Tuning the selectivity in the hydrogenation of aromatic ketones catalyzed by similar ruthenium and rhodium nanoparticles

Article abstract of DOI:10.1002/cctc.201402524

Ru and Rh nanoparticles (NPs) RuI, RuII, RhI and RhII, stabilised by triphenylphosphine (PPh₃) and diphenylphosphinobutane (dppb) were synthesised, characterised and applied as catalysts in the hydrogenation of several aromatic ketones. The effects of the nature of the metal and of the stabilising agent on the aryl versus ketone hydrogenation were studied. For RhNPs, the coordination of arene dominates the interaction of the substrate with the NP, whereas the coordination of the ketone group was not evidenced. For RuNPs, however, the results show that both arene and ketone coordinate to the NPs surface in a competitive manner. The properties of the stabilising ligands have a clear influence on the outcome of the reaction, and for the Rh-catalysed reactions, products of hydrogenolysis were only formed if PPh3 was used as the stabiliser. The structure of the substrate was also a key factor for the selectivity.

Full text of DOI:10.1002/cctc.201402524

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201402524
Tuning the Selectivity in the Hydrogenation of Aromatic
Ketones Catalyzed by Similar Ruthenium and Rhodium
Nanoparticles
Jessica Llop Castelbou,[a] Emma Bresꢀ-Femenia,[b, c] Pascal Blondeau,[b] Bruno Chaudret,[c]
Sergio Castillꢀn,*[b] Carmen Claver,[a] and Cyril Godard*[a]
Ru and Rh nanoparticles (NPs) RuI, RuII, RhI and RhII, stabilised
by triphenylphosphine (PPh3) and diphenylphosphinobutane
(dppb) were synthesised, characterised and applied as catalysts
in the hydrogenation of several aromatic ketones. The effects
of the nature of the metal and of the stabilising agent on the
aryl versus ketone hydrogenation were studied. For RhNPs, the
coordination of arene dominates the interaction of the sub-
strate with the NP, whereas the coordination of the ketone
group was not evidenced. For RuNPs, however, the results
show that both arene and ketone coordinate to the NPs sur-
face in a competitive manner. The properties of the stabilising
ligands have a clear influence on the outcome of the reaction,
and for the Rh-catalysed reactions, products of hydrogenolysis
were only formed if PPh3 was used as the stabiliser. The struc-
ture of the substrate was also a key factor for the selectivity.
Introduction
Over the last decades, transition-metal nanoparticles (MNPs)
have received a great deal of attention in various areas of re-
search such as optoelectronics, sensing, medicine and catalysis.
As catalysts, they potentially combine the advantages of heter-
ogeneous and homogeneous catalysts, and exhibit high activi-
ties in addition to retaining their tunability and selectivity
through their well-defined composition with a narrow size dis-
tribution.[1–3] The stabilisation of MNPs can be achieved in the
presence of polymers, surfactants or ligands, which allow the
control of their size, shape and dispersion as well as their sur-
face state. The choice of an appropriate stabiliser for the MNPs
is thus of critical importance to tailor their catalytic
performance.[2,3]
In the hydrogenation of aromatic ketones catalysed by solu-
ble NPs (Scheme 1), the use of Ru and RhNPs was reported
mainly, although a few examples of Pd and IrNPs were also
described.[4]
Scheme 1. Hydrogenation of aromatic ketones.
[a] Dr. J. Llop Castelbou,+ Prof. Dr. C. Claver, Dr. C. Godard
Departament de Quꢀmica Fꢀsica i Inorgꢁnic
Universitat Rovira I Virgili
Recently, some of us reported the use of RuNPs stabilised
by N-heterocyclic carbenes[5] and large-bite-angle diphosphine
ligands[6] as stabilisers with interesting ligand effects in the hy-
drogenation of a series of ketones. Indeed, the presence of ar-
omatic groups in the ligands was shown to form more stable
catalysts, and the more electron-donating the ligand, the faster
the hydrogenation of aromatic compounds. Furthermore,
a clear thermodynamic preference for arene coordination over
ketone coordination was observed for Ru, which led to inter-
esting selectivity in the case of aryl ketones. For acetophenone,
however, lower selectivity was obtained, which was explained
by the possibility for the substrate to coordinate to the NP sur-
face through both phenyl and carbonyl groups.
C/Marcel.li Domingo s/n
43007 Tarragona (Spain)
Fax: (+34)977-55-9563
E-mail: cyril.godard@urv.cat
[b] E. Bresꢂ-Femenia,+ Dr. P. Blondeau, Prof. Dr. S. Castillꢂn
Departament de Quꢀmica Analꢀtica i Orgꢁnica
Universitat Rovira I Virgili
C/Marcel.li Domingo s/n
43007 Tarragona (Spain)
Fax: (+34)977-55-8446
E-mail: sergio.castillon@urv.cat
[c] E. Bresꢂ-Femenia,+ Prof. Dr. B. Chaudret
LPCNP (Laboratoire de Physique et Chimie de Nano-Objets)
Universitꢃ de Toulouse, INSA, UPS, LPCNO, CNRS-UMR5215
135 Avenue de Rangueil
31077 Toulouse (France)
Fax: (+33)0561559697
[+] Both authors contributed equally to this work.
Stable RuNPs stabilised by a sulfonated diphosphine ligand
and mixtures of this ligand together with cyclodextrin were
also reported and tested in the hydrogenation of acetophe-
none in water.[7] Under these conditions, the hydrogenation of
the ketone function of the substrate was favoured over the
phenyl hydrogenation (90 vs. 10%), the positive effect of cyclo-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201402524.
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3160

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
dextrin was observed as at higher cyclodextrin content, and in-
creased turnover number (TON) and turnover frequency (TOF)
values were obtained. Interestingly, Dyson and co-workers
showed recently that the addition of water-soluble phosphine
ligands to RhNPs of approximately 3 nm in diameter synthes-
ised previously in the presence of polyvinylpyrrolidone (PVP)
improves the selectivity of the hydrogenation reactions in
favour of the arene hydrogenation.[8] They postulated that the
selective coordination of these ligands could modify the steric
hindrance at specific sites of the NPs and hence generate
selectivity.
In recent years, our research group has been interested in
the use of P-donor ligands as NPs stabilisers in selective cataly-
sis. We reported the hydrogenation of disubstituted arenes
using Ru, Rh and IrNPs stabilised by diphosphite ligands.[9]
These catalytic systems were active in the hydrogenation of
methylanisoles in pentane with high cis/trans selectivities.
More recently, we reported the ligand effect observed using
RhNPs stabilised by phosphine and phosphite ligands in the
partial hydrogenation of arenes and the selective hydrogena-
tion of styrene into ethylbenzene.[10]
Figure 1. TEM micrographs and the corresponding size histograms of RuI,
RuII, RhI and RhII.
The TEM micrographs of these NPs revealed in all cases the
formation of small nanoparticles with a spherical shape,
narrow size distribution and similar diameter (ꢀ 1.5 nm;
Figure 1). The RuNPs stabilised by PPh3, RuI, showed the small-
est size of the series (ꢀ1.3 nm diameter), whereas the others
displayed a size near 1.5 nm. The mean diameter and size dis-
tribution of the RuI and RuIINPs are in agreement with the
size of RuNPs stabilised with these ligands reported previously
Herein, we report the synthesis and characterisation of
a series of Ru and RhNPs stabilised by P-donor ligands and
their application in the hydrogenation of aromatic ketones.
The objective of this work is the comparison of the catalytic
behaviour of Ru and RhNPs stabilised by the same ligands
under identical conditions. The selectivity of the reaction,
namely, the hydrogenation of the arene versus that of the
ketone function, has been looked at in particular by using sev-
eral series of aromatic ketones as substrates.
[11]
at lower ligand-to-metal ratios (0.1 equivalent of PPh3 and
dppb).[12] Therefore, it was concluded that for these systems,
the amount of ligand used to stabilise the nanoparticles does
not affect their size significantly.
Diffuse peaks were observed in the XRD pattern of these
NPs, as expected for a homogeneous distribution of very small
particles with a hexagonal close-packing (hcp) lattice structure.
No reflections caused by ruthenium oxide were observed, and
coherence lengths in agreement with TEM analysis were ob-
tained. TGA of the RuI and RuII systems showed that these
NPs contained approximately 2% solvent, 25% phosphine li-
gands and 70% Ru, in agreement with previous reports.[11,12]
Similar results were obtained with the Rh systems (see details
in the Experimental Section). The characterisation data for the
RhI and RhIINPs were reported recently by our group.[10]
Results and Discussion
Soluble Ru and RhNPs stabilised by the phosphorus donor li-
gands triphenylphosphine (PPh3) and diphenylphosphinobu-
tane (dppb; P:Ru/Rh=0.4; Scheme 2) were synthesised by de-
composition of the organometallic precursors [Ru(COD)(COT)]
(COD=cyclooctadiene, COT=cyclooctatetraene) and [Rh(h3-
C3H5)3], respectively, in THF under H2 pressure. The NPs were
isolated as black powders after precipitation with pentane and
characterised by TEM, XRD, X-ray photoelectron spectroscopy
(XPS), wide-angle X-ray scattering (WAXS), elemental analysis
(EA) and thermogravimetric analysis (TGA).
Acetophenone reduction
Acetophenone 1 was first used to evaluate the selectivity of
the hydrogenation (aryl group versus ketone group) using
these catalysts. Three main products were expected in this pro-
cess (Scheme 3): cyclohexylmethylketone (1a), which results
from the selective reduction of the aromatic ring, phenyletha-
nol (1b), which results from the selective reduction of the
ketone group, and cyclohexylethanol (1c), formed by complete
hydrogenation of the substrate.
Initially, the RuNPs RuI stabilised by triphenylphosphine (I),
were tested in the hydrogenation of 1 in various solvents at
308C and under 20 bar H2. In THF, full conversion (TON=77)
was obtained if the reaction was performed for 5 h (Table 1,
Scheme 2. P-donor ligands used in this study to stabilise the RuNPs and
RhNPs.
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3161

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
If the RhII system, which bears dppb (II) as the stabilising
ligand, is used under the same conditions, full conversion
(TON=83) was obtained after 5 h (entry 15). Interestingly, in
terms of selectivity, relevant differences between the two Rh
systems were observed. Indeed, if RhII is used as the catalyst,
the hydrogenolysis products 1d and 1e were not detected, 1-
phenylethanol 1b (46%) was the main product and 1a and 1c
were obtained in similar yields of 28 and 26%, respectively.
Therefore, these results indicated strong differences that
depend on the metal and stabilising agents used. In terms of
activities, no important changes were observed, although the
RuII system was slightly less active than the other NPs. Howev-
er, significant differences were observed in terms of selectivity.
With the Ru systems, the selectivity towards 1a varied from
13% with RuI (57% at shorter reaction times) to 43% using
RuII. With the Rh systems, the same trend was observed, al-
though in this case the main difference in selectivity was the
formation or not of hydrogenolysis products, as products 1e
and 1 f were detected if the RhI system was used but not with
RhII.
Scheme 3. Expected products for the hydrogenation of 1.
Table 1. Hydrogenation of 1 catalysed by RuI, RuII, RhI and RhII.[a]
Entry Substrate Catalyst Solvent t [h] TON[b] 1a[c] 1b[c] 1c[c]
To further investigate this reaction, the hydrogenation of
1 catalysed by RuI and RhI was repeated and monitored by
GC–MS.
1
2
3
4
5
6
7
8[d]
9[e]
10
11
12
13[f]
14
15
1
1
1
1
RuI
RuI
RuI
RuI
RuII
RuI
RuI
RhI
RhI
RhI
RhI
RhI
RhI
RhI
RhII
THF
THF
pentane
CH3CN
THF
THF
THF
THF
pentane
CH3CN
THF
pentane
THF
5
77
13
57
23
–
43
–
–
15
6
–
–
–
1
–
4
53
–
41
–
95
23
1
–
–
–
26
–
87
39
24
–
16
100
5
33
8
–
2.5 69
5
5
19
0
59
77
4
70
78
–
–
–
78
–
83
If the RuI system was used, full conversion was obtained
after approximately 4 h (Figure 2). During the first 30 min, the
conversion reached approximately 20%, and rapid formation
1
5
1a
1b
1
1
1
1a
1a
1b
1c
1
16
16
5
5
5
16
5
16
16
5
–
–
–
–
THF
THF
–
28
46
26
[a] General conditions: 2 mol% catalyst,1.24 mmol of substrate, T=308C,
P=20 bar H2. [b] TON was defined as the number of moles of substrate
converted per mol of surface Ru/Rh. [c] Yield [%] determined by GC.
[d] 13% of 1d and 17% of 1e were obtained. [e] 57% of 1d and 27% of
1e were obtained. [f] 63% of 1e was obtained.
entry 1). Under these conditions, 87% of the total hydrogenat-
ed product 1c was obtained. If the reaction was repeated for
2.5 h under the same conditions, 90% conversion (TON=69)
was reached with up to 57% selectivity for 1a (Table 1,
entry 2), and 39% for the fully hydrogenated 1c. In this experi-
ment, only 4% of 1b was obtained.
Figure 2. Monitoring of the catalytic hydrogenation of 1 using RuI as the
catalyst. (Conditions: 1.24 mmol of substrate, 2 mol% RuI, solvent=THF,
T=308C, P=20 bar H2).
Next, the RhNPs RhI and RhII were used as catalysts in the
hydrogenation of 1 under the same reaction conditions. If RhI
was used as catalyst in this reaction for 5 h (entries 8–10), THF
again provided high conversion and selectivities (entry 8).
However, with this catalyst, in addition to the expected prod-
ucts 1a–c obtained in 15, 23 and 33% yield, respectively, the
hydrogenolysis products ethylbenzene (1d) and ethylcyclohex-
ane (1e) were detected in 13 and 17% yield, respectively. In
previous reports on the hydrogenation of acetophenone by
soluble RhNPs, these products were at most formed as
traces.[4b,13]
of all three products 1a, 1b and 1c was observed, of which
1a was the major product (ꢀ60%). If we compare the initial
rates of formation of 1a and 1b, the hydrogenation of the
arene ring was three times faster than that of the ketone func-
tion. The concentration of 1a remained constant for approxi-
mately 1 h before it decreased to 15% after 5 h. At 20% con-
version, the product 1b reached a maximum of approximately
20% selectivity, which decreased progressively at longer reac-
tion times, until it disappeared fully after 5 h. During the reac-
tion, 2-cyclohexylethanol 1c became the main product
through the hydrogenation of 1a and 1b.
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3162

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
To compare the reactivity of the intermediates 1a and 1b
under hydrogenation conditions, these compounds were used
as substrates under the same reaction conditions as 1 (Table 1,
entries 6 and 7). Surprisingly, if the reaction was performed
with 1b as the substrate, only 5% conversion (TON=4) into
the totally hydrogenated product 1c was obtained, even after
16 h.
no conversion was achieved even at 808C. In contrast, if 1b
was used as the substrate, complete conversion (TON=78)
was achieved with up to 66% selectivity for 1d and 26% for
1c. Therefore, these results show that during the hydrogena-
tion of 1, the formation of 1c only arises from the hydrogena-
tion of the aryl group of 1b and not from the hydrogenation
of the ketone group of 1a (Scheme 4). Furthermore, the forma-
tion of the hydrogenolysis product 1e necessarily involves the
This result is in contrast with that of the kinetic study in
which 1b disappeared fully after 5 h, although in this case, the
conversion of 1b was very low. It was deduced that the rela-
tive concentration of 1b in solution could explain the differ-
ence observed. Recently, Kꢂhn and co-workers[14] reported that
alcohols such as 1b can deactivate the catalyst by forming
stable adducts with the active species. However, if 1a was
used, total conversion (TON=77) was observed (Table 1,
entry 6). This indicated that 1c is formed mainly through 1a.
The same study was performed by using RhI. In the early
stages of the reaction, the product 1b is formed rapidly to
reach a maximum selectivity of approximately 60% (Figure 3).
Scheme 4. Reaction pathway for the hydrogenation of 1 using the RhI
system.
product 1b as an intermediate, whereas the transformation of
1a and 1c into 1e did not proceed under these conditions.
To summarise, the results obtained in the hydrogenation of
acetophenone 1 show that, surprisingly, distinct reaction path-
ways are employed that depend on the metal and ligand
used:
*
RuNPs favour the reduction of the aromatic ring over that
of the keto group and are able to reduce 1a to produce
1c. However, with these catalysts, 1b is reduced very
slowly. Selectivity is influenced by the stabilising ligand.
*
RhNPs favour the reduction of the keto group over that of
the aromatic ring to produce 1b, which is further hydro-
genated to form 1c or hydrogenolysed to 1d exclusively if
RhI is used as the catalyst. Interestingly, the RhNPs do not
reduce 1a.
Figure 3. Monitoring of the catalytic hydrogenation of 1 using RhI as the
catalyst. (Conditions: 1.24 mmol of substrate, 2 mol% RhI, solvent=THF,
T=308C, P=20 bar H2).
The selectivity trends observed for the hydrogenation of 1 are
summarised in Scheme 5. Based on these results, the RuINPs
are more active than RuII, RhI and RhII. Therefore, to obtain
similar conversions and to facilitate a comparison of the selec-
tivity, the hydrogenation of other substrates 2–9 was per-
If we compare the initial rates of formation of 1a and 1b, the
hydrogenation of the arene ring was four times slower than
that of the ketone function. For longer reaction
times, the concentration of 1b decreased steadily.
The selectivity to 1a reached 20% after 40 min and
remained practically unchanged at higher conver-
sions. As expected, the formation of 1c increased
progressively until the end of the reaction. The hy-
drogenolysis product 1d followed a similar pattern
to 1a, with rapid formation at the start of the reac-
tion and a constant concentration during the rest of
the hydrogenation process. The concentration of
product 1d was also practically constant throughout
the experiment, which suggests that this product is
formed at a similar rate to that at which it is convert-
ed into 1e. Next, the products 1a–c were used as
substrates in the hydrogenation reaction using RhI as
Scheme 5. Schematic representation of the selectivity trends observed in the hydrogena-
the catalyst. Interestingly, if 1a or 1c were reacted, tion of 1.
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3163

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
formed over 2.5 h using RuI as the catalyst and 5 h if RuII, RhI
and RhII were employed.
proximately 75% of the cyclohexylketone derivative 2a and up
to 94% of product 3a were achieved. The selectivities ob-
tained for both substrates were similar for both RhI and RhII.
High selectivity towards arene reduction in the presence of dis-
tant carbonyl groups has been already observed.[4a,5,8] Notably,
low selectivity for the fully hydrogenated products was ob-
tained in these experiments.
The behaviour of Ru and RhNPs in the reduction of aceto-
phenone seems to be complementary: RuNPs mainly reduce
the arene ring, whereas RhNPs first reduce the keto group.
However, RuNPs reduce the keto group of 1a, whereas RhNPs
do not. These results prompted us to explore the behaviour of
these nanocatalysts in the hydrogenation of non-conjugated
aryl ketones.
For substrates 2 and 3, the nature of the metal affects the
catalytic results significantly. Indeed, with RhNPs, the selectivity
for arene reduction increases if the separation between the
arene and carbonyl group increases until almost total selectivi-
ty for arene reduction is reached. Moreover, the stabilising
ligand does not influence the conversion or the selectivity.
Thus the selectivities obtained with RhI and RhII are identical.
This suggests that substrate coordination takes place through
the aromatic ring as the carbonyl group is pushed away from
the NP progressively. That the ligand does not affect the activi-
ty and selectivity of the catalysts indicates that both ligands
leave enough space at the surface of these NPs for arene
coordination.
Reduction of non-conjugated aryl ketones
The reduction of non-conjugated aryl ketones and the influ-
ence of the alkyl chain length between the phenyl and the
ketone (substrates 2 and 3) groups was studied using RuI,
RuII, RhI and RhII as catalysts (Table 2). If we used RuI or RuII,
similar activities were obtained (TON=50–60) for both sub-
Table 2. Hydrogenation of 2 and 3 catalysed by RuI, RuII, RhI and RhII.[a]
In the case of the Ru catalysts, the stabilising ligand does
not influence the activity but it does affect the selectivity of
the catalysts, and the level of ketone reduction increases if
dppb is the stabiliser (RuII). Notably, even for 3, 37% ketone
reduction was obtained. However, this is not surprising as we
observed earlier that Ru is able to reduce alkyl ketones such as
1a (Table 1). These results suggest that the stronger coordina-
tion and higher steric hindrance of the dppb ligand limits the
coordination of the arene more efficiently than the coordina-
tion of ketone. This may be caused by the higher bonding
energy of phosphines towards Ru than Rh and/or to the differ-
ent geometries of these small polyhedral NPs as a result of
their different crystal structure, namely, hcp and face-centred
cubic (fcc).[15]
Entry
Substrate
Catalyst
TON[c]
a[d]
b[d]
c[d]
1[b]
2
3[b]
4
5
6
2
2
3
3
2
2
3
3
RuI
RuII
RuI
RuII
RhI
RhII
RhI
RhII
52
56
51
57
61
53
57
66
58
30
74
47
76
74
93
94
39
70
17
37
24
26
3
3
–
9
16
–
–
7
8
4
4
2
[a] General conditions: 2 mol% catalyst, 1.24 mmol of substrate, 10 mL
THF, T=308C, P=20 bar H2, t=5 h. [b] t=2.5 h. [c] TON was defined as
the number of moles of substrate converted per mol of surface Ru/Rh.
[d] Yield [%] determined by GC.
If we compare the results shown in Tables 1 and 2, we can
observe that the RhNPs reduce the arene in 2 and 3 selectively
as these catalysts do not reduce alkyl ketones such as 1a
(Scheme 6). However, in the case of 1 the reduction of the
ketone group is preferred. With regard to RuNPs, the situation
is different as the reduction of arene and carbonyl groups are
strates (Table 2, entries 1 and 2, and 3 and 4). For substrate 2,
58% selectivity for the arene hydrogenated product 2a was
obtained (entry 1) using RuI,
whereas only 30% was achieved
using RuII (entry 2). For sub-
strate 3, the selectivity to 3a is
higher than for 2a, similar to the
trend observed in the hydroge-
nation of 2 for which RuI provid-
ed a higher selectivity to 3a
than RuII.
If the reaction was performed
with RhI or RhII as the catalyst,
good activities (TONs up to 66)
were obtained in all cases. In
these reactions, high to excellent
selectivity for the arene hydro-
genated products were ach-
ieved, and yields of up to ap- Scheme 6. Schematic representation of the selectivity trends observed in the hydrogenation of 2 and 3.
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3164

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
competitive for all substrates, which includes non-conjugated
aryl ketones. However, the selectivity towards arene reduction
increases if the aryl and keto groups become more separated.
These results suggest that 1, which is often selected to test
competition between the reduction of arene and carbonyl
groups, is in fact a particular case that deserves additional at-
tention. In this context, we decided to enlarge the study to the
reduction of the acetophenone derivatives 4–9 that contain
different substituents in the alkyl and phenyl moieties in
search of information about the particular behaviour of 1.
The results of the reduction of 4–9 in the presence of RhI
and RhII are collected in Table 3. Results obtained for the re-
duction of 1 under the same conditions are included for com-
parison. Moreover, as demonstrated by the results shown in
Table 1, RhI and RhII do not reduce cyclohexyl alkyl ketones,
so compounds c, d and e are generated from b, and a new
value that results from the sum of the yields of all these com-
pounds is included in Table 3.
RhI was used, was not affected by the substitution. In contrast
with previous studies that determined that substrates with
electron-donating groups are hydrogenated faster than sub-
strates with electron-withdrawing groups,[5] no clear effect of
the substituents on the aromatic ring on the activity was ob-
served in this study (Table 3, entries 11–14).
With regard to the selectivity, products obtained from the
reduction of the ketone group were obtained preferentially in
all cases (>70%), and the phenylethanol derivatives 6b–9b
were the major products (Table 3, entries 7–14), in agreement
with the results observed for 1. In general, ketone reduction
decreases slightly if RhII (with the dppb stabiliser) is used. The
selectivity is not influenced by the modification of the alkyl
moiety, but is affected more significantly if substituents are
present on the arene moiety. This is particularly true if the sub-
stituent is located in the para position as arene reduction de-
creases notably. Interestingly, if RhII is used as the catalyst, the
exclusive reduction of the carbonyl group was obtained in the
reduction of 9, which incorporates the electron-withdrawing
group CF3. However, a general trend of the influence of the
stabilising ligand on the selectivity cannot be established.
Similar to the results obtained for 1, hydrogenolysis of 1-
phenylethanol derivatives, produced by the reduction of the
carbonyl group of the substrates to provide 1d and then 1e
was only observed with the RhNPs stabilised with PPh3. These
results can be related to the lability of the monophosphine in
contrast with the more strongly coordinating diphosphine.
Indeed, PPh3 can dissociate from the particle and, therefore,
liberate the sites responsible for the hydrogenolysis reaction.
These sites are, therefore, expected to be the apex and edge
sites that are left under-coordinated.
Table 3. Hydrogenation of 4–9 catalysed by RhI and RhII.[a]
Entry
Substrate
Catalyst
TON[b]
a[c]
(b+c+d+e)[c]
Notably, the hydrogenolysis process stopped if the steric
hindrance of the alkyl side chain increased, but not if the sub-
stituents are present in the ortho or para position of the aro-
matic ring.
1
2
3
4
5
6
7
8
1
1
4
4
5
5
6
6
7
7
8
8
9
9
RhI
RhII
RhI
RhII
RhI
RhII
RhI
RhII
RhI
RhII
RhI
RhII
RhI
RhII
70
83
70
75
39
44
66
70
83
70
75
39
44
66
15
28
13
21
18
27
24
32
20
11
17
16
8
85[d] (23+33+12+17)
72[d] (46+26+0+0)
87[d] (69+18+0+0)
79[d] (63+16+0 +0)
82[d] (71+11+0+0)
73[d] (66+7+0+0)
76[d] (59+0+14+3)
68[d] (59+9+0+0)
80[d] (50+0+26+4)
89[d] (73+16+0+0)
83[d] (43+9+24+7)
84[d] (48+36+0+0)
92[d] (57+0+23+12)
100[d] (87+13+0+0)
To conclude, in the case of RhNPs, the coordination of the
arene group dominates the interaction of the substrate with
the catalyst surface, whereas the coordination of the carbonyl
group with the NP is not evidenced by the results of this study
as only the aromatic ring is reduced in substrates such as alkyl
ketones and aromatic compounds that contain a carbonyl
group far from the ring. The case of acetophenone derivatives
is singular, as although coordination to the NP takes place
through the aromatic ring, the carbonyl group remains in a po-
sition very favourable for its reduction, in such a way that it
takes place faster than the arene reduction. The selectivity of
this process is scarcely affected by substitution in the alkyl or
arene sides, and only substitution in the para position drives
the reaction completely towards the reduction of the ketone.
No significant electronic effect was observed (Table 3 , cf. en-
tries 11 and 12 vs. 13 and 14).
9
10
11
12
13
14
0
[a] General conditions: 2 mol% catalyst,1.24 mmol of substrate, 10 mL
THF, T=308C, P=20 bar H2, t=5 h. [b] TON was defined as the number
of moles of substrate converted per mol of surface Rh. [c] Yield [%] deter-
mined by GC. [d] Sum of the yields [%] of b–e.
The activity of the catalysts is affected significantly by the
steric hindrance of the alkyl substituent of the ketone (cf. 4
and 5, entries 3 and 5) but the ratio of products a:b+c+d+e
remains almost unchanged.
Once the carbonyl group is reduced, further aromatic ring
reduction can take place and if the reaction is catalysed by RhI
(which is stabilised by PPh3), hydrogenolysis is also observed.
The results obtained in the hydrogenation of the acetophe-
none derivatives 4–9 using RuI and RuII as catalysts are sum-
marised in Table 4. In general, the activities with RuI and
Interestingly, the presence of substituents on the aromatic
ring has little influence, and high activities were obtained in all
cases. Notably, the transformation of compounds with sub-
stituents in the para position (8 and 9), and more generally if
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3165

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
quence of the affinity of the arene for the NPs. However, the
results obtained with 4–9 indicate that small variations in the
steric properties of the substituents of the aromatic ring and
of the alkyl keto group cause important shifts in selectivity.
Table 4. Hydrogenation of 4–9 catalysed by RuI and RuII.[a]
Conclusions
A series of Ru and Rh nanoparticles (NPs) stabilised by P-based
ligands were synthesised and characterised. Comparable re-
sults in terms of size and stabiliser content were obtained. A
comparative study of the reduction of aryl ketones using these
NPs was performed, which provided some general clues about
this process.
Entry
Substrate
Catalyst
TON[b]
a[c]
b[c]
c[c]
1[e]
2
1
1
4
4
5
5
6
6
7
7
8
8
9
9
RuI
RuII
RuI
RuII
RuI
RuII
RuI
RuII
RuI
RuII
RuI
RuII
RuI
RuII
69
59
30
30
27
31
34
21
39
17
46
36
48
42
57 (96)[d]
43 (59)[d]
54 (67)[d]
47 (59)[d]
38 (46)[d]
35 (41)[d]
64
36
39
22
34 (50)[d]
18 (37)[d]
16
4
41
33
41
54
59
36
64
61
78
50
63
84
86
39
16
13
12
8
6
–
–
–
In general, it can be considered that in the case of RhNPs,
the coordination of the arene dominates the interaction of the
substrate with the NP, whereas the coordination of the ketone
group with the NP was not evidenced, and for alkyl ketones
and aromatic compounds that contain a carbonyl group far
from the ring, the carbonyl remains unaltered and only the ar-
omatic ring is reduced.
3[e]
4
5[e]
6
7[e]
8
9[e]
10
11[e]
12
13[e]
14
–
The case of acetophenone derivatives is singular, as al-
though coordination to the NP takes place through the aro-
matic ring, the carbonyl group remains in a position very fa-
vourable for its reduction, in such a way that it is hydrogenat-
ed faster than the arene moiety using RhNPs. The selectivity of
this process is scarcely affected by substitution in the alkyl or
arene groups, and only the substitution in the para position of
the aromatic ring drives the reaction completely towards the
reduction of ketone.
16
19
–
14
–
[a] General conditions: 2 mol% catalyst, 1.24 mmol of substrate, 10 mL
THF, T=308C, P=20 bar H2, t=5 h. [b] TON was defined as the number
of moles of substrate converted per mol of surface Ru. [c] Yield [%] deter-
mined by GC. [d] Selectivity in arene reduction considering that c is gen-
erated from a. [e] t=2.5 h.
For RuNPs, however, the results show that both arene and
ketone coordinate to the NPs surface competitively. The selec-
tivity is affected strongly by the steric hindrance in the ligand
and in the substrate, which in both cases increases the ketone
reduction.
RuIINPs were lower those than with RhI and RhII, and were af-
fected by the increase of substitution in the alkyl chain (en-
tries 3–6) or in the aromatic ring (entries 7–14), particularly if
the substituents are located in the ortho position (entries 7–
10). For substrates 6, 7 and 9, the products that result from
the total reduction were not observed (entries 7–10, 13, 14).
The stabilising ligand also affected the activity of these cata-
lysts significantly and, in general, if RuI was used (entries 3, 5,
7, 9, 11, 13), higher or similar activity than that with RuII was
obtained (entries 4, 6, 8, 10, 12, 14). This indicates that in this
case, the coordination of both aromatic ring and keto groups
are affected strongly by steric factors in the substrate or in the
ligand. Consequently, low TONs of 21 and 17 were achieved
for substrates 6 and 7, respectively (entries 8 and 10).
As a whole, this study that uses two similar transition-metal
NPs of similar sizes that bear two ligands that have very similar
electronic and steric effects but with different dissociation abil-
ities leads to different results in terms of activities, selectivities
and even capacity to perform a given reaction, here hydroge-
nolysis. This, therefore, demonstrates the interest to tailor the
composition and surface of these organometallic NPs to
modify their reactivity, comparable to that achieved for molec-
ular complexes.
The selectivity was also affected by the same factors. Indeed,
the increase of steric hindrance in the substrate or the use of
dppb as the stabiliser shifts the selectivity towards ketone re-
duction, which indicates that the rate of arene reduction is
more affected than that of ketone reduction. Furthermore, in
this case, 9 has a particular behaviour as the reaction affords
high percentages of carbonyl group reduction. This probably
confirms the slow arene reduction in the 1-arylethanol deriva-
tive as in this case product 9c was not observed.
Experimental Section
Reagents and general procedures
All syntheses were performed using standard Schlenk techniques
under a N2 or Ar atmosphere. Chemicals were purchased from Al-
drich Chemical Co, Fluka and Strem. All solvents were distilled over
drying reagents and were deoxygenated before use. The precursor
[Ru(COD)(COT)] was purchased from Nanomeps. The precursor Rh-
(h3-(C3H5)3 was prepared following methods described previous-
ly.[1,2] The synthesis of the NPs was performed by using a 1 L
Fisher–Porter reactor pressurised on a high-pressure line.
If we compare the results obtained with substrates 1 and 4–
9, the reduction of the arene group is preferred in 1 as a conse-
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3166

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Fischer–Porter reactor was then pressurised under 3 bar of H2, and
the reaction mixture was stirred for 24 h at RT. The initial yellow so-
lution became black after 20 min. After elimination of excess H2,
a small amount (approximately five drops) of the solution was de-
posited on a carbon-covered copper grid under an Ar atmosphere
for TEM analysis. The rest of the solution was concentrated under
reduced pressure to 40 mL. Precipitation and washing with pen-
tane (3ꢇ15 mL) was performed to obtain a black precipitate.
Characterisation techniques
The TEM experiments were performed at the Unitat de Microscopia
dels Serveis Cientificotꢃcnics de la Universitat Rovira I Virgili (TEM-
SCAN) in Tarragona by using a Zeiss 10 CA electron microscope
that operated at 100 kV with resolution of 3 ꢄ. The particle size dis-
tributions were determined by the manual analysis of enlarged
images. At least 300 particles on a given grid were measured to
obtain a statistical size distribution and a mean diameter.
High-resolution TEM experiments were performed at the Unitat de
Microscopia dels Seveis Cientꢅfics i Tecnolꢆgics de la Universitat de
Barcelona by using a JEOL 1010 electron microscope at 200 kV
with a resolution of 2.5 ꢄ. The particle size distributions were de-
termined by the manual analysis of enlarged images. At least 300
particles on a given grid were measured to obtain a statistical size
distribution and a mean diameter.
General procedure for the synthesis of the RhNPs
In a typical procedure, [Rh(h3-(C3H5)3] (400 mg, 1.746 mmol) was
placed into a Fischer–Porter reactor at À110 8C (acetone/N2 bath)
with dry THF (64 mL), which was deoxygenated by freeze–pump–
thaw cycles, in the presence of the appropriate ligand (0.2 equiv.
for dppb and 0.4 equiv. for PPh3). The Fischer–Porter reactor was
then pressurised under 6 bar of H2, and the reaction mixture was
stirred for 30 min at RT. The solution was then heated to 408C and
stirred at this temperature for 24 h. The initial colourless solution
became black after 1 h. A small amount (approximately 20 drops)
of the solution was deposited on a carbon-covered copper grid
under an Ar atmosphere for TEM analysis. The rest of the solution
was concentrated under reduced pressure. Precipitation and wash-
ing with pentane (3ꢇ15 mL) was then performed to obtain a black
precipitate.
XRD data were measured by using a Siemens D5000 diffractometer
(Bragg–Brentano parafocusing geometry and vertical q-q goniome-
ter) fitted with a curved graphite diffracted-beam monochromator,
incident and diffracted-beam Soller slits, a 0.068 receiving slit and
a scintillation counter as a detector. The angular 2q diffraction
range was between 26 and 958. The data were collected with an
angular step of 0.058 with 16 s per step and sample rotation. A low
background Si(510) wafer was used as the sample holder. CuKa ra-
diation was obtained from a copper X-ray tube operated at 40 kV
and 30 mA.
General procedure for the hydrogenation reactions
XPS experiments were performed by using a PHI 5500 Multitechni-
que System (from Physical Electronics) with a monochromatic X-
ray source (AlKa line of 1486.6 eV energy and 350 W), placed per-
pendicular to the analyser axis and calibrated using the 3d5/2 line
of Ag with a full width at half maximum (FWHM) of 0.8 eV. The an-
alysed area was a circle of 0.8 mm diameter, and the selected reso-
lution for the spectra was 187.5 eV pass energy and 0.8 eV/step for
the general spectra and 23.5 eV pass energy and 0.1 eV/step for
the spectra of the different elements in the depth profile spectra.
A low-energy electron gun (<10 eV) was used to discharge the
surface if necessary. All measurements were performed under an
ultra-high vacuum (UHV) chamber pressure between 5ꢇ10–9 and
2ꢇ10–8 Torr.
A Par 477 autoclave equipped with a proportional-integral-deriva-
tive (PID) temperature controller and reservoir for kinetic measure-
ments and a HEL 24 Cat reactor for substrate scope were used as
reactors for the hydrogenation reactions. In a typical experiment,
the autoclave was charged in the glove-box with Ru or RhNPs
(3 mg of RuNPs; 3.5 mg of RhNPs; the catalyst concentration was
calculated based on the total number of metal atoms in the NPs)
and the substrate (1.24 mmol, approximate substrate-to-metal
ratio=55) in THF (10 mL). H2 was then introduced until the desired
pressure was reached. The reaction was stirred for the appropriate
time at 308C. The autoclave was then depressurised. The solution
was filtered through silica and analysed by GC. The conversion and
the selectivities of the product were determined by using a Fisons
instrument (GC 9000 series) equipped with an HP-5MS column.
WAXS analyses were performed at CEMES-CNRS. Samples were
sealed in 1 mm diameter Lindemann glass capillaries. The samples
were
irradiated
with
graphite-monochromatised
MoKa
The conversion and selectivity were determined by GC–MS. The re-
tention times (tr) for the main products detected for each sub-
strate are detailed below:
(0.071069 nm) radiation, and the X-ray intensity scattered measure-
ments were performed by using a dedicated two-axis diffractome-
ter. Radial distribution functions (RDF) were obtained after Fourier
transform of the reduced intensity functions.
Substrate 1: tr1 =16.70 min, tr1a =15.0 min, tr1b =16.51 min,
tr1c =15.82 min, tr1e =6.3 min.
TGA experiments were performed by using the furnace of a Mettler
Toledo TGA/SDTA851 instrument.
Substrate 2: tr2 =19.80 min, tr2a =19.07 min, tr2b =20.01 min,
tr2c =20.14 min.
GS–MS spectroscopy was performed by using a HP 6890A spec-
trometer with an achiral HP-5 column (0.25 mmꢇ30 mꢇ0.25um).
The method used consisted of an initial isothermal period at 408C
for 3 min followed by a 38CminÀ1 temperature ramp to 1208C and
Substrate 3: tr3 =25.24 min, tr3a =24.79 min, tr3b =25.87 min,
tr3c =25.05 min.
Substrate 4: tr4 =23.44 min, tr4a =21.4 min, tr4b =23.65 min,
tr4c =23.06 min, tr4d =24.61 min.
a hold time of 12 min, and the flow rate was 1.3 mLminÀ1
.
Substrate 5: tr5 =24.82 min, tr5a =23.04 min, tr5b =26.28 min,
tr5c =25.36 min, tr5d =25.71 min.
General procedure for the synthesis of the RuNPs
Substrate 6: tr6 =20.18 min, tr6a =18.20 min, tr6b =22.14 min,
tr6c =19.70 min, tr6d =12.51 min, tr6e =10.11 min.
In a typical procedure, [Ru(COD)(COT)] (400 mg, 1.268 mmol) was
placed in a Fischer–Porter reactor with dry THF (400 mL), which
was deoxygenated by freeze–pump–thaw cycles, in the presence
of the ligand (0.2 equiv. for dppb and 0.4 equiv. for PPh3). The
Substrate 7: tr7 =27.54 min, tr7a =22.80 min, tr7b =27.09 min,
tr7c =23.06 min, tr7d =17.60 min, tr7e =13.70 min.
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3167

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Conejero, R. Poteau, K. Philippot, B. Chaudret, Angew. Chem. 2011, 123,
12286–12290; Angew. Chem. Int. Ed. 2011, 50, 12080–12084.
[3] P. Lara, K. Philippot, B. Chaudret, ChemCatChem 2013, 5, 28–45.
[4] a) G. S. Fonseca, J. D. Scholten, J. Dupont, Synlett 2004, 9, 1525–1528;
b) F. Jutz, J. M. Andanson, A. Baiker, J. Catal. 2009, 268, 256–266.
[5] D. Gonzalez-Galvez, P. Lara, O. Rivada-Wheelaghan, S. Conejero, B. Chau-
dret, K. Philippot, P. W. N. M. van Leeuwen, Catal. Sci. Technol. 2013, 3,
99–105.
Substrate 8: tr8 =30.21 min, tr8a =24.72 min, tr8b =28.33 min,
tr8c =24.85 min, tr8d =15.53 min, tr8e =14.72 min.
Substrate 9: tr9 =16.65 min, tr9a =17.89 min, tr9b =19.16 min,
tr9c =18.30 min, tr9d =9.45 min, tr9e =9.62 min.
Acknowledgements
[6] D. Gonzalez-Galvez, P. Nolis, K. Philippot, B. Chaudret, P. W. N. M. van
Leeuwen, ACS Catal. 2012, 2, 317–321.
We are grateful to the Spanish Ministerio de Economꢀa y Compet-
itividad (Ramꢂn y Cajal to C.G., CTQ2010-15835, CTQ2013-46404-
R and CTQ-2011-22872) for financial support. E.B. thanks the Min-
isterio de Educaciꢂn Cultura y Deporte for a grant.
[7] M. Guerrero, Y. Coppel, N. T. T. Chau, A. Roucoux, A. Denicourt-Nowicki,
E. Monflier, H. Bricout, P. Lecante, K. Philippot, ChemCatChem 2013, 5,
1–11.
[8] D. J. M. Snelders, N. Yan, W. Gan, G. Laurenczy, P. J. Dyson, ACS Catal.
2012, 2, 201–207.
[9] a) A. Gual, M. R. Axet, K. Philippot, B. Chaudret, A. Denicourt-Nowicki, A.
Roucoux, S. Castillꢀn, C. Claver, Chem. Commun. 2008, 2759–2761; b) A.
Gual, C. Godard, K. Philippot, B. Chaudret, A. Denicourt-Nowicki, A. Rou-
coux, S. Castillꢀn, C. Claver, ChemSusChem 2009, 2, 769–779.
[10] J. Llop, A. Gual, E. Mercadꢈ, C. Claver, C. Godard, Catal. Sci. Technol.
2013, 3, 2828–2833.
[11] M. V. Escꢉrcega-Bobadilla, C. Tortosa, E. Teuma, C. Pradel, A. Orejꢀn, M.
Gꢀmez, A. M. Masdeu-Bultꢀ, Catal. Today 2009, 148, 398–404.
[12] J. Garcꢅa-Antꢀn, M. R. Axet, S. Jansat, K. Philippot, B. Chaudret, T. Pery,
G. Buntkowsky, H. Limbach, Angew. Chem. 2008, 120, 2104–2108;
Angew. Chem. Int. Ed. 2008, 47, 2074–2078.
[13] G. S. Fonseca, A. P. Umpierre, P. F. P. Fichtner, S. R. Teixeira, J. Dupont,
Chem. Eur. J. 2003, 9, 3263–3269.
[14] D. Jantke, M. Cokoja, M. Drees, W. A. Herrmann, F. E. Kꢂhn, ChemCatCh-
em 2013, 5, 3241–3248.
[15] Y. Xia, Y. Xiong, B. Lim, S. E. Skrabalak, Angew. Chem. 2009, 121, 62–108;
Angew. Chem. Int. Ed. 2009, 48, 60–103.
Keywords: hydrogenation
rhodium · ruthenium
·
ketones
·
nanoparticles
·
[1] a) Nanoscale Materials in Chemistry (Eds.: K. J. Klabunde, R. M. Richards),
Wiley-Interscience, New York, 2001; b) Metal Nanoparticles: Synthesis
Characterization and Applications (Eds.: L. Feldheim, C. A. Foss, Jr.),
Marcel Dekker, New York, 2002; c) G. Schmid in Clusters and Colloids
From Theory to Applications, Wiley-VCH, Weinheim, 2004; d) G. Schmid
in Nanoparticles, From Theory to Application, Wiley-VCH, Weinheim,
2004; e) K. J. Klabunde, C. Mohs, Chemistry of Advanced Materials, An
Overview, Wiley-VCH, New York, 1998; f) A. Roucoux, Top. Organomet.
Chem. 2005, 16, 261–279.
[2] a) K. Philippot, B. Chaudret, C. R. Chim. 2003, 6, 1019–1034; b) P. J. De-
bouttiꢃre, V. Martinez, K, Philippot, B. Chaudret, Dalton Trans. 2009,
10172–10174; c) M. Guerrero, J. Garcꢅa-Antꢀn, M. Tristany, J. Pons, J.
Ros, K. Philippot, P. Lecante, B. Chaudret, Langmuir 2010, 26, 15532–
15540; d) M. R. Axet, K. Philippot, B. Chaudret, M. Cabi, S. Giorgio, C. R.
Henry, Small 2011, 7, 235–241; e) P. Lara, O. Rivada-Wheelaghan, S.
Received: July 7, 2014
Published online on September 18, 2014
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 3160 – 3168 3168