2 (2.3 wt.% metal loading)
catalysts was prepared to assess for the effect of noble metals on
the catalytic performance. Fig. 2 collects the results obtained in
m
terms of aniline conversion and yields to PBI and PBA under N
2
2
atmosphere. Fig. 2 also includes a blank test, indicating a very
small yet detectable conversion of aniline into PBI (<3%) in the ab-
sence of catalyst. In line with the results reported by Suib and co-
workers [65], despite the notable activity endowed by the raw K-
OMS-2 (ca. 47% aniline conversion), only the imine product (i.e.,
PBI) was formed, suggesting an inefficient hydrogen transfer from
the catalyst to the imine. A similar type of effect was observed on
Pt/ and Au/K-OMS-2 with no amine formation. A different picture
was observed for the Ru-, Ag-, and Pd-doped K-OMS-2 and
Pd/K,H-OMS-2 samples, showing a partial PBA formation and
7
(
0
0
a
b
c
Fig. 1. Model amination reaction of aniline and benzyl alcohol for catalyst screening (a = PBI, b = PBA, c = DBA).

4
42
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
According to this body of results, we decided to concentrate our
further attention on ion exchange and IWI as strategies for the de-
sign of Pd-modified/supported amination catalysts.
Table 1 compiles the main physical and textural characteristics
of the most active and selective Pd-exchanged manganese oxides
and Pd-impregnated formulations found in this study in the alkyl-
ation reaction of aniline with benzyl alcohol (PBA yield > 30%). For
the sake of comparison, Table 1 also includes the results reported
by Shi and co-workers [37] and Corma and co-workers [38],
2 3
respectively, over Pd/Fe O and Pd/MgO catalysts taken as bench-
mark systems. Regardless of the acid–base and redox nature of
the supports, the different catalytic formulations offered partial
selectivity to the secondary amine (PBA) at reasonable aniline con-
versions. The tertiary amine (DBA) and other by-products (e.g.,
benzene and toluene) were also detected, but at very low yields.
In all cases, the different formulations offered a competitive perfor-
Fig. 2. Summary of aniline conversions and PBI and PBA yields obtained for the
reaction of BA with A over modified X/Y-OMS-2 catalysts, where X = Pd, Pt, Au, Ru,
Ag or – and Y = K and/or H. For comparison, a blank experiment is also included.
Conditions: 1:3 A/BA molar mixture, 160 °C, 3-h reaction, N atmosphere, 60 mg
2
catalyst (2 wt.% X loading).
2 3
mance compared to the benchmark Pd/Fe O and Pd/MgO formu-
lations at similar Pd loading (ca. 2 wt.%).
As a general trend, Pd-impregnated samples based on supports
with a moderate basic or amphoteric behavior (i.e., K-OMS-2,
2 3 2 3 2 3
Mn O , Al O , Fe O ) offered higher PBA yields, but with no
straightforward correlation with neither their redox nature nor
specific surface (Table 1). Among the different samples, Pd/K-
OMS-2 offered the best aniline conversion and PBA yields. Encour-
aged by these results, we decided to conduct a complete study on
this catalyst not only to assess its performance, but also to rational-
ize the nature of the active sites with special insight into the subtle
interplay between Pd moieties and basic sites. These aspects are
tackled in detail below.
comparatively higher aniline conversions. Among these formula-
tions, Pd/K-OMS-2 offered the best catalytic performance, reaching
9
9% aniline conversion and a PBA yield as high as 96% with the sole
formation of PBI (3%) as by-product (neither DBA nor benzene/tol-
uene was detected after reaction). The catalytic performance was,
however, affected by the acid–base properties of OMS-2. Indeed,
partial acidification of the sample decreased the PBA yield down
to 70% while keeping the aniline conversion essentially
unmodified.
Given the promising performance of the Pd/K-OMS-2 formula-
tion, we decided to carry out a comparative study on other Pd-
exchanged manganese oxides (i.e., amorphous MnO and Mn O )
2 2 3
3
3
.2. Catalytic performance of Pd/K-OMS-2
.2.1. Reaction kinetics
and on standard Pd-supported catalysts prepared both by the IWI
and CP methods. In most cases, Pd incorporation by impregnation
afforded the genesis of samples with improved aniline conversions
and PBA yields than the co-precipitated counterparts (ESI, Fig. S1),
most likely due to a better dispersion of the Pd phase on the
support after calcination. Only two exceptions were found for the
Additional tests were carried out to investigate the effect of the
reaction time on the reactivity of the Pd/K-OMS-2 catalyst. The
tests were carried out at standard reaction conditions (vide supra)
for 1, 3, and 6 h (Fig. 3). On the one hand, the aniline conversion
increased from 86% after 1 h to reach 100% after 3 h and beyond.
On the other hand, the PBI yield decreases from a value of 12% after
2 3
co-precipitated Pd/Fe O and Pd/CuO systems, the former offering
1
h reaction down to 3% after 3 h and vanished after 6 h, whereas
full aniline conversion and a PBA yield as high as 96% in good
agreement with the results reported by Shi and co-workers [37].
The former sample displayed well-dispersed Pd nanoparticles
around 10 nm with some larger particles about 50 nm (not shown).
the PBA yield increased from 73% to 93% when the reaction time
evolved from 1 to 3 h and became stable at 99% after 6 h. These
trends suggest the possibility of full PBI transformation into PBA
by simply adjusting the reaction time when operating at 160 °C.
Table 1
Textural and physical characteristics properties of the different Pd-supported catalysts (2.3 wt.%) prepared in this study and performance in the alkylation reaction of aniline with
benzyl alcohol. Reaction conditions: 1 A : 3 BA, 60 mg catalyst, 160 °C, 3 h, 600 rpm, 1 bar N .
2
2
Samples
BET surface area (m /g)
Acid–base nature
Pd doping
Max. Pd size (nm)
Aniline conversion (%)
Yield (%)
PBI
PBA
DBA
Others
Pd/K-OMS-2
Pd/H-K-OMS-2
96
84
94
11
200
150
200
60
Acid/basic (+)
Acid (+)
Acid (+)
Basic (+)
Acid (+)
Acid (+)
Amphoteric
Basic (+)
Neutral
Ion exchange
NP
NP
NP
NP
<5
–
99
56
64
76
42
73
87
100
60
3
23
25
26
13
41
26
24
24
96
33
39
50
27
34
61
74
35
–
–
–
–
1
–
–
2
–
–
–
–
–
1
–
–
–
1
Pd/MnO
2
(amorph.)
Pd/Mn
Pd/SiO
2
O
3
IWIb
2
Pd/TiO
Pd/Al
2
2
O
3
–
Pd/Fe
2
O
3
10–50
5
Pd/Ca
>1000
Pd/Fe
Pd/MgO [38]
2
O
3
c[37]
60
670
Basic (+)
Basic (++)
Co-precip.
IWI
3
<5
99
96
3
12
96
80
0
0
0
0
d
à
a
b
c
5
wt.% Pd.
Incipient wetness impregnation.
Prepared by co-precipitation at RT.
d
Prepared by IWI (0.8 wt.% Pd); reactions conditions: benzyl alcohol (1 mmol), aniline (3 mmol), n-dodecane (0.1 mmol), Pd (0.0075 mmol), trifluorotoluene (1 mL),
T = 180 °C, time = 6 h.

M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
443
3.2.3. Influence of the atmosphere
The nature of the atmosphere in direct contact with the reaction
mixture was found to affect the selectivity of our Pd/K-OMS-2 cat-
alyst (Fig. 5). As a matter of fact, when air was used instead of N at
the very same reaction conditions, the PBA selectivity decreased to
6% from a value of 93%. In both cases, almost full aniline conver-
2
5
sion was reached. The severe impact of the atmosphere on the cat-
alyst selectivity can be explained by the relative competition
between molecular oxygen and PBI as electron acceptors, mitigat-
ing the capacity of the catalyst to borrow stored ‘‘H
sence of strictly inert conditions.
2
’’ in the ab-
3.2.4. Activity and selectivity in successive runs
The reusability of the Pd/K-OMS-2 catalyst was tested in three
consecutive runs in the alkylation reaction of aniline with benzyl
alcohol. After each run, the catalyst was recovered, washed with
a 50:50 water acetone solution, and dried at 120 °C before use
for the next run. After the first run, the catalyst kept a selectivity
>90% to PBA with an aniline conversion about 70%. We attribute
this hampered activity after the first run to a phase conversion of
the original Pd/K-OMS-2 catalyst. For the sake of clarity, this new
phase will be termed as Pd/MMO (MMO for ‘‘modified manganese
oxide’’) in the remainder of this study.
Fig. 3. Influence of the reaction time on the catalytic activity and selectivity for the
Pd/K-OMS-2 catalyst in the alkylation reaction of aniline with benzyl alcohol under
inert atmosphere. Reaction conditions: 1 A:3 BA, 60 mg catalyst, 160 °C, 600 rpm,
1
2
bar N atmosphere.
At lower temperatures (i.e., 120 and 140 °C, results not shown), a
maximum PBI yield about 85% could only be reached after 3 h reac-
tion time. In all cases, irrespective of the reaction time and temper-
ature, neither the tertiary amine (DBA) nor other reduced by-
products such as toluene were detected even after 6 h of reaction.
3
3
.3. Structure–activity relationship in Pd/K-OMS-2
.3.1. PXRD patterns
Fig. 6 plots the XRD patterns of the regular K-OMS-2, the Pd/K-
OMS-2, and the Pd/MMO phase after reaction. For both the K-OMS-
and Pd/K-OMS-2 samples, the XRD patterns showed characteris-
tic (110), (200), (220), (310), (211), (301), (411), (600), (521),
002), (541), (312), (402), and (332) reflections belonging,
respectively, to 2-theta angles 12.7°, 18.0°, 22.0°, 28.7°, 37.4°,
2.0°, 49.8°, 56.1°, 60°, 65.3°, 69.3°, 73.1°, 77.5°, and 78.7° which
can be attributed to the synthetic cryptomelane (KMn 16, JCPDS
No: 42-1348, I4/m tetragonal unit cell) with no indication of other
phase impurities attributed to manganese oxides. Furthermore, in
the case of the Pd/K-OMS-2, no reflections indicative of palladium
oxide formation were observed (the more intense peak of PdO
should appear at about 2-theta = 33.5°), suggesting a high disper-
sion of Pd on K-OMS-2 or substitution into the framework. The
average crystallite size calculated by applying Scherrer equation
2
3
.2.2. Influence of the catalyst weight
To survey the effect of the catalyst weight on the performance
(
of the Pd/K-OMS-2 catalyst, a series of devoted catalytic tests were
carried out at standard reaction conditions, with the catalyst
weight varied in the range 7.5–90 mg (Fig. 4). Within the limits
of the experimental error, the aniline conversion and PBA yield
were kept at high values and essentially unchanged with the cata-
lyst weight in the range 30–60 mg. At lower weights (i.e., 7.5 and
4
8
O
1
5 mg), the conversion declined with a relative increase in the
PBI yield, whereas at 90 mg, the aniline conversion increased at
the expense of a decline in the PBA selectivity.
Fig. 4. Influence of the catalyst weight effect on the activity and selectivity of the
Pd/K-OMS-2 catalyst in the alkylation reaction of aniline with benzyl alcohol under
N atmosphere. Other reaction conditions as in Fig. 3.
2
Fig. 5. Influence of the reaction atmosphere on the catalytic activity and selectivity
of the Pd/K-OMS-2 catalyst in the alkylation reaction of aniline with benzyl alcohol.
Other reaction conditions are shown in Fig. 3.

4
44
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
A deeper inspection of the microstructure of the Pd/K-OMS-2
sample using HRTEM (Fig. 7B) showed the presence of lattice
fringes separated by a distance of 0.48 nm. This distance matches
the (200) d-spacing of the K-OMS-2 phase (reflection appearing
at 2-theta = 18.0, Fig. 6) [68], confirming the formation of a crypto-
melane phase with no visible structural modification after Pd
incorporation. Although ICP-AES and EDX analyses confirmed the
presence of Pd in the sample (ca. 2. wt.%), the TEM and HRTEM
micrographs did not evidence supported Pd nanoparticles on K-
OMS-2 after ion exchange. This observation, matching the general
picture obtained on Pd-exchanged amorphous MnO
2 2 3
and Mn O
(ESI, Fig. S3), suggests Pd incorporation either in the cryptomelane
channels by preferential K-substitution, or in the framework via
exchange of octahedrally coordinated Mn ions. Both scenarios are
compatible with the very similar XRD patterns obtained for the
Pd-doped and parent K-OMS-2 samples (Fig. 6), revealing an insig-
nificant change in both the symmetry and unit cell parameters of
the cryptomelane phase after Pd incorporation.
Fig. 6. XRD patterns of (a) K-OMS-2, (b) Pd/K-OMS-2, and (c) Pd/MMO after the
amination reaction.
3.3.3. Surface area and porosity
to the strongest reflection (2-theta = 37.4°) was 23 nm for both K-
OMS-2 and Pd/K-OMS-2.
After one reaction run, the original Pd/K-OMS-2 phase was con-
verted into a pure hausmannite phase showing a spinel structure
2
Fig. S3 (see ESI) plots the N adsorption/isotherms at 77.4 K for
the fresh Pd/K-OMS-2 catalyst and the same sample after reaction
(i.e., Pd/MMO). Both samples showed a characteristic Type II sorp-
tion pattern matching the general behavior observed for the raw K-
OMS-2 (not plotted in Fig. S3 for the sake of clarity) indicative of
multilayer adsorption. Furthermore, the isotherms showed the
presence of a distinct H3-type hysteresis loop between the adsorp-
(
hausmannite = Mn
reflections at 2-theta = 18, 31.4, 32.5, 36.1, 38.3, 44.5, 51, 54.1,
6.1, and 64.7. An additional peak at about 19° was observed cor-
responding probably to the MnO phase. No peaks corresponding
to metallic Pd, PdO or PdO or K O/KOH were observed, suggesting
3 4
O , JCPDS No: 24-0734) with characteristic
5
2
0
tion and desorption branches for P/P > 0.6 that can be attributed
2
2
to intercrystalline mesopores between nearby particles with non-
uniform size or shapes. The hysteresis loop closed at a higher pres-
sure for the Pd/MMO sample, reflecting a slight increase of
intercrystalline pore size during the amination reaction.
a good dispersion or partial incorporation of Pd and K onto/into the
hausmannite phase. The estimated crystallite size using Scherrer
equation applied to the peak centered at 2-theta = 37.4 was about
2
3 nm, matching the sizes estimated on the parent K-OMS-2 and
The BET-specific surface and total pore volume of the K-OMS-2,
Pd/K-OMS-2, and Pd/MMO samples showed in all cases similar val-
the fresh Pd/K-OMS-2 samples, suggesting no modification of the
crystallite size after the amination reaction.
2
3
ues in the range 80–100 m /g and 0.4–0.5 cm /g, respectively (Ta-
ble 2). The BJH pore size estimated from the adsorption curve was
about 12 nm for the three samples. The application of the t-plot
method reflected a major contribution of the external surface area
at a level higher than 60% for the former two samples, whereas the
BET-specific surface of the Pd/MMO sample corresponded com-
pletely to the external surface. Indeed, the measured BET surfaces
are in good agreement with the theoretical values that can be de-
3.3.2. TEM/HRTEM
The morphology of the fresh Pd/K-OMS-2 and used Pd/MMO
catalysts was inspected by TEM and HRTEM (Figs. 7 and 8). Both
samples kept a needle-like morphology of the raw K-OMS-2 (ESI,
Fig. S2), showing fibers with 50–250 nm length and 5–20 nm width
matching the dimensions obtained on the parent K-OMS-2 sample.
Transposing these dimensions into an equivalent spherical size, the
maximum crystallite size can be estimated at about 20 nm for both
K-OMS-2 and Pd/K-OMS-2, matching the values obtained by direct
application of Scherrer equation to the (211) reflection (23 nm).
2
duced from the crystallite size, about 60 m /g assuming a spherical
3
shape and taking a density of 4.3 g/cm for all the samples [69]. The
estimated micropore volume for the former two samples was
3
about 0.43 cm /g, whereas the last sample displayed a negligible
value as expected for a non-porous solid.
Fig. 7. TEM micrographs of the fresh Pd/K-OMS-2 sample: (a) low resolution, and (b) high-resolution showing (200) lattice fringes.

M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
445
Table 2
Textural properties, phase, and bulk chemical composition of Pd-modified catalysts.
2
Crystallite size (nm)a
Catalysts
BET surface area External surface area (m /g) Pore volume Phases (XRD)
2 2
H consumption (TPR-H ) (mmol/g)
2
3
(m /g)
(cm /g)
K-OMS-2
96
82
79
93
0.52
0.35
0.43
Cryptomelane
-MnO
Cryptomelane
-MnO
Hausmannite (Mn
24
24
23
11.8
12.2
7.8
(
c
2
)
Pd/KOMS-2 93
Pd-MMO 93
(c
2
)
3 4
O )
a
Computed using Scherrer equation from the more intense reflection in the XRD pattern (2-theta = 37.4°).
The results stated above suggest a major contribution of N
2
adsorption at the external surface of the particles, but with negli-
gible adsorption in the OMS-2 channels in the former two samples.
Accordingly, N
micropores, such behavior being at the origin of the Type II form
of the N isotherm instead of the rather expected Type I as in the
2
adsorption at 77.4 K appears to be ‘‘invisible’’
2
case of zeolites. Such observation, already reported by Suib and
co-workers [70] on a variety of OMS-type frameworks, is attributed
to a hindered diffusion of N
genic temperatures. As a way out of this shortcoming, CO
2
into the OMS-2 micropores at cryo-
physi-
2
sorption at 273 K was claimed to offer a proper measurement of
the microporosity of OMS-2 samples due to the improved diffusion
of CO
ization to N
pected to be located at the external surface of Pd/K-OMS-2 crystals
vide infra).
2
into the channels. In our case, we restricted the character-
2
adsorption, since the active sites for amination are ex-
(
3.3.4. IR spectra
Fig. 9. IR spectra of K-OMS-2, Pd/K-OMS-2, and Pd/MMO.
The lattice vibrational behavior of the different samples was
studied using FTIR spectroscopy to probe the effect of Pd substitu-
tion on the spectral features of K-OMS-2 and assess the evolution
of the Pd/K-OMS-2 sample after reaction. Fig. 9 plots the different
indication of neither adsorbed benzyl alcohol (in excess during
the reaction) nor aniline (fully consumed during the reaction).
IR spectra obtained. Both K-OMS-2 and Pd/K-OMS-2 showed four
ꢂ1
characteristic bands around 721, 606, 530, and 471 cm
with
3.3.5. H
The H
MMO catalysts are presented in Fig. 10, while the corresponding
2
H consumption is listed in Table 2. The raw K-OMS-2 displayed
2
-TPR profiles
comparable relative intensities that can be assigned to Mn–O lat-
tice vibration modes in [MnO ] octahedra [71]. Furthermore, no
peak ascribed to adsorbed water in the tunnels was observed
2
-TPR profiles of the raw K-OMS-2, Pd/K-OMS-2 and Pd/
6
ꢂ1
(
about 1627 cm ), suggesting a proper outgassing of the samples
two characteristic peaks appearing in the temperature range
and low hydration kinetics under ambient atmosphere.
255–300 °C, which are attributed to the two-step reduction of
After the reaction, the Pd/MMO sample exhibited two bands
centered at 614 and 513 cm that can be attributed, respectively,
to Mn–O vibrations in tetrahedral and octahedral environments
72]. The former two bands are compatible with the crystalline
K-OMS-2 to Mn
ture. Although no information regarding the oxidation state of Mn
could be unambiguously inferred from the H consumption, the
value measured (11.8 mmol/g) is in good agreement with the
3 4 3 4
O and further Mn O to MnO at higher tempera-
ꢂ1
2
[
data obtained on the XRD patterns, pointing out the presence of
two different environments for Mn (tetrahedral and octahedral)
as expected for the hausmannite phase (spinel structure). Even
after washing, characteristic bands belonging to adsorbed N-phe-
expected value for complete reduction of MnO
(11.5 mmol/g).
The reduction behavior of K-OMS-2 showed a drastic change
after doping with Pd (i.e., Pd/K-OMS-2), the two characteristic
2
to MnO
nyl benzylamine could be detected (i.e., 1596, 1504, 1452, 1429,
peaks being displaced to 45–120 °C, but with a comparable H
sumption as for the parent K-OMS-2 (Table 2). This observation is
2
con-
ꢂ1
1
374, 1065, 1028, 989, 868, 714, and 510 cm ), but with no
Fig. 8. TEM micrographs of the Pd/MMO sample after the amination tests: (a) low resolution and (b) high resolution.

4
46
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
2
Fig. 10. TPR-H profiles of the K-OMS-2, Pd/K-OMS-2, and Pd/MMO.
classically linked to the formation of hydrides over noble metal is-
lands (in our case Pd), transferring fast to the MnO phase by spill-
x
over and favoring accordingly its reduction [73]. Note, however,
that the temperature range for reduction found in this study is
remarkably lower than the range reported by Liu et al. [74] on sup-
ported PdO/K-OMS-2 (110–190 °C for 2.5 wt.% Pd), most likely due
to a better dispersion of Pd in our case. No distinguishable peak
attributed to Pd was detected, being most probably included in
the main peak appearing at ca. 45 °C. In such situation, the larger
2
H consumption for the Pd/K-OMS-2 sample compared to the value
measured for K-OMS-2 (about 3.5%) could be tentatively ascribed
to Pd(II) or Pd(IV) moieties either incorporated or supported over
K-OMS-2.
After the amination reaction, the Pd/MMO phase showed a
markedly different reduction profile compared to the fresh Pd/K-
OMS-2 sample. Indeed, a sole reduction peak was observed cen-
tered at 180 °C with a H
agreement with the theoretical value of 8.3 mmol/g for hausman-
nite (Mn = MnO + Mn ), involving a distribution of Mn cat-
ions as 1/3 Mn(II) and 2/3 Mn(III).
2
consumption about 7.8 mmol/g, in good
3
O
4
2 3
O
3.3.6. TGA analyses
The thermal stability of the different samples was assessed
using TGA/DTG analysis. Fig. 11 plots the different thermal pat-
terns obtained under N atmosphere using a heating ramp of
°C/min. The K-OMS-2 and Pd/K-OMS-2 samples showed qualita-
tively similar thermal patterns characterized by four main steps:
i) 50–200 °C (2 wt.% loss) attributed to physisorbed water, (ii)
50–460 °C (4 wt.% loss) ascribed to chemisorbed O and water,
iii) 490–570 °C (3 wt.% loss) involving the liberation of lattice oxy-
gen from the cryptomelane phase of deoxygenation of MnO to
, and (iv) 630–730 °C (2 wt.%) due to the conversion of
into Mn (hausmannite). However, the K-OMS-2 frame-
2
5
(
3
2
(
2
Mn
Mn
2
O
O
3
3
2
3 4
O
work showed a higher thermal stability after Pd incorporation with
no additional weight loss steps.
Fig. 11. TGA/DTG thermal profiles for the K-OMS-2, Pd/K-OMS-2 and Pd/MMO
samples (upper and middle plots), and evolution of the O concentration in the gas
2
A different thermal pattern was observed for the Pd/MMO sam-
ple after reaction. The TGA/DTG profiles were obtained after wash-
ing the sample to remove the adsorbed organic compounds. As a
matter of fact, this sample showed three main weight loss steps:
stream (lower plot). Additional plots on the evolution of the CO
concentration in the gas stream can be found in the ESI.
2 2
, H O and NO
3.3.7. Elemental analyses
Table 3 lists the bulk Mn, K, and Pd composition (weight basis),
(
i) 50–200 °C (2 wt.% loss) attributed to adsorbed water and some
remaining organic products according to the evolution of the H
and CO concentration in the gas stream (Fig. S4) and to the FTIR
as well as the average oxidation state (AOS) of the different sam-
ples prepared in this study as inferred from EDX elemental analy-
sis. First, Pd incorporation to the K-OMS-2 framework triggered a
decrease in the manganese composition from 94.3 wt.% to
2
O
2
spectra (Fig. 9), (ii) 250-430 °C (14.3 wt.% loss) due to elimination
of organic products, and (iii) 520–650 °C (1.2 wt.% loss) due to re-
moval of residual organic fragments.
9
1.1 wt.%, whereas K only decreased about 12% to afford the

M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
447
Table 3
Bulk Mn, K, and Pd composition measured by EDX in the K-OMS-2, Pd/K-OMS-2 and Pd/MMO samples.
Catalysts
Molar fraction (%)
Weight fraction (wt.%)
Molar ratios (R, –)
AOSa
Mn
K
Pd
Mn
K
Pd
K/Mn
Pd/Mn
Pd/K
K-OMS-2
Pd/K-OMS-2
Pd/MMO
92.2
91.0
95.6
7.83
7.03
2.77
–
1.97
1.68
94.3
91.1
94.8
5.70
5.02
1.96
–
0.0849
0.0774
0.0290
–
–
3.96
3.97
–
3.82 (2.3 wt.%)b
3.23
0.0216
0.0176
0.28
0.606
a
Values measured from the K/Mn and K/Pd molar ratios in K-OMS-2 and Pd/K-OMS-2 samples.
Value measured by ICP-AES.
b
molecular formula K0.63Mn7.83Pd0.17
lar ratios showed a modest decrease from 0.0849 in K-OMS-2 to
.0774 in Pd/K-OMS-2. Taking into account the large amount of
O
16. As a result, the K/Mn mo-
distinction about the oxidation state of Pd due to the potential sub-
tle interplay between substituted Pd cations and Mn(III)/Mn(IV)
species. In contrast, in the absence of Pd or if this is dispersed over
0
Pd incorporated into the K-OMS-2 sample (3.82 wt.% in Mn, K,
and Pd basis) overcoming largely the K removal (0.70 wt.% in the
same basis), we might attribute the decline in the K composition
to a slight yet measurable increased oxidation state of the sample.
Assuming that no Pd occupied VIII-coordinated channel sites (see
justification below), the AOS of both samples was estimated using
the following expression obtained by imposing the condition of
charge balance in the OMS-2 framework
the external surface of K-OMS-2, RK/(Mn+Pd) transforms into RK/Mn.
More details on this point supporting also our choice for Eq. (1)
for inferring the AOS of the Pd/K-OMS-2 sample is provided below.
After the reaction, the Pd/MMO phase showed a markedly high-
er Mn bulk concentration (ca. 95 wt.%), but with a strong decrease
in the K concentration from the initial value of 5.02 wt.% in the Pd/
K-OMS-2 phase to 1.96 wt.% in Pd/MMO most probably due to par-
tial leaching. In contrast, the Pd concentration kept essentially
unchanged.
AOS ¼ 4 ꢂ RK=ðMnþPdÞ
ð1Þ
where RK/(Mn+Pd) is the K/(Mn + Pd) molar ratio showing limiting
values between 0 (AOS = 4) and 1 (AOS = 3). This ratio can be com-
puted from the experimental K/Mn and K/Pd ratios listed in Table 3
using the expression
3
.3.8. XPS
A complete XPS study was conducted to gain insight into the
oxidation state, surface composition, and atomic environment of
Mn, K, and Pd species in the different samples under study. Figs. 12
and 13 show, respectively, the experimental and simulated spectra
obtained for the corresponding Mn3s/Mn2p/Pd3d and K2p/O1s
core levels, while Tables 4 and 5 compile, respectively, the binding
energies (BE) and surface composition of the Mn, K and Pd species
after peak deconvolution. Detailed spectra for the range 0–1200 eV
1
1
1
¼ R
þ R
ð2Þ
R
K=ðMnþPdÞ
K=Mn
K=Pd
In the presence of Pd, Eq. (1) applies only if Pd replaces octahedral
Mn sites in OMS-2 either as Pd(II) or Pd(IV), but with no clear
Fig. 12. XPS spectra of (a–c) Mn3s, (d–f) Mn2p, (g and h) Pd3d, (i–k) O1s, and (i–n) K2p core levels for K-OMS-2, Pd/K-OMS-2, and Pd/MMO.

4
48
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
Fig. 13. XPS spectra of the O1s core levels for (a) K-OMS-2, (b) Pd/K-OMS-2, and (c) Pd/MMO.
Table 4
Mn, K, and Pd binding energies (main peak values, eV), spin–orbit energies (eV), and relative peak intensities (–) in the K-OMS-2, Pd/K-OMS-2, and Pd/MMO samples.
Catalyst
Mn3s
Mn2p
BE Mn3s BE
K2p
Pd3d
BE Mn3s(1) BE Mn3s(2)
D
BE
D
BE Mn2p BE K2p1/2
BE K2p3/2
D
K2p BE Pd3p3/2
BE Pd3p5/2
D
Pd3d
Mn2p1/2
Mn2p3/2
K-OMS-2
Pd/K-OMS-2 88.6
Pd/MMO 88.6
88.7
84.0
84.0
83.0
4.7
4.6
5.6
653.9
654.1
653.7
641.9
641.8
641.0
12.0
12.3
12.7
294.0
293.8
295.0
291.0
290.8
292.0
3.0
3.0
3.0
–
–
–
5.4
342.6
337.2
342.9/340.9 337.8/335.7 5.1/5.2
Table 5
Surface composition measured from XPS and relative abundance of the different oxidation states in the K-OMS-2, Pd/K-OMS-2, and Pd/MMO samples.
Catalyst
Molar fraction (%)
Molar ratio (R, –)
Mn species (%)
Pd species (%)
Pd(Ox)
AOSa
Mn
K
Pd
K/Mn
Pd/Mn
Pd/K
Mn(IV)
Mn(III)
Mn(II)
Pd(Red)
K-OMS-2
Pd/K-OMS-2
Pd/MMO
91.3
89.2
89.1
8.66
8.70
9.74
–
2.12
1.20
0.0948
0.0976
0.109
–
–
100
100
0
ND
ND
67
0
0
33
–
100
50
–
0
50
3.97
3.96
2.70
0.0239
0.0133
0.245
0.122
a
Values measured from the K/Mn and Pd/Mn molar ratios in K-OMS-2 and Pd/K-OMS-2 samples and from the relative surfaces of Mn(III) and Mn(II) moieties in Mn2p
peaks in the Pd/MMO sample (Figs. 12 and 13).
can be found in the ESI (Fig. S5). For all the samples, the Mn3s
Mn2p and Mn3s unpaired electrons and the 5dN electrons (i.e.,
2g and e for O symmetry in high-spin configuration) in the va-
(Fig. 12A–C) and Mn2p (Fig. 12D–F) core levels were split into
t
g
h
two peaks (Mn3s(1)/Mn3s(2) and Mn2p3/2/Mn2p5/2, respectively)
due to symmetric/antisymmetric spin correlations between the
lence shell of Mn (N = 3, 4, and 5 for Mn(IV), Mn(III), and Mn(II),
respectively). An additional smaller peak appeared at higher BE

M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
Table 6
449
that can attributed to a shake-up satellite for the Mn2p1/2 level
Binding energy and relative abundance of the different oxygen species in K-OMS-2,
Pd/K-OMS-2, and Pd/MMO samples.
encompassing a concomitant 2-electron transfer [75].
Only Mn(IV) could be unambiguously discriminated in both the
K-OMS-2 and Pd/K-OMS-2 samples, as inferred from the character-
istic BEs of the Mn3s and Mn2p doublets, at about 88.6/84.0 eV and
Catalyst
O1s (O
I
)
O1s (OII
BE (eV)
)
BE (eV)
Abundance (%)
Abundance (%)
6
50/641.8 eV (Fig. 12D and E), respectively, with a relative inten-
K-OMS-2
Pd/K-OMS-2
Pd/MMO
531.5
530.9
530.8
34
28
72
529.5
529.5
529.5
66
72
28
sity of 1.8:1 and 1.4:1. However, this fact does not exclude the
presence of a small and hardly detectable amount of Mn(III) cat-
ions in the cryptomelane framework. The spin–orbit splitting or
energy difference between the Mn2p1/2 and Mn2p3/2 levels was
slightly shifted from 12.0 to 12.3 eV after Pd incorporation,
whereas the spin–orbit energy for the Mn3s level was kept essen-
tially unchanged at a value of 4.6–4.7 eV, suggesting either the ab-
sence or lack of detectable variation in the AOS of Mn after Pd
incorporation [76]. The surface Mn composition in the cryptomel-
ane framework evolved from an initial value of 93.7 wt.% in K-
OMS-2 to 89.6 wt.% in Pd/K-OMS-2 with a change in the K/Mn mo-
lar ratios from 0.0948 to 0.0979, respectively, suggesting partial
Mn substitution by Pd. This reduction is similar to that measured
for the bulk Mn composition, reflecting a reasonable homogeneity
of the Pd distribution within the particle thickness.
to 1.96 wt.% in Pd/MMO after reaction, suggesting partial removal,
whereas the surface K composition exhibited the opposite trend,
increasing from 6.22 to 7.05 wt.%.
Finally, it is worth noting an outstanding change between the
O1s patterns for K-OMS-2 and Pd/K-OMS-2 and that of Pd/MMO.
Fig. 13I–K plots the corresponding spectra. In all cases, the O1s
spectrum clearly showed two characteristic bands after deconvolu-
tion that are classically attributed to framework O species (OII
and surface O , O , and OH species and oxygen vacancies (O
An additional band corresponding to adsorbed H
deconvoluted for K-OMS-2, but with a very modest surface (<2%).
The OII BE was kept unchanged for all the samples at a value of
2ꢂ
)
).
2ꢂ
ꢂ
I
2
O could be also
After the amination reaction, the Pd/MMO phase exhibited a
moderate decrease in the Mn2p and Mn3s BEs with values of
8
5
I
29.5 eV, whereas the O BE decreased noticeably after Pd incorpo-
ration (from 531.5 to 530.9 eV) and kept a similar value after
reaction.
8.6/83.0 eV (Fig. 12C) and of 653.7/641.0 eV (Fig. 12F), respec-
tively, for the corresponding doublets, whereas the spin–orbit
energies decreased in each case to values of 5.7 and 12.7 eV. The
relative intensity of Mn3s(1)/Mn3s(2) and Mn2p3/2/Mn2p1/2 peaks
kept unmodified at 1.8:1 and 1.5:1, respectively. These observa-
tions are consistent with a reduction in the AOS of the Pd/MMO
phase after reaction (Table 5). The XPS patterns for each level could
be successfully deconvoluted into two primary bands attributed to
Mn(III) and Mn(II) cations with the relative intensity 2:1 matching
the expected stoichiometry of hausmannite.
The Pd3d spectrum of Pd/K-OMS-2 sample (Fig. 12G) also
showed a characteristic doublet that can be assigned to the
Pd3d3/2 and Pd3d5/2 core levels. The corresponding BEs showed
unexpectedly high values, about 342.6 and 337.2 eV (Table 4),
respectively, with a spin–orbit energy of 5.4 eV and relative inten-
sity of 1:1.7 between the peaks. After the reaction, the Pd3d spec-
trum of the as-generated Pd/MMO phase showed the presence of
two additional bands appearing at lower energies (i.e., 340.9 and
The relative intensity of OII and O
fected by the introduction of Pd into the K-OMS-2 framework,
showing a decline of O moieties (34% in K-OMS-2 vs. 28% in Pd/
K-OMS-2). In contrast, after reaction, the Pd/MMO sample became
enriched in O moieties (28% in Pd/K-OMS-2 vs. 56% in Pd/MMO).
These results point out an outstanding increase in surface defects
after the reaction, this factor being most likely linked to the pro-
moted activity and selectivity of the Pd/K-OMS-2 for amination.
The impact of such defects on the catalytic activity is discussed
below.
I
species (Table 6) was also af-
I
I
4. Discussion
4.1. Location and oxidation state of Pd centers in Pd/K-OMS-2
3
35.7 eV, Fig. 12H). This result suggests the formation of reduced
The body of data collected in the Results section provides clear
evidence of a very good Pd dispersion over K-OMS-2 in the
Pd/K-OMS-2 sample before reaction and also in the as-generated
Pd/MMO hausmannite phase after reaction with no observable
formation of supported palladium oxide nanoparticles. Pd can be
localized a priori in three possible environments: (1) inside the
K-OMS-2 channels as Pd(II), encompassing partial substitution of
Pd species after reaction. The relative intensity between the oxi-
dized and reduced Pd species was 1:1. The bulk Pd composition
was only slightly reduced after the synthesis, confirming a lack
of leaching from the support. In contrast, the surface Pd composi-
tion measured by XPS showed a remarkable decrease after the
reaction. Such observation seems to indicate a segregation of Pd
within the particle thickness in Pd/MMO, involving also a possible
segregation of oxidation states.
+
the original K , (2) in the K-OMS-2 framework as Pd(II) or Pd(IV)
via a partial substitution of octahedral Mn(IV) or Mn(III), and (3)
The K2p spectra of the different samples (Fig. 13L–N) also com-
prised two peak components with no observable satellites. The
corresponding BEs for the different peaks increased after reaction
from the values of ca. 294 and 291 eV (K2p3/2 and K2p1/2) in both
K-OMS-2 and Pd/K-OMS-2 to 295.0 and 292.0 eV (K2p3/2 and
K2p1/2) after reaction (i.e., Pd/MMO). However, the corresponding
spin–orbit energies remained unchanged at a value of 3.0 eV with
an intensity ratio between the K2p3/2/K2p1/2 peaks at 2:1. This
body of data supports a dramatic change in the K environment
after the reaction, evolving from a highly dispersed eight-coordi-
nated environment inside the cryptomelane tunnels into a sur-
x
finely dispersed over or embedded into K-OMS-2 as PdO nanopar-
ticles. The former option appears to be excluded in light of the bulk
and surface Pd and K composition in Pd/K-OMS-2. Indeed, even if
the bulk K loading decreased after Pd loading compared to the
initial value in K-OMS-2, this reduction was much lower than the
loaded Pd (compare the values of 7.83–7.03 = 0.80 mol% for K
change with 1.97 mol% for Pd loading in Mn, K and Pd basis,
Table 3). Moreover, the surface K concentration kept unchanged
after Pd incorporation (Table 5), confirming the lack of significant
K-substitution by Pd even at sublayers near the external surface
of K-OMS-2.
face-segregated K
2
O or KOH phase in hausmannite (the K2p BEs
The third option also seems discouraged, since no evidence of
in K O are about 295.7 and 292.9 eV [77]). This hypothesis can
2
the formation of PdO
micrographs for neither Pd/K-OMS-2 (Fig. 7) nor Pd/MnO
Pd/Mn (Fig. S2). Moreover, as stated above, the similar bulk
and surface Pd concentrations suggest a homogeneous distribution
x
nanoparticles was found in the HRTEM
be reinforced by the difference in the bulk and surface K composi-
tions in Pd/MMO (Table 5). As a matter of fact, the bulk K compo-
sition decreased from an initial value of 5.02 wt.% in Pd/K-OMS-2
2
and
2 3
O

4
50
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
of Pd within the K-OMS-2 particle thickness, suggesting no embed-
ding of PdO nanoparticles by cryptomelane. The first observation
opposes to the results reported by Liu et al. [74], where supported
PdO nanoparticles on K-OMS-2 could be unambiguously identified
by the presence of a characteristic reflection at 2-theta = 30° in the
XRD patterns for samples with Pd loading P 2.5 wt.%. Moreover,
AOS of the cryptomelane framework after Pd substitution, involv-
x
ing in turn a decrease in the K composition and a reduction of O
oxygen species (Table 6, entries 1 and 2).
I
According to Eqs. (3) and (4), the molecular formula for Pd/K-
OMS-2 proposed in the Results subsection 3.3.7 (vide supra) can
be resolved to the final form K0.63Mn(III)0.63Mn(IV)7.20Pd0.17(IV)O16
.
2
the H -TPR tests performed on those samples at comparable exper-
imental conditions and for similar particle sizes reflected much
higher reduction temperatures than our Pd/K-OMS-2 sample
4.2. Reaction scheme: interplay of Pd and K centers
(
Fig. 10), suggesting in their case a less efficient formation of
hydrides.
According to the above-stated reasons, only the second option,
After elucidating the nature of Pd in the cryptomelane frame-
work, we need to address the interplay of Pd and basic centers
on the catalytic activity of Pd/K-OMS-2 for amination reactions.
namely Pd substitution into the cryptomelane framework, appears
to be feasible. The fact that the bulk and surface Mn concentrations
decreased after Pd incorporation in a similar degree as loaded Pd
2
The XRD (Fig. 6), FTIR (Fig. 9), and TPR-H (Fig. 10) patterns provide
clear evidence of the conversion of the initial Pd/K-OMS-2 into a
reduced hausmannite phase (Pd/MMO) after the amination reac-
tion, keeping still high selectivity to the secondary amine (i.e.,
PBA). The formation of this phase can be described by the following
set of redox reactions in the presence of benzyl alcohol as reducing
agent
(Table 3) provides a direct evidence of such hypothesis. Partial
Pd substitution into the cryptomelane framework encompasses a
slight yet detectable increase in the AOS of Mn from 3.95 in the
parent K-OMS-2 to 3.97 in Pd/K-OMS-2. This increase might be
at the origin of the observed reduction in the bulk K concentration
after Pd deposition (compare the bulk K/Mn molar ratio of 0.0849
in K-OMS-2 with the value of 0.0774 in Pd/K-OMS-2). The high AOS
approaching 4 might explain why only Mn(IV) could be unequivo-
cally distinguished in the XPS spectra for both Mn3s and Mn2p
core levels (Fig. 12A, B, D and E).
Based on this rationale, the next point to elucidate is the oxida-
tion and local environment of Pd. This question is crucial to further
unveil the nature of the active sites for amination. The existence of
a high-valent Pd species, most likely Pd(IV), in place of Mn in the
cryptomelane framework appears to be a priori the best option
due to symmetry reasons (Pd(II) prefers planar IV-fold environ-
ments, whereas Pd(IV) can accept octahedral coordination), and
to the similar ionic radii of Pd(IV) and Mn(IV)/Mn(III) under a
VIII-fold environment (6.2 Å for Pd(IV) vs. 5.3 Å for Mn(IV) and
1
2
1
2
VI
VI
VI
MnðIVÞ þ O2ꢂ
OH ! MnðIIIÞ þ HOꢂ
VI
þ
C
7
H
7
þ
C
7
H
6
O
ð5Þ
ð6Þ
ð7Þ
OMS
OMS
MnðIVÞ þ O2ꢂ þ C
H
OH ! MnðIIÞ þ H
IV
O þ C
7
7
2
7 6
H O
OMS
1
2
1
2
MnðIIIÞ þ O2ꢂ
þ
C
7
H
7
OH ! MnðIIÞ þ HOꢂ
IV
þ
7 6
C H O
OMS
OMS
The genesis of the hausmannite phase keeps and probably even
intensifies the dispersion of Pd(IV) centers in light of the slight but
detectable increase in the BEs for Pd3d5/2 and Pd3d3/2 core levels
(
342.9 and 337.8 eV, Table 4), generating about 50% of reduced
Pd centers most likely based Pd(II) species, as inferred from the
new bands appearing at lower BEs in the XPS spectra plotted in
Fig. 12H (Pd3d5/2 BE = 340.9 eV, Pd3d3/2 BE = 335.7 eV, Table 4).
The formation of Pd(II) species can be described by the reaction
5
.8 Å for Mn(III) compared to 8.6 Å for Pd(II) [78]). This hypothesis
is supported by the abnormally high Pd3d5/2 BE measured on Pd/K-
OMS-2 (337.2 eV, Table 4), becoming even larger after reaction (i.e.,
3
37.8 eV), compared to the values that would be expected for stan-
dard PdO and Pd(0). As a general rule, the Pd3d5/2 BEs show values
in the range 335.2–337.1 eV for PdO, and 334.1–336.4 for Pd(0)
1
2
1
2
VI
PdðIVÞ þ 2O2ꢂ
C H OH ! PdðIIÞ þ 2HOꢂ
IV
þ
7
7
þ
C H O ð8Þ
7
6
OMS
OMS
[
79,80], increasing inversely with the nanoparticle size [81]. The
Pd3d5/2 BE measured in our study is also slightly larger than the
values reported on Pd/MnO [82], Pd/MnCeO [83], and Pd/K-
Given the high dispersion of Pd after reaction, we can anticipate
x
x
that Pd(IV) and Pd(II) species share, respectively, octahedral and
tetrahedral sites together with Mn(III) and Mn(II) cations in the
spinel structure of hausmannite at a ratio 1:1 (Table 5) to give
OMS-2 [74] (up to 337.1 eV) based on highly dispersed supported
PdO nanoparticles (<4.5 nm), appearing within the common range
for bulk and dispersed PdO
2
(337.6–338.6 eV [37,84,85]).
the molecular formula Mn(II)0.98Pd(II)0.026Mn(III)1.96Pd(IV)0.026
O
4-
Assuming that Pd(IV) was the dominant species and imposing
charge balance in the cryptomelane framework, the following
two-cation exchange (redox) processes can be proposed account-
ing for Mn substitution by Pd
ꢁ0.085K
2
O. These centers might provide optimal interaction with
K being expelled from the cryptomelane tunnels during the amina-
tion reaction, favoring hydrogen borrowing. In the end, part of the
2
original K in Pd/K-OMS-2 is expected to segregate as surface K O or
KOH moieties near surface OH groups generated from lattice oxy-
gen (OII) (Table 6, entries 1 and 3), as deduced from the values of
the K2p3/2 and K2p1/2 BEs (295.0 and 292.0 eV, Table 4). The facile
release of OH groups is confirmed by the TGA/DTG patterns on Pd/
K-OMS-2 plotted in Fig. 11, favoring the adsorption of benzyl alco-
hol and therefore the first dehydrogenation step of the hydrogen
borrowing mechanism.
In light of the above-stated comments, a tentative scheme of
the amination process during the conversion of Pd/K-OMS-2 into
Pd/MMO is represented in Fig. 14. Highly dispersed Pd(IV) and
Pd(II) species might promote hydrogen borrowing via the genera-
2
ðaqÞ
þ
VI
MnðIVÞ ! PdðIVÞ þ Mn2þ
VI
Pd
þ
ð3Þ
ðaqÞ
1
2
VI
MnðIIIÞ þ HOꢂ
VIII
þ
VI
2ꢂ
þ
K ! MnðIVÞ þ O
þ
H
2
OMS
OMS
VIIIH þ Kþ
"
þ
ð4Þ
ðaqÞ
where VIII
H
° represent vacant VIII-fold tunnel octahedral sites.
On the one hand, Eq. (3) involves the direct substitution of
Mn(IV) by Pd(IV) with no change in the K+ distribution within
the cryptomelane channel. This process is expected to be favored
given the positive standard potential of the process under neutral
tion of stable intermediates, probably involving HO-Pd(IV)-(H)
2
2
ðaqÞ
þ
conditions:
D
E ꢃ EðMnO
2
=Mn Þ ꢂ EðPdO
2
=PdOÞ ¼ 1:23 V ꢂ 0:73
and HO-Pd(II)-H hydride species due to the low energy of forma-
tion of the latter species from Pd(II)O hydration in step III (about
7.8 kcal/mol, DFT [37]). This process would be performed near K-
OH moieties, favoring the adsorption of benzyl alcohol.
V ¼ 0:50 V [78,86]. On the other hand, Eq. (3) provides a basis for
choice of Eq. (1) for inferring the AOS of the Pd/K-OMS-2 sample.
Furthermore, Eq. (4) supports the observed enhancement of the

M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
451
Fig. 14. Hydrogen borrowing mechanism for amine alkylation with alcohols with concomitant Pd/K-OMS-2 conversion into Pd/hausmannite.
5
. Conclusions
We have shown along this paper the promising potentials of Pd/
References
[
1] H.A. Wittcoff, B.G. Reuben, J.S. Plotkin, Industrial Organic Chemicals, second
ed., Wiley, NY, 2004.
cryptomelane formulations for performing amination reactions
operating via the borrowing hydrogen mechanism. A proof of con-
cept was demonstrated for the model alkylation reaction of aniline
with benzyl alcohol, affording a PBA yield as high as 96% at 160 °C
during 3 h with no formation of neither the tertiary amine nor tol-
uene. The improved performance of this formulation was attrib-
uted to a highly active intermediate phase during the reaction
including highly dispersed Pd(IV)/Pd(II) species near K and defec-
tive OH moieties, favoring hydrogen transfer. After the reaction,
the original cryptomelane phase evolved into a Pd/hausmannite
phase still showing a high selectivity to PBA (>90%). This study
confirms the necessary correlation between surface basicity and
hydrogen transfer to perform amination reactions.
[
2] V.A. Veefkind, J.A. Lercher, J. Catal. 180 (1998) 258–269.
[3] K.S. Hayes, Appl. Catal. A: Gen. 221 (2001) 187–195.
[
4] A.A.N. Magro, G.R. Eastham, D.J. Cole-Hamilton, Chem. Commun. (2007) 3154–
156.
5] G.E. Dobereiner, R.H. Crabtree, Chem. Rev. 110 (2010) 681–703.
3
[
[6] G. Guillena, D.J. Ramon, M. Yus, Chem. Rev. 110 (2010) 1611–1641.
[
[
[
7] J.L. Klinkenberg, J.F. Hartwig, Angew. Chem. Int. Ed. 50 (2011) 86–95.
8] Fluzard, Bo, Huzard, Internal Rhodia, Report No. 2174, 1954.
9] G. Bo, Ph. Perras, Internal Rhodia, Report No. 2932, 1958.
[10] D.A. Gardner, R.T. Clark, US Patent 4,255,357 to Pennwalt Corporation, 1981.
[11] J.V. Martinez de Pinillos, R.L. Fowlkes, US Patent 4,314,084 to Air Products and
Chemicals Inc., 1982.
12] T.L. Renken, H.P. Klein, WO Patent 2009/114438 (A2) to Huntsman
Petrochemical Corporation.
[13] A.K. Rausch, E. van Steen, F. Roessner, J. Catal. 253 (2008) 111–118.
14] G.A. Vedage, L.A. Emig, H.-X. Li, J.N. Armor, US Patent 5,917,092 to Air Products
and Chemicals Inc., 1998.
[
[
A perspective of this study encompasses operando studies on
the Pd/K-OMS-2 phase under reaction conditions to ascertain
the nature of the intermediate species formed during conversion
into hausmannite. These studies will be communicated in due
time.
[
15] G.A. Vedage, K.S. Hayes, M. Leeaphon, J.N. Armor, US Patent 5,932,769 to Air
Products Inc., 1999.
16] A. Fischer, T. Mallat, A. Baiker, J. Mol. Catal. A: Chem. 149 (1999) 197–204.
17] A. Fischer, T. Mallat, A. Baiker, J. Catal. 182 (1999) 289–291.
18] G. Jenzer, T. Mallat, A. Baiker, Catal. Lett. 61 (1999) 111–114.
[
[
[
[19] M. Imabeppu, K. Kiyoga, S. Okamura, H. Shoho, H. Kimura, Catal. Commun. 10
2009) 753–757.
20] A. Fischer, M. Maciejewski, T. Bürgi, T. Mallat, A. Baiker, J. Catal. 183 (1999)
73–383.
21] F. Santoro, R. Psaro, N. Ravasio, F. Zaccheria, ChemCatChem 4 (2012) 1249–
254.
[22] K.-I. Shimizu, N. Imaiida, K. Kon, S.M.A. Hakim Siddiki, A. Satsuma, ACS Catal. 3
2013) 998–1005.
(
[
[
3
Acknowledgments
1
This study was funded by Solvay and CNRS. The authors would
like to express their gratitude to Prof. Feng Shi from the Lanzhou
Institute of Chemical Physics (CAS) for insightful discussions on
Pd-catalyzed amination mechanisms. The authors are also grateful
to Dr. A. Beaurain at UCCS Lille for carrying out the XPS
measurements.
(
[
[
23] X. Cui, X. Dai, Y.Q. Deng, F. Shi, Chem. Eur. J. 19 (2013) 3665–3675.
24] P.R. Likhar, R. Arundhathi, M.L. Kantam, P.S. Prathima, Eur. J. Org. Chem. (2009)
5383–5389.
25] J. Sun, X. Jin, F.W. Zhang, W. Hu, J. Liu, R. Li, Catal. Commun. 24 (2012) 30–33.
26] Y. Zhang, S.N. Riduan, Chem. Soc. Rev. 41 (2012) 2083–2094.
27] N. Zotova, F.J. Roberts, G.H. Kelsall, A.S. Jessiman, K. Hellgardt, K.K. Hii, Green
Chem. 14 (2012) 226–232.
[
[
[
[
28] T. Ichida, H. Kawakita, T. Akita, M. Haruta, Gold Bull. 42 (2009) 267–274.
Appendix A. Supplementary material
[29] Y.S. Demidova, I.L. Simakova, M. Estrada, S. Beloshapkin, E.V. Suslov, D.V.
Korchagina, K.P. Volcho, N.F. Salakhutdinov, A.V. Simakov, D.Y. Murzin, Appl.
Catal. A: Gen. 464–465 (2013) 348–356.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.10.003.
[
30] X. Cui, Y. Zhang, F. Shi, Y.Q. Deng, Chem. Eur. J. 17 (2011) 1021–1028.
[31] H. Liu, G.-K. Chuah, S. Jaenicke, J. Catal. 292 (2012) 130–137.

4
52
M. Ousmane et al. / Journal of Catalysis 309 (2014) 439–452
[
32] J.-F. Soulé, H. Miyamura, S. Kobayashi, J. Am. Chem. Soc. 133 (2011) 18550–
8553.
[62] R. Ghosh, Y.C. Son, V.D. Makwana, S.L. Suib, J. Catal. 224 (2004) 288–296.
[63] G.D. Yadav, R.K. Mewada, Catal. Today 198 (2012) 330–337.
[64] H.G. Manyar, B. Yang, H. Daly, H. Moor, S. McMonagle, Y. Tao, G.D. Yadav, A.
Goguet, P. Hu, C. Hardacre, ChemCatChem 5 (2013) 506–512.
[65] S. Sithambaram, R. Kumar, Y.C. Son, S.L. Suib, J. Catal. 253 (2008) 269–277.
[66] K. Yamaguchi, H. Kobayashi, T. Oishi, N. Mizuno, Angew. Chem. Int. Ed. 51
(2012) 544–547.
1
[
[
[
[
[
33] K. Uehara, JP Patent 5,148,191 to Idemitsu Petrochemical Co., 1993.
34] K. Uehara, JP Patent 6,001,758 to Idemitsu Petrochemical Co., 1994.
35] J.W. Kim, K. Yamaguchi, N. Mizuno, J. Catal. 263 (2009) 205–208.
36] K. Yamaguchi, J. He, T. Oishi, N. Mizuno, Chem. Eur. J. 16 (2010) 7199–7207.
37] Y. Zhang, X. Qi, X. Cui, F. Shi, Y.Q. Deng, Tetrahedron Lett. 52 (2011) 1334–
1
338.
[67] G.G. Xia, W. Tong, E.N. Tolentino, N.-G. Duan, S.L. Brock, J.-Y. Wang, S.L. Suib, T.
Ressler, Chem. Mater. 13 (2001) 1585–1592.
[68] N.N. Opembe, C.K. King’ondu, A.E. Espinal, C. Chen, E.K. Nyutu, V.M.
Crisostomo, S.L. Suib, J. Phys. Chem. C 114 (2010) 14417–14426.
[69] S. Dharmarathna, C.K. King’ondu, W. Pedrick, L. Pahalagedara, S.L. Suib, Chem.
Mater. 24 (2012) 705–712.
[
[
[
[
[
[
38] A. Corma, T. Rodenas, M.J. Sabater, Chem. Eur. J. 16 (2010) 254–260.
39] S.L. Suib, Acc. Chem. Res. 41 (2008) 479–487.
40] J. Cai, J. Liu, W.S. Willis, S.L. Suib, Chem. Mater. 13 (2001) 2413–2422.
41] Q. Feng, H. Kanoh, Y. Miyai, K. Ooi, Chem. Mater. 7 (1995) 148–153.
42] R. Kumar, S. Sithambaran, S.L. Suib, J. Catal. 262 (2009) 304–313.
43] J. Liu, V. Makwana, J. Cai, S.L. Suib, M. Aindow, J. Phys. Chem. B 107 (2003)
[70] J. Luo, Q. Zhang, J. Garcia-Martinez, S.L. Suib, J. Am. Chem. Soc. 130 (2008)
3198–3207.
9
185–9194.
[
[
44] G.G. Xia, Y.G. Yin, W.S. Willis, J.Y. Wang, S.L. Suib, J. Catal. 185 (1999) 91–105.
45] W. Gac, G. Giecko, S. Pasieczna-Patkowska, T. Borowiecki, L. Kepinski, Catal.
Today 137 (2008) 397–402.
[71] C.K. King’ondu, N. Opembe, C.-H. Chen, K. Ngala, H. Huang, A. Iyer, H.F. Garcés,
S.L. Suib, Adv. Funct. Mater. 21 (2011) 312–323.
[72] M. Ishii, M. Nakahira, Solid State Commun. 11 (1972) 209–212.
[73] S.J. Gentry, N.W. Hurst, A. Jones, J. Chem. Soc., Faraday Trans. 1977 (1981) 603–
619.
[74] X. Liu, J. Lu, W. Wang, M. Luo, Chin. J. Catal. 31 (2010) 181–185.
[75] M. Oku, K. Hirokawa, S. Ikeda, J. Electron Spectrosc. Relat. Phenom. 7 (1975)
465–473.
[76] V.R. Galakhov, M. Demeter, S. Bartkowski, M. Neumann, N.A. Ovechkina, E.Z.
Kurmaev, N.I. Lobachevskaya, Y.M. Mukovskii, J. Mitchell, D.L. Ederer, Phys.
Rev. B 65 (2002) 1131021–1131024.
[
[
[
[
[
46] X. Yang, J. Han, Z. Du, H. Yuan, F. Jin, Y. Wu, Catal. Commun. 11 (2010) 643–
6
46.
47] S. Ching, P.F. Driscoll, K.S. Kielyka, M.R. Marvel, S.L. Suib, Chem. Commun.
2001) 2486–2487.
(
48] W.Y. Hernandez, M.A. Centeno, F. Romero-Sarria, S. Ivanova, M. Montes, J.A.
Odriozola, Catal. Today 157 (2010) 160–165.
49] M. Sun, L. Yu, F. Ye, G. Diao, Q. Yu, Z. Hao, Y. Zheng, L. Yuan, Chem. Eng. J. 220
(
2013) 320–327.
50] W.Y. Hernandez, M.A. Centeno, S. Ivanova, P. Eloy, E. Gaigneaux, J.A. Odriozola,
Appl. Catal. B: Environ. 123–124 (2012) 27–35.
[77] B.V. Crist, Handbook of Monochromatic XPS Spectra: The Elements and Native
Oxides, vol. 1, John Wiley & Sons, Chichester, 2000.
[
[
51] H. Nur, F. Hayati, H. Hamdan, Catal. Commun. 8 (2007) 2007–2011.
52] L. Sun, Q. Cao, B. Hua, J. Li, J. Hao, G. Jing, X. Tang, Appl. Catal. A: Gen. 393
[78] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 84th ed., CRC Press,
Boca Raton, 2003.
(
2011) 323–330.
[79] A.V. Naumkin, A. Kraut-Vass, S.W. Gaarenstroom, C.J. Powell, NIST X-ray
Photoelectron Spectroscopy Database, NIST Standard Reference Database 20,
Version 4.1, <http://srdata.nist.gov/xps/selEnergyType.aspx> (retrieved
13.06.13).
[
[
53] C. Calvert, R. Joersten, K. Ngala, J. Villegas, A. Morey, X. Shen, S.L. Suib, Chem.
Mater. 20 (2008) 6382–6388.
54] X.-S. Liu, Z.-N. Jin, J.-Q. Lua, X.-X. Wang, M.-F. Luo, Chem. Eng. J. 162 (2010)
1
51–157.
[80] M. Peuckert, J. Phys. Chem. 89 (1985) 2481–2486.
[81] T. Wu, W.E. Kaden, W.A. Kunkel, S.L. Anderson, Surf. Sci. 603 (2009) 2764–
2770.
[82] H.T. Tan, Y. Chen, C. Zhou, X. Jia, J. Zhu, J. Chen, X. Rui, Q. Yan, Y. Yang, Appl.
Catal. B: Environ. 119–120 (2012) 166–174.
[83] Y. Chen, H. Zheng, Z. Guo, C. Zhou, C. Wang, A. Borgna, Y. Yang, J. Catal. 283
(2011) 34–44.
[84] D.H. Kim, S.I. Woo, J.M. Lee, O.B. Yang, Catal. Lett. 70 (2000) 35–41.
[85] L.S. Kibis, A.I. Stadnichenko, S.V. Koscheev, V.I. Zaikovskii, A.I. Boronin, J. Phys.
Chem. C 116 (2012) 19342–19348.
[
[
55] H. Tian, J. He, L. Liu, D. Wang, Ceram. Int. 39 (2013) 315–321.
56] O. Sanz, J.J. Delgado, P. Navarro, G. Arzamendi, L.M. Gandía, M. Montes, Appl.
Catal. B: Environ. 110 (2011) 231–237.
57] T. Martínez, L.M. Arzamendi, F. Romero-Sarria, W.Y. Hernández, M.A. Centeno,
J.A. Odriozola, Appl. Catal. A: Gen. 423–424 (2012) 137–145.
58] Z.-Y. Cai, M.-Q. Zhu, J. Chen, Y.-Y. Shen, J. Zhao, Y. Tang, X.-Z. Chen, Catal.
Commun. 12 (2010) 197–201.
[
[
[
[
[
59] Y.C. Son, V.D. Makwana, A.R. Howell, S.L. Suib, Angew. Chem. Int. Ed. 9 (1999)
3
19–333.
60] F. Schurz, J.M. Bauchert, T. Merker, T. Schleid, H. Hasse, R. Gläser, Appl. Catal. A:
Gen. 355 (2009) 42–49.
61] V.D. Makwana, Y.C. Son, A.R. Howell, S.L. Suib, J. Catal. 210 (2002) 46–52.
[86] A.J. Bard, R. Parsons, R.J. Jordan, Standard Potentials in Aqueous Solutions,
IUPAC (Marcel Dekker), NY, 1985.