Co-based heterogeneous catalysts from well-defined Α-diimine complexes: Discussing the role of nitrogen

Article abstract of DOI:10.1016/j.jcat.2017.04.014

Ar-BIANs and related α-diimine Co complexes were wet impregnated onto Vulcan XC 72 R carbon black powder and used as precursors for the synthesis of heterogeneous supported nanoscale catalysts by pyrolysis under argon at 800?°C. The catalytic materials feature a core-shell structure composed of metallic Co and Co oxides decorated with nitrogen-doped graphitic layers (NGr). These catalysts display high activity in the liquid phase hydrogenation of aromatic nitro compounds (110?°C, 50 bar H₂) to give chemoselectively substituted aryl amines. The catalytic activity is closely related to the amount and type of nitrogen atoms in the final catalytic material, which suggests a heterolytic activation of dihydrogen.

Full text of DOI:10.1016/j.jcat.2017.04.014

Journal of Catalysis 351 (2017) 79–89
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Co-based heterogeneous catalysts from well-defined
complexes: Discussing the role of nitrogen
a-diimine
a,b,2
a,2
b,1
b
b
Dario Formenti
, Francesco Ferretti , Christoph Topf , Annette-Enrica Surkus , Marga-Martina Pohl ,
b
b
b
b,
⇑
a,
⇑
Jörg Radnik , Matthias Schneider , Kathrin Junge , Matthias Beller , Fabio Ragaini
a
Dipartimento di Chimica – Università degli Studi di Milano, Via C. Golgi 19, 20133 Milano, Italy
Leibniz-Institut für Katalyse e.V. an der Universität Rostock, Albert-Einstein-Straße 29a, 18059 Rostock, Germany
b
a r t i c l e i n f o
a b s t r a c t
Ar-BIANs and related a
-diimine Co complexes were wet impregnated onto VulcanÒ XC 72 R carbon black
powder and used as precursors for the synthesis of heterogeneous supported nanoscale catalysts by
pyrolysis under argon at 800 °C. The catalytic materials feature a core-shell structure composed of metal-
lic Co and Co oxides decorated with nitrogen-doped graphitic layers (NGr). These catalysts display high
Article history:
Received 21 December 2016
Revised 4 April 2017
Accepted 12 April 2017
2
activity in the liquid phase hydrogenation of aromatic nitro compounds (110 °C, 50 bar H ) to give
chemoselectively substituted aryl amines. The catalytic activity is closely related to the amount and type
of nitrogen atoms in the final catalytic material, which suggests a heterolytic activation of dihydrogen.
Ó 2017 Elsevier Inc. All rights reserved.
Keywords:
Cobalt
Hydrogenation
Nitro compounds
Kinetic
Heterolytic activation
1
. Introduction
Many of the industrially relevant catalytic processes in the fine
the support and the supported metal [17–19]. Consequently, it is
possible to modify the activity and adjust the selectivity of the final
catalytic material toward the desired transformation. Among the
various dopants, nitrogen attracted most interest [15,20–22] and
applications of transition metal/N-doped graphene (NGr) based
catalytic systems ranges from oxygen reduction reactions (ORR)
[23–26], hydrogen evolving reactions (HER) [27,28], photocatalysis
[29,30], oxidation [31–38], and reduction [39–44] reactions of
organic molecules, CAC bond formation [45] to many others [46–
51]. Hence, merging rationally-designed suitable modified sup-
ports with cheap transition metals permits access to the design
of active, selective and cheap catalytic materials thus allowing to
match or to outperform noble-metal-based catalysts. During the
last three years, some of us reported the preparation of active
and selective NGr-decorated Co-based catalysts from the pyrolysis
of in situ generated 1,10-phenanthroline (Phen) metal complexes
chemical industry are still based on expensive and rare late transi-
tion metals such as Pd, Pt, Rh, Ru and Ir [1–5]. Although these met-
als exhibit very good catalytic performance for advanced organic
substrates, the decreasing availability of these elements requires
the development of efficient alternative metal-based catalysts
[
6]. In this regard, the design of catalytic systems based on abun-
dant and biocompatible metals is an important goal for the imple-
mentation and progress of green and sustainable chemistry. Hence,
transition metals such as Fe, Co and Cu are ideal candidates, which
meet these requirements [7–10]. As a matter of fact, these ele-
ments are among the most abundant metals in the Earth’s upper
crust, thus being readily accessible [11]. Recently, heteroatom-
doped carbon materials attracted major interest in the field of
metal supported heterogeneous catalysts for both chemical syn-
thesis [12–14] and/or energy-relevant transformations [15,16].
Indeed, doping graphene with heteroatoms such as N, B, P, or S
leads to a radical modification of the electronic properties of both
using VulcanÒ XC 72 R carbon [52–60], ceria [61] or
a-alumina
as supports [62]. Some of the obtained nanoscale catalysts exhibit
a core-shell architecture in which a Co metallic core is enveloped
by an oxidic sheath composed of Co O . In addition, this oxidic
3 4
shell is augmented by layers of NGr derived from the thermal
decomposition of the Phen ligand. Nevertheless, from an
economical point of view, phenanthroline is expensive and its
functionalization requires multistep fair-yielding transformation
procedures. Thus, we were interested in using other, more
easily tunable, nitrogen compounds. In this respect, chelating
⇑
Corresponding authors.
E-mail addresses: matthias.beller@catalysis.de (M. Beller), fabio.ragaini@unimi.it
(
F. Ragaini).
1
Present address: Institut für Katalyse
– Johannes Kepler Universitat Linz,
Altemberger Straße 69, 4040 Linz, Austria.
2
These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.jcat.2017.04.014
0
021-9517/Ó 2017 Elsevier Inc. All rights reserved.

8
0
D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
a
-diimines, especially Ar-BIAN ligands, are widely used in
PerkinElmer AAS Analyst 300 after fusion melts and acidic dissolv-
ing of the sample.
TPR-H measurements were conducted using a Micrometrics
Autochem II 2920 instrument equipped with a TCD detector. The
experiment run was carried out from 36 °C to 700 °C in a 5% H
transition-metal catalyzed reactions [63]. As the starting materials
are inexpensive (acenaphthenequinone and a variety of aromatic
or aliphatic amines), it is possible to prepare a large number of
ligands with different electronic and steric properties. This allows
for the synthesis of tailor-made transition-metal based catalysts.
Nevertheless, apart from a traditional immobilized complexes
2
2
/
3
À1
À1
Ar flow (20 cm min ) with a heating rate of 10 K min
.
[
64], to the best of our knowledge, no examples of their use as
2.3. Catalyst preparation
heterogeneous catalysts precursors are known. Herein, we report
for the first time that Ar-BIANs and related ligands are able to
generate efficient Co/NGr catalysts for the hydrogenation of
aromatic nitro compounds.
The procedure was adapted from that reported for the synthesis
of Co/Phen based catalysts [58] (see Fig. 1). Cobalt(II) acetate
tetrahydrate was added to absolute ethanol (40 mL of EtOH for
1
mmol of Co(OAc)
tion (10 min., formation of a clear purple solution). Then the ligand
2 mmol) was added (color change to deep red) and the resulting
2
Á4H
2
O) and stirred until complete solubiliza-
2
. Experimental
(
solution was stirred at 60 °C for 2 h. Owing to their scarce solubil-
ity in EtOH, ligands L2 and L7 were initially solubilized in the min-
2.1. Synthesis of the ligands
imum amount of inhibitor-free THF and then dropwise added to a
Concerning the synthesis of the ligands, all the reactions were
Ò
solution of Co(OAc)
1.392 g for 1 mmol of Co(OAc)
during about 30 min and the suspension was stirred at 25 °C for
8 h. Then, the solvent was removed and the obtained solid was
2
Á4H
2
O in EtOH. After that, VULCAN XC 72R
carried out under a nitrogen atmosphere using standard Schlenk
techniques. All glassware and magnetic stirring bars were kept in
an oven at 120 °C for at least two hours and were cooled to room
(
2
2
Á4H O) was portionwise added
1
3
temperature under vacuum prior to use. CDCl used for the NMR
dried for 4 h under vacuum, grinded to a very fine powder and
finally transferred into a ceramic crucible, equipped with a lid,
and placed in the pyrolysis oven. The oven was evacuated to ca.
experiments was filtered on basic alumina and stored under nitro-
gen over 4 Å molecular sieves. Chemicals and solvents were pur-
chased from Sigma-Aldrich, Alfa Aesar or Tokyo Chemical
Industry. Seven different ligands were prepared (L1–L7) following
protocols previously reported by some of us (for the preparation of
L1, L2, L4, L5 see Supporting Information) [65]. Ligand L3 has been
known for more than one century [66] and its synthesis was per-
formed adapting a procedure previously described in the literature
5
8
mbar and then flushed with argon. Afterward, it was heated to
00 °C at a rate of 25 °C per minute and held at 800 °C for 2 h under
Ar atmosphere. Finally, heating was stopped and the oven was
cooled down to room temperature. During the whole process, a
constant flux of argon through the oven was maintained. The ele-
mental analyses of the prepared materials are reported in the Sup-
porting Information.
[
67]. Concerning the preparation of ligand L6, its synthesis was
adapted from that previously reported by some of us [68]. Finally,
L7 was synthesized adapting a procedure reported many years ago
2.4. General methods for catalytic reactions in the autoclave
[
66]. The detailed protocols for the preparation of the seven ligands
are reported in the Supporting Information.
In an 8 mL glass vial fitted with a magnetic stirring bar and a
septum cap, the catalyst (the amount depends on the catalyst)
was added followed by the nitroarene (0.5 mmol), the internal
standard (hexadecane, 20 mg) and the solvent (2 mL). A needle
was inserted in the septum cap, which allows dihydrogen to enter.
The vials (up to 7) were placed into a 300 mL steel Parr autoclave
which was flushed twice with dihydrogen at 20 bar and then pres-
surized to 50 bar. Then the autoclave was placed into an aluminum
block pre-heated at 110 °C. At the end of the reaction, the autoclave
was quickly cooled down at room temperature with an ice bath
and vented. Finally, the samples were removed from the autoclave,
diluted with a suitable solvent, filtered using a Pasteur pipette
filled with CeliteÒ (6 cm pad) and analyzed by GC using n-
hexadecane as an internal standard. Control experiments showed
that the position of the vial inside the autoclave is not influential.
The same outcome was obtained when the reaction was repeated
by moving a vial from a peripheral to a central position.
2
.2. General analysis and characterization methods
TEM measurements were performed at 200 kV with an
aberration-corrected JEM-ARM200F (JEOL, Corrector: CEOS). The
microscope is equipped with a JED-2300 (JEOL) energy-dispersive
X-ray spectrometer (EDXS) for chemical analysis. The sample was
deposited without any pre-treatment on a holey carbon supported
Cu-grid (mesh 300) and transferred to the microscope. The High-
Angle Annular Dark Field (HAADF) and Annular Bright Field (ABF)
images were recorded with a spot size of approximately 0.1 nm,
a probe current of 120 pA and a convergence angle of 30–36°.
The collection semi-angles for HAADF and ABF were 70–170 mrad
and 11–22 mrad, respectively.
XPS data were obtained with a VG ESCALAB220iXL (Thermo
Scientific) with monochromatic Al K
a (1486.6 eV) radiation. The
electron binding (EB) energies were obtained without charge
compensation. For quantitative analysis, the peaks were deconvo-
luted with Gaussian-Lorentzian curves, and the peak area was
divided by a sensitivity factor obtained from the element specific
Scofiled factor and the transmission function of the spectrometer.
XRD patterns of the materials were recorded on a Panalytical
2.5. Procedure for quantitative determination of the reaction products
All nitroarenes employed and all anilines reported in Fig. 8 are
commercially available compounds. Their amount was determined
by GC analysis (HP 6890 series GC system) using n-hexadecane as
internal standard and calibrating the response factor by using pure
compounds (Aldrich, Alfa-Aesar, Tokyo Chemical Industry). For
product 2aa, at the end of the reaction the catalyst was separated
using a Pasteur pipette filled with CeliteÒ and the Celite pad was
washed with EtOH. The solvent was evaporated and the desired
product isolated using column chromatography (AcOEt:hep-
tane = 1:1). The product was obtained as a light brown solid.
Regarding product 2ab, after the reaction was complete, the cata-
X’Pert Pro diffractometer in reflection mode with Cu K
k = 1.5406 Å) and a silicon strip detector (X’Celerator).
NMR spectra of ligands and isolated anilines were recorded on a
a radiation
(
Ò
Bruker Avance DRX 300 or on a Bruker Avance DRX 400 operating
at 300 and 400 MHz, respectively.
CHN analyses were performed using a Leco Microanalysator
TruSpec or a PerkinElmer 2400 CHN. Metal content of the catalysts
was determined by atom absorption spectroscopy using
a

D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
81
Fig. 1. Catalyst preparation.
lyst was separated using a Pasteur pipette filled with CeliteÒ and
Ò
the Celite pad was washed with AcOEt. The solvent was removed
Fig. 2. XRD patterns of the selected catalysts.
in vacuo affording the product as a pale orange solid without the
need of purification procedures. For the spectroscopic characteri-
zation of both the products see the Supporting Information.
correspond to (111), (200) and (220) crystal planes of cobalt in
the zero oxidation state [69]. Minor peaks corresponding to 2h val-
ues of 37.3, 42.5, 59.5 and 65.2 were further detected and assigned
2.6. Procedure for recycling experiments
to CoO and Co
h = 25 is ascribed to the amorphous carbon-based support.
Next, XPS was employed to investigate the surface composition
3 4
O [70]. In addition, the broad peak at around
2
For the catalyst recycling experiments, sixfold scaled up reac-
tions were carried out. All the reactions were performed in glass
vials set up according to previously described procedure. After
completion of the reaction the content of the vial was quantita-
tively transferred into a centrifuge tube. Hereafter, the reaction
mixture was centrifugated and the catalyst was separated from
the supernatant. The catalyst was washed three times with EtOH
and dried under vacuum overnight. This material was then used
for the next catalytic reaction.
of the materials. Fig. 3 proves that nitrogen is incorporated into the
carbonaceous matrix indicating that the employed ligands partici-
pate in the generation of these materials.
Deconvolution of the N 1s spectrum gives rise at least to two
individual peaks for each catalyst. Detected peaks ranging from
3
97.9 eV to 398.1 eV are ascribed to pyridinic-type N atoms while
peaks around 400 eV can be ascribed to pyrrolic-type N atoms or to
N coordinated to Co (Co-N centers) [71]. Because of the expected
x
proximity of the latter signals, it is impossible to unequivocally
quantify the two independent contributions. Catalysts Co/L1 and
Co/L3 exhibited two additional peaks at 404.4 eV and 403.8 eV,
respectively. These can be assigned to N-oxides of the pyridinic N
modifications. In the case of Co/L3, N-oxide formation upon pyrol-
ysis is explained by the presence of oxygen directly bound to the
chelating nitrogen atom in the precursor whereas in the thermally
treated Co/L1 system the N-O signals are likely to originate from
traces of adventitious oxygen inside the pyrolysis oven. Co/L7
showed slightly different peak features with three states at
2
.7. Maitlis’ hot filtration test
The procedure for setting up a standard catalytic experiment
was followed, but the reaction was interrupted after 2 h. Hereafter,
the autoclave was cooled down to room temperature and vented. A
small amount of the vial content was filtered through a Pasteur
pipette filled with CeliteÒ (6 cm pad) and an aliquot of the filtrate
was analyzed by GC. Then the remaining reaction mixture was
quantitatively transferred into a Schlenk flask and heated up to
9
0 °C for 1 h. The mixture was rapidly filtered while hot using a
3
98.1 eV, 399.7 eV and 400.6 eV. Whereas the first state can be
attributed to pyridinic N as for the other samples, the both latter
ones can be correlated with the Co-N centers and pyrrolic N. Co
p region was further examined (see Figs. S1–S4 in the Supporting
Ò
Pasteur pipette filled with Celite (6 cm pad). The filtrate was
quantitatively transferred into a clean reaction vial and subjected
to the standard reaction conditions. In the latter case, the reaction
time was prolonged to 24 h in order to take into account any sol-
uble Co species that exhibit only low catalytic activity. The reaction
mixture was eventually analyzed by GC.
x
2
Information). Catalysts Co/L1 and Co/L2 displayed similar peaks.
For the Co/L1 and Co/L2 two peaks were observed, one at 780 eV
corresponding to the 2p3/2 state and the other one at 795 eV corre-
lated with the 2p1/2 state. Due to the low amount of Co it is not
possible to distinguish between di- and trivalent Co. Peaks at
3
. Results and discussion
7
86 eV and 800 eV (typical for CoO) were found for the other
two samples [26]. Finally, Co/L7 presented an additional peak at
78.8 eV representative of Co atoms in zero oxidation state [72].
3.1. Characterization of the catalytic materials
7
Four catalytic materials (Co/L1, Co/L2, Co/L3 and Co/L7) were
fully characterized by using X-ray diffraction (XRD), X-ray photo-
electron spectroscopy (XPS), transmission electron microscopy
TEM) and temperature programmed reduction (TPR) techniques.
We commenced the catalyst characterization with XRD measure-
ments in order to identify the constituting phases of the sample
Fig. 2).
The XRD patterns revealed that the main species present is
metallic Co. In fact, peaks at 2h values of 44.4, 54.7 and 76.0
In addition, carbon and oxygen XPS regions were studied (see
Supporting Information). Concerning the former, the four catalysts
displayed an analogous pattern. The peaks at 283.8, 284.9 and
285.2–286.3 eV can be attributed to the presence of C@C, C@N
and CAN bonds confirming the inclusion of the N atoms into the
(
graphitic matrix [26]. Every C 1s spectra showed the shake-up
⁄
(
feature around 290 eV arising from the
p
to
p
transition typical
of graphitic like compounds. The interpretation of the oxygen
region is not straightforward due to a multitude of possible O

8
2
D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
Fig. 3. High-resolution N 1s spectra of Co/L1 (a), Co/L2 (b), Co/L3 (c) and Co/L7 (d). For XPS survey, O 1s, C 1s and Co 2p of the selected catalysts see SI (Figs. S1–S4 and
corresponding tables).
containing compounds, e.g. CoO, Co(OH)
pounds with CAO and C@O bounds. The peaks between 530.6 and
2
and several organic com-
top of metallic Co fractions intimately in contact with the particles.
For Co/L1 no graphitic envelops were found since the metallic
cobalt is completely embedded by cobalt oxide. From Co/L2 to
Co/L7 the cobalt oxidic shell decreases and it can be observed
(Co/L2) that only free metallic Co is in tight contact to the graphitic
or graphene layers.
Finally, TPR measurements were performed to get insight into
the reducibility of the catalytic materials. As shown in Fig. 5, Co/
L1, Co/L2 and Co/L3 showed a very similar behavior. In contrast,
Co/L7 displayed a peak area much smaller than the previous ones.
This is ascribed to a greater metallic contribution in the NP spot.
This result is in agreement with that obtained from the XPS and
STEM studies for the same catalysts.
5
33 eV can be ascribed to all these compounds [73].
To evaluate morphological differences caused by the use of dif-
ferent ligands, high resolution scanning transmission electron
microscopy (STEM) technique was exploited. Parallel high angle
annular dark field (HAADF) and annular bright field (ABF) images
were taken. The HAADF images are sensitive to differences in the
atomic number (Z) of the atoms present in the sample. In the cur-
rent case, oxidic cobalt phases show less contrast than metallic
cobalt. In combination with EDX measurements and indexing of
high resolution images, it is possible to describe the Co containing
phase nature. Since HAADF prefers the imaging of heavy elements,
ABF has to be used for light elements like carbon. By using this
technique, graphitic structures at the top of nanoparticles were
observed. Fig. 4 shows typical images for catalysts Co/L1, Co/L2,
Co/L3 and Co/L7. The different appearance of particles with metal-
lic core and oxidic shells can be easily seen. Whereas Co/L1 con-
tains completely covered particles, the oxidic coverage decreases
for Co/L2 and Co/L3. Co/L7 shows a different morphology. Here,
the Co-containing phase is split into agglomerates of small oxidic
and bigger nearly uncovered metallic particles. As shown by the
ABF images graphitic structures (graphene layers) appear at the
In conclusion, most of the prepared materials exhibit a core-
shell structure [74] in which metallic and oxidic species (CoO)
coexist. In addition, nitrogen-doped graphitic and graphene-type
layers in contact with these NP are observed.
3.2. Evaluation of the catalytic performances
The catalytic performance of the prepared materials was
primarily explored in the hydrogenation of nitrobenzene to aniline.
Aniline and its derivatives are key intermediates for the

D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
83
Co/L1
Co/L2
Co/L3
Co/L7
Co/L7
Co/L2
Co/L3
Co/L1
Fig. 4. STEM images of the selected catalysts. HAADF and ABF images are depicted in the upper and lower row, respectively. For supplementary STEM images and EDX
patterns see SI (Figs. S5–S14).
previous findings, organic bases are beneficial for achieving higher
conversions [79]. Thus, different base additives were tested
(Table S7) in combination with catalyst Co/L2. Inorganic bases
led to full conversion but the selectivity into the desired aniline
dropped owing to the formation of azobenzene and azoxybenzene
as side products. On the contrary, when Et N was employed, com-
3
plete conversion and full selectivity were achieved, thus avoiding
the formation of undesired side products. For a complete investiga-
tion of the catalytic system, acidic additives were also tested
4 3
(Table S8). Two Brønsted (HBF and CF COOH) and two Lewis acids
(
Al(OTf) and Zn(OTf) ) were used at various concentrations. In all
3
2
cases a slight decrease in the conversion and a drop in the selectiv-
ity were detected. The decline in the selectivity is attributed to the
formation of 2-ethoxy- and 4-ethoxyaniline. Control experiments
carried out under a nitrogen atmosphere without the catalyst
4
and in the presence of HBF (2 equivalents) revealed that nitroben-
zene, nitrosobenzene or aniline do not react. On the contrary, N-
phenylhydroxylamine under the same reaction conditions led to
the formation of 2-ethoxyaniline, 4-ethoxyaniline and azoxyben-
zene in 49, 9, and 29% yield, respectively. This behavior was previ-
ously reported in the case of both Brønsted [80,81] or Lewis [82]
acids. Having good conditions in our hands, the seven prepared
catalysts were tested in the benchmark hydrogenation of nitroben-
Fig. 5. TPR of selected catalysts.
pharmaceutical, agrochemical and dye industries [75]. In the last
century, classic stoichiometric manufacturing processes (e.g.
Béchamp reduction or sulfur-based reduction reactions) were
replaced by more economical catalytic protocols [76,77]. Although
the hydrogenation of simple substituted nitro compounds does not
pose significant selectivity problems, the situation is more chal-
lenging for substrates carrying reducible or poisoning-capable
functional groups [78]. Therefore, the development of efficient,
robust, cheap and green catalysts for the selective reduction of
nitro compounds continues to attract the attention of academic
and industrial researchers. Initially, Co/L1 was employed for the
hydrogenation of nitrobenzene in the presence of different sol-
vents. While apolar solvents proved to be detrimental for the reac-
tivity, polar ones allowed for higher conversions (Table S6). In
particular, EtOH was found to be the best solvent. The addition
of a small amount of water (1:20 by volume) further increased
nitrobenzene conversion. We were pleased to obtain complete
selectivity toward the desired aniline (Fig. S15). In agreement with
zene to aniline with and without the addition of Et N (Fig. 6).
3
More specifically, Co/L7 displayed the maximum activity fol-
lowed by Co/L3, Co/L2, Co/L6, Co/L5, Co/L4 and finally Co/L1. Inter-
estingly, this trend is retained even if Et N was added, which
3
indicates that the base acts as promoter but leaves the catalytic
material unaffected. In all cases selectivities were very high
(Fig. S15b and d) and no side products such as nitrosobenzene,
azobenzene or azoxybenzene were detected at the end of the cat-
alytic experiments. This observation, coupled with the fact that
selectivities into aniline increased with increasing reaction time,
clearly demonstrates that phenylhydroxylamine is an intermediate
in the present catalytic transformation.
Next, the reuse of Co/L7 was investigated in a sixfold scaled up
reaction showing a good recyclability (Fig. 7).
A slight decrease in the conversion was observed after the first
and second recycling. However, after the second run the catalytic

8
4
D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
Fig. 6. Variation of [PhNO
O, 50 bar H , 110 °C.
2 3 2
] during the reaction without (a) and with (b) 1 equiv. of Et N. Reaction conditions: 0.5 mmol PhNO , 0.7 mol% Co, solvent: 2 mL EtOH + 100 lL
H
2
2
were well tolerated. Other reduction-labile functional groups such
as double (2j) or triple (2k) carbon-carbon bonds, nitriles (2l),
ketones (2m), simple and conjugated esters (2n, 2r) and amides
(2o) are well tolerated affording the corresponding anilines in very
high yields. Heteroaromatic nitro compounds were hydrogenated
furnishing the corresponding anilines in good (2s) and moderate
(2q) yields. An excellent selectivity was also achieved in the reduc-
tion of 2,4-dinitrotoluene (1t) into the corresponding diamine.
Such hydrogenation is currently employed yearly on a multi-
million tons scale as
toluendiisocyanate [4]. As mentioned earlier, the addition of 1
equivalent of Et N boosts the activity of the catalytic system (see
l and 1q). The pharmaceutically important substrates nitroresor-
a step in the preparation of 2,4-
3
1
cinol (1aa) and Flutamide (1ab) were also investigated in the cat-
alytic hydrogenation (Fig. 9). The corresponding reduction
products are key-intermediates in the preparation of biologically
active compounds and are valuable starting molecules in total syn-
theses [84–90].
Fig. 7. Recycling experiments of Co/L7. Reaction conditions: 3 mmol PhNO
.7 mol% Co, solvent: 12 mL EtOH + 600 L H O, 8 h, 50 bar H , 110 °C.
2
,
2
In the case of 1aa, the reaction was slow using EtOH-H O 20:1
0
l
2
2
as solvent mixture but proceeded faster in a reaction media almost
exclusively composed of water. This is likely to be attributed to the
very strong intramolecular H-bond between the partially positively
charged H of the OH groups and the two negatively-charged oxy-
gen atoms of the adjacent nitro group [91]. However, it was not
possible to conduct the catalytic transformation in pure water as
the solubility of the substrate was insufficient in this case. This
reaction is the first example of a non-noble metal catalyzed reduc-
tion of 2-nitroresorcinol to 2-aminoresorcinol. Similarly, hydro-
genation of 1ab proceeded with excellent chemoselectivity and
its purification does not require expensive and waste-producing
chromatographic separation techniques. In fact, just filtration over
CeliteÒ and consecutive solvent evaporation were needed to obtain
the product in a pure form at the end of the reaction.
performance remained constant. In addition, the selectivity was
completely retained at a value of >99% throughout the 5 recycling
experiments. ICP analysis of the liquid phase after each recycle
indicated no leaching of Co (detection limit 0.5 ppm). In addition,
Maitlis’ hot filtration test was carried out in order to detect
whether soluble active Co species were present to catalyze the
reaction (Table S10). The results confirmed that no active cobalt
was leached. Thus, we conclude that the employed catalyst is
inherently heterogeneous.
Subsequently, the reaction scope was explored employing vari-
ous substituted aromatic nitro compounds as substrates (Fig. 8).
The catalytic system is excellently tolerant to halogen contain-
ing substrates. The presence of Br, Cl or F did not affect the selec-
tivity and the various nitro compounds were smoothly converted
into corresponding anilines (2a, 2b and 2c). Only in the case of
3.3. Kinetic studies
4
-iodonitrobenzene (1d), dehalogenation occurred producing ani-
In order to understand the different catalytic behaviors of the
seven employed catalysts, kinetic experiments were carried out.
The collected data, summarized in Fig. 10, are in accordance with
a first order kinetics in PhNO2.
Despite the structural and morphological similarity of the pre-
pared materials, the bulk N content (wt.%) varies from catalyst to
catalyst (see Table S1). Since control experiments confirmed that
the catalyst prepared in the absence of the nitrogen ligand is totally
line and N-ethylaniline in 87% and 13% yield, respectively. The for-
mation of N-ethylaniline is likely to be caused by in situ formation
of EtI from ethanol (solvent) and HI derived from the dehalogena-
tion followed by alkylation of the amino group [83]. Notably, 3-
iodonitrobenzene (1e), did not show this undesired pathway
affording the desired product in excellent yield. Electron donating
substituents (2f, 2g, 2u, 2x, 2y, 2z) and steric hindrance (2h, 2i)

D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
85
Fig. 8. Hydrogenation of nitroarenes: substrate scope. Reaction conditions: 0.5 mmol nitroarene, 2 mL EtOH + 100
using n-hexadecane as internal standard; 1.4 mol% of Co/L7 was used; 1 equiv. (with respect to the substrate) of Et N was added.
3
l
L H
2
O. Complete conversions were observed. GC yields
b
c

8
6
D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
Fig. 9. Co/L7 catalyzed synthesis of biologically important molecules. Yields refer to isolated compounds.
Fig. 10. First order kinetic plot of the seven catalysts without (a) and with (b) 1 equiv. of Et
3
N. Reaction conditions are reported in the caption of Fig. 6.
inactive (Table S11, entry 5), we assumed that the activities of the
catalysts correlate with the bulk nitrogen content in the final cat-
alytic material. In fact, plotting the kinetic constant vs the total
nitrogen content in the catalytic material, a clear trend is observed
(
Fig. 11): the higher the nitrogen content, the better the activity of
the catalyst. The same trend is retained if the reactions were
carried out in the presence of 1 equivalent of Et N (see also
3
Fig. S16). In this case, kinetic constants were dramatically
improved (Table S9).
This general trend is observed for all catalysts except for Co/L5.
This can be attributed to the steric hindrance of the naphthyl
groups that probably negatively affect the coordination ability of
the ligand to the metal center. Thus, the N atoms are equally incor-
porated into the carbon matrix but not close to the catalytically
active site. In fact, the catalyst prepared using a 1:1 M ratio of Co
(
OAc)
2
and L5 exhibited a similar activity to Co/L5. Conversely,
and L1 (less sterically
the catalyst prepared by combining Co(OAc)
2
hindered) in a 1:1 M ratio showed a much lower activity (Fig. S19).
In addition, for the selected characterized materials, a correlation
with the free pyridinic N content measured by XPS was established
(Figs. 12 and S17).
Since no similar correlations were observed in the case of
Fig. 11. Correlation between first-order kinetic constant and total nitrogen content
pyrrolic/Co-N
x
configurations (Fig. S18), we conclude that
in the catalyst material.

D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
87
uncoordinated pyridinic N atoms play a pivotal role in the catalytic
activity (graphitic or N-oxidic configurations were not taken into
account since they were not found in all the catalysts). A related
observation has been described for NGr-based catalysts for electro-
chemical applications [24,92–94]. Investigations of the reaction
order with respect to Et
the model catalyst indicate a first-order kinetics in Et
Figs. 13 and 14).
First order kinetics with respect to nitrobenzene, dihydrogen
3
N and hydrogen pressure using Co/L2 as
3
N and H
2
(
and, if present, triethylamine cannot be explained by a single slow
step without the existence of at least one kinetically relevant equi-
librium stage before the r.d.s. Given that activation of nitrobenzene
is very unlikely to be reversible (nitroarenes are very poor ligands
and their metal complexes are extremely labile), the most likely
explanation for the observed kinetic data is that reversible forma-
tion of an activated form of dihydrogen first occurs, the equilibrium
being strongly shifted to the reagents side (if this were not true the
2 3
kinetics would deviate from first order for both H and Et N. See the
Supporting Information for a detailed discussion). The activated
dihydrogen then reacts in the r.d.s. with nitrobenzene. Based on
reactivity data, we postulate that this kind of hydrogenations pro-
ceeds by a heterolytic activation of dihydrogen [95–97]. The kinetic
data obtained in this work not only confirm the general validity of
this explanation for this class of catalysts, but also add further. In
particular, the linear dependence of the reaction rate on triethy-
lamine concentration with a nonzero intercept points to a reaction
scenario where two independent and competing mechanisms
occur, in which either a nitrogen atom of the support or that of
Fig. 12. Correlation between first order kinetic constant and free pyridinic N
content in the final material.
3
Et N is involved, but not both. Based on the kinetic investigations,
we propose the pyridinic N content is crucial for the catalyst
activity. The fact that the dihydrogen activation equilibrium is
shifted toward the left is not only consistent with the kinetic orders
observed, but also explains why the rate of the Et N-free reaction is
3
not affected by triethylamine concentration. Indeed, at any given
moment the active sites are virtually all available independent of
the Et N concentration. A schematic representation of the initial
3
steps of the reaction is shown in Scheme 1.
Recent both experimental [25,45,98–101] and theoretical [102]
works dealing with transition metal/NGr catalysts indicate that the
nature of the active sites can be a M-N -Cy network in which the
x
metal is present as a single atom (single site heterogeneous
catalyst). According to our structural characterization we postulate
the involvement of either single-site Co centers or Co NP
coordinated with pyridinic N atoms whose configurations are
not definitely elucidated. The established correlation between
non-coordinated pyridinic N configurations and activity (Fig. 10)
suggests an internal base role by the latter. The relationship
between activity and amount of pyridinic N can be ascribed to its
basicity. In agreement with this proposal, a recent study described
that in NGr-based materials, the pyridinic N exhibits the maximum
basicity among the three types of N (the other two are pyrrolic and
graphitic) [103]. These conclusions are corroborated by control
experiments. Indeed, it must be stressed that the contemporary
presence of both a cobalt source and a nitrogen ligand before the
pyrolysis step is required to obtain active catalysts. The materials
prepared either without the addition of the N-ligand or the metal
3
Fig. 13. First order kinetic plot for Et N (Co/L2 as catalyst).
(Table S11, entries 6–8) were inactive, whether Et
3
N was also
present or not [104].
Finally, the most active catalyst (Co/L7) was compared with
commercially available noble-metal based heterogeneous catalysts
(
Tables S13 and S14). Two model substrates were employed for
that purpose: (E)-4-nitrostilbene (1j) and 3-nitroacetophenone
1m). In both cases, Co/L7 showed superior chemoselectivity
(
toward the formation of the desired products, while several side-
and over-reduction products were detected when the reactions
were carried out with standard noble-metal based catalysts.
Fig. 14. First order kinetic plot for hydrogen pressure (Co/L2 as catalyst).

8
8
D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
Scheme 1. Proposed initial reaction steps.
[
2] C. Descorme, P. Gallezot, C. Geantet, C. George, ChemCatChem 4 (2012) 1897–
906.
4
. Conclusions
1
[
3] M. Beller, A. Renken, R.A. Van Santen, Catalysis: From Principles to
Applications, Wiley VCH-Verlag, 2012.
In conclusion, an array of Co-based core-shell NGr decorated
carbon-supported catalysts was prepared from the pyrolysis of
well-defined Ar-BIAN and -diimine Co complexes. The catalyti-
[4] H.-J. Arpe, Industrial Organic Chemistry, Wiley-VCH, 2010.
[
[
5] J. Hagen, Industrial Catalysis: A Practical Approach, Wiley-VCH Verlag, 2015.
6] M.L.C.M. Henckens, P.P.J. Driessen, E. Worrell, Resour. Conserv. Recyc. 93
a
cally active materials were successfully applied to the reduction
of nitrobenzene to aniline using molecular hydrogen as the reduc-
tant. Furthermore, excellent chemoselectivities were achieved
employing these catalysts in the hydrogenation of nitro com-
pounds bearing reduction-labile functional groups. Straightfor-
(
2014) 1–8.
[7] R.M. Bullock, Catalysis without precious metals, Wiley-VCH Verlag, 2010.
[
[
8] M.S. Holzwarth, B. Plietker, ChemCatChem 5 (2013) 1650–1679.
9] J.-P.D. Zhongwei Chen, Jiujun Zhang, Non-Noble Metal Fuel Cell Catalysts,
Wiley-VCH, 2014.
[10] B. Su, Z.-C. Cao, Z.-J. Shi, Acc. Chem. Res. 48 (2015) 886–896.
[
11] K.H. Wedepohl, The composition of earth’s upper crust, natural cycles of
elements, natural resources, Elements and Their Compounds in the
Environment, Wiley-VCH Verlag GmbH, 2008, pp. 2–16.
ward kinetic measurements suggest
a
direct and linear
correlation between the intrinsic activity of the catalysts (in terms
of the kinetic constant) and the total amount of nitrogen in the
support. An additional correlation between the activity and the
pyridinic nitrogen content in the support was further demon-
strated. For all employed catalysts, a dramatic increase in the activ-
ity (up to almost three times) was observed when an organic base
was added to the reaction mixture. All our findings are in accor-
dance with a polar reaction mechanism in which a heterolytic acti-
vation of the hydrogen molecule occurs.
We believe that the findings in this paper constitute a valuable
source of inspiration for the design of new high performance mate-
rials that enable the hydrogenation of challenging substrates such
as carboxylic acids and amides.
[12] A. Schaetz, M. Zeltner, W.J. Stark, ACS Catal. 2 (2012) 1267–1284.
[
[
[
13] B.F. Machado, P. Serp, Catal. Sci. Technol. 2 (2012) 54–75.
14] J. Albero, H. Garcia, J. Mol. Catal. A: Chem. 408 (2015) 296–309.
15] R. Zhou, M. Jaroniec, S.-Z. Qiao, ChemCatChem 7 (2015) 3808–3817.
[16] C.N.R. Rao, K. Gopalakrishnan, A. Govindaraj, Nano Today 9 (2014) 324–343.
[
[
[
17] X. Fan, G. Zhang, F. Zhang, Chem. Soc. Rev. 44 (2015) 3023–3035.
18] S.R. Stoyanov, A.V. Titov, P. Král, Coord. Chem. Rev. 253 (2009) 2852–2871.
19] H. Liu, Y. Liu, D. Zhu, J. Mater. Chem. 21 (2011) 3335–3345.
[20] H. Wang, T. Maiyalagan, X. Wang, ACS Catal. 2 (2012) 781–794.
[21] J. Masa, W. Xia, M. Muhler, W. Schuhmann, Angew. Chem. Int. Ed. 54 (2015)
1
0102–10120.
22] L. He, F. Weniger, H. Neumann, M. Beller, Angew. Chem. Int. Ed. 55 (2016)
2582–12594.
[23] Q. Lai, Q. Gao, Q. Su, Y. Liang, Y. Wang, Z. Yang, Nanoscale 7 (2015) 14707–
4714.
24] F. Jaouen, S. Marcotte, J.-P. Dodelet, G. Lindbergh, J. Phys. Chem. B 107 (2003)
376–1386.
[25] J. Han, Y.J. Sa, Y. Shim, M. Choi, N. Park, S.H. Joo, S. Park, Angew. Chem. Int. Ed.
4 (2015) 12622–12626.
26] Y. Su, Y. Zhu, H. Jiang, J. Shen, X. Yang, W. Zou, J. Chen, C. Li, Nanoscale 6
2014) 15080–15089.
[27] H.-W. Liang, S. Bruller, R. Dong, J. Zhang, X. Feng, K. Mullen, Nat. Commun. 6
2015).
[
1
1
[
1
Acknowledgments
5
[
The work was supported by the state of Mecklenburg-
Vorpommern (Germany), Ministero dell’Università e della Ricerca
(
(
(
MIUR, Italy, PRIN 2015_20154X9ATP) and F. Hoffmann-La Roche.
The authors thank Mrs. Astrid Lehmann and Mrs. Anja Simmula
Analytic Department, LIKAT) for excellent technical assistance.
[
28] C.H. Choi, C. Baldizzone, J.-P. Grote, A.K. Schuppert, F. Jaouen, K.J.J. Mayrhofer,
Angew. Chem. Int. Ed. 54 (2015) 12753–12757.
[29] D. Chauhan, P. Kumar, C. Joshi, N. Labhsetwar, S.K. Ganguly, S.L. Jain, New J.
Chem. 39 (2015) 6193–6200.
30] J. Xi, Y. Xia, Y. Xu, J. Xiao, S. Wang, Chem. Commun. 51 (2015) 10479–10482.
31] C.E. Chan-Thaw, A. Villa, P. Katekomol, D. Su, A. Thomas, L. Prati, Nano Lett. 10
(2010) 537–541.
32] X. Ning, H. Yu, F. Peng, H. Wang, J. Catal. 325 (2015) 136–144.
33] J. Deng, H.-J. Song, M.-S. Cui, Y.-P. Du, Y. Fu, Chemsuschem 7 (2014) 3334–
(
Special thank is due to Dr. Hanan Atia (LIKAT) for theoretical and
technical assistance regarding the TPR measurements. D.F. and F.
F. thank Erasmus+ Programme and Milan University for a fellow-
ship, respectively.
[
[
[
[
3340.
Appendix A. Supplementary material
[34] Y. Chen, L. Fu, Z. Liu, Chem. Commun. 51 (2015) 16637–16640.
[35] Y. Chen, S. Zhao, Z. Liu, PCCP 17 (2015) 14012–14020.
[36] A.V. Iosub, S.S. Stahl, Org. Lett. 17 (2015) 4404–4407.
[37] C. Bai, A. Li, X. Yao, H. Liu, Y. Li, Green Chem. 18 (2016) 1061–1069.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2017.04.014.
[38] C. Yang, L. Fu, R. Zhu, Z. Liu, PCCP 18 (2016) 4635–4642.
[
39] S. Sahin, P. Mäki-Arvela, J.-P. Tessonnier, A. Villa, S. Reiche, S. Wrabetz, D. Su,
R. Schlögl, T. Salmi, D.Y. Murzin, Appl. Catal. A 408 (2011) 137–147.
40] P. Chen, F. Yang, A. Kostka, W. Xia, ACS Catal. 4 (2014) 1478–1486.
41] J. Long, Y. Zhou, Y. Li, Chem. Commun. 51 (2015) 2331–2334.
References
[
[
[
1] R.A. Sheldon, I. Arendes, U. Hanefeld, Green Chemistry and Catalysis, Wiley-
VCH Verlag, 2007.
[42] Z. Wei, J. Wang, S. Mao, D. Su, H. Jin, Y. Wang, F. Xu, H. Li, Y. Wang, ACS Catal. 5
(2015) 4783–4789.

D. Formenti et al. / Journal of Catalysis 351 (2017) 79–89
89
[
[
43] Z. Li, J. Li, J. Liu, Z. Zhao, C. Xia, F. Li, ChemCatChem 6 (2014) 1333–1339.
44] Y. Marco, L. Roldán, S. Armenise, E. García-Bordejé, ChemCatChem 5 (2013)
[77] D.J. Dixon, O. Pando Morejón, Recent developments in the reduction of nitro
and nitroso compounds, in: P. Knochel (Ed.), Comprehensive Organic
Synthesis II, second ed., Elsevier, Amsterdam, 2014, pp. 479–492.
[78] H.-U. Blaser, H. Steiner, M. Studer, ChemCatChem 1 (2009) 210–221.
[79] D. Formenti, C. Topf, K. Junge, F. Ragaini, M. Beller, Catal. Sci. Technol. 6 (2016)
4473–4477.
[80] E. Bamberger, J. Lagutt, Chem. Ber. 31 (1898) 1500–1508.
[81] H. Lund, Cathodic reduction of nitro and related compounds, in: O.H. Henning
Lund (Ed.), Organic Electrochemistry, CRC Press, New York, 2000.
[82] E. Müller, E. Lindemann, Angew. Chem. 46 (1933) 681–685.
[83] Y.-S. Choi, Y.-J. Kim, L.-L. Shen, Y.S. Lee, J.-H. Jeong, Synlett 26 (2015) 970–974.
[84] A. Urios, M. Largeron, M.-B. Fleury, M. Blanco, Free Radical Biol. Med. 40
(2006) 791–800.
3
829–3834.
[
45] L. Zhang, A. Wang, W. Wang, Y. Huang, X. Liu, S. Miao, J. Liu, T. Zhang, ACS
Catal. (2015) 6563–6572.
46] D. Li, D. Shi, Z. Chen, H. Liu, D. Jia, Z. Guo, RSC Adv. 3 (2013) 5003–5008.
47] H. Kim, A.W. Robertson, S.O. Kim, J.M. Kim, J.H. Warner, ACS Nano 9 (2015)
[
[
5
947–5957.
48] W. Bao, B. Yu, W. Li, H. Fan, J. Bai, Z. Ren, J. Alloys Compd. 647 (2015) 873–
79.
49] L. Duan, Y. Chen, K. Zhang, H. Luo, J. Huang, A. Xu, RSC Adv. 5 (2015) 84303–
4310.
50] D. Su, J. Wang, H. Jin, Y. Gong, M. Li, Z. Pang, Y. Wang, J. Mater. Chem. A 3
2015) 11756–11761.
[
[
[
[
8
8
(
[85] H. Gao, J. Kawabata, Bioorg. Med. Chem. Lett. 18 (2008) 812–815.
[86] E. Kato, K. Oikawa, K. Takahashi, J. Kawabata, Biosci. Biotechnol. Biochem. 76
(2012) 1044–1046.
51] L. Jia, D.A. Bulushev, O.Y. Podyacheva, A.I. Boronin, L.S. Kibis, E.Y. Gerasimov,
S. Beloshapkin, I.A. Seryak, Z.R. Ismagilov, J.R.H. Ross, J. Catal. 307 (2013) 94–
1
02.
[87] E. Kato, H. Tsuji, J. Kawabata, Tetrahedron 71 (2015) 1419–1424.
[88] B. Wen, K.J. Coe, P. Rademacher, W.L. Fitch, M. Monshouwer, S.D. Nelson,
Chem. Res. Toxicol. 21 (2008) 2393–2406.
[
52] R.V. Jagadeesh, H. Junge, M.-M. Pohl, J. Radnik, A. Brückner, M. Beller, J. Am.
Chem. Soc. 135 (2013) 10776–10782.
[
53] F.A. Westerhaus, R.V. Jagadeesh, G. Wienhöfer, M.-M. Pohl, J. Radnik, A.-E.
Surkus, J. Rabeah, K. Junge, H. Junge, M. Nielsen, A. Brückner, M. Beller, Nat.
Chem. 5 (2013) 537–543.
[89] W. Herath, I.A. Khan, Chem. Pharm. Bull. 58 (2010) 562–564.
[90] C. Udagawa, S. Fukuyoshi, S. Morimoto, Y. Tanimoto, R. Nakagaki, J.
Photochem. Photobiol., A 226 (2011) 57–63.
[
54] D. Banerjee, R.V. Jagadeesh, K. Junge, M.-M. Pohl, J. Radnik, A. Brückner, M.
Beller, Angew. Chem. Int. Ed. 53 (2014) 4359–4363.
55] R.V. Jagadeesh, H. Junge, M. Beller, Nat. Commun. 5 (2014).
56] T. Stemmler, F.A. Westerhaus, A.-E. Surkus, M.-M. Pohl, K. Junge, M. Beller,
Green Chem. 16 (2014) 4535–4540.
57] R.V. Jagadeesh, D. Banerjee, P.B. Arockiam, H. Junge, K. Junge, M.-M. Pohl, J.
Radnik, A. Bruckner, M. Beller, Green Chem. 17 (2015) 898–902.
58] R.V. Jagadeesh, T. Stemmler, A.-E. Surkus, M. Bauer, M.-M. Pohl, J. Radnik, K.
Junge, H. Junge, A. Bruckner, M. Beller, Nat. Protoc. 10 (2015) 916–926.
59] S. Pisiewicz, T. Stemmler, A.-E. Surkus, K. Junge, M. Beller, ChemCatChem 7
[91] K.B. Borisenko, I. Hargittai, J. Phys. Chem. 97 (1993) 4080–4084.
[92] P.H. Matter, L. Zhang, U.S. Ozkan, J. Catal. 239 (2006) 83–96.
[93] G. Faubert, R. Côté, J.P. Dodelet, M. Lefèvre, P. Bertrand, Electrochim. Acta 44
(1999) 2589–2603.
[94] D. Guo, R. Shibuya, C. Akiba, S. Saji, T. Kondo, J. Nakamura, Science 351 (2016)
361–365.
[95] M. Fang, N. Machalaba, R.A. Sanchez-Delgado, Dalton Trans. 40 (2011)
10621–10632.
[96] M. Fang, R.A. Sánchez-Delgado, J. Catal. 311 (2014) 357–368.
[97] A. Sánchez, M. Fang, A. Ahmed, R.A. Sánchez-Delgado, Appl. Catal. A 477
(2014) 117–124.
[98] U.I. Kramm, I. Herrmann-Geppert, J. Behrends, K. Lips, S. Fiechter, P.
Bogdanoff, J. Am. Chem. Soc. 138 (2016) 635–640.
[99] A. Zitolo, V. Goellner, V. Armel, M.-T. Sougrati, T. Mineva, L. Stievano, E. Fonda,
F. Jaouen, Nat. Mater. 14 (2015) 937–942.
[100] T. Cheng, H. Yu, F. Peng, H. Wang, B. Zhang, D. Su, Catal. Sci. Technol. 6 (2016)
1007–1015.
[101] W. Liu, L. Zhang, W. Yan, X. Liu, X. Yang, S. Miao, W. Wang, A. Wang, T. Zhang,
Chem. Sci. 7 (2016) 5758–5764.
[
[
[
[
[
[
[
[
[
[
(
2015) 62–64.
60] F.A. Westerhaus, I. Sorribes, G. Wienhöfer, K. Junge, M. Beller, Synlett 26
2015) 313–317.
(
61] T. Stemmler, F. Chen, S. Pisiewicz, A.E. Surkus, M.M. Pohl, C. Topf, M. Beller, J.
Mater. Chem. A 3 (2015) 17728–17737.
62] F. Chen, A.-E. Surkus, L. He, M.-M. Pohl, J. Radnik, C. Topf, K. Junge, M. Beller, J.
Am. Chem. Soc. 137 (2015) 11718–11724.
63] M. Viganò, F. Ferretti, A. Caselli, F. Ragaini, M. Rossi, P. Mussini, P. Macchi,
Chem. Eur. J. 20 (2014) 14451–14464.
64] M. Viganò, F. Ragaini, M.G. Buonomenna, R. Lariccia, A. Caselli, E. Gallo, S.
Cenini, J.C. Jansen, E. Drioli, ChemCatChem 2 (2010) 1150–1164.
65] M. Gasperini, F. Ragaini, S. Cenini, Organometallics 21 (2002) 2950–2957.
66] C. Graebe, E. Gfeller, Justus Liebigs Ann. der Chem. 276 (1893) 1–20.
67] H.C. Rai, B.K. Rai, Asian J. Chem. 13 (2001) 264–268.
[102] C.E. Szakacs, M. Lefevre, U.I. Kramm, J.-P. Dodelet, F. Vidal, PCCP 16 (2014)
13654–13661.
[103] B. Li, X. Sun, D. Su, PCCP 17 (2015) 6691–6694.
[
[
[
[
[104] It should be stressed that the conclusions on the identity of the active site
reached in this paper holds for the presently discussed system, although they
may be more general. However, they are not supposed to be universally true
for any catalyst derived by a thermal degradation of a metal complex
precursor. Graphitic catalysts have also been generated by other groups with
different procedures (see cited references), but their structure can be very
different from that of ours. For example, very recently, Corma and coworkers
investigated a catalyst prepared by reduction under dihydrogen of a material
obtained by reacting cobalt nitrate with EDTA [L. Liu, P. Conception, A. Corma,
J. Catal., 140 (2016) 1–9]. The catalytic activity was attributed to metallic
cobalt and the role of carbon was identified as that of preventing metal over-
oxidation, a conclusion opposite to that reached by us. However, Corma’s
catalyst and ours are markedly different. Most notable differences are: (a)
Corma’s catalyst contains much lesser carbon than ours (cobalt constitutes
95% of the mass of his catalyst, whereas it is ꢀ3% in ours); (b) the carbon is
mostly graphitic in our catalyst and mostly amorphous in Corma’s catalyst;
(c) most importantly, no nitrogen is present in Corma’s catalyst after it has
been activated, whereas nitrogen plays a key role in our system. Thus, it is
clear that the two catalysts are not comparable, even if similarities exist in
their preparation, and it is not surprising that the catalytically active sites
differ.
68] F. Ragaini, M. Gasperini, P. Parma, E. Gallo, N. Casati, P. Macchi, New J. Chem.
3
0 (2006) 1046–1057.
69] N.A.M. Barakat, B. Kim, S.J. Park, Y. Jo, M.-H. Jung, H.Y. Kim, J. Mater. Chem. 19
2009) 7371–7378.
70] L. Zhang, P. Hu, X. Zhao, R. Tian, R. Zou, D. Xia, J. Mater. Chem. 21 (2011)
8279–18283.
[
[
[
(
1
71] F. Jaouen, J. Herranz, M. Lefèvre, J.-P. Dodelet, U.I. Kramm, I. Herrmann, P.
Bogdanoff, J. Maruyama, T. Nagaoka, A. Garsuch, J.R. Dahn, T. Olson, S.
Pylypenko, P. Atanassov, E.A. Ustinov, ACS Appl. Mater. Interf. 1 (2009) 1623–
1
639.
[
[
[
[
72] Q. He, Q. Li, S. Khene, X. Ren, F.E. López-Suárez, D. Lozano-Castelló, A. Bueno-
López, G. Wu, J. Phys. Chem. C 117 (2013) 8697–8707.
73] M.C. Biesinger, B.P. Payne, A.P. Grosvenor, L.W.M. Lau, A.R. Gerson, R.S.C.
Smart, Appl. Surf. Sci. 257 (2011) 2717–2730.
74] M.B. Gawande, A. Goswami, T. Asefa, H. Guo, A.V. Biradar, D.-L. Peng, R. Zboril,
R.S. Varma, Chem. Soc. Rev. 44 (2015) 7540–7590.
75] T. Kahl, K.-W. Schröder, F.R. Lawrence, W.J. Marshall, H. Höke, R. Jäckh,
Aniline, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag
GmbH & Co. KGaA, 2000.
[
76] A.R. Cartolano, G.A. Vedage, Amines by reduction, Kirk-Othmer Encyclopedia
of Chemical Technology, John Wiley & Sons Inc, 2000.