DOI: 10.1039/C5CC00211G
Page 3 of 4
Journal Name
ChemComm
COMMUNICATION
under air at 80 °C for 4h. After this treatment we observed the
release of molecules in solution that correspond to ICy.(p-tol)NCN
ligand, as determined by ElectroSpray Ionization Mass Spectrometry
†
(
ESI-MS) (S5, ESI ). Thus, it can be assumed that the betaine ligand
remains intact when coordinated at the metal surface. As a
consequence, the absence of signals at 144 ppm and 150 ppm in the
1
3
C MAS-NMR spectrum of the NPs is most probably due to a line
broadening arising from the coordination of the ligand at the
ruthenium surface through the NCN nitrogen atoms which may be
reinforced by the sigma-donor ability of the carbene moiety. This
coordination mode could be evidenced by doing a blank test
consisting in using the carbodiimide ligand ( NCN (0.2 equiv.) to
Figure 3. MAS C{ H} NMR spectra of Ru-ICy.(p-tol)NCN0.1 (left)
1
3
1
and Ru-ICy.(p-tol)NCN0.2 after exposure to CO (1 bar, 20h, at r.t.).
13
p-tol)
prepare RuNPs. In these conditions, ca. 1.6 (0.4) nm were obtained
†
while COt are. This observation is in agreement with the lack of for
mobility observed terminal CO which can result from the nearby
coordination of betaine ligands thus limiting their moving. As the
main difference between the two RuNP samples is the decrease in
(
S6, ESI ), thus proving the ability of the NCN motif to coordinate at
the metallic surface.
The next step was to probe the surface state of the particles by
measuring their capacity to bind CO by Fourier transform infrared
.
(p-
size with increased ligand quantity (ca.1.0 nm for Ru-ICy
tol)
.(p-tol)
NCN0.2 against ca. 1.3 nm for Ru-ICy
NCN0.1), a plausible
1
3
1
13
(
1
(
FT-IR) and MAS
C{ H} and CP MAS
C{ H} NMR.
explanation could be that the coordination of CO in a bridging mode
is hindered by a lack of available faces on the surface for the smaller
p-tol)
Interestingly, FT-IR spectra of purified NPs Ru-ICy
NCN0.2 showed a band displaying the characteristic of a
CO absorption between 1900 and 2000 cm before their exposure to
NCN0.1 and
.
(p-tol)
Ru-ICy
.
(p-tol)
(p-tol)
13
.(p-tol)
Ru-ICy
NCN0.2
.
Nevertheless, either in Ru-ICy
NCN0.1 or
-
1
.
Ru-ICy
NCN0.2 NPs, CO is mainly coordinated in terminal
.
(p-tol)
CO. Moreover, for the Ru-ICy
NCN0.1 sample, the intensity of
mode. This can result from the very small size of the NPs, where the
face surfaces is limited and most ruthenium atoms concentrate on
edges and apexes, where coordination in a terminal mode is more
favorable.
the band did not change after CO exposure (1 bar; r.t.), suggesting
that the NP surface was already saturated by CO after synthesis. Ru-
.
(p-tol)
ICy
NCN0.2 showed a different band intensity suggesting that in
this case the Ru surface was not totally covered by CO, (S7 and S8,
†
ESI ). Then, the question was: where does CO come from?
The small size and specific electronic configuration of MNPs make
them interesting systems for many catalytic reactions. In particular
RuNPs are able to catalyze both the hydrogenation of alkenes and
arenes under mild conditions while molecular complexes are just
Considering the reaction conditions for the RuNPs synthesis, the
only source of CO could be THF, through a decarbonylation process
which would release propane. We thus analyzed the composition of
the reaction sky resulting from the NP synthesis mixture by gas
phase NMR and Mass Spectrometry (MS). After the synthesis of Ru-
1
4
active in alkene hydrogenation. Styrene hydrogenation can thus be
applied as a model reaction to probe the surface reactivity of MNPs
because it offers the possibility of measuring the selectivity between
vinyl and aromatic hydrogenation. Thus, we investigated the
.
(p-tol)
ICy
NCN0.1, we observed two peaks at 0.75 and 1.20 ppm by gas
†
phase NMR corresponding to propane (S9, ESI ). By MS both
compounds, CO and propane, were detected (S10, ESI ). In addition,
following a previously reported procedure, the RuNPs were reacted
with 2-norbornene and the amount of norbornane formed was
†
.
(p-tol)
1
3
performance of Ru-ICy
NCN NPs in styrene hydrogenation in
mild reaction conditions. Table 1 reports the results obtained in
.
(p-tol)
.(p-
.
(p-tol)
styrene hydrogenation with Ru-ICy
NCN0.1 and Ru-ICy
measured by G.C. analysis (see ESI†). Ru-ICy
NCN0.1 and Ru-
tol)
.(p-tol)
.
(p-tol)
NCN0.2 at r.t. in THF. Ru-ICy NCN0.1 hydrogenates styrene
first to ethylbenzene and then to ethylcyclohexane. The Ru-ICy
NCN0.1 catalyst allowed the total hydrogenation of styrene into
ICy
NCN0.2 NPs display, 0.0 and 0.6 H per surface Ru (Table
.
(p-
S1, ESI†); These values are lower than usual ones for other RuNPs
and evidence elimination of hydrides from the surface probably due
to a full or partial coverage of the Ru surface by CO, respectfully.
Altogether, these data demonstrate that CO is coordinated at the
RuNP surface, as the result of a decarbonylation process of THF
used as solvent during their synthesis, hence evidencing the high
tol)
.
(p-tol)
2
4h. While using Ru-ICy
NCN0.2 as nanocatalyst, the
hydrogenation rate of the vinyl group remained similar to that of Ru-
.
(p-tol)
ICy
NCN0.1, the hydrogenation rate of the arene moiety
decreased. The selectivity for vinyl hydrogenation was therefore
lightly improved: 74 % of the ethylbenzene was hydrogenated to
ethylcyclohexane after 24 h. As the presence of available faces are
1
3
1
surface reactivity of these NPs. The MAS C{ H} NMR spectrum
.
(p-tol)
of solid samples of Ru-ICy
NCN0.1 NPs exposed to 1 bar of
CO at r.t. (Figure 3 left) displays two signals attributed to adsorbed
CO molecules: a broad peak at δ ~ 229 ppm that can be assigned to
1
5
1
3
necessary to hydrogenate aromatic rings, these catalysis results
.
(p-tol)
corroborate the fact that less faces are present in Ru-ICy
NCN0.2
.
(p-tol)
.(p-
1
3
NPs than in Ru-ICy
NCN0.1 ones. This makes Ru-ICy
NCN0.2 NPs to have a behavior in catalysis more similar to that of
CO coordinated in a bridging mode (COb) and a sharp resonance at
tol)
1
3
δ = 199 ppm due to CO coordinated in a terminal mode (COt).
From previous works, bridging CO molecules are expected to be
coordinated onto faces and terminal ones onto corners and edges.13
1
6
molecular complexes than to facetted nanoparticles.
Moreover, the presence of spinning bands (*) suggests that COt are Conclusions
static on the surface. The limitation of CO mobility on the metallic
surface can be explained by the coordination of betaine ligands in
A betaine adduct of NHC and carbodiimide (NHC-NCN) has been
close proximity of the ruthenium sites where CO are coordinated. On
identified as a new ligand capable of stabilizing very small Ru
1
3
1
†
the CP MAS C{ H} NMR spectrum (S11, ESI ), the intensity of
the signal for bridging CO is decreased compared to that of the
terminal ones. This means that bridging CO groups are not affected
by cross polarization from close hydrogen carriers while terminal
COs are. It can be thus assumed that bridging CO groups are not
located in the vicinity of hydrogen carriers (namely, betaine ligands)
(p-tol)
nanoparticles. Specifically, ICy·
NCN gave rise to ultra-small
and crystalline Ru NPs. A clear relationship was observed between
(p-tol)
the ICy·
NCN/Ru ratio used in the NP synthesis and the mean
size of the obtained NPs. In particular ca. 1.3 nm NPs were produced
(p-tol)
when 0.1 equiv. of ICy·
NCN per Ru were used and ca. 1.0 nm
NPs in the case of 0.2 equiv. This difference in size leads to different
This journal is © The Royal Society of Chemistry 2012
J. Name., 2012, 00, 1-3 | 3