Characterization of secondary phosphine oxide ligands on the surface of iridium nanoparticles

Article abstract of DOI:10.1039/c7cp03439c

The synthesis of iridium nanoparticles (IrNPs) ligated by various secondary phosphine oxides (SPOs) is described. This highly reproducible and simple method via H₂ reduction produces well dispersed, small nanoparticles (NPs), which were characterized by the state-of-the-art techniques, such as TEM, HRTEM, WAXS and ATR FT-IR spectroscopy. In particular, multinuclear solid state MAS-NMR spectroscopy with and without cross polarization (CP) enabled us to investigate the different binding modes adopted by the ligand at the nanoparticle surface, suggesting the presence of three possible types of coordination: as a purely anionic ligand Ir-P(O)R₂, as a neutral acid R₂P-O-H and as a monoanionic bidentate H-bonded dimer R₂P-O-H?O=PR₂. Specifically, the higher basicity of the dicyclohexyl system leads to the formation of IrNPs in which the bidentate binding mode is most abundant. Such cyclohexyl groups are bent towards the edges, as is suggested by the study of ¹³CO coordination on the NP surface. This study also showed that the number of surface sites on faces available for bridging CO molecules is higher than the number of sites for terminal CO species on edges and apices, which is unexpected taking into account the small size of the nanoparticles. In addition, the IrNPs present a high chemoselectivity in the hydrogenation of cinnamaldehyde to the unsaturated alcohol.

Full text of DOI:10.1039/c7cp03439c

PCCP
View Article Online
View Journal
PAPER
Characterization of secondary phosphine oxide
ligands on the surface of iridium nanoparticles†
Cite this:DOI: 10.1039/c7cp03439c
a
bc
Israel Cano, *a Luis M. Martınez-Prieto,
Pier F. Fazzini,a Yannick Coppel,
´
Bruno Chaudreta and Piet W. N. M. van Leeuwen
*
a
The synthesis of iridium nanoparticles (IrNPs) ligated by various secondary phosphine oxides (SPOs) is
described. This highly reproducible and simple method via H2 reduction produces well dispersed, small
nanoparticles (NPs), which were characterized by the state-of-the-art techniques, such as TEM, HRTEM,
WAXS and ATR FT-IR spectroscopy. In particular, multinuclear solid state MAS-NMR spectroscopy with
and without cross polarization (CP) enabled us to investigate the different binding modes adopted by
the ligand at the nanoparticle surface, suggesting the presence of three possible types of coordination:
as a purely anionic ligand Ir–P(O)R2, as a neutral acid R2P–O–H and as a monoanionic bidentate
H-bonded dimer R2P–O–HÁ Á ÁOQPR2. Specifically, the higher basicity of the dicyclohexyl system leads
to the formation of IrNPs in which the bidentate binding mode is most abundant. Such cyclohexyl
groups are bent towards the edges, as is suggested by the study of 13CO coordination on the NP
surface. This study also showed that the number of surface sites on faces available for bridging CO
molecules is higher than the number of sites for terminal CO species on edges and apices, which is
unexpected taking into account the small size of the nanoparticles. In addition, the IrNPs present a high
chemoselectivity in the hydrogenation of cinnamaldehyde to the unsaturated alcohol.
Received 22nd May 2017,
Accepted 27th July 2017
DOI: 10.1039/c7cp03439c
rsc.li/pccp
depend on the hydrocarbyl substituents at phosphorus. Aromatic
SPOs ligated to AuNPs present a strong polarity in their PQO
Introduction
The use of secondary phosphine oxides (SPOs) as stabilizing bonds and they coordinate to the surface in their anionic form, B.
ligands for iridium nanoparticles (IrNPs) has been scarcely In contrast, the more basic aliphatic phosphine oxides hold on
explored. Our group has a longstanding experience in the use more strongly to the protons and the actual coordination mode
of SPOs as multifunctional ligands, both in organometallic1–6 is at least in part the neutral acid A or the H-bonded dimer C.
and MNP6–10 catalysis. SPOs may coordinate as a neutral acid The AuNPs prepared with aryl-substituted SPOs were active
(A),11,12 a deprotonated phosphinito anion (B),13,14 or a hydrogen catalysts for the highly chemoselective hydrogenation of sub-
bonded pair of the two (C),4–6,11 thus obtaining a monoanionic, stituted aldehydes. On the contrary, AuNPs containing alkyl
supramolecular bidentate ligand (Scheme 1).15 In addition, the acid SPOs showed low activity and selectivity. In a subsequent
can be deprotonated to generate the O-metalated species D.16,17
publication we described the formation of IrNPs stabilized with
We demonstrated the presence of P–O–H groups in ruthenium chiral SPOs,10 in which we suggested a binding mode of the
nanoparticles (RuNPs) ligated by SPOs through H2/D2 isotope ligand on the IrNPs surface similar to that previously described
exchange reactions.7 Along this line, our group developed the for AuNPs ligated by aryl SPOs, i.e. a purely anionic SPO would
synthesis of gold nanoparticles (AuNPs) stabilized by SPO be present. However, definite proofs were not obtained. These
ligands.8,9 An exhaustive characterization demonstrated that the IrNPs were active catalysts for the asymmetric hydrogenation
nature of the ligand is a key feature, since important differences of ketones. Furthermore, we recently reported a comparative
in the size, morphology and catalytic behavior were found to catalytic study in which a specific SPO ligand (tert-butyl-
(phenyl)phosphine oxide) was employed to generate an Ir–SPO
complex and SPO-stabilized IrNPs.6 These Ir–SPO based catalysts
showed markedly different selectivities and rates in the hydrogena-
tion of various substituted aldehydes. This ligand coordinates as the
H-bonded dimer C in the Ir–SPO complex, but the actual coordina-
tion mode of the SPO on the IrNP surface was not established.
Therefore, studies focusing on the location and possible
a Laboratoire de Physique et Chimie des Nano Objets, LPCNO,
´
UMR5215 INSA-UPS-CNRS, Institut National des Sciences Appliquees,
135 Avenue de Rangueil, Toulouse 31077, France.
E-mail: israel.canorico@nottingham.ac.uk, vanleeuw@insa-toulouse.fr
b CNRS, LCC (Laboratoire de Chimie de Coordination) 205 Route de Narbonne,
BP44099, F-31077 Toulouse Cedex 04, France
c Universite’ de Toulouse, UPS, INPT, F-31077 Toulouse Cedex 04, France
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7cp03439c types of coordination by the ligands present at the nanoparticle
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
Scheme 1 Tautomeric forms of SPOs and possible coordination modes to a metal center.
surface are of great importance, but the determination
TEM micrographs revealed the presence of well distributed
of the real binding mode remains difficult. Thus, inspired by these and small NPs (Fig. 2 and Section 6, ESI†) with a size in the
works, herein we report Magic Angle Spinning NMR (MAS-NMR) range 1.3–1.5 nm (1, 1.33(0.28) nm; 2, 1.43(0.38) nm; 3,
spectroscopy as a powerful tool that can provide insight in the 1.39(0.50) nm), although 3 displayed a broader dispersion than
morphology and coordination mode of SPO ligands at the surface 1 and 2 and some aggregation of NPs.
and hence about the reactivity of the nanoparticles. To this end, a
HRTEM images showed crystalline IrNPs compatible with
group of SPOs with different electronic and steric properties was a cubic close-packed structure (ccp). This was confirmed by
selected for the generation of a series of IrNPs (Fig. 1).
Fast Fourier Transformation (FFT) analysis carried out on the
HRTEM images as shown in Fig. S18 (ESI†). WAXS analysis
confirmed the ccp structure and indicated a coherence length
close to ca. 1.5 nm (Section 4, ESI†). These data are in agree-
ment with the sizes determined by TEM and corroborate the
crystallinity of the IrNPs.
Elemental analyses (EA) gave an Ir content in the range
48–50% and were consistent with a metal–ligand ratio of 1.85 : 1
for 1, 1.66:1 for 2, and 1.67 : 1 for 3 (Table 1). It is worth noting
that the Ir/L ratio accounts for a large amount of SPO coordinated
to the surface of the NP, since the approximate number of ligands
is only slightly lower than the calculated number of surface atoms
(estimated from IrNP mean size as observed by TEM).22
Fig. 1 Secondary phosphine oxides employed in this study.
The IrNPs were characterized by the use of a wide range of
techniques, including Transmission Electron Microscopy
(TEM), High Resolution TEM (HRTEM), Wide Angle X-Ray Table 1 Analytical data of IrNPs 1–3
Scattering (WAXS) and Attenuated Total Reflectance Fourier-
a
b
IrNP
Diameter (nm)
Ir content (%)
Ir/L ratio
Irx/Ly
Ns
Transform Infrared (ATR FT-IR) spectroscopy. Specifically,
MAS-NMR spectroscopy enabled us to explore the possible
coordination modes adopted by the SPO ligand at the NP
surface. Multinuclear solid state NMR techniques have been
used to characterize the structure of ligands on MNPs and their
surface chemistry as summarized in several recent reviews.18–21
1
2
3
1.33 Æ 0.28
1.43 Æ 0.38
1.39 Æ 0.50
47.9
50.4
48.6
1.85 : 1
1.66 : 1
1.67 : 1
117/63
145/87
133/80
78
91
89
a
The approximate composition is based on the Ir/L ratio and the
b
number-average diameter measured. Number of surface atoms.
Approximate values obtained from the graphs of Van Hardeveld and
Hartog.23
Results and discussion
Synthesis and characterization of SPO-stabilized Ir
The ATR FT-IR spectra of free SPO ligands showed the P–H
stretching absorption in the range 2280–2370 cmÀ1 (Fig. S3, S5
and S8, ESI†). This P–H band was not visible in the ATR FT-IR
spectra of 1–3 (Fig. S4, S6 and S9, ESI†), suggesting coordination
of the ligand to the NP surface through the phosphorus atom.
Additionally, no absorption was observed in the region corres-
ponding to the (P–O)–H stretching band (ca. 3100–3500 cmÀ1).
This lack of protons suggests that the phosphine adopts a
coordination mode similar to that previously described for aryl
SPO-ligated AuNPs (B).8,9 However, this hypothesis could not be
confirmed by ATR FT-IR spectroscopy, since the O–H–O vibra-
tion of the third possible coordination mode of SPO ligands
to metals, that is as a monoanionic bidentate ligand with
an intramolecular hydrogen bond (C), is usually not observed
in IR.24
nanoparticles
We chose the same protocol that was previously used for the
synthesis of SPO-stabilized IrNPs.6,10 The IrNPs 1–3 were prepared
by H2 reduction of (1,5-cyclooctadiene)(methoxy)iridium(I) dimer
([Ir(OMe)(COD)]2) in the presence of 0.5 eq. (based on Ir) of SPO L
(Scheme 2, for further details see Section 2, ESI†).
Finally, an intense band at ca. 1980–2005 cmÀ1 was found in
the spectra of IrNPs and recognized as CO coordinated to the
Scheme 2 Synthesis of IrNPs 1–3 from [Ir(OMe)(COD)]2 and L1–L3.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
Fig. 2 Representative TEM micrographs of IrNPs 1–3 and size distributions determined from TEM images by counting 4300 non-touching particles
obtained from images captured from distinct quadrants of the grid.
NP surface, which probably results from a decarbonylation Solid state NMR studies
process of THF and/or methoxide during the NP formation.10,25
The intensity of this band increased after bubbling CO
modes of the ligands on the nanoparticle surface that can be
through solutions of 1–3 in THF (R.T., 5 min, Fig. 3 and
Section 5, ESI†), confirming the presence of coordinated
CO at the Ir surface in the purified NPs. This band is
identified as CO coordinated in a terminal mode (COt), while
Given the relatively limited information about the coordination
obtained by FT-IR, we found solid state NMR to be a useful
spectroscopic tool for this purpose. The coordination of ligands
to the nanoparticle surface was demonstrated by solid state
MAS NMR spectroscopy. 31P MAS NMR spectra of IrNPs 1–3
another stretching band appears at lower frequency shift
(Fig. S19–S22, ESI†) showed broad heterogeneous signals
(ca. 1800 cmÀ1), which is assigned to CO coordinated in a
centered around 68 ppm for 1, 29 ppm for 2 and 94 ppm for
bridging mode (COb).26,27 The ATR FT-IR spectra of 1–3 before
3, with spectral range between 80 and 120 ppm. These broad
bubbling with CO display a slight displacement in the COt
signals are clearly different from the sharp ones of the corres-
band to lower frequency following an increase in the electron-
ponding free ligands (polymorphic crystal form signals centered
donor character of the substituents in the ligand, and thus in
at 42 ppm for L1, 17 ppm for L2, and 48 ppm for L3) with
on average a resonance shift to higher frequency. The high
frequency shift can be associated to the coordination of the
ligand to the metal through the phosphorus atom, as is observed
the presence of a more electron-rich iridium surface (Fig. S11,
ESI†).28,29 Conversely, an increase in the amount of CO on the
NP surface involves a shift of the COt stretching absorption to
for molecular Ir–SPO complexes.6,11,14,34 The spinning frequency
higher frequency (Fig. 3), as was earlier reported for metal
surfaces.30–33
for 31P NMR experiments was set to 16 kHz (or above) to average
the chemical anisotropy and the 31P homonuclear dipolar couplings
to zero. With this fast MAS NMR, the broadness and heterogeneity
of 31P signals corresponding to the ligands coordinated on the
IrNP surface can be mostly associated to chemical shift dis-
tributions that come from chemical and structural heterogeneities
(or disorder).
Broadening contributions coming from magnetic field
inhomogeneity, magnetic susceptibility and chemical shift
distribution can be removed with a Carr–Purcell–Meiboom–Gill
(CPMG) NMR experiment.35–44 The 31P resonances of IrNPs 1–3
showed strong sharpening in a CPMG experiment, confirming
the presence of structural and chemical heterogeneities for SPO
ligands at the IrNP surface (Fig. S23, ESI†). 1H/31P HETCOR
experiments with short contact time (0.3 ms) associated with
CPMG detection to improve sensitivity showed that the different
31P resonances of IrNPs 1–3 are not spatially close to the same
hydrogen atoms (Fig. 4 and Fig. S24, S25, ESI†). In addition
Fig. 3 Overlay of ATR-IR spectra of 1 before and after bubbling with CO
in the region 1000–2400 cmÀ1. In view of the short contact times the
spectra do not represent equilibrium states.
to the correlations with protons of organic substituents
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
(phenyl with dH B 7.3 ppm, cyclohexyl with dH B 1.4 ppm and sometimes observed. This 31P signal has an unusual chemical
tert-butyl with dH B 1.2 ppm groups), the 31P resonances shift for purely anionic or monoanionic H-bonded dimer
showed dipolar correlations with 1H in the range between species, with a short TP1 relaxation time of 60(Æ20) ms com-
6 and 14 ppm, indicating the presence of different OH groups pared to the TP1 of the main 31P signals of IrNPs 1–3 (average TP1
close to 31P atoms, engaged in weak or strong hydrogen bonds. of 5.7(Æ0.7) s, 7.8(Æ1.1) s and 10.7(Æ2.0) s, for 1, 2 and 3,
This proves that, in addition to the deprotonated anionic respectively). Due to its short TP1, this resonance can be edited
form B, monoanionic H-bonded dimer C (correlations with in a 31P experiment using a TP1 filter (Fig. S26, ESI†). We
OH signal above 7 ppm) and neutral acid A (correlations with assigned this signal to Knight-shifted 31P resonances of ligand
OH signal in the range 3–6 ppm) ligands are also present. For coordinated to coalesced IrNPs.45–51 In transition metals such
IrNPs 1 and 2 (Fig. S24 and S25, ESI†), the monoanionic as iridium, the 31P NMR resonances of surface-adsorbed mole-
bidentate coordination mode showed on average lower fre- cules on large MNPs experience a large paramagnetic shift, a
quency shifts (heterogeneous signals centered around 42 ppm Knight-shift,52 due to the mixing of molecular orbitals with
for 1 and 20 ppm for 2, respectively) than the purely anionic metal conduction bands of electrons.45,46 Indeed, we have
form (heterogeneous signals centered around 80 ppm for 1 and previously proved that the presence of conduction electrons
50 ppm for 2, respectively), a trend generally observed for at the surface of MNPs is related to their size, leading to the
31P resonances. Furthermore, the concentration of neutral acid Knight-shift found in large NPs.52,53
species seems very low. For IrNPs 3 (Fig. 4), the heterogeneous
The 13C MAS NMR spectra of 1–3 exhibit the characteristic
signals centered around 94 ppm correspond mainly to signals belonging to the respective ligands present. 13C resonance
H-bonded dimer species, while the neutral acid binding mode broadenings after the SPO coordination to the IrNP surface are
(dipolar correlations at dP B 50 ppm and dH B 6.1 ppm) is significantly less than the ones observed for the 31P resonances
observed in a small ratio and anionic ligands seem to be a (Fig. S27–S29, ESI†), confirming the binding of the SPO ligands
minority as is to be expected for the most basic ligand. The to the IrNPs through the phosphorus atoms. In addition, the
1H/31P CPMG-HETCOR experiment clearly illustrated the struc- 13C MAS NMR spectrum of 2 (Fig. S28, ESI†) presents a signal at
tural and chemical heterogeneity of the IrNPs samples.
ca. 28 ppm that corresponds to cyclohexyl groups obtained by a
partial hydrogenation of phenyl groups during the synthesis
under H2. The small peak at ca. 67 ppm observed for 1–3 is
likely due to MeOH.10
13CO was employed as a probe molecule to investigate the
surface chemistry and the location of the open sites of these
particles by solid state 13C MAS-NMR. Thus, solid samples of
1–3 were exposed to 1 bar of 13CO at R.T. for 20 h. 13C MAS spectra
of 1–3 display new signals above 150 ppm corresponding to CO
coordinated to the nanoparticle surface (Fig. S30, ESI†).10,54–58
Two set of signals were identified (Table 2): (i) a broad signal
centered at 210 Æ 10 ppm with spectral range of 115 Æ 15 ppm
that belongs to CO coordinated in a bridging mode (COb); (ii) a
narrow signal centered at 171 Æ 3 ppm with spectral range of
35 Æ 5 ppm assigned to CO coordinated in a terminal mode
(COt).10 Quantification of CO species was then performed in a
Hahn-echo experiment with long recovery time (100 s, Fig. S30,
ESI†). For the three samples, bridging CO molecules were the
most abundant species with a CO population of 83, 75 and 58
Fig. 4 31P CPMG-HETCOR MAS NMR spectrum of IrNPs 3. Lines are given
as guide to the eyes.
A broad 31P signal at higher frequency (centered around (Æ4)% of total CO for IrNPs 1, 2 and 3, respectively. Previous
130(Æ20) ppm with half-height linewidth of 35(Æ5) kHz) was investigations on CO coordination over ruthenium nanoparticles
Table 2 13C NMR data of CO species in IrNPs 1–3
1
2
3
IrNPs
COb
COt
COb
COt
COb
COt
d (ppm)
222 Æ 3
120 Æ 5
83 Æ 2
173 Æ 1
37 Æ 2
208 Æ 3
127 Æ 6
75 Æ 2
170 Æ 1
38 Æ 2
200 Æ 3
103 Æ 4
58 Æ 2
169 Æ 1
31 Æ 2
Linewidth (ppm)
Ratioa (%)
TC1 (s)
17 Æ 2
25 Æ 2
42 Æ 2
1.7 Æ 0.5
2.5 Æ 0.3
13 Æ 2
21.7 Æ 2.9
2.2 Æ 0.3
14 Æ 2
1.4 Æ 0.2
2.8 Æ 0.3
11 Æ 2
29.5 Æ 2.1
2.4 Æ 0.6
13 Æ 2
29.3 Æ 7.4
2.4 Æ 0.6
6 Æ 2
32.3 Æ 2.8
1.0 Æ 0.3
7 Æ 2
TIS (ms)
TH1r (ms)
a
Ratio of bridging CO to terminal CO obtained by deconvolution of the Hahn-echo spectra (Fig. S31, ESI).
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
proved that COb molecules coordinate onto faces and COt onto given that in a densely packed situation the O atoms of SPO
edges and apices.25,59,60 The higher relative amount of COb ligands look at the faces, whereas the cyclohexyl substituents
molecules indicates that the number of surface sites available point to the edges and thus are closer to COt molecules. The
for these on faces is large in comparison with the corresponding proton density and internal dynamics also affect the T H1r, which
sites for COt on edges and apices, which is unexpected due to the is the shortest for 3 (6.5 Æ 2 ms) while it is longer for 1 and 2
small size of the IrNPs. Interestingly, the relative population of (13.5 Æ 2 and 12.0 Æ 2 ms, respectively). Note that the similar
COb is the smallest for 3 (58 Æ 2% vs. 83 Æ 2 and 75 Æ 2% for T H1r values observed for COt and COb were expected. Indeed, in
1 and 2), which points to a steric effect of cyclohexyl groups that a dense proton network, spin diffusion phenomena usually
restricts CO coordination on faces.
average the TH1r over few nm distances.
T1 can be affected by motion on a time-scale of the Larmor
The collected NMR data demonstrated that IrNPs 1–3 are
frequency (in the ps to ms range).61 The T1C of the two types similar, but IrNPs 3 showed slight differences due to the
of CO resonances were found very different for IrNPs 1 and 2 presence of two cyclohexyl substituents in the ligand L3. The
with respective values of 1.7 Æ 0.5 s, 1.4 Æ 0.2 s for COb and higher basicity or electron-donor character of the dicyclohexyl
21.7 Æ 2.9 s, 29.5 Æ 2.1 s for COt species. For IrNPs 3, the T C1 of system (compared to diphenyl and tert-butyl(phenyl)) favors the
COb and COt species were more similar, with respective values coordination of the SPO ligand as the H-bonded dimer C, as
of 29.3 Æ 7.4 and 32.3 Æ 2.8 s. The variations in TC1 observed for was previously observed for AuNPs ligated by alkyl SPOs.9 Their
COb and COt of 1 and 2 suggest a notable difference of local bulkiness also decreases the ratio of CO coordinated in a bridging
mobility for these species at the surface of IrNPs. The most mode to CO coordinated in a terminal mode and limits the
likely explanation is a faster diffusion of the COb molecules.62,63 mobility of COb molecules.
For IrNPs 3, the steric effect of the cyclohexyl groups and/or the
presence of mainly monoanionic bidentate ligands could limit
the COb diffusion.
Catalytic hydrogenation of substituted aldehydes with 1–3
SPOs are an interesting alternative to common tertiary phos-
phorus ligands in catalysis because of their singular properties.
Indeed, their corresponding metal complexes exhibit a particular
ability to cleave H2 heterolytically through the metal and the
oxygen atom.1–6 Our group applied this concept for hydrogena-
tion reactions with Ru,7 Au8,9,64 and Ir6,10 NPs stabilized by SPOs,
for which the ligand plays a triple role, as stabilizer for nano-
particles, as modifying ligand, and as functional ligand supplying
the oxygen function that participates in the heterolytic cleavage of
dihydrogen. Along this line, we decided to evaluate the catalytic
To get more information about the CO species, we studied
their CP kinetics with 1H (Fig. 5 and Fig. S32, ESI†). The CP
kinetics should follow the classical I–S model,61 since the H–CO
heteronuclear dipolar interactions are expected to be weak
(spatial remoteness) while the H–H homonuclear dipolar inter-
actions should be moderate or strong. The CP kinetics in the
I–S model is described by eqn (1):61
I = I(0)(1 À TIS/T1Hr
)
À1[exp(Àt/T1Hr)] À exp(Àt/TIS)
(1)
in which t is the contact time, I(0) is the signal amplitude, TIS is activity of IrNPs 1–3 in the chemoselective hydrogenation of
the CP time constant and TH1r is the proton spin–lattice relaxa- cinnamaldehyde and p-nitrobenzaldehyde (Table 3). The optimal
tion time in the rotating frame. For IrNPs 1–3 the TIS of COt and reaction conditions were found to be THF as the solvent, R.T. and
COb show differences (Table 2). For IrNPs 1 and 2 the TIS 18 h as reaction time. The IrNPs were very selective to the
difference between the two CO species is small (2.2 Æ 0.3 ms vs. carbonyl functionality in cinnamaldehyde and the corresponding
2.5 Æ 0.3 ms and 2.4 Æ 0.6 ms vs. 2.8 Æ 0.3 ms for 1 and 2, cinnamyl alcohol was obtained with high chemoselectivity
respectively). However, for 3 the TIS difference is more pro- (entries 1–3). As shown, the smallest NPs, 1, are the most active
nounced (1.0 Æ 0.3 ms vs. 2.4 Æ 0.6 ms). The short TIS observed ones, followed by 2 and 3. The former ones are the largest NPs
for COt of 3 can be explained by the high density of protons and for the latter aggregation was observed which could lead
close to the terminal CO due to presence of the cyclohexyl to a decrease in activity. On the other hand, we observed a loss
groups, since TIS is governed by internuclear distances (1/r3 factor) of selectivity in the hydrogenation of p-nitrobenzaldehyde
and the number of neighbouring protons. This is not surprising, (entries 4–6) and both aldehyde and nitro groups were
Fig. 5 13C CP/MAS kinetic curves for IrNPs 1–3.
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
Table 3 Catalytic hydrogenation of aldehydes with 1–3a
Entry
Substrate
IrNP
Product
Conversionb (%)
Selectivityc (%)
1
2
3
1
2
3
99
72
61
99 : 1 (UA : A)
98 : 2 (UA : A)
98 : 2 (UA : A)
4
5
6
1
2
3
499
499
499
95
50d
95
a
Reagents and conditions: 1–3 (0.0025 mmol Ir assuming % of Ir from elemental analysis), substrate (1 mmol), THF (0.75 mL), 18 hours, 293 K,
b
1
10 (entries 1–3) or 20 (entries 4–6) bar H2. Conversions and product identities were determined by H NMR spectroscopy (average of two runs).
c
d
UA = unsaturated alcohol, A = saturated alcohol. 4-Aminobenzaldehyde, azoarenes and other hydrogenation compounds as by-products.66,73
hydrogenated, affording 4-aminobenzyl alcohol as main apices, which is unexpected in view of the small size of the
product and other by-products resulting from nitro group IrNPs. The IrNPs 3 exhibit the lowest relative population of COb
hydrogenation, such as azoarenes and 4-aminobenzaldehyde. molecules, since the bulkiness of cyclohexyl groups and the
The reduction of nitroarenes to anilines has been extensively presence of mainly monoanionic bidentate SPOs decrease the
described with the use of MNPs.65–72
ratio of COb to COt and limits the mobility of both species.
In conclusion, thanks to solid state NMR we have elucidated
The IrNPs described herein do not exhibit any ligand effect
and, except for slight variations, 1–3 display a similar catalytic the coordination modes adopted by SPO ligands at the NP surface,
behaviour. This is unsurprising since the ionic form B is not the obtaining a clear picture of the surface of SPO-stabilized IrNPs.
exclusive binding mode of the SPOs for the IrNPs, and also Therefore, multinuclear solid state NMR spectroscopy has shown
the neutral acid A and the H-bonded dimer C exist at the its merit as a powerful technique for the comprehensive char-
nanoparticle surface. As a consequence, the Ir surface contains acterization of MNPs, which can aid the design of ligands that
different active sites, which leads to a reduction in selectivity, in provide better catalytic properties.
a similar way to the previously reported alkyl SPO AuNPs.9
In this context, AuNPs ligated by SPOs showed drastic differ-
ences in the properties depending on the substituents in the
Conflicts of interest
ligand.8,9 SPOs with aromatic substituents present R2PQO
anions as the single type of ligand coordination, and thus There are no conflicts of interest to declare.
electron-withdrawing SPOs could possibly do this for IrNPs.
Acknowledgements
Conclusions
We thank UPS-Toulouse and INSA for financial support.
In summary, we have successfully used a series of secondary The authors acknowledge Dr Pierre Lacante for WAXS analyses
phosphine oxide ligands for the synthesis of iridium nano- (CEMES-CNRS). Prof. P. W. N. M. van Leeuwen thanks the
´ ´ ´
´ ´
particles. The IrNPs showed a high chemoselectivity in the Universite Federale Toulouse Midi-Pyrenees for an IDEX Chaire
´
hydrogenation of cinnamaldehyde, although we observed a loss d’Attractivite.
of selectivity for p-nitrobenzaldehyde and both aldehyde and
nitro groups were hydrogenated. An extensive characterization
of these nanoparticles was performed by several procedures.
Notes and references
Specifically, we have proven that multinuclear solid state NMR
spectroscopy is a powerful tool for the characterization of
phosphine oxides directly attached to the IrNP surface, reveal-
ing structural and chemical heterogeneity of ligands at the
surface. This study showed that the SPOs are bound to the NP
surface by three coordination modes: as a purely anionic ligand
Ir–P(O)R2, as the neutral acid R2P–O–H and as a monoanionic
bidentate H-bonded dimer R2P–O–HÁ Á ÁOQPR2. Interestingly,
the bidentate binding mode is the most favored for 3 due to the
higher basicity of the dicyclohexyl system.
1 P. W. N. M. van Leeuwen and C. F. Roobeek, Eur. Pat. Appl.
EP 82576, 1983 (Chem. Abstr., 1983, 99, 121813.
2 P. W. N. M. van Leeuwen, C. F. Roobeek, R. L. Wife and
J. H. G. Frijns, J. Chem. Soc., Chem. Commun., 1986, 31–33.
3 P. W. N. M. van Leeuwen, C. F. Roobeek, J. H. G. Frijns and
A. G. Orpen, Organometallics, 1990, 9, 1211–1222.
4 P. W. N. M. van Leeuwen and C. F. Roobeek, New J. Chem.,
1990, 14, 487–493.
5 P. M. Castro, H. Gulyas, J. Benet-Buchholz, C. Bo, Z. Freixa
and P. W. N. M. van Leeuwen, Catal. Sci. Technol., 2011, 1,
401–407.
In addition, the surface state of the nanoparticles was
studied by 13CO coordination. Solid state 13C MAS-NMR experi-
ments showed that the number of open sites on faces available
for COb is higher than the number of sites for COt on edges and
´
6 I. Cano, L. M. Martınez-Prieto, B. Chaudret and P. W. N. M.
van Leeuwen, Chem. – Eur. J., 2017, 23, 1444–1450.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
7 E. Rafter, T. Gutmann, F. Low, G. Buntkowsky, K. Philippot, 30 R. P. Eischens and W. A. Pliskin, Adv. Catal., 1958, 10, 1–56.
B. Chaudret and P. W. N. M. van Leeuwen, Catal. Sci. 31 A. M. Bradshaw and F. M. Hoffmann, Surf. Sci., 1978, 72, 513–535.
Technol., 2013, 3, 595–599.
32 S. Kinayyigit, P. Lara, P. Lecante, K. Philippot and
8 I. Cano, A. M. Chapman, A. Urakawa and P. W. N. M. van
Leeuwen, J. Am. Chem. Soc., 2014, 136, 2520–2528.
9 I. Cano, M. A. Huertos, A. M. Chapman, G. Buntkowsky,
T. Gutmann, P. B. Groszewicz and P. W. N. M. van Leeuwen,
J. Am. Chem. Soc., 2015, 137, 7718–7727.
B. Chaudret, Nanoscale, 2014, 6, 539–546.
´
´
33 L. M. Martınez-Prieto, I. Cano, A. Marquez, E. A. Baquero,
S. Tricard, L. Cusynato, I. del Rosal, R. Poteau, Y. Coppel,
K. Philippot, B. Chaudret, J. Campora and P. W. N. M. van
Leeuwen, Chem. Sci., 2017, 8, 2931–2941.
´
10 I. Cano, M. J.-L. Tschan, L. M. Martınez-Prieto, K. Philippot, 34 J. A. S. Duncan, D. Hedden, D. M. Roundhill, T. A. Stephenson
B. Chaudret and P. W. N. M. van Leeuwen, Catal. Sci.
Technol., 2016, 6, 3758–3766.
and M. D. Walkinshaw, Angew. Chem., 1982, 94, 463–464.
35 J. T. Cheng and P. D. Ellis, J. Phys. Chem., 1989, 93, 2549–2555.
11 D. Martin, D. Moraleda, T. Achard, L. Giordano and 36 B. A. Cowans and J. B. Grutzner, J. Magn. Reson., Ser. A, 1993,
G. Buono, Chem. – Eur. J., 2011, 17, 12729–12740. 105, 10–18.
12 F. Schroder, C. Tugny, E. Salanouve, H. Clavier, L. Giordano, 37 F. H. Larsen, H. J. Jakobsen, P. D. Ellis and N. C. Nielsen,
¨
`
D. Moraleda, Y. Gimbert, V. Mouries-Mansuy, J.-P. Goddard
J. Phys. Chem. A, 1997, 101, 8597–8606.
and L. Fensterbank, Organometallics, 2014, 33, 4051–4056.
13 C. Scriban, D. K. Wicht, D. S. Glueck, L. N. Zakharov, J. A. Golen
and A. L. Rheingold, Organometallics, 2006, 25, 3370–3378.
14 M. Liniger, B. Gschwend, M. Neuburger, S. Schaffner and
A. Pfaltz, Organometallics, 2010, 29, 5953–5958.
15 P. Sutra and A. Igau, Coord. Chem. Rev., 2016, 308, 97–116.
16 D. E. Fogg, N. J. Taylor, A. Meyer and A. J. Carty, Organo-
metallics, 1987, 6, 2252–2254.
38 F. H. Larsen, H. J. Jakobsen, P. D. Ellis and N. C. Nielsen,
Mol. Phys., 1998, 95, 1185–1195.
39 A. S. Lipton, J. A. Sears and P. D. Ellis, J. Magn. Reson., 2001,
151, 48–59.
40 The CPMG experiment has been used lately to improve
NMR spectra of spin 1/2 of heterogeneous materials where
distribution of frequencies, not refocused by MAS, can be
refocused by rf pulses to compress the range of chemical
shifts into spikelets and quicken signal acquisition. See
ref. 41–44.
17 D. E. Fogg and A. J. Carty, Polyhedron, 1988, 7, 2285–2295.
18 B. Zhang and B. Yan, Anal. Bioanal. Chem., 2010, 396,
973–982.
41 R. Siegel, T. T. Nakashima and R. E. Wasylishen, J. Phys.
Chem. B, 2004, 108, 2218–2226.
19 T. Gutmann, I. del Rosal, B. Chaudret, R. Poteau, H.-H.
Limbach and G. Buntkowsky, ChemPhysChem, 2013, 14, 42 J. W. Wiench, V. S.-Y. Lin and M. Pruski, J. Magn. Reson.,
3026–3033. 2008, 193, 233–242.
20 T. Gutmann, A. Gru¨nberg, N. Rothermel, M. Werner, 43 M. Dechamps, C. Roiland, B. Bureau, G. Yang, L. Le Polles and
M. Srour, S. Abdulhussain, S. Tan, Y. Xu, H. Breitzke and D. Massiot, Solid state Nucl. Magn. Reson., 2011, 40, 72–77.
G. Buntkowsky, Solid State Nucl. Magn. Reson., 2013, 55–56, 44 K. J. Harris, Z. E. M. Reeve, D. Wang, X. Li, X. Sun and
1–11. G. R. Goward, Chem. Mater., 2015, 27, 3299–3305.
21 L. E. Marbella and J. E. Millstone, Chem. Mater., 2015, 27, 45 W. D. Knight, Phys. Rev., 1949, 76, 1259–1260.
`
2721–2739.
46 J. J. van der Klink and H. B. Brom, Prog. Nucl. Magn. Reson.
Spectrosc., 2000, 36, 89–201.
22 In addition, the presence of nanoparticles too small
(o13 atoms) to be observed by the TEM equipment is likely, 47 S. E. Shore, J. P. Ansermet, C. P. Slichter and J. H. Sinfelt,
since very small spots were sometimes observed. This would
explain the high ratio of SPO ligands.
23 R. Van Hardeveld and F. Hartog, Surf. Sci., 1969, 15,
189–230.
Phys. Rev. Lett., 1987, 58, 953–956.
48 J. S. Bradley, J. M. Millar and E. W. Hill, J. Am. Chem. Soc.,
1991, 113, 4016–4017.
49 L. R. Becerra, C. P. Slichter and J. H. Sinfelt, J. Phys. Chem.,
1993, 97, 10–12.
24 P. Mastrorilli, M. Latronico, C. F. Nobile, G. P. Suranna,
F. P. Fanizzi, U. Englert and G. Ciccarella, Dalton Trans., 50 L. R. Becerra, C. P. Slichter and J. H. Sinfelt, Phys. Rev. B:
2004, 1117–1119. Condens. Matter Mater. Phys., 1995, 52, 11457–11461.
25 L. M. Martınez-Prieto, C. Urbaneja, P. Palma, J. Campora, 51 K. W. Zilm, L. Bonneviot, D. M. Hamilton, G. G. Webb and
´
´
K. Philippot and B. Chaudret, Chem. Commun., 2015, 51,
4647–4650.
G. L. Haller, J. Phys. Chem., 1990, 94, 1463–1472.
´
52 J. M. Asensio, S. Tricard, Y. Coppel, R. Andres, B. Chaudret
26 J. S. Bradley, J. M. Millar, E. W. Hill, S. Behal, B. Chaudret
and A. Duteil, Faraday Discuss., 1991, 92, 255–268.
and E. de Jesu´s, Angew. Chem., Int. Ed., 2017, 56, 865–869.
53 Small MNPs (o100 metal atoms) tend to behave as mole-
cular species without free electrons and no Knight shift is
observed. See e.g. J. S. Bradley, J. M. Millar, E. W. Hill and
S. Behal, J. Catal., 1991, 129, 530–539.
´
27 A. Duteil, R. Queau and B. Chaudret, Chem. Mater., 1993, 5,
341–347.
28 F. Dassenoy, K. Philippot, T. O. Ely, C. Amiens, P. Lecante,
E. Snoeck, A. Mosset, M.-J. Casanove and B. Chaudret, 54 See ref. 55–58 for a variety of examples related to coordination
New J. Chem., 1998, 703–711. of CO at iridium complexes in a bridging and terminal mode.
29 E. Ramırez, L. Erades, K. Philippot, P. Lecante and 55 B. E. Mann, B. T. Pickup and A. K. Smith, J. Chem. Soc.,
´
`
B. Chaudret, Adv. Funct. Mater., 2007, 17, 2219–2228.
Dalton Trans., 1989, 889–893.
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
56 S. Moreton, Inorg. Chim. Acta, 1992, 202, 225–230.
57 L. Garlaschelli, F. Greco, G. Peli, M. Manassero, M. Sansoni,
R. Gobetto, L. Salassa and R. D. Pergola, Eur. J. Inorg. Chem.,
2003, 2108–2112.
65 See ref. 66–69 for a variety of reviews related to the
reduction of nitroarenes catalyzed by MNPs and ref. 70–72
for several examples concerning the reduction of nitroarenes
catalyzed by IrNPs.
58 A. Tassan, M. Mozzon, G. Facchin, A. Dolmella and S. Detti, 66 H.-U. Blaser, H. Steiner and M. Studer, ChemCatChem, 2009,
Inorg. Chim. Acta, 2015, 424, 91–102. 1, 210–221.
59 F. Novio, K. Philippot and B. Chaudret, Catal. Lett., 2010, 67 P. Lara and K. Philippot, Catal. Sci. Technol., 2014, 4,
140, 1–7. 2445–2465.
´
60 L. M. Martınez-Prieto, A. Ferry, P. Lara, C. Richter, K. Philippot, 68 H. Chong and M. Zhu, ChemCatChem, 2015, 7, 2296–2304.
F. Glorius and B. Chaudret, Chem. – Eur. J., 2015, 21, 69 T. Aditya, A. Pal and T. Pal, Chem. Commun., 2015, 51,
17495–17502.
61 W. Kolodziejski and J. Klinowski, Chem. Rev., 2002, 102, 70 G.-Y. Fan, L. Zhang, H.-Y. Fu, M.-L. Yuan, R.-X. Li, H. Chen
613–628. and X.-J. Li, Catal. Commun., 2010, 11, 451–455.
62 P.-K. Wang, J.-P. Ansermet, C. P. Slichter and J. H. Sinfelt, 71 C. Campos, C. Torres, M. Oportus, M. A. Pena, J. L. G. Fierro
Phys. Rev. Lett., 1985, 55, 2731–2734. and P. Reyes, Catal. Today, 2013, 213, 93–100.
63 L. R. Becerra, C. A. Klug and C. P. Slichter, J. Phys. Chem., 72 A. B. Dongil, C. Rivera-Carcamo, L. Pastor-Perez, A. Sepu´lveda-
1993, 97, 12014–12019. Escribano and P. Reyes, Catal. Today, 2015, 249, 72–78.
64 N. Almora-Barrios, I. Cano, P. W. N. M. van Leeuwen and 73 L. Hu, X. Cao, L. Chen, J. Zheng, J. Lu, X. Sun and H. Gu,
9410–9431.
˜
´
´
N. Lopez, ACS Catal., 2017, 7, 3949–3954.
Chem. Commun., 2012, 48, 3445–3447.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017