Chemoselective reduction of quinoline over Rh-C60 nanocatalysts

Article abstract of DOI:10.1039/c9cy02025j

The design and engineering of heterogeneous nanocatalysts that are both highly active and selective for hydrogenation reactions constitute a crucial challenge. In that context, herein a series of Rh-C₆₀ nanocatalysts have been synthesized via the decomposition of an organometallic rhodium complex in the presence of fullerene C₆₀ under a H₂ atmosphere. Rhodium atomically dispersed or rhodium nanoparticles on Rh-C₆₀ spherical fulleride particles were produced by tuning the Rh/C₆₀ molar ratio. Significant charge transfer between rhodium and C₆₀ was evidenced through Raman and X-ray photoelectron spectroscopy, which indicates electron-deficient Rh species. The resulting heterostructured nanomaterials were applied successfully in the catalytic hydrogenation of quinoline, exhibiting excellent activity and producing selectively the partially hydrogenated product, 1,2,3,4-tetrahydroquinoline. Density functional theory (DFT) calculations show that the hydride coverage of the Rh NPs plays a key role in the adsorption modes of quinoline and 1,2,3,4-tetrahydroquinoline on the surface of the NPs, and that these adsorption modes are modulated by the presence of fullerene C₆₀, thus affecting the activity and selectivity obtained with this rhodium based catalyst.

Full text of DOI:10.1039/c9cy02025j

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Chemoselective reduction of quinoline over Rh–
C60 nanocatalysts†
Cite this: Catal. Sci. Technol., 2019,
9, 6884
Zhishan Luo,a Yuanyuan Min,a Divya Nechiyil,b Wolfgang Bacsa,b Yann Tison,c
d
Hervé Martinez, c Pierre Lecante, b Iann C. Gerber,
a
Philippe Serpa and M. Rosa Axet
*
The design and engineering of heterogeneous nanocatalysts that are both highly active and selective for
hydrogenation reactions constitute a crucial challenge. In that context, herein a series of Rh–C60
nanocatalysts have been synthesized via the decomposition of an organometallic rhodium complex in the
presence of fullerene C60 under a H2 atmosphere. Rhodium atomically dispersed or rhodium nanoparticles
on Rh–C60 spherical fulleride particles were produced by tuning the Rh/C60 molar ratio. Significant charge
transfer between rhodium and C60 was evidenced through Raman and X-ray photoelectron spectroscopy,
which indicates electron-deficient Rh species. The resulting heterostructured nanomaterials were applied
successfully in the catalytic hydrogenation of quinoline, exhibiting excellent activity and producing
selectively the partially hydrogenated product, 1,2,3,4-tetrahydroquinoline. Density functional theory (DFT)
calculations show that the hydride coverage of the Rh NPs plays a key role in the adsorption modes of
quinoline and 1,2,3,4-tetrahydroquinoline on the surface of the NPs, and that these adsorption modes are
modulated by the presence of fullerene C60, thus affecting the activity and selectivity obtained with this
rhodium based catalyst.
Received 8th October 2019,
Accepted 30th October 2019
DOI: 10.1039/c9cy02025j
rsc.li/catalysis
particular,
quinolines
allow
the
preparation
of
Introduction
tetrahydroquinoline products that are broadly found in drug
molecules.14,16 A simple and effective synthesis strategy is the
direct regioselective hydrogenation of quinoline (Scheme 1).
However, 5,6,7,8-tetrahydroquinoline (3) and the fully
hydrogenated product decahydroquinoline (4) may be also
produced.
A variety of heterogeneous catalysts have been developed
to improve activity and selectivity for this reaction in recent
years. Au,17 Rh,18–25 Ru,26,27 Pd (ref. 28) and Pt (ref. 29)-based
catalysts have been revealed to be very active and selective.
Specifically, Rh nanoparticles (NPs) show high selectivity
towards 1,2,3,4-tetrahydroquinoline (2). For instance,
rhodium NPs immobilized in a Lewis acidic ionic liquid (IL)
catalyse a large scope of heteroarenes, being very selective to
the arene ring without hydrogenating other functional
groups.23 In the specific case of quinoline, it was selectively
hydrogenated to 2 (95% yield) at 30 bar of H2 and 80 °C after
15 h of reaction. Other Rh catalysts have produced the partial
Heterogeneous catalytic hydrogenation reactions are broadly
known as one of the most fundamental chemical
transformations in both the laboratory and industry.1–5 Over
the years, numerous heterogeneous nanocatalysts, in
particular, carbon-based hybrid nanostructures, have shown
high catalytic activity and selectivity in hydrogenation
reactions,6–12 owing to their high surface area to volume ratio,
exclusive electronic properties to favour charge transfer, high
chemical and thermal stability, ease of recovery from the
reaction mixture and high mechanical strength. Catalytic
reduction of quinoline (1) is a key organic reaction where one
of the products, 1,2,3,4-tetrahydroquinoline (2), is an
important intermediate for pharmaceuticals, agrochemicals,
alkaloids and various biological active molecules.13–16 In
a LCC-CNRS, Université de Toulouse, CNRS, INPT, Toulouse, France.
E-mail: rosa.axet@lcc-toulouse.fr
b Centre d'Élaboration des Matériaux et d'Études Structurales UPR CNRS 8011, 29
Rue Jeanne-Marvig, BP 4347, 31055 Toulouse, France
c CNRS/UNIV PAU & PAYS ADOUR/E2S UPPA, Institut des Sciences Analytiques et
de Physicochimie pour l'Environnement et les Matériaux, UMR5254, 64000, Pau,
France
d LPCNO, Université de Toulouse, INSA-CNRS-UPS, 135 Avenue de Rangueil,
31077 Toulouse, France
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c9cy02025j
Scheme 1 Hydrogenation of quinoline.
6884 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
hydrogenated product
2
very selectively under similar
which evidenced charge transfer from rhodium to fullerene.
The nanostructures were used as catalysts in the
hydrogenation of quinoline, showing excellent activity (TOF =
488.0 h−1) and selectivity to 2 (up to 98%). Density functional
theory (DFT) calculations were also performed to support
these findings.
reaction conditions: Rh NPs over reduced graphene oxide
(99% yield; 30 bar of H2, 80 °C, 12 h);25 Rh NPs stabilized
with polyethylene glycol (99% selectivity; 30 bar of H2, 100
°C, 3 h);22 Rh NPs stabilized with a phosphine-functionalized
IL (99% selectivity; 30 bar of H2, 50 °C, 5 h);21 Rh NPs
stabilized with tannin (94% selectivity; 20 bar of H2, 80 °C,
0.5 h).19 Interestingly, the Rh/AlO(OH)18 catalyst and Rh NPs
stabilized with an N-heterocyclic carbene (NHC)24 were able
to hydrogenate quinoline under milder reaction conditions.
Results and discussion
A series of Rh–C60 compounds were synthesised through the
decomposition of [RhIJη3-C3H5)3] under 3 bars of H2 in the
presence of C60 using 1,2-dichlorobenzene (DCB) as a solvent
at 50 °C. In this series, the molar Rh/C60 ratio has been
varied ranging from [Rh]/[C60] = 1 to 20. Fig. 1 shows the
representative TEM images of the as-synthesised products.
Nanospheres with a size of about 100–170 nm were observed,
which were decorated with Rh NPs (Fig. 1 and S1† and
Table 1). The NP size was in the range of 2–3 nm, presenting
a broad size distribution (polydispersities of around 30%),
which displays a bimodal distribution for samples 5/1 and
10/1. These species are structurally similar to the ruthenium
nanocatalysts stabilized with fullerene C60 described in a
previous study by some of us.30 In the above-mentioned
work, the decomposition of [RuIJCOD)IJCOT)] under
dihydrogen in the presence of fullerene C60 provided a Ru
based catalyst composed of Ru–C60 molecular spheres with
their surface decorated with Ru NPs, as pointed out by a
combination of characterization techniques, including TEM,
WAXS, and extended X-ray absorption fine structure (EXAFS),
and supported by DFT calculations. Also, Ru–C60 fulleride
nanospheres synthesised in DCB displayed a size of 200.0
38.5 nm, which is similar to the ones of Rh–C60 materials.
Furthermore, HRTEM analyses were performed for the Rh–
C60 1/1 and 10/1 samples. The HAADF-STEM mode allowed
the detection of a few Rh NPs on the surface of the
nanospheres in the Rh–C60 1/1 compound, together with
isolated atoms on the surface, which was in contrast with the
Rh/AlO(OH) was active at room temperature using
a
hydrogen balloon and produced 2 with a selectivity of 94%
after 6 h of reaction. Rh–NHC was also active at low
temperature (30 °C) under 20 bar of H2 pressure reaching full
conversion after 140 min of reaction, but in this case, only
75% selectivity to 2 was observed (TOF = 238 h−1). Full
conversion after only 30 min of reaction was achieved by
increasing the temperature to 60 °C and quinoline was
completely consumed with 75% selectivity towards 2; after 24
hours of reaction, the fully reduced product 4 was obtained
(TOF = 496 h−1). These last two examples describe very active
catalysts able to produce the fully hydrogenated
decahydroquinoline (Rh/AlOIJOH)) under slightly harsher
reaction conditions (8 bar of H2, 100 °C, 5 h),20 which is in
opposition to the findings using Rh NPs immobilized in a
Lewis acidic IL.23 These results suggest that electron-rich Rh
NPs are very active, but, in general, less selective to the
partially hydrogenated product 2.
On the other hand, electron-poor Ru NPs can be produced
straightforwardly from the decomposition of (1,5-
cyclooctadiene)IJ1,3,5-cyclooctatriene)rutheniumIJ0) [RuIJCOD)-
IJCOT)] with H2 in the presence of fullerene C60.30 This
nanocatalyst shows excellent activity and selectivity in the
hydrogenation of nitrobenzene.31,32 The same methodology
has been used to obtain other metal NPs.33 Also, fullerene
C60 based catalysts with other metals have been described,34
including Pd,35 Ni (ref. 36) and Ru.37,38 With the ability of
fullerene to accommodate various metals,39 as well as the
effectiveness of the organometallic approach to produce
almost any kind of noble metal NP, we decided to synthesize
Rh–C60 catalysts with the aim of producing electron-poor Rh
NPs, which can be very interesting for selective quinoline
hydrogenation.23,25 In this work, we demonstrate that
rhodium NPs can be synthesized straightforwardly via the
decomposition of the [RhIJη3-C3H5)3] complex in the presence
of fullerene C60 in 1,2-dichlorobenzene (DCB) under a H2
atmosphere. Rh atomically dispersed or Rh NPs on Rh–C60
fulleride spherical particles were produced by tuning the
ratio of the Rh precursor and fullerene. These Rh compounds
were characterized using several techniques such as
transmission electron microscopy (TEM), wide angle X-ray
scattering (WAXS), inductively coupled plasma optical
emission spectroscopy (ICP-OES), attenuated total reflection
infrared spectroscopy (ATR-IR), solid-state NMR (SSNMR),
and thermal gravimetric analysis (TGA), including X-ray
photoelectron spectroscopy (XPS) and Raman spectroscopy,
Fig. 1 TEM micrographs of Rh–C60 1/1, 5/1, 10/1, and 20/1 (scale bar
200 nm).
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6885

View Article Online
Paper
Catalysis Science & Technology
Table 1 Mean size distributions and rhodium content of the Rh–C60 nanocatalyst
Rh–C60
Rh contenta (%)
NP mean sizeb (nm)
Nanosphere mean sizeb (nm)
1/1
5/1
10/1
20/1
9.2
23.3
35.9
46.2
2.0 0.4
158.2 27.5
103.0 15.5
167.9 27.3
96.3 14.4
1.1 0.3 (21%)/3.0 0.9 (79%)c
1.9 0.5 (69%)/4.7 1.1 (31%)c
2.3 0.6
a
c
By ICP. b Manual measurements from enlarged TEM micrographs of at least 150 objects. In brackets is the percentage of each population.
Fig. 2 STEM-HAADF images of Rh–C60 1/1 (top) and Rh–C60 10/1IJbottom) together with the EDX line scanning profile of Rh signals extracted from
cumulated EDX spectra on the line scans (right) (scale bars from left to right: 50 nm, 50 nm, 10 nm and 50 nm).
analyses carried out for the Rh–C60 10/1 sample, for which
the Rh NPs on the surface were more abundant (Fig. 2). In
addition, the distribution of the Rh signal by EDX analysis
along the line scan is uniform for sample 1/1 (Fig. 2) while
this confirms the presence of Rh NPs on the surface of Rh–
C60 10/1, as a higher Rh content was detected on the surface
of the sphere. ICP analyses pointed out that the rhodium
content was lower than expected, which suggested the
presence of residual DCB on the samples, further
corroborated by IR, SSNMR and TGA, as detailed below.
In order to obtain more insight into the evolution of these
Rh–C60 species during the synthesis, the reaction was
monitored by ex situ TEM analyses. For that, the
decomposition of [RhIJη3-C3H5)3] under hydrogen in the
presence of fullerene C60 (Rh/C60 = 10/1) was performed at
low temperature (−20 °C). The reaction mixture was sampled
over time and the TEM images are displayed in Fig. S2†
together with their respective size histograms. The evolution
of the nanostructures over time showed that at 2 min of
reaction, worm-like structures composed of nanospheres
were already formed, with a nanosphere mean size of 218.7
23.9 nm, which evolved to independent nanospheres after 5
min of reaction displaying a larger mean size, 274.4 33.1
nm (Table S1†). Rh nanoparticles were observed after 15 min
of reaction on the surface of the nanospheres and dispersed
on the TEM grid, 1.7 0.5 and 3.4 1.3 nm, respectively. The
reaction was allowed to proceed overnight at −20 °C and the
last TEM analysis performed showed that the size of the
nanospheres remained similar, 255.7 39.5 nm, while the Rh
NPs on the surface presented a bimodal distribution, 2.4
0.7 and 4.7
0.4 nm and isolated NPs were no longer
observed. These observations make it feasible to propose
that, as previously observed during the synthesis of Ru–C60
compounds,30 the nanospheres are composed of molecular
species, which form this particular shape due to the stirring
and the different viscosities between the as-synthesized
species and the solvent.40 For the previously described Ru–
C60 species, lower solvent viscosity leads to smaller
nanospheres. Here, the decrease of the temperature to −20 °C
increased the size of the nanospheres when compared to the
synthesis performed at 50 °C (255.7 39.5 nm for −20 °C,
103.0
15.5 nm for 50 °C). Taking into account that the
viscosity increases when the temperature decreases, an
increase of the diameter is expected, as observed here. Also,
the detection of isolated NPs at 15 min could indicate that
double NP nucleation occurs during the synthesis: one on
the surface of the Rh–C60 nanospheres and a second one in
solution. It is possible that the rhodium precursor reacts with
the solvent, as arene–rhodium complexes have been
described in the past.41–43 The arene ring of DCB, being
electro-deficient, should coordinate weakly when compared
to other arene compounds such as benzene or toluene,43
6886 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
producing also NPs under reaction conditions. In addition,
the fact that the Rh NPs are stabilized by a weakly stabilizing
ligand (DCB) could also explain that the isolated NPs
observed first are then adsorbed on the surface of the
nanospheres, explaining in turn the bimodal size
distributions and the polydispersity (around 30%) of the
samples.
WAXS analyses were performed on the Rh nanostructures.
After corrections and Fourier transform, the radial
distribution function was obtained (RDF, Fig. S3†). For the
lowest Rh amounts (i.e. 1/1 and 2/1), the observed RDFs
present only weak features actually dominated by the
distances in C60. From 5/1, the distances are consistent with
metallic Rh in the fcc structure and both amplitude and
coherence length increase in relation to the Rh amount.
Coherence lengths could be estimated to be 3.3 nm for 5/1,
3.5 nm for 10/1 and 3.8 nm for 20/1. It should be noted that
the scattering phenomenon strongly emphasizes the
contribution of larger particles, which does not make it very
sensitive to size distribution. This behaviour has been
previously observed for Ru–C60 species.30
To gain further insight into the interactions between Rh
and fullerene, XPS analyses of Rh–C60 1/1 and 10/1 were
conducted (Fig. 3 and S4 and Table S2†). Peaks in the full
XPS spectra indicate the presence of Rh, C and O elements
on both samples (Fig. S4†). Rh 3d peaks could be fitted into
a unimodal doublet, pointing to the presence of only one Rh
chemical state: metallic Rh0. In the Rh 3d region of the Rh–
C60 1/1 XPS spectrum, a single doublet is observed at 307.8
eV (3d5/2) and 312.4 eV (3d3/2). These binding energies may
evidence charge transfer from the metal to the C60 molecules,
since they are significantly shifted compared to that of
metallic rhodium (Rh 3d5/2 component at 307.2 eV).44 The Rh
3d core peaks of Rh–C60 10/1 are asymmetric and appear at
relatively lower binding energies i.e. 307.2 eV (3d5/2) and
311.9 eV (3d3/2), which are typical of metallic Rh0. Such an
asymmetry could be explained by the presence of a second
contribution at slightly higher binding energies, consistent
with the charge transfer mentioned above. In agreement with
the EDX results presented earlier, the atomic percentage
(Table S2†) for rhodium is higher for Rh–C60 10/1 (8.0%) than
that for Rh–C60 1/1 (1.5%). For both samples, the C 1s peaks
(Fig. S4 and Table S2†) are dominated by a peak at 284.3 eV
attributed to fullerene C60, with less intense peaks associated
with contaminating carbon (285.0 eV), oxygenated carbon
groups (between 286 and 288.5 eV) and a shake-up peak (at
ca. 290.0 eV). It is to be noted that the binding energy of
284.3 eV is significantly lower than the value reported for the
C 1s peak of fullerene films thicker than one monolayer
(approximately 284.8 eV).45 This shift is consistent with the
charge transfer mentioned above, as observed for
a
1-monolayer-thick film of C60 on Cu(111).45 The O 1s peaks
(Fig. S4 and Table S2†) for both samples correspond to
oxygen atoms bound to carbon.
Raman spectroscopy is an efficient technique to study C60
as it can provide important information on various
vibrational modes, and thereby one can understand the
crystallinity, structure, intramolecular bond formation and
molecular deformation. The C60 molecule has 46
intramolecular modes, among which 10 are Raman active (2
Ag + 8 Hg). The Ag (1) mode at ∼492 cm−1 is known as the
breathing mode, which originates from the in phase radial
movement of all sixty carbon atoms in the icosahedral
structure. The second Ag mode Ag (2), situated at 1458–1468
cm−1, is also known as the pentagonal pinch mode
corresponding to the in phase tangential displacement of
carbon atoms, leading to an out of phase breathing mode of
pentagon and hexagon faces. Similarly, the low frequency Hg
modes (<700 cm−1) mainly arise from the radial molecular
movement, whereas higher frequency Hg modes are mainly
due to the tangential displacements.46,47 The Raman
spectrum of C60 displays intense tangential Raman modes of
vibrations such as Ag (2), Hg (7), and Hg (8) at 1459 cm−1,
1429 cm−1, and 1561 cm−1 respectively. Fig. 4 shows the
Raman spectra of Rh–C60 samples with different Rh/C60
ratios and Table 2 lists the peak position, peak shift and full
width at half maximum (FWHM) as a function of the Rh/C60
ratios, which are obtained by peak fitting using the
Lorentzian function. It is observed that after Rh introduction,
the peaks are broadened with a decrease in intensity as well
as an increase in luminescence background. The Ag (2) and
Hg (7) peaks show downshifts, whereas Hg (8) shows an
upshift as compared to the corresponding C60 peak positions.
The FWHM values of the rhodium containing samples are
considerably larger compared to that of the C60 fullerene. For
the Rh–C60 samples, the low frequency vibrations are not
clearly visible because of a large luminescence background
and the softening of these modes. The shifts in the tangential
modes at a higher wavenumber have been assigned in the
literature to the electron transfer from dopants to the C60
molecule.48 The broadening of these prominent modes,
mainly those of Hg (7) and Hg (8), are explained by the
electron–phonon coupling due to doping. The broadening of
the Ag (2) mode is explained by the expansion of the C60
molecule, which reduces the force constants.48 The reported
Fig. 3 XPS Rh 3d spectra of a) Rh–C60 1/1 and b) Rh–C60 10/1.
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6887

View Article Online
Paper
Catalysis Science & Technology
Fig. 4 a) Raman spectra of C60 fullerene together with Rh–C60 nanostructures and b) shift in peak position vs. Rh/C60 ratio.
Table 2 Peak position, shift of peaks and FWHM of C60 and Rh–C60
Ag (2)
Hg (7)
Hg (8)
Position (cm−1
)
Shift (cm−1
)
FWHM Position (cm−1
)
Shift (cm−1
)
FWHM Position (cm−1
)
Shift (cm−1
)
FWHM
C60
Rh–C60 1/1
Rh–C60 5/1
Rh–C60 10/1 1432.8
Rh–C60 20/1 1435.6
1459.1
1439.9
1434. 0
0
15.1
65
105.6
71
1423.6
1378.4
1360.3
1366.1
1374.6
0
17.2
206.4
181.5
267.1
239.8
1563.2
1577.3
1578.0
1571.1
1569.4
0
20.1
84.6
81.2
82.85
77
−19.2
−25.1
−26.3
−23.5
−45.2
−63.3
−57.5
−49
14.1
14.8
7.9
6.2
61.1
electron transfer for the downshift of tangential modes of
vibrations at a higher wavelength are −6 cm−1 per electron for
Ag (2), −7 cm−1 per electron for Hg (7) and −15 cm−1 per
electron for Hg (8).49 The downshift of the pentagonal pinch
mode is commonly used to quantify the charge transfer from
the metal to the antibonding electron orbitals of the C60
molecule since it is the most intense peak in the Raman
spectra of C60. Meanwhile, Rh–C60 10/1 shows a higher
downshift of −26.3 cm−1 corresponding to an electron
transfer of ∼4.4 e− and Rh–C60 20/1 exhibits a downshift of
−23.5 cm−1 corresponding to an electron transfer of ∼3.9 e−.
For Rh doping, the negative shift and broadening increases
for up to 10/1, but decreases for 20/1, and from this, we can
predict that there is saturation with respect to Rh doping.
The ATR-IR spectra of the Rh–C60 nanostructures were
recorded in the solid state (Fig. S5-a†). Characteristic peaks
of fullerene C60 are observed at 518, 577, and 1182 cm−1,
which confirm the presence of the stabilizing molecules in
the samples. Other intense peaks are attributed to the
presence of DCB (657, 745, 1033, 1124, 1432 and 1454 cm−1).
The SSNMR of Rh–C60 10/1 (Fig. S6†) confirmed the presence
of fullerene C60, whose signal appears at 147.1 ppm. The
MAS 13C{1H} NMR also show two signals at 130.9 and 128.1
ppm, which were attributed to remaining DCB and
corroborated the presence of the solvent in the samples as
seen in the ATR-IR spectra. The CP-MAS 13C NMR (CP =
1H–13C cross polarization) allowed the observation of a broad
signal centered around 43 ppm attributed to a mixture of
hydrogenated fullerene species. The hydrogenation of
fullerene C60 at ambient temperature and pressure has
already been reported for the Rh/Al2O3 catalyst,50 and the
formation of these hydrogenated species were also observed
for Ru–C60 materials.30 Adsorbed DCB was removed by
annealing the samples at 200 °C under Ar. The ATR-IR
spectra (Fig. S5-b†) recorded for the heat-treated samples
show that the DCB peaks vanished. Additionally, TGA
analyses (Fig. S7†) confirmed the efficiency of the thermal
treatment. No appreciable structural change of the
nanospheres was observed from TEM analyses on the Rh–C60
5/1 sample (denoted as Rh–C60 5/1 TT) (Fig. S8†); however,
the size distribution of the NPs was unimodal, centred at 1.8
0.5 nm, which could be attributed to slight sintering of the
small nanoparticles.
To evaluate the catalytic performance of the Rh–C60
catalysts, we selected the quinoline hydrogenation, as
rhodium based catalysts have shown very good performances
in this reaction (Scheme 1 and Tables 3 and 4).20,22,24 In
addition, the fact that Rh is electron deficient, as evidenced
by Raman and XPS, can be an advantage to obtain higher
selectivity, as Rh–NHC carbene species,24 bearing strong
donating groups on the Rh NP surface, are very active for this
reduction reaction working at 30 °C but not very selective to
the synthesis of 2. Theoretical calculations supported our
first analysis, as Rh-based catalysts intrinsically coordinate
6888 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
Table 3 Selective hydrogenation of quinoline using the Rh–C60 10/1 nanocatalysta
Selectivityc (%)
T
(°C)
TOFb
(h−1
Time
(h)
Conversionb,c
(%)
Entry
Solvent
THF
)
2
3
4
1
2
3
4
5
6
7
8
9
10
60
19.5
1
4
1
4
1
4
1
4
1
4
7
100
98
97
97
98
97
97
98
98
97
—
1
1
1
1
1
1
1
1
1
—
1
2
2
1
2
2
1
1
2
44
38
65
75
100
57
90
78
99
80
THF
47.8
100
100
100
THF
118.6
99.8
Toluene
Isopropanol
123.4
a
Reaction conditions: 0.02 mmol Rh, 404 mg (3.1 mmol) of quinoline, 75 mg (0.41 mmol) of dodecane (internal standard), 20 bar H2, 25 ml of
b
c
solvent. TOFs calculated at 1 h of reaction related to the % of Rh. Conversion and selectivity determined by GC analyses using the internal
standard technique.
Table 4 Selective hydrogenation of quinoline using the Rh–C60 nanocatalysta
Selectivityc (%)
TOFb
(h−1
Time
(h)
Conversionb,c
(%)
Entry
Catalyst
)
2
3
4
1
2
3
4
5
6
7
8
9
Rh–C60 1/1
94.3
1
4
1
4
1
1
4
1
4
54
80
98
99
99
78
99
97
99
100
100
96
94
97
98
97
92
90
—
—
1
1
1
1
1
3
3
—
—
3
5
3
1
2
5
7
Rh–C60 5/1
164.3
Rh–C60 5/1 TT
Rh–C60 10/1
157.1
123.4
Rh–C60 20/1
157.7
a
Reaction conditions: 0.02 mmol Rh, 404 mg (3.1 mmol) of quinoline, 75 mg (0.41 mmol) of dodecane (internal standard), 100 °C, 20 bar H2,
b
c
25 ml of solvent. TOFs calculated at 1 h of reaction related to the % of Rh. Conversion and selectivity determined by GC analyses using the
internal standard technique.
the N-bearing ring more efficiently (Table 5, Rh13H20 N vs.
aromatic ring coordination), therefore, giving preferentially
the partially hydrogenated product 2, as observed in most of
the reviewed reported studies. Concerning the coordination
of 1,2,3,4-tetrahydroquinoline (2) to rhodium clusters, this
coordinate the N-bearing aromatic ring in two distinct modes
(by N or by two carbon bonds) with a similar Eads. This is in
contrast with the coordination of quinoline to Rh13H20, in
which the coordination of the N atom is largely favourable.
This could explain the high TOF observed in the
difference is less pronounced (Table 6, Rh13H20
N
vs.
hydrogenation of quinoline using Rh–C60 as a catalyst
aromatic ring coordination), as both coordination modes are
possible with similar adsorption energies (Eads) and a priori
the coordination of the aromatic group could lead to the
complete hydrogenation of the molecule. As this is not so
usual in the existing literature, probably steric effects should
be taken into consideration, as further coordination of
another 2 molecule would give extra stabilization, leading to
poisoning of the Rh surface by the product, an argument
which is pointed out in some cases to explain the high
selectivity of Rh systems.24 We recall that full hydrogenation
is possible too.51,52 The present DFT study also supported
our hypothesis that the use of Rh electron deficient species
should be an advantage to obtain higher selectivity. Indeed,
by comparing the coordination of quinoline to systems, such
as Rh13H20 (to mimic bare Rh NPs under hydrogenation
conditions) and C60–Rh13H20–C60 (Table 5), it is observed that
the hydrogenated Rh NPs interacting with fullerenes can
discussed below. The analyses of the coordination of
compound 2 to the C60–Rh13H20–C60 system (Table 6), in
comparison with Rh13H20, show that in this case, only the
coordination by the N atom is possible, in contrast with
Rh13H20 for which two coordinating modes close in
adsorption energies are possible. This observation, the strong
coordination of the partially hydrogenated product, could
explain the high selectivity of Rh–C60.
Experimentally, Rh–C60 catalysts were active and highly
selective for the hydrogenation of the pyridine ring of
quinoline (Tables 3 and 4). Using the best conditions, the
selectivity towards 1,2,3,4-tetrahydroquinoline (2) at full
conversion was in the range of 92–98% for all catalysts. The
Rh–C60 10/1 catalyst was used to optimise the reaction
conditions (Table 3). As expected, the catalytic activity
increased with the temperature, and the conversion of
quinoline at 4 h of reaction was 44%, 65% and 100% at 60
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6889

View Article Online
Paper
Catalysis Science & Technology
Table 5 DFT results of quinoline adsorption on several Rh13 clusters; comparisons between adsorption modes (N vs. aromatic ring) are highlighted and
charge transfera is given
Support
Rh13
Adsorption mode
Eads (kcal mol−1
−24
)
Charge transfer (e−)
+0.01
−64
−9
+0.40
Rh13H20
−0.01
−25
−25
−0.10
−0.09
C60–Rh13–C60
−59
−26
+0.24
C60–Rh13H20–C60
−0.14
−33
+0.03
a
Positive charge transfer means that the substrate receives electronic density from the support.
°C, 80 °C and 100 °C, respectively. The catalytic activity was
also sensitive to the solvent, and the conversion of quinoline
increased, following the order toluene < THF ≤ isopropanol,
which is in agreement with previous studies where highly
polar solvents are more efficient in hydrogenation reactions
with regard to low-polarity ones.53–56 The reactions
performed at 100 °C reached full conversion at 3 h of
reaction when using isopropanol or THF as a solvent and 7 h
in the case of toluene, with high selectivity to 2, 97%, 97%
and 95%, respectively. Reactions carried out in isopropanol
and toluene were allowed to react for longer times in order to
produce the fully hydrogenated product. Nevertheless, after
24 h of reaction, the composition of the reaction mixture was
very similar; for toluene, the selectivity towards 2 at 7 h (full
conversion) and 24
h
was the same (97%), while for
isopropanol, a slight increase of the selectivity to fully
hydrogenated compound 4 was observed, from 1% at 3 h to
5% at 24 h. Time–concentration curves for the quinoline
hydrogenation with Rh–C60 10/1 in toluene and isopropanol
are shown in Fig. S9 and S10,† respectively. The high
selectivity of Rh–C60 towards 2 is independent of the solvent
used, which is in contrast with some reported systems in
which the selectivity is sensitive to the solvent.26,57
Additionally, the hydrogenation of
2
and 5,6,7,8-
tetrahydroquinoline (3) were evaluated in independent
catalytic tests using Rh–C60 10/1 as the catalyst in
isopropanol. The hydrogenation of 2 was very slow (TOF = 4.8
h−1), reaching a conversion of 11% at 5 h, which remained
6890 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
Table 6 DFT results of 1,2,3,4-tetrahydroquinoline adsorption on various supports; comparisons between adsorption modes (N vs. aromatic ring) are
highlighted and charge transfera is given
Support
Rh13
Adsorption mode
Eads (kcal mol−1
−27
)
Charge transfer (e−)
−0.08
−54
+0.10
Rh13H20
−17
−18
−0.19
−0.13
C
60–Rh13–C60
−64
−20
−0.08
−0.23
C60–Rh13H20–C60
a
Positive charge transfer means that the substrate receives electronic density from the support.
almost unchanged at 24 h of reaction (13% conversion). On
the other hand, the hydrogenation of 3 was slightly more
efficient (TOF = 23.1 h−1), displaying 15% conversion at 5 h
and 26% conversion at 24 h of reaction.
364.2
27.8 nm (Fig. S13†) were observed (the mean size
distribution of the as-synthesised compound was 158.2 27.5
nm), indicating that these molecular species can evolve
under the catalytic reaction conditions. Also, the Rh NP mean
size distribution for Rh–C60 5/1 and Rh–C60 10/1 changed,
displaying a monomodal distribution but is very polydisperse
(Fig. S13†). In addition, a few isolated NPs were observed in
the TEM grid on the Rh–C60 10/1 sample, indicating that the
Rh NPs are not stable on the nanosphere surface during
catalysis.
At first sight, no remarkable effect of the thermal
treatment (TT), performed to remove adsorbed DCB, was
noticed for sample Rh–C60 5/1 vs. Rh–C60 5/1 TT (Table 4,
entry 3 vs. entry 5), although it is difficult to compare as the
reaction was complete within one hour. The catalytic
performances of both catalysts, Rh–C60 5/1 and Rh–C60 5/1
TT, were compared at a lower temperature, 80 °C, using
smaller time intervals to analyze the reaction mixtures. Under
these conditions, Rh–C60 5/1 TT showed a higher TOF at 15
minutes of reaction, 302.2 h−1, than Rh–C60 5/1, 253.8 h−1.
Quinoline was completely consumed within 1 h and 1.5 h,
respectively. When the reaction was performed at 100 °C
using Rh–C60 5/1 TT as the catalyst, the quinoline conversion
The Rh–C60 series was tested as catalysts in the
hydrogenation of quinoline under the optimized conditions,
i.e. using isopropanol as the solvent at 100 °C under 20 bar
of H2 pressure (Table 4). The time–concentration curves for
the quinoline hydrogenation with the Rh–C60 series are
shown in Fig. S11.† In all cases, except for Rh–C60 1/1, full
conversion of quinoline was achieved within an hour, with
high selectivity towards 2 (92–98%). The TOFs were very
similar among them, indicating that the concentration of
rhodium has no effect on both activity and selectivity. This
result is in line with Raman analyses, in which similar charge
transfer was evidenced for samples 5/1 to 20/1. In contrast,
the Rh–C60 1/1 catalyst displayed the lowest activity in the
series, and it is probably because only a few nanoparticles
are available on the surface of the nanospheres as
ascertained by HRTEM analyses. The TEM analyses of the
catalysts after the hydrogenation reaction showed that the
nanospheres remained almost unchanged except for Rh–C60
1/1 where two populations of sizes centred at 75.0 23.6 and
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6891

View Article Online
Paper
Catalysis Science & Technology
was 94% at 15 min, reaching full conversion within 30 min
(98% selectivity towards 2), thus the calculated TOF at 15
min is 488.0 h−1, which is as high as those of some other
Rh–C60 5/1 could be reused three times without a loss of
catalytic activity; nevertheless, at the fourth recycling, the
conversion dropped to 70%. On the other hand, Rh–C60 5/1
TT was revealed to be more robust and was recycled five
times without a loss of catalytic activity. Also, TEM analyses
of the Rh–C60 TT 5/1 catalyst after the recycling test showed
no appreciable change in the nanosphere size while the Rh
NP mean size increased to 2.7 0.7 nm.
representative
heterogeneous
catalysts
(Table
S3†).17,22–24,27–29,58
The recycling of the catalysts was investigated using Rh–
C60 5/1 and Rh–C60 5/1 TT due to their higher activity and
selectivity when compared to the rest of the series (Fig. 5).
Fig. 5 Conversion and selectivity after the recycling tests using Rh–C60 5/1 and Rh–C60 5/1 TT together with the TEM images and size distribution
histograms of Rh–C60 5/1 TT after the recycling test (scale bar from left to right: 100 nm and 50 nm).
6892 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
A DFT study has been performed in order to explore the
coordination thermodynamics of quinoline (1) and 1,2,3,4-
tetrahydroquinoline (2) on several Rh13 molecular models
(Tables 5 and 6). The four molecular models, denoted as
supports in the tables, included a bare Rh13 cluster, a hydride
covered Rh13 cluster (Rh13H20), and the analogous systems
with two coordinated fullerene C60 onto the Rh surface (C60–
Rh13–C60 and C60–Rh13H20–C60). A coverage value of 1.7 H
atom per surface Rh atom was slightly higher than the value
previously used by us for Ru systems (1.5 H in Ru13H18),31,32
since one can assume that extra hydrides with small
adsorption energies are probably involved in the
hydrogenation reactions. Hydrogenated Rh NP models
simulate the surface of the NPs under hydrogenation
conditions.59 The use of two fullerene C60 was chosen by
analogy to our previous DFT calculations on Ru–C60
systems.31,32 The adsorption energies and charge transfer of
compounds 1 and 2 on those models are given after carefully
checking the nature of the most stable geometries. The
presence of the surface hydrides highly influences the
coordination mode and the adsorption energy in both
compounds. The bare Rh13 cluster very efficiently coordinates
quinoline through π-coordination of the N-bearing aryl ring
(−64 kcal mol−1 Rh13-1, Table 5), while Rh13H20 prefers to
coordinate through the N atom (−25 kcal mol−1 Rh13H20-1,
Table 5) and π-coordination is calculated to be very
unfavourable, but possible (−9 kcal mol−1 Rh13H20-1, Table 5).
Similarly, the coordination mode of 2 over Rh13 and Rh13H20
is largely influenced by the surface hydrides. While the bare
Rh13 cluster accommodates compound 2 in a very stable way
through π-coordination of the aryl ring (−54 kcal mol−1 Rh13-
2, Table 6) after the addition of hydrides onto the Rh surface,
this coordination is not possible leading to coordination
through the N atom (−18 kcal mol−1 Rh13H20-2, Table 6) or
through the coordination of two C atoms of the aryl ring in
the η2 position with Rh–C distances of 2.27 and 2.42 Å, which
is very close in adsorption energy (−17 kcal mol−1 Rh13H20-2,
Table 6).
Likewise, this phenomenon is also observed in the
adsorption of compound 2 on Rh13 clusters bearing fullerene
C60 on their surface (π-coordination on C60–Rh13–C60 (−64
kcal mol−1 C60–Rh13–C60-2, Table 6) vs. N-coordination on
C60–Rh13H20–C60 (−20 kcal mol−1 C60–Rh13H20–C60-2, Table 6)).
The coordination of quinoline 1 on C60–Rh13H20–C60 followed
a slightly different trend. Even if similar, π-coordination is
not possible over the hydrogenated cluster in contrast with
the highly stable π-coordination observed on C60–Rh13–C60
(−59 kcal mol−1 C60–Rh13–C60-1, Table 5) and N-coordination
of 1 is favourable over C60–Rh13H20–C60 (−26 kcal mol−1 C60–
Rh13H20–C60-1, Table 5); the coordination of two carbon
atoms of the N-aryl ring to one rhodium atom displays a
more stable adsorption energy (−33 kcal mol−1 C60–Rh13H20–
C60-1, Table 5). As a general trend for these calculations, the
hydride coverage changes the coordination mode stability.
Highly stable π-coordination is observed on bare Rh13 clusters,
whereas at high coverage (Rh13H20), N-coordination prevails.
The effect of the hydride coverage could be attributed to two
factors, first, they pull electronic densities from Rh atoms and
their presence on the surface impedes the π-coordination
through steric effects. This effect has been already observed
by us in previous studies dealing with Ru NPs.31,32
The hydride coverage governs the coordination mode of
the substrate and the partially hydrogenated product to the
Rh surface; nevertheless, whether fullerene C60 on the surface
is present or not, it also modulates the stability of the
adsorbed species. In general, the addition of fullerene C60
decreased the charge transfer, as experimentally observed.
Quinoline prefers to coordinate to the surface through a
π-coordination mode of the N-bearing aryl ring in bare Rh13
and C60–Rh13–C60, with a slight decrease of the adsorption
energy and a change in the charge transfer (Table 5, −64 kcal
mol−1 +0.40 e− vs. −59 kcal mol−1 +0.24 e−, respectively).
Likewise, the coordination of compound 2 on bare Rh13 and
C60–Rh13–C60 displayed
a
similar trend, but the
π-coordination on C60–Rh13–C60 was found to be more stable
(Table 6, −54 kcal mol−1 +0.10 e− (Rh13) vs. −64 kcal mol−1
−0.08 e− (C60–Rh13–C60)). The comparison of the effect of the
fullerene C60 coordination on hydrogenated clusters supports
the trends observed experimentally, i.e. the enhanced activity
observed with the presence of fullerene C60 on the surface
compared to other reported systems. Quinoline coordinates
to the surface of Rh13H20 and C60–Rh13H20–C60 through the
N-atom displaying similar adsorption energies (−25 kcal
mol−1 vs. −26 kcal mol−1, Table 5), but the presence of
fullerene C60 allowed the coordination of quinoline through
the two carbon atoms of the N-aryl ring in a more stable way
(−33 kcal mol−1, Table 5), which could explain the higher
activity of Rh–C60 compared to those of other reported
systems
(see
Table
S3†).
Also,
C60–Rh13H20–C60
accommodates compound 2 through the N-atom in a more
stable manner compared to Rh13H20, which could explain
that the Rh–C60 system is highly selective. This is also
supported by the fact that the coordination through the
aromatic ring was found to be not possible in contrast with
Rh13H20. It is also noteworthy to point out that the product 2
did not poison the system for the hydrogenation of quinoline,
as quinoline is coordinated more efficiently than 2 to
hydrogenated rhodium surfaces (−33 kcal mol−1 C60–Rh13H20–
C60-1, Table 5, vs. 20 kcal mol−1 C60–Rh13H20–C60-2, Table 6).
Conclusions
A series of Rh–C60 catalysts were produced via a H2 assisted
solution synthesis route. Rh atomically dispersed or Rh NPs
were prepared by tuning the molar ratio of the Rh precursor
and fullerene C60. A significant interaction between Rh and
C60 was evidenced by XPS and Raman analyses, which
consisted of charge transfer from the metal to fullerene C60
.
These electron-deficient Rh species are excellent catalysts for
the selective reduction of quinoline to the partially
hydrogenated product with a selectivity of up to 98%.
Thermal annealing of the catalyst was highly beneficial to
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6893

View Article Online
Paper
Catalysis Science & Technology
improve the stability of these species, as seen in the recycling
test, and also activity (TOF 488.0 h−1), which is in between
the most remarkable heterogeneous catalysts for this
reaction. In addition, DFT calculations allowed the
determination of the origin of the intrinsic selectivity of Rh
NP-based systems when applied as catalysts in the
hydrogenation of quinoline. The hydride coverage in Rh–C60
systems plays a pivotal role and explains the origin of the
high selectivity of fullerene modified Rh NP systems when
applied as a catalyst in the hydrogenation of quinoline.
General procedure for the hydrogenation of quinoline
The hydrogenation of quinoline was performed in a 40 ml
stainless steel high-pressure batch BR-25 Berghof reactor.
Typically, a mixture of Rh–C60 catalysts (0.02 mmol Rh),
dodecane (75 mg; 0.44 mmol) and quinoline (404 mg, 3.12
mmol) in 25 ml of the desired solvent was loaded into the
autoclave in the glove box. The autoclave was purged three
times with H2 to remove the inert atmosphere, heated to 100
°C and charged with 20 bar of H2. The stirring rate was fixed
at 1500 rpm. Samples of the reaction mixture were taken at
different time intervals and analysed by gas chromatography.
Quantitative analyses of the reaction mixtures were
performed via GC using calibration solutions of commercially
available products.
Experimental section
General methods
All operations were carried out under an argon atmosphere
using standard Schlenk techniques or in a MBraun glove box.
Solvents were purified by standard methods or by an MBraun
SPS-800 solvent purification system. [RhIJη3-C3H5)3] was
purchased from Nanomeps Toulouse, fullerene C60 from
Bucky USA, and H2 from Air Liquide. All these reactants were
used as received.
Catalytic recycling tests
A mixture of quinoline (404 mg, 3.10 mmol), dodecane (75
mg, 0.41 mmol), Rh–C60 5/1 catalysts (0.02 mmol Rh) and 25
ml of isopropanol was prepared in the glove box and
introduced into the autoclave. The autoclave was purged three
times with H2, heated to 100 °C and pressurized with 20 bar
of H2 during 1 h keeping the stirring rate at 1500 rpm. After
the reaction, the catalyst was recovered into a Schlenk tube by
filtering with a cannula and washing several times with EtOH
under argon, and then dried under vacuum. The catalyst was
reused in another five cycles of the hydrogenation reaction
following the procedure described above.
Synthesis of Rh–C60
In
a
typical synthesis, the [RhIJη3-C3H5)3] complex and
fullerene C60 were placed in a Fisher-Porter bottle in a glove
box and dissolved in 1,2-dichlorobenzene (DCB). The
resulting purple solution was stirred for 30 min at room
temperature, pressurized with 3 bar of H2 and then placed in
a 50 °C oil bath. The solution turned black in a few minutes
and was maintained at this temperature under stirring
overnight. After cooling down to room temperature, pentane
was added to the reaction mixture to precipitate Rh–C60
which was cleaned twice with pentane. The Rh nanoparticles
were obtained as a black solid powder after drying under
vacuum overnight. For each ratio studied, the quantities of
reactants are detailed hereafter:
TEM analyses
TEM and HRTEM analyses were performed at the “Centre de
microcaracterisation Raimond Castaing, UMS 3623,
Toulouse” by using
a JEOL JEM 1011 CXT electron
microscope operating at 100 kV with a point resolution of 4.5
Å. The high resolution analyses were conducted using a JEOL
JEM 2100F equipped with a field emission gun (FEG)
operating at 200 kV with a point resolution of 2.3 Å and a
JEOL JEM-ARM200F Cold FEG operating at 200 kV with a
point resolution of >1.9 Å. The approximation of the particle
mean size was established through a manual analysis of
enlarged micrographs by measuring at least 200 particles on
a given grid. Other TEM micrographs were acquired with a
Rh–C60 1/1: 50 mg (0.22 mmol) of [RhIJη3-C3H5)3], 160 mg
(0.22 mmol) of fullerene C60 and 100 mL of DCB. Yield: 160
mg. Rh content: 9.2%.
Rh–C60 5/1: 25 mg (0.11 mmol) of [RhIJη3-C3H5)3], 16 mg
(0.02 mmol) of fullerene C60 and 10 mL of DCB. Yield: 20
mg. Rh content: 23.3%.
JEOL 2100F S/TEM microscope equipped with
a FEG
Rh–C60 10/1: 100 mg (0.44 mmol) of [RhIJη3-C3H5)3], 32 mg
(0.04 mmol) of fullerene C60 and 20 mL of DCB. Yield: 100
mg. Rh content: 35.9%.
operating at 200 kV, a spherical aberration probe corrector
and a GATAN Tridiem energy filter. The resolutions attained
are 2 Å and 1.1 Å under parallel TEM and scanning STEM
modes, respectively. For STEM-HAADF analyses, the spot size
was 0.13 nm, the current density was 140 pA, and the camera
focal length was 10 cm, corresponding to the inner and outer
detection angles of the annular detector of about 60 mrad
and 160 mrad.
Rh–C60 20/1: 200 mg (0.86 mmol) of [RhIJη3-C3H5)3], 32 mg
(0.04 mmol) of fullerene C60 and 20 mL of DCB. Yield: 75
mg. Rh content: 46.2%.
Growth mechanism
The reaction was performed at −20 °C following the standard
procedure: 50 mg (0.22 mmol) of [RhIJη3-C3H5)3], 16 mg (0.02
mmol) of fullerene C60 and 10 mL of DCB. The reaction was
followed by sampling the mixture over time.
WAXS analyses
Wide angle X-ray scattering measurements were performed at
CEMES on a diffractometer dedicated to pair distribution
6894 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
function
(PDF)
analysis:
graphite-monochromatized
GC analyses
molybdenum radiation (0.07169 nm), solid-state detection
and low background setup. Samples were sealed in
Lindemann glass capillaries (diameter of 1 mm) to avoid any
oxidation after filling in a glove box. For all samples, data
were collected on an extended angular range (129 degrees in
2 theta) with counting times of typically 150 s for each of the
457 data points, thus allowing PDF analysis. Classic
corrections (polarization and absorption in cylindrical
geometry) were applied before reduction and Fourier
transform.
GC analyses were performed on a PerkinElmer Autosystem
GC equipped with an Elite-5MS capillary column (30 m × 0.25
mm × 0.25 μm). Split (20 : 1); carrier gas flow: He, 20 ml
min−1; injector temperature: 250 °C; detector (FID)
temperature: 250 °C; oven program: 110 °C (hold 12 min) to
230 °C at 20 °C min−1 (hold 3 min) for a total run time of 21
min; retention time: dodecane, 5.3 min; 1, 7.0 min; 2, 11.3
min; 3, 6.5 min; 4, 4.4 min.
Computational details
DFT calculations were performed using the Vienna ab initio
simulation package VASP,61–64 based on the full-potential
projector augmented wave framework.65,66 Exchange–
correlation effects have been approximated using the spin-
polarized version of the PBE functional.67 A kinetic-energy
cutoff of 400 eV was found to be sufficient to achieve a total-
energy convergence within several meV, considering the
k-point sampling in Gamma-point only calculations for
isolated molecules and complexes, in conjunction with
Gaussian smearing with a width of 0.05 eV. During geometry
optimization runs, all the atoms were fully relaxed until
ICP analyses
The rhodium content was established by inductively coupled
plasma optical emission spectroscopy (ICP-OES) performed at
the LCC on a Thermo Scientific ICAP 6300 instrument after
acid mineralization of the samples.
Raman analyses
Raman measurements were carried out with
a Horiba
XPLORA-MV2000 spectrometer. For the measurements, an
excitation wavelength of 532 nm and a laser power of 0.084
mW were used. The samples were kept under vacuum and
exposed to atmospheric air shortly before measurements.
forces on individual atoms were smaller than 0.01 eV Å−1
.
Calculation cells for isolated molecules and complexes were (25
× 26 × 27) Å3 to avoid spurious interactions between periodic
images. Figures of the different geometries were produced
thanks to the 3D visualization program VESTA.68 Bader charge
analyses were performed using Henkelman's group code.69 The
optimal geometries upon H2 adsorption were constructed,
following the results of ref. 70, i.e. all available μ3 sites were
occupied and then the top sites and, if needed, some bridge
sites were used to build the starting geometries.
XPS analyses
XPS measurements were performed on
a Thermo Kα
spectrometer working at a base pressure of 5 × 10−9 mbar
and equipped with a monochromatic Al Kα X-ray source
(1486.7 eV). The spectra presented here were recorded with a
pass energy of 20 eV. The data were processed with CasaXPS
using Gaussian–Lorentzian combinations and
a Shirley
background. Scofield photoionization cross-sections60
corrected for the transmission function of the analyzer and
the analysis depth were used for quantifications.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
ATR-IR
This work was supported by the Centre National de la
Recherche Scientifique (CNRS), which we gratefully
acknowledge. I. C. G. acknowledges the Calcul en Midi-
Pyrénées initiative CALMIP (Project p0812) for allocations of
computer time as well as GENCI-CINES and GENCI-IDRIS for
Grant No. 2018-A006096649.
ATR-IR spectra were recorded on a Perkin-Elmer GX2000
spectrometer available in a glove box in the range of 4000–
400 cm−1.
SSNMR
Solid state NMR (MAS-NMR) with and without 1H–13C cross
polarization (CP) were performed at the LCC on a Bruker References
Avance 400WB instrument equipped with a 4 mm probe with
1 R. A. Johnstone, A. H. Wilby and I. D. Entwistle,
the sample rotation frequency being set at 12 kHz unless
otherwise indicated. Measurements were carried out in a 4
mm ZrO2 rotor.
Heterogeneous catalytic transfer hydrogenation and its
relation to other methods for reduction of organic
compounds, Chem. Rev., 1985, 85, 129–170.
2 H. U. Blaser, C. Malan, B. Pugin, F. Spindler, H. Steiner and
M. Studer, Selective hydrogenation for fine chemicals: recent
trends and new developments, Adv. Synth. Catal., 2003, 345,
103–151.
TGA
Thermal analysis under air was performed on
thermobalance Perkin Elmer Diamond TG.
a
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6895

View Article Online
Paper
Catalysis Science & Technology
3 M. J. Gilkey and B. Xu, Heterogeneous catalytic transfer
hydrogenation as an effective pathway in biomass
upgrading, ACS Catal., 2016, 6, 1420–1436.
4 F. Meemken and A. Baiker, Recent progress in
heterogeneous asymmetric hydrogenation of C=O and C=C
bonds on supported noble metal catalysts, Chem. Rev.,
2017, 117, 11522–11569.
5 J. Jia, C. Qian, Y. Dong, Y. F. Li, H. Wang, M. Ghoussoub,
K. T. Butler, A. Walsh and G. A. Ozin, Heterogeneous
catalytic hydrogenation of CO2 by metal oxides: defect
engineering–perfecting imperfection, Chem. Soc. Rev.,
2017, 46, 4631–4644.
6 G. G. Wildgoose, C. E. Banks and R. G. Compton, Metal
nanoparticles and related materials supported on carbon
nanotubes: methods and applications, Small, 2006, 2,
182–193.
7 X.-H. Li and M. Antonietti, Metal nanoparticles at
mesoporous N-doped carbons and carbon nitrides:
functional Mott–Schottky heterojunctions for catalysis,
Chem. Soc. Rev., 2013, 42, 6593–6604.
8 Y. Zhai, Z. Zhu and S. Dong, Carbon-based nanostructures
for advanced catalysis, ChemCatChem, 2015, 7, 2806–2815.
9 X. Cui, A.-E. Surkus, K. Junge, C. Topf, J. Radnik, C.
Kreyenschulte and M. Beller, Highly selective hydrogenation
of arenes using nanostructured ruthenium catalysts
modified with a carbon–nitrogen matrix, Nat. Commun.,
2016, 7, 11326pp.
18 I. S. Park, M. S. Kwon, K. Y. Kang, J. S. Lee and J. Park,
Rhodium and iridium nanoparticles entrapped in aluminum
oxyhydroxide nanofibers: catalysts for hydrogenations of
arenes and ketones at room temperature with hydrogen
balloon, Adv. Synth. Catal., 2007, 349, 2039–2047.
19 H. Mao, X. Liao and B. Shi, Amphiphilic tannin-stabilized
Rh nanoparticles: A highly active and reusable catalyst in
biphasic aqueous-organic system, Catal. Commun., 2011, 16,
210–214.
20 G.-Y. Fan and J. Wu, Mild hydrogenation of quinoline to
decahydroquinoline over rhodium nanoparticles entrapped
in aluminum oxy-hydroxide, Catal. Commun., 2013, 31,
81–85.
21 H.-Y. Jiang and X.-X. Zheng, Phosphine-functionalized ionic
liquid-stabilized rhodium nanoparticles for selective
hydrogenation of aromatic compounds, Appl. Catal., A,
2015, 499, 118–123.
22 M. Niu, Y. Wang, P. Chen, D. Du, J. Jiang and Z. Jin, Highly
efficient and recyclable rhodium nanoparticle catalysts for
hydrogenation of quinoline and its derivatives, Catal. Sci.
Technol., 2015, 5, 4746–4749.
23 A. Karakulina, A. Gopakumar, İ. Akçok, B. L. Roulier, T.
LaGrange, S. A. Katsyuba, S. Das and P. J. Dyson, A rhodium
nanoparticle–lewis acidic ionic liquid catalyst for the
chemoselective reduction of heteroarenes, Angew. Chem.,
2016, 128, 300–304.
24 F. Martinez-Espinar, P. Blondeau, P. Nolis, B. Chaudret, C.
Claver, S. Castillón and C. Godard, NHC-stabilised Rh
nanoparticles: Surface study and application in the catalytic
hydrogenation of aromatic substrates, J. Catal., 2017, 354,
113–127.
10 S. Navalon, A. Dhakshinamoorthy, M. Alvaro and H. Garcia,
Metal nanoparticles supported on two-dimensional
graphenes as heterogeneous catalysts, Coord. Chem. Rev.,
2016, 312, 99–148.
11 E. Perez-Mayoral, V. Calvino-Casilda and E. Soriano, Metal-
supported carbon-based materials: opportunities and
challenges in the synthesis of valuable products, Catal. Sci.
Technol., 2016, 6, 1265–1291.
25 A. Karakulina, A. Gopakumar, Z. Fei and P. J. Dyson,
Chemoselective reduction of heteroarenes with a reduced
graphene oxide supported rhodium nanoparticle catalyst,
Catal. Sci. Technol., 2018, 8, 5091–5097.
12 H. Zhang, X. Yan, Y. Huang, M. Zhang, Y. Tang, D. Sun, L.
Xu and S. Wei, Cyanogel-derived N-doped C nanosheets
immobilizing Pd-P nanoparticles: One-pot synthesis and
enhanced hydrogenation catalytic performance, Appl. Surf.
Sci., 2017, 396, 812–820.
26 L. Zhang, X. Wang, Y. Xue, X. Zeng, H. Chen, R. Li and S.
Wang, Cooperation between the surface hydroxyl groups of
Ru–SiO2@ mSiO2 and water for good catalytic performance
for hydrogenation of quinoline, Catal. Sci. Technol., 2014, 4,
1939–1948.
13 R. T. Shuman, P. L. Ornstein, J. W. Paschal and P. D.
Gesellchen, An improved synthesis of homoproline and
derivatives, J. Org. Chem., 1990, 55, 738–741.
27 H. Konnerth and M. H. G. Prechtl, Selective hydrogenation
of N-heterocyclic compounds using Ru nanocatalysts in ionic
liquids, Green Chem., 2017, 19, 2762–2767.
14 A. R. Katritzky, S. Rachwal and B. Rachwal, Recent progress
in the synthesis of 1,2,3,4,-tetrahydroquinolines, Tetrahedron,
1996, 52, 15031–15070.
15 V. Sridharan, P. A. Suryavanshi and J. C. Menéndez,
Advances in the chemistry of tetrahydroquinolines, Chem.
Rev., 2011, 111, 7157–7259.
16 J. D. Keay, in Comprehensive Organic Synthesis, ed. B. M.
Trost and I. Fleming, Pergamon, Oxford, 1991, vol. 8, pp.
579–601.
28 M. M. Dell'Anna, V. F. Capodiferro, M. Mali, D. Manno, P.
Cotugno, A. Monopoli and P. Mastrorilli, Highly selective
hydrogenation of quinolines promoted by recyclable polymer
supported palladium nanoparticles under mild conditions in
aqueous medium, Appl. Catal., A, 2014, 481, 89–95.
29 L. Bai, X. Wang, Q. Chen, Y. Ye, H. Zheng, J. Guo, Y. Yin and
C. Gao, Explaining the size dependence in platinum-
nanoparticle-catalyzed hydrogenation reactions, Angew.
Chem., Int. Ed., 2016, 55, 15656–15661.
17 D. Ren, L. He, L. Yu, R.-S. Ding, Y.-M. Liu, Y. Cao, H.-Y. He
and K.-N. Fan, An unusual chemoselective hydrogenation of
quinoline compounds using supported gold catalysts, J. Am.
Chem. Soc., 2012, 134, 17592–17598.
30 F. Leng, I. C. Gerber, P. Lecante, W. Bacsa, J. Miller, J. R.
Gallagher, S. Moldovan, M. Girleanu, M. R. Axet and P. Serp,
Synthesis and structure of ruthenium-fullerides, RSC Adv.,
2016, 6, 69135–69148.
6896 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019

View Article Online
Catalysis Science & Technology
Paper
31 F. Leng, I. C. Gerber, P. Lecante, S. Moldovan, M. Girleanu,
M. R. Axet and P. Serp, Controlled and chemoselective
hydrogenation of nitrobenzene over Ru@C60 catalysts, ACS
Catal., 2016, 6, 6018–6024.
of C60 adsorption on Cu(111) surfaces, Phys. Rev. B: Condens.
Matter, 1997, 56, 15412–15420.
46 M. S. Dresselhaus, G. Dresselhaus and P. C. Eklund, Raman
scattering in fullerenes, J. Raman Spectrosc., 1996, 27,
351–371.
32 M. R. Axet, S. Conejero and I. C. Gerber, Ligand effects on
the
selective
hydrogenation
of
nitrobenzene
to
47 X. H. Chen, X. J. Zhou and S. Roth, Raman scattering in
calcium-doped C60, Phys. Rev. B: Condens. Matter, 1996, 54,
3971–3975.
cyclohexylamine using ruthenium nanoparticles as catalyst,
ACS Appl. Nano Mater., 2018, 1, 5885–5894.
33 C. Amiens, D. Ciuculescu-Pradines and K. Philippot,
Controlled metal nanostructures: Fertile ground for
coordination chemists, Coord. Chem. Rev., 2016, 308, 409–432.
34 B. Coq, J. M. Planeix and V. Brotons, Fullerene-based
materials as new support media in heterogeneous catalysis
by metals, Appl. Catal., A, 1998, 173, 175–183.
48 R. A. Jishi and M. S. Dresselhaus, Mode softening and mode
stiffening in C60 doped with alkali metals, Phys. Rev. B:
Condens. Matter, 1992, 45, 6914–6918.
49 S. Sakai, H. Naramoto, P. V. Avramov, T. Yaita, V. Lavrentiev,
K. Narumi, Y. Baba and Y. Maeda, Comparative study of
structures and electrical properties in cobalt-fullerene
mixtures by systematic change of cobalt content, Thin Solid
Films, 2007, 515, 7758–7764.
35 H. Veisi, R. Masti, D. Kordestani, M. Safaei and O. Sahin,
Functionalization of fullerene (C60) with metformine to
immobilized palladium as
a
novel heterogeneous and
50 L. Becker, T. P. Evans and J. L. Bada, Synthesis of
reusable nanocatalyst in the Suzuki–Miyaura coupling
reaction at room temperature, J. Mol. Catal. A: Chem.,
2014, 385, 61–67.
[hydrogenated fullerene]
C60H2 by rhodium-catalyzed
hydrogenation of C60, J. Org. Chem., 1993, 58, 7630–7631.
51 C. Chaudhari, H. Imatome, Y. Nishida, K. Sato and K.
Nagaoka, Recyclable Rh-PVP nanoparticles catalyzed
hydrogenation of benzoic acid derivatives and quinolines
under solvent-free conditions, Catal. Commun., 2019, 126,
55–60.
52 M. Mateen, K. Shah, Z. Chen, C. Chen and Y. Li, Selective
hydrogenation of N-heterocyclic compounds over rhodium-
copper bimetallic nanocrystals under ambient conditions,
Nano Res., 2019, 12, 1631–1634.
36 H. Keypour, M. Noroozi and A. Rashidi, Schiff base complex
method for the preparation of gullerene-based Ni
nanocatalyst used in the hydrogenation of benzene in
gasoline, Synth. React. Inorg. Met.-Org. Chem., 2015, 45,
1701–1709.
37 T. Braun, M. Wohlers, T. Belz and R. Schlögl, Fullerene-
based ruthenium catalysts: a novel approach for anchoring
metal to carbonaceous supports. II. Hydrogenation activity,
Catal. Lett., 1997, 43, 175–180.
53 Z. Wei, Y. Chen, J. Wang, D. Su, M. Tang, S. Mao and Y.
Wang, Cobalt encapsulated in N-doped graphene layers: an
efficient and stable catalyst for hydrogenation of quinoline
compounds, ACS Catal., 2016, 6, 5816–5822.
38 M. Lashdaf, A. Hase, E. Kauppinen and A. O. I. Krause,
Fullerene-based ruthenium catalysts in cinnamaldehyde
hydrogenation, Catal. Lett., 1998, 52, 199–204.
39 M. R. Axet, O. Dechy-Cabaret, J. Durand, M. Gouygou and P.
Serp, Coordination chemistry on carbon surfaces, Coord.
Chem. Rev., 2016, 308, 236–345.
40 C. Mabille, F. Leal-Calderon, J. Bibette and V. Schmitt,
Monodisperse fragmentation in emulsions: mechanisms and
kinetics, Europhys. Lett., 2003, 61, 708–714.
41 J. A. Osborn and R. R. Schrock, Coordinatively unsaturated
cationic complexes of rhodiumIJI), iridiumIJI), palladiumIJII),
and platinumIJII). Generation, synthetic utility, and some
catalytic studies, J. Am. Chem. Soc., 1971, 93, 3089–3091.
42 M. Green and T. A. Kuc, Cationic transition-metal
complexes. II. Reaction of arenes and olefins with bisIJ1,5-
cyclooctadiene or norbornadiene)rhodium tetrafluoroborate,
J. Chem. Soc., Dalton Trans., 1972, 832–839.
54 Y.-P. Sun, H.-Y. Fu, D.-l. Zhang, R.-X. Li, H. Chen and X.-J.
Li, Complete hydrogenation of quinoline over hydroxyapatite
supported ruthenium catalyst, Catal. Commun., 2010, 12,
188–192.
55 Y. Gong, P. Zhang, X. Xu, Y. Li, H. Li and Y. Wang, A novel
catalyst Pd@ ompg-C3N4 for highly chemoselective
hydrogenation of quinoline under mild conditions, J. Catal.,
2013, 297, 272–280.
56 R. A. Sánchez-Delgado and N. Machalaba, Hydrogenation of
quinoline by ruthenium nanoparticles immobilized on poly
(4-vinylpyridine), Catal. Commun., 2007, 8, 2115–2118.
57 B. Sun, F.-A. Khan, A. Vallat and G. Süss-Fink, NanoRu@
hectorite:
a
heterogeneous catalyst with switchable
selectivity for the hydrogenation of quinoline, Appl. Catal., A,
2013, 467, 310–314.
43 R. Uson, L. A. Oro, C. Foces-Foces, F. H. Cano, S. Garcia-
Blanco and M. Valderrama, Arene-rhodiumIJI) complexes
with trimethyltetrafluorobenzobarrelene. Crystal structure of
58 H. Mao, J. Ma, Y. Liao, S. Zhao and X. Liao, Using plant
tannin as natural amphiphilic stabilizer to construct an
aqueous-organic biphasic system for highly active and
selective hydrogenation of quinoline, Catal. Sci. Technol.,
2013, 3, 1612–1617.
59 L. M. Martinez-Prieto and B. Chaudret, Organometallic
ruthenium nanoparticles: synthesis, surface chemistry, and
insights into ligand coordination, Acc. Chem. Res., 2018, 51,
376–384.
[(Me3TFB)RhIJp-Me2C6H4)]ClO4,
1982, 229, 293–304.
J.
Organomet.
Chem.,
44 Y. Abe, K. Kato, M. Kawamura and K. Sasaki, Rhodium and
rhodium oxide thin films characterized by XPS, Surf. Sci.
Spectra, 2001, 8, 117–125.
45 K.-D. Tsuei, J.-Y. Yuh, C.-T. Tzeng, R.-Y. Chu, S.-C. Chung
and K.-L. Tsang, Photoemission and photoabsorption study
This journal is © The Royal Society of Chemistry 2019
Catal. Sci. Technol., 2019, 9, 6884–6898 | 6897

View Article Online
Paper
Catalysis Science & Technology
60 J. H. Scofield, Hartree-Slater subshell photoionization cross-
sections at 1254 and 1487 eV, J. Electron Spectrosc. Relat.
Phenom., 1976, 8, 129–137.
66 G. Kresse and D. Joubert, From ultrasoft pseudopotentials to
the projector augmented-wave method, Phys. Rev. B,
1999, 59, 1758–1775.
61 G. Kresse and J. Furthmüller, Efficiency of ab-initio total
energy calculations for metals and semiconductors using a
plane-wave basis set, Comput. Mater. Sci., 1996, 6, 15–50.
62 G. Kresse and J. Hafner, Ab initio molecular dynamics for
liquid metals, Phys. Rev. B, 1993, 47, 558–561.
63 G. Kresse and J. Furthmüller, Efficient iterative schemes for
ab initio total-energy calculations using a plane-wave basis
set, Phys. Rev. B, 1996, 54, 11169–11186.
64 G. Kresse and J. Hafner, Ab initio molecular-dynamics
simulation of the liquid-metal–amorphous-semiconductor
transition in germanium, Phys. Rev. B, 1994, 49,
14251–14269.
65 P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B,
1994, 50, 17953–17979.
67 J. P. Perdew, K. Burke and M. Ernzerhof, Generalized
gradient approximation made simple, Phys. Rev. Lett.,
1996, 77, 3865–3868.
68 K. Momma and F. Izumi, VESTA 3 for three-dimensional
visualization of crystal, volumetric and morphology data,
J. Appl. Crystallogr., 2011, 44, 1272–1276.
69 W. Tang, E. Sanville and G. Henkelman, A grid-based Bader
analysis algorithm without lattice bias, J. Phys.: Condens.
Matter, 2009, 21, 084204/1–084204/7.
70 I. del Rosal, L. Truflandier, R. Poteau and I. C. Gerber, A
density functional theory study of spectroscopic and
thermodynamic properties of surfacic hydrides on Ru (0001)
model surface: the influence of the coordination modes and
the coverage, J. Phys. Chem. C, 2011, 115, 2169–2178.
6898 | Catal. Sci. Technol., 2019, 9, 6884–6898
This journal is © The Royal Society of Chemistry 2019