Supramolecular cluster catalysis: Facts and problems

Article abstract of DOI:10.1016/j.jorganchem.2003.12.032

By checking the chemistry underlying the concept of "supramolecular cluster catalysis" we identified two major errors in our publications related to this topic, which are essentially due to contamination problems. (1) The conversion of the "closed" cluster cation [H₃Ru₃(C₆H₆) (C₆Me₆)₂(O)]⁺ (1) into the "open" cluster cation [H₂Ru₃ (C₆H₆)(C₆Me₆)₂ (O)(OH)]⁺ (2), which we had ascribed to a reaction with water in the presence of ethylbenzene is simply an oxidation reaction which occurs in the presence of air. (2) The higher catalytic activity observed with ethylbenzene, which we had erroneously attributed to the "open" cluster cation [H₂Ru₃ (C₆H₆)(C₆Me₆)₂ (O)(OH)]⁺ (2), was due to the formation of RuO₂·nH₂O, caused by a hydroperoxide contamination present in ethylbenzene.

Full text of DOI:10.1016/j.jorganchem.2003.12.032

Journal of Organometallic Chemistry 689 (2004) 1362–1369
www.elsevier.com/locate/jorganchem
Supramolecular cluster catalysis: facts and problems
a,
a
a
a
Georg Suss-Fink *, Bruno Therrien , Ludovic Vieille-Petit , Mathieu Tschan ,
€
a
Vladimir B. Romakh , Thomas R. Ward , Massoud Dadras , Gabor Laurenczy
a
b
c
a
Institut de Chimie, Universitꢀe de Neuch^atel, Case postale 2, CH-2007 Neuch^atel, Switzerland
b
Institut de Microtechnique, Universitꢀe de Neuch^atel, Case postale 2, CH-2007 Neuch^atel, Switzerland
c
Institut de Chimie Molꢀeculaire et Biologique, Ecole Polytechnique Fꢀedꢀerale de Lausanne, BCH, CH-1015 Lausanne, Switzerland
Received 8 October 2003; accepted 9 December 2003
Abstract
By checking the chemistry underlying the concept of ‘‘supramolecular cluster catalysis’’ we identified two major errors in our
publications related to this topic, which are essentially due to contamination problems. (1) The conversion of the ‘‘closed’’ cluster
cation [H3Ru3(C6H6)(C6Me6)2(O)]þ (1) into the ‘‘open’’ cluster cation [H2Ru3(C6H6)(C6Me6)2(O)(OH)]þ (2), which we had as-
cribed to a reaction with water in the presence of ethylbenzene is simply an oxidation reaction which occurs in the presence of air. (2)
The higher catalytic activity observed with ethylbenzene, which we had erroneously attributed to the ‘‘open’’ cluster cation
[H2Ru3(C6H6)(C6Me6)2(O)(OH)]þ (2), was due to the formation of RuO2 ꢀ nH2O, caused by a hydroperoxide contamination present
in ethylbenzene.
Ó 2004 Elsevier B.V. All rights reserved.
Keywords: Catalysis; Ruthenium; Clusters; Hydrogenation benzene
1. Introduction
Only recently catalytic mechanisms without coordi-
nation of the substrate to the metal centre of the catalyst
molecule have been considered [6,7], based on accu-
mulating evidence for hydrogen transfer within merely
hydrogen-bonded substrate–catalyst complexes in the
case of catalytic ketone transfer hydrogenation reac-
tions [8] and for oxygen transfer via direct olefin attack
to the oxo ligand of the catalyst in epoxidation reactions
[9].
Organometallic catalysis is generally supposed to
proceed through a catalytic cycle that involves the co-
ordination of the substrate, either by ligand substitution
or by oxidative addition, transformation of the coordi-
nated substrate, and the liberation of the product, either
by decoordination or by reductive elimination [1].
Classical examples that have been studied in great detail
are the hydrogenation of olefins with WilkinsonÕs cata-
lyst [2] and the carbonylation of methanol with rhodium
iodide (Monsanto Process) [3]. The complete charac-
terisation of the intermediates of the latter process and
the proposal of a well-established catalytic cycle repre-
sents one of the triumphs of organometallic chemistry
[4]. In all these reactions, the elementary steps of the
catalytic process are believed to occur within the first
coordination sphere of the organometallic catalyst [5].
C6H6 þ 3H2 ! C6H12
We discovered in 1999 that the cluster cation
[H3Ru3(C6H6)(C6Me6)2(O)]þ (1), employed as the wa-
ter-soluble tetrafluoroborate salt, efficiently catalyses the
hydrogenation of benzene to cyclohexane under bi-
phasic conditions [10]. From mass spectroscopic mea-
surements and from molecular modelling studies, we
concluded that the substrate molecule is incorporated in
the hydrophobic pocket spanned by the three arene li-
gands in 1 (see Scheme 1), suggesting the catalytic re-
action to occur within this host–guest complex without
prior coordination of the substrate (‘‘supramolecular
cluster catalysis’’) [11].
* Corresponding author. Tel.: +41-32-718-24-00; fax: +41-32-718-25-
11.
€
E-mail address: georg.suess-fink@unine.ch (G. Suss-Fink).
0022-328X/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2003.12.032

G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
1363
other hand, rhodium colloids obtained from a mixture
of [(CnH2nþ1)NMe2(CH2CH2OH)]Br with rhodium
powder in water give a reusable suspension for the cat-
alytic hydrogenation of benzene [29]. Homogeneous
arene hydrogenation catalysts have been reviewed re-
cently by Finke and Widegren [30a] and by Dyson [30b],
the question of the true nature of the catalytic species in
arene hydrogenation with soluble metal complexes has
been critically addressed by Finke and Widegren [31].
The water-soluble cluster cations 1 and 2, which we
had reported to catalyse efficiently the hydrogenation of
benzene and benzene derivatives, seem to be molecular
catalysts that work under biphasic conditions and show
molecular recognition for the substrate. This observa-
tion led us to suggest catalysis by the intact trinuclear
cluster system, the catalytic transformations taking
place inside the hydrophobic pocket [11], an interpre-
tation which has been questioned recently [31,32].
However, last year the reported synthesis of the ‘‘open’’
cluster cation 2 was not longer reproducible in our
hands, which forced us to check all our results in order
to eliminate errors which we had made and to find evi-
dence for the catalytic concept we had proposed. Herein
we report the results of this re-investigation and a re-
vised version of supramolecular cluster catalysis.
+
+
O
H
O
H
Ru
Ru
H
Ru
H
Ru
H
Ru
Ru
OH
(1)
(2)
Scheme 1. The ‘‘closed’’ cluster cation [H3Ru3(C6H6)(C6Me6)2(O)]þ
(1) (48 electron species) and the ‘‘open’’ cluster cation
[H2Ru3(C6H6)(C6Me6)2(O)(OH)]þ (2) (50 electron species). In both
cases, the three arene ligands span a hydrophobic pocket underneath
the triruthenium plane.
With ethylbenzene as the substrate, we observed the
catalytic activity of 1 to be higher than with benzene or
other benzene derivatives. As we isolated from the
aqueous phase the cluster cation [H2Ru3(C6H6)-
(C6Me6)2(O)(OH)]þ (2), which derives from 1 by
replacing a hydrido bridge by a hydroxo bridge, we
attributed the higher catalytic activity to this ‘‘open’’
trinuclear cluster which contains only two ruthenium–
ꢁ
ruthenium bonds [2.780(1) A] in accordance with the
electron count of 50 [11–13].
The hydrogenation of benzene and of other aromatics
has remained predominantly a field of heterogeneous
catalysis [14]. Most organometallic complexes known to
catalyse the homogeneous hydrogenation of olefins,
ketones or acetylenes fail in the case of benzene or other
aromatics; the first putative example of a homogeneous
benzene hydrogenation catalyst being (C3H5)Co-
[P(OMe)3]3 reported in 1974 by Muetterties [15]. Other
examples include (C6Me6)2Ru [16], H2Ru2(C6Me6)2Cl2
[17], H2Ru(H2)2(PCy3)2 [18], [H4Ru4(C6H6)4]Cl2 in
water [19] or ionic liquids [20], Ru2(C6H6)2Cl2 [21],
Rh2(C5Me5)2Cl2 in combination with NEt3 [22]. The
most significant development in the field of homoge-
neous benzene hydrogenation resulted in the Dimersol
2. Results and discussion
2.1. The ‘‘closed’’ cluster cation [H3Ru3(C6H6)-
(C6Me6)2(O)]þ (1): confirmation of the synthetic
procedure and of the catalytic properties
The synthesis of the closed cluster [H3Ru3(C6H6)-
(C6Me6)2(O)]þ (1) works as reported [10] without
problems in aqueous solution: The dinuclear percursor
[H3Ru2(C6Me6)2]þ and the mononuclear percursor
[Ru(C6H6)(H2O)3]2þ are formed in situ from Ru2-
(C6Me6)2Cl4 and NaBH4 and from Ru2(C6H6)2Cl4 in
water; cation 1 can be isolated as the tetrafluoroborate
salt with the reported yield.
ꢀ
process by the Institut Franßcais du Petrole, based on
nickel or cobalt alkoxides, acetylacetonates or carb-
oxylates and trialkylaluminium activators [23], which
appears to be a viable homogeneous alternative to the
heterogeneous technology of industrial benzene hydro-
genation [24]. A new generation of homogeneous arene
hydrogenation catalysts on the basis of niobium and
tantalum hydride derivatives have been developed by
Rothwell [25], e.g. Nb(OC6HPh4-2,3,5,6)2Cl3 in combi-
nation with BuLi.
2þ
½H3Ru2ðC6Me6Þ ꢁþ þ ½RuðC6H6ÞðH2OÞ ꢁ
2
3
! ½H3Ru3ðC6H6ÞðC6Me6Þ ðOÞꢁþ þ 2H2O þ 2Hþ
2
We then checked the catalytic properties of the water-
soluble cluster cation 1: It serves as an efficient catalyst
(or possibly catalyst precursor) for the hydrogenation of
benzene to give the corresponding cyclohexane under
biphasic conditions (catalyst/substrate ratio 1:1000) with
the catalytic activity reported [10,11]. In the case of
methyl-substituted benzene derivatives, the TON and
TOF values obtained under rigorously controlled con-
ditions (Table 1) are slightly lower than the previously
reported values [10]. In all cases, 1[BF4] can be recov-
ered almost quantitatively >95%) from the aqueous
For organometallic benzene hydrogenation catalysts
the true nature of the catalytic species remained a de-
batable point (‘‘is it homogeneous or heterogeneous
catalysis?’’) [26]. In the case of the putative homoge-
neous [(C8H17)3Me][RhCl4] ion pair catalyst [27], Finke
and co-workers [28] was able to demonstrate in a pio-
neering paper that rhodium(0) nanoclusters are the true
catalysts (‘‘soluble heterogeneous catalysts’’). On the

1364
G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
Table 1
Hydrogenation of benzene and benzene derivatives catalysed by [H3Ru3(C6H6)(C6Me6)2(O)]þ (1) under biphasic conditionsa
Substrate
Product
Time (h)
Yieldb (%)
TONc
TOFd (hꢂ1
)
Benzene
Toluene
Cyclohexane
2.5
2.5
2.5
2.5
2.5
2.5
2.5
74
50
45
43
44
5
740
500
450
430
440
50
296
200
180
172
176
20
Methylcyclohexane
p-Xylene
m-Xylene
o-Xylene
Pseudo-Cumene
Durene
1,4-Dimethylcyclohexanee
1,3-Dimethylcyclohexanee
1,2-Dimethylcyclohexanee
1,2,4-Trimethylcyclohexanee
1,2,4,5-Tetramethylcyclohexanee
0
0
0
a Conditions: H2O 10 ml, catalyst [1][BF4] 0.01 mmol, substrate 10 mmol, H2 pressure 60 bar, temperature 110 °C, stirring frequency 900 minꢂ1
.
b Determined by gas chromatography.
c Catalytic Turnover Number: mol product formed per mol catalyst.
d (Mean) Catalytic Turnover Frequency: mol product formed per mol catalyst and per time unit.
e Mixture of cis and trans isomers.
phase after a catalytic run; the trinuclear cluster is intact
and can be reused for further runs.
small amount of the cluster [H3Ru3(C6H5Me)-
(C6Me6)2(O)]þ, in which the benzene ligand is ex-
changed against a toluene ligand; given that this cluster
is not visible in the NMR spectrum, we estimate its
content to less than 5%. We believe that the exchange of
the benzene ligand in 1 by the more electron-rich toluene
is a side-reaction which takes place in parallel to the
catalytic toluene hydrogenation inside the hydrophobic
pocket of 1. And indeed, we observed the same exchange
(less than 5%) by heating a mixture of toluene with an
aqueous solution of 1 (1000:1) for 2 h to 110 °C without
hydrogen pressure.
The catalytic activity of 1 for the hydrogenation of
benzene derivatives under biphasic conditions depends
predominantly on steric factors: Although methyl sub-
stituents increase the electronic density of the aromatic
system, which should favour the catalytic activity, the
increasing bulk of the substrate slows the catalytic re-
action down: Thus, the highest catalytic activity is ob-
served with the parent benzene substrate (TOF 290 hꢂ1).
The catalytic activity is reduced for toluene with one
methyl substituent (TOF 200 hꢂ1) and for the xylenes
with two methyl substituents (TOF 172–180 hꢂ1). The
system still tolerates three methyl substituents (TOF 20
hꢂ1 for pseudo-cumene), but with four methyl substitu-
ents (durene), no catalytic activity is observed. This
shape-selectivity with respect to the substrate suggests
molecular recognition of the substrate by the catalyst.
We studied in particular the question of arene ligand
exchange between cluster 1 and the aromatic substrate
during the catalytic reaction, since the mechanistic
scheme of supramolecular cluster catalyst we had pro-
posed [11,13] does not involve coordination of the aro-
matic substrate to a ruthenium atom of the catalyst. If
the catalytic reaction is carried out with hexadeutero-
benzene as the substrate under the usual biphasic con-
ditions (catalyst/substrate ratio 1:1000, 110 °C, 60 bar),
the reaction yields as expected 1, 2, 3, 4, 5, 6-hexadeu-
terocyclohexane, while cluster 1 still contains one C6H6
and two C6Me6 ligands (and no coordinated C6D6) after
the catalytic reaction, according to NMR and MS data.
This observation rules out an g6-coordination of the
substrate to a metal centre in this catalytic reaction,
being consistent with only supramolecular effects for the
interaction between the catalyst and the substrate.
In the case of toluene hydrogenation (catalyst/sub-
strate ratio 1:1000, 110 °C, 60 bar), the situation was not
We also found the catalytic reaction to work only in
the presence of water (biphasic conditions); in homo-
geneous phase with both, catalyst ([1][BF4]) and sub-
strate (benzene) dissolved in tetrahydrofuran/ethanol
(5:1), there is no reaction. These findings are consistent
with the formation of a supramolecular catalyst–sub-
strate host–guest complex being a key species in the
catalytic process.
2.2. The ‘‘open’’ cluster cation [H2Ru3(C6H6)(C6Me6)2-
(O)(OH)]þ (2): elucidation of the synthetic problems
and rectification of the catalytic properties
The synthetic problem of the ‘‘open’’ cluster cation
[H2Ru3(C6H6)(C6Me6)2(O)(OH)]þ (2) turned out to be
not trivial, the synthesis reported [11,12] being not
longer reproducible. Variations of the temperature (20–
110 °C) or the pH value (3–10) of the aqueous solution
did not open the ‘‘closed’’ cluster cation 1 to give 2; in all
cases 1 could be recovered quantitatively, unless com-
plete degradation of the organometallic species was
observed. However, we found the solution to the prob-
lem, when we admitted air into the reaction system: The
conversion of 1 into 2, which we had formulated as a
dehydrogenating hydrolysis reaction [11,12] turned out
to be a simple aerobic oxidation reaction; the hydroxo
bridge in 2 comes from dioxygen (air) and not from
water. The reaction is best carried out in aqueous so-
lution under atmospheric pressure of air; the complete
1
so clear: While the H NMR spectrum of the organo-
metallic residue of the aqueous phase dissolved in D2O
showed only the expected signals of 1, the electrospray
mass spectrum of the organometallic residue revealed a

G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
1365
conversion of 1 into 2 takes about 4 days. An NMR
study of the conversion of 1 into 2, carried out at 27 °C
in D2O solution with approximately 10 equivalents of
O2 (with respect to 1), shows no indication of interme-
diary species.
tion in the ethylbenzene we had used for the synthesis of
2. The original batch of ethylbenzene showed indeed an
enhanced catalytic activity of 1 (or 2), when it was now
employed as a substrate. This catalytic activity (TOF
3600 hꢂ1) could not be reproduced with rigorously pu-
rified ethylbenzene, with which we observed only the
‘‘normal’’ catalytic activity of [1][BF4] of 290 hꢂ1 under
the same biphasic conditions (catalyst/substrate ratio
1:1000, 110 °C, 60 bar). Thus, it became clear that the
ethylbenzene we had originally used contained a con-
tamination which served as an activator for the catalytic
reaction.
The search for this activator in the contaminated
ethylbenzene sample proved to be non-trivial: A GC/MS
analysis of the sample revealed two major contaminants:
Phenylmethylketone and 1-phenylethan-1-ol. However,
neither of these two contaminants turned out to be the
activator: When we added either of these compounds to
the catalytic hydrogenation of benzene under biphasic
conditions (catalyst/activator/substrate ratio 1:100:1000,
110 °C, 60 bar), there was essentially no effect.
However, when we checked the catalytic activity of
the ‘‘open’’ cluster 2 (dissolved as the tetrafluoroborate
salt in water) for the hydrogenation of benzene under
biphasic conditions (catalyst/substrate ratio 1:1000, 110
°C, 60 bar), we could not reproduce the higher activity
we had claimed (around 3000 instead of 300 cycles per
hour) [11,12]: Indeed, by using [2]BF4 as a catalyst,
benzene is hydrogenated to give cyclohexane, the cata-
lytic activity, however, being almost the same as for
[1]BF4. This observation finds its explanation in the
instability of 2 under the catalytic conditions: Under
hydrogen pressure (60 bars) at 110 °C, 2 converts back
quantitatively into 1 (see Fig. 1); cation 1 can be re-
covered at the end of the catalytic reaction as the tet-
rafluoroborate salt. It is therefore not surprising that,
even if [2][BF4] is used as catalyst precursor for the
hydrogenation of benzene under biphasic conditions,
the catalytic activity of the ‘‘closed’’ cluster 1 is observed
after some hours, because 1 seems to be the stable spe-
cies present during the catalytic run.
C6H5–CH2–CH3 þ O2 ! C6H5–CHðOOHÞ–CH3
2C6H5–CHðOOHÞ–CH3 ! C6H5–CHðOHÞ–CH3
þ C6H5–CðOÞ–CH3
2.3. The contamination problem: elucidation of the
increased catalytic activity by using contaminated ethyl-
benzene
Given that both contaminants found by GC/MS
analysis are oxidation products of ethylbenzene [33], we
considered the mechanism of the aerobic ethylbenzene
oxidation [34]: Ethylbenzene is attacked by molecular
oxygen from air at the methylene group to give the
hydroperoxide intermediate C6H5–CH(OOH)–CH3
which will then decompose to give phenylmethylketone
It turned out very soon that the higher catalytic ac-
tivity, which we had observed before and erroneously
attributed to cluster 2 [11,12], was due to a contamina-
+
+
O
O
H2
H2 O
H
Ru
H
Ru
Ru
H
Ru
H
OH
Ru
Ru
H
1/2 O2
(2)
(1)
-11.0
-12.0
-13.0
-14.0
-15.0
-16.0
-17.0
-18.0
-19.0
-20.0
-21.0
-22.0
-23.0
-24.0
(ppm)
Fig. 1. Conversion of the ‘‘closed cluster 1 with molecular oxygen into the ‘‘open’’ cluster 2 and conversion of 2 into 1 with molecular hydrogen. The
1H NMR spectrum shows a D2O solution of both 1 and 2 in equilibrium. (The enlargement reveals the hydride signal of 2 being a singlet at
d ¼ ꢂ13:77 ppm, while the hydride signals of 1 show the expected multiplicities [d ¼ ꢂ19:35 ppm (doublet)] and d ¼ ꢂ20:03 ppm (triplet) in a 2:1
integral ratio.)

1366
G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
and 1-phenylethan-1-ol. We reasoned that, if this hy-
droperoxide intermediate was present in the contami-
nated ethylbenzene, it would not be detected by GC/MS
analysis, because it would decompose on the GC column
to give the two compounds observed.
Based on the idea that the activating component in
the contaminated ethylbenzene is this hydroperoxide
intermediate, we tested the commercially available tert-
butylhydroperoxide as an activator for the hydrogena-
tion of benzene, catalysed under biphasic conditions by
[1][BF4], and we did indeed find an enhanced activity: By
using an aqueous phase (10 ml) obtained from the re-
action (30 min) of [1][BF4] (in 10 ml H2O) with 10
t
equivalents of BuOOH (in decane) after washing three
t
times with 20 ml of ether (to remove the excess of Bu-
OOH) for the hydrogenation of benzene (catalyst/sub-
strate ratio 1:1000, 110 °C, 60 bar), we observed a
catalytic activity (TOF 3700 hꢂ1) 10 times greater than
without activator. From this result, it can be concluded
that the contamination, which had caused the enhanced
catalytic activity, was indeed the primary ethylbenzene
oxidation product C6H5–CH(OOH)–CH3.
Fig. 2. Scanning electron micrograph of the aqueous phase recovered
from the biphasic hydrogenation of hydroperoxide-contaminated
ethylbenzene catalysed by [1][BF4], showing solid micro-scale particles
(presumably RuO2 ꢀ nH2O).
decomposition (<5%). After a catalytic run, the clear,
red, aqueous phase containing 1 can be reused for fur-
ther runs.
However, the aqueous phase obtained by treatment
with tert- butylhydroperoxide (or with the contaminated
ethylbenzene) did not contain 1 or other soluble orga-
nometallic species any more. It was cloudy and con-
tained a finely dispersed dark solid, which could be
filtered off and which proved to be a highly reactive
heterogeneous catalyst for the hydrogenation of ben-
zene. A second catalytic run with this solid catalyst,
suspended in water, reproduced the same catalytic ac-
tivity (TOF 3700 hꢂ1). The isolated solid contained 21%
oxygen, suggesting the presence of hydrated ruthenium
dioxide, RuO2 ꢀ nH2O, in accordance with energy dis-
persive X-ray spectroscopy (EDS). Scanning electron
microscopy (SEM) of the aqueous phase revealed a
particle size on the micro- (not nano) scale (Fig. 2).
The high catalytic activity of ruthenium dioxide, a
simple inorganic solid, for the hydrogenation of benzene
and benzene derivatives is due to the fact that we did the
catalytic reaction under biphasic conditions. In fact,
while commercial RuO2 ꢀ nH2O shows approximately
the same activity for the hydrogenation of benzene
without or with additional water (TOF around 3000
As originally proposed [11], 1 can accommodate the
substrate molecule benzene in its hydrophobic pocket
formed by the three arene ligands and to place it in a
perfect position underneath the Ru3 face opposite to the
oxo cap of the cluster. This is possible as, under biphasic
conditions, the hydrophobic substrate tries to escape
from the aqueous medium. In the resulting supramo-
lecular catalyst–substrate host–guest complex, the sub-
strate molecule is not coordinated to ruthenium but
interacts with the Ru3 surface only through weak in-
termolecular interactions.
The catalytic hydrogenation is believed to occur
within this supramolecular host–guest complex in a
three-step mechanism (Scheme 2): Transfer of two hy-
drogen atoms from the cluster molecule to the substrate
within the cluster–benzene complex leads to a cluster–
cyclohexadiene complex, in which the unsaturated
cluster [HRu3(C6H6)(C6Me6)2(O)]þ would react with
molecular hydrogen to regenerate [H3Ru3(C6H6)-
(C6Me6)2(O)]þ (1), capable of transferring two hydrogen
atoms to the cyclohexadiene molecule. In the resulting
cluster–cyclohexene complex, the unsaturated cluster
[HRu3(C6H6)(C6Me6)2(O)]þ would again react with H2
to regenerate 1, which would again transfer two hy-
drogen atoms to the substrate. The cyclohexane formed
would leave the hydrophobic pocket, while 1 is regen-
erated with molecular hydrogen. At the end of the cat-
alytic reaction, 1 is recovered unchanged as the
tetrafluoroborate salt.
h
ꢂ1), whereas anhydrous RuO2 was almost completely
inactive (TOF 2 hꢂ1) under the same mild conditions
(catalyst/substrate ratio 1:1000, 110 °C, 60 bar).
2.4. Supramolecular cluster catalysis revisited: isolation
and characterisation of supramolecular catalyst–substrate
host–guest complexes
In the absence of hydroperoxide contaminations, the
‘‘closed’’ cluster cation [H3Ru3(C6H6)(C6Me6)2(O)]þ (1)
catalyses the hydrogenation of benzene and benzene
derivatives under biphasic conditions essentially without
Recently, we prepared two ‘‘closed’’ cluster cations
analogous to 1, [H3Ru3{C6H5(CH2)2OH}(C6Me6)2-
(O)]þ (3), and [H3Ru3{C6H5(CH2)3OH}(C6Me6)2(O)]þ
(4), by analogy from the dinuclear precursor

G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
1367
Ru
H
Ru
H
Ru
H
H2
Ru
H
H
H
H
H
H
H
H
H
Ru
Ru
Ru
Ru
Ru
Ru
Ru
Ru
Ru
H
Ru
Ru
H
Ru
H
H
H
H
H
H
H
H
H
Ru
Ru
Ru
Ru
Ru
Ru
Ru
Ru
H2
H2
H
Scheme 2. Proposed mechanism for the stepwise catalytic hydrogenation of benzene to cyclohexane within the hydrophobic pocket (dashed line)
spanned by the three arene ligands of the intact trinuclear cluster 1–4 (arene ligands and oxo cap omitted for clarity).
[H3Ru2(C6Me6)2]þ and the corresponding mononuclear
arene ruthenium triaqua complex [35].
½C6H6 ꢃ 3ꢁþ, the angle formed by the C6 plane and the
Ru3 plane is 66.78(7)°, while it is 88.63(9)° in
½C6H6 ꢃ 4ꢁþ, the shortest distances between the metal-
bound hydrogen atoms and the closest carbon atoms of
2þ
½H3Ru2ðC6Me6Þ ꢁþ þ ½RufC6H5ðCH2Þ OHgðH2OÞ ꢁ
2
n
3
! ½H3Ru3fC6H5ðCH2Þ OHgðC6Me6Þ ðOÞꢁþ
ꢁ
the benzenþe guest molecule being 3.49 and 3.69 A in
n
2
þ
ꢁ
þ 2H2O þ 2Hþ ðn ¼ 2; 3Þ
½C6H6 ꢃ 3ꢁ , and 3.26 and 3.77 A in ½C6H6 ꢃ 4ꢁ (Figs.
5 and 6).
The direct observation of the catalyst–substrate host–
guest complexes was a missing link in the concept of
supramolecular cluster catalysis. The isolation and the
X-ray crystallographic characterisation of ½C6H6 ꢃ 3ꢁ-
½PF6ꢁ and ½C6H6 ꢃ 4ꢁ½BF4ꢁ demonstrate that the hy-
pothesis of catalyst–substrate host–guest complexes as
key species in the catalytic hydrogenation of benzene is
not unreasonable. However, on the basis of the data
obtained so far, we cannot exclude catalysis by ‘‘solu-
ble’’ metallic species (nanoclusters or colloids), and we
are presently engaged in a collaborative effort to refutte
or support this alternative hypothesis [37].
In these two cluster cations containing a (CH2)nOH
(n ¼ 2, 3) side-arm at the benzene ligand, we were able
to isolate the catalyst–substrate host–guest complexes
½C6H6 ꢃ 3ꢁ½PF6ꢁ and ½C6H6 ꢃ 4ꢁ½BF4ꢁ and to characte-
rise these postulated supramolecular intermediates by
single-crystal X-ray structure analysis [35]. A structural
comparison of 3 and ½C6H6 ꢃ 3ꢁþ shows almost identi-
cal geometrical parameters, differences appear only at
the periphery (Figs. 3 and 4).
In the two catalyst–substrate complexes isolated, the
substrate guest is accommodated by the catalyst host in
an inclined fashion inside the hydrophobic pocket: In
Fig. 3. Molecular structure of cation 3 and ½C6H6 ꢃ 3ꢁþ at 25% probability level, H atoms and PF6 omitted for clarity.

1368
G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
3. Conclusions
By checking the chemistry underlying the concept of
‘‘supramolecular cluster catalysis’’ we identified two
major errors in our publications related to this topic
[11–13,36], which are essentially due to contamination
problems.
(1) The conversion of the ‘‘closed’’ cluster cation
[H3Ru3(C6H6)(C6Me6)2(O)]þ (1) into the ‘‘open’’ cluster
cation [H2Ru3(C6H6) (C6Me6)2(O)(OH)]þ (2), which we
had ascribed to a reaction with water in the presence of
ethylbenzene [11–13,36], is simply an oxidation reaction
which occurs in the presence of air.
(2) The higher catalytic activity observed with ethyl-
benzene, which we had erroneously attributed to the
‘‘open’’ cluster cation [H2Ru3(C6H6)(C6Me6)2(O)-
(OH)]þ (2), was due to the formation of RuO2 ꢀ nH2O,
caused by a hydroperoxide contamination present in
ethylbenzene.
Fig. 4. Molecular structure of ½C6H6 ꢃ 4ꢁþ at 25% probability level, H
atoms, H2O and BF4 omitted for clarity.
Additionally, we have been able to isolate and un-
equivocally characterise by X-ray crystallography the
novel benzene host–guest complexes ½C6H6 ꢃ 3ꢁþ and
½C6H6 ꢃ 4ꢁþ.
The authors are grateful to the Fonds National Suisse
de la Recherche Scientifique for financial support. A
generous loan of ruthenium chloride hydrate from the
Johnson Matthey Technology Centre is gratefully ac-
knowledged.
References
Fig. 5. Space filling representation of the host–guest complexes
½C6H6 ꢃ 3ꢁþ and ½C6H6 ꢃ 4ꢁþ, based on the X-ray data, showing the
benzene host (at the top) penetrating the hydrophobic pocket of 3
and 4.
[1] (a) S. Bhaduri, D. Mukesh, Homogeneous Catalysis, Wiley–
Interscience, New York, 2000;
(b) B. Cornils, W.A. Herrmann, Applied Homogeneous Catalysis
with Organometallic Compounds, VCH, Weinheim, 1996;
(c) G.W. Parshall, S.D. Ittel, Homogeneous Catalysis, Wiley–
Interscience, New York, 1992.
[2] B.R. James, Adv. Organomet. Chem. 17 (1979) 319.
[3] D. Forster, Adv. Organomet. Chem. 17 (1979) 255.
[4] P.M. Maitlis, A. Haynes, G.J. Sunley, M.J. Howard, J. Chem.
Soc., Dalton Trans. (1996) 2187.
[5] W.A. Herrmann, B. Cornils, Angew. Chem., Int. Ed. Engl. 36
(1997) 1048.
[6] (a) R. Noyori, T. Ohkuma, Angew. Chem. 113 (2001) 41;
Angew. Chem. Int. Ed. 40 (2001) 40;
(b) D.A. Alonso, P. Brandt, S.J.M. Nordin, P.G. Andersen, J.
Am. Chem. Soc. 121 (1999) 9580.
[7] E.N. Jacobsen, M.H. Wu, in: E.N. Jacobsen, A. Pfaltz, H.
Yamamoto (Eds.), Comprehensive Asymmetric Catalysis, vol. II,
Springer, Berlin, 1999, p. 650.
[8] (a) M. Yamakawa, H. Ito, R. Noyori, J. Am. Chem. Soc. 122
(2000) 1466;
(b) K. Abdur-Rashid, A.J. Lough, R.H. Morris, Organometallics
19 (2000) 2655;
(c) D.G.I. Petra, J.N.H. Reek, J.-W. Handgraaf, E.J. Meijer, P.
Dierkes, P.C.J. Kamer, J. Brussee, H.E. Schoemaker, P.W.N.M.
van Leeuwen, Chem. Eur. J. 6 (2000) 2818.
Fig. 6. Space filling representation of the host–guest complexes
½C6H6 ꢃ 3ꢁþ and ½C6H6 ꢃ 4ꢁþ with the benzene guest (top) ap-
proaching the Ru3 plane, the arene ligands being omitted for clarity.
€
[9] M. Palucki, N.S. Finney, P.J. Pospisil, M.L. Guler, T. Ishida,
E.N. Jacobsen, J. Am. Chem. Soc. 120 (1998) 948.

G. S€uss-Fink et al. / Journal of Organometallic Chemistry 689 (2004) 1362–1369
1369
€
[10] M. Faure, M. Jahncke, A. Neels, H. Steckli-Evans, G. Suss-Fink,
Polyhedron 18 (1999) 2679.
[22] M.J. Russell, C. White, P.M. Maitlis, J. Chem. Soc., Chem.
Commun. (1977) 427.
€
[11] G. Suss-Fink, M. Faure, T.R. Ward, Angew. Chem. 114 (2002)
[23] D. Durand, G. Hillion, C. Lassau, L. Sajus, US Patent 4,271,323,
1981.
105;
Angew. Chem. Int. Ed. 41 (2002) 99.
€
[12] M. Faure, A. Tesuro Vallina, H. Steckli-Evans, G. Suss-Fink, J.
[24] H. Brunner, Hydrogenation, VCH, Weinheim, 1996, p. 216.
[25] (a) I.P. Rothwell, Chem. Commun. (1997) 1331;
(b) V.M. Visciglio, J.R. Clark, M.T. Nguyen, D.R. Mulford, P.E.
Fanwick, I.P. Rothwell, J. Am. Chem. Soc. 119 (1997) 3490.
[26] Y. Lin, R.G. Finke, Inorg. Chem. 33 (1994) 4891.
[27] (a) K.R. Januszkiewicz, H. Alper, Organometallics 2 (1983) 1055;
(b) J. Blum, I. Amer, K.P.C. Vollhardt, H. Schwarz, G. Hoehne,
J. Org. Chem. 52 (1987) 2804.
Organomet. Chem. 621 (2001) 103.
[13] G. Laurenczy, M. Faure, L. Vieille-Petit, G. Suss-Fink, T.R.
€
Ward, Adv. Synth. Catal. 344 (2002) 1073.
[14] A literature search (SciFinder connecting to CAS) of benzene
hydrogenation reveals 1498 references (April 2003), including only
54 references for homogeneous benzene hydrogenation..
[15] (a) E.L. Muetterties, F.J. Hirsekorn, J. Am. Chem. Soc. 96 (1974)
4063;
[28] K.S. Weddle, J.D. Aiken III, R.G. Finke, J. Am. Chem. Soc. 120
(1998) 5653.
(b) F.J. Hirsekorn, M.C. Rakowski, E.L. Muetterties, J. Am.
Chem. Soc. 97 (1975) 237;
[29] (a) J. Schulz, A. Roucoux, H. Patin, Chem. Commun. (1999) 28;
(b) J. Schulz, A. Roucoux, H. Patin, Chem. Eur. J. 6 (2000) 618.
[30] (a) J.A. Widegren, R.G. Finke, J. Mol. Catal. A 191 (2003) 187;
(b) P.J. Dyson, Dalton Trans. (2003) 2964.
(c) F.J. Hirsekorn, M.C. Rakowski, E.L. Muetterties, W.D.
Larson, V.J. Basus, F.A.L. Anet, J. Am. Chem. Soc. 97 (1975)
1266;
[31] J.A. Widegren, R.G. Finke, J. Mol. Catal. A 198 (2003) 317.
[32] J.A. Widegren, M.A. Bennett, R.G. Finke, J. Am. Chem. Soc. 125
(2003) 10301.
(d) M.C. Rakowski, F.J. Hirsekorn, L.S. Stuhl, E.L. Muetterties,
Inorg. Chem. 15 (1976) 2379;
(e) L.S. Stuhl, M. Rakowski DuBois, F.J. Hirsekorn, J.R. Bleeke,
A.E. Stevens, E.L. Muetterties, J. Am. Chem. Soc. 100 (1978)
2405.
[33] C. Guo, Q. Peng, Q. Liu, G. Jiang, J. Mol. Catal. A 192 (2003) 295.
[34] A.E. Shilov, G.B. ShulÕpin, Activation and Catalytic Reactions of
Saturated Hydrocarbons in the Presence of Metal Complexes,
Kluwer Academic Publishers, Dordrecht, 2000.
[16] J.W. Johnson, E.L. Muetterties, J. Am. Chem. Soc. 99 (1977)
7395.
€
[35] L. Vieille-Petit, B. Therrien, G. Suss-Fink, T.R. Ward, J.
[17] M.A. Bennett, T.-N. Huang, T.W. Turney, J. Chem. Soc., Chem.
Commun. (1979) 312.
Organomet. Chem. 684 (2003) 117.
€
[36] G. Suss-Fink, in: W.A. Herrmann (Ed.), Synthetic Methods of
Organometallic and Inorganic Chemistry (Herrmann/Brauer), vol.
10, Thieme, Stuttgart, 2001, p. 121.
[18] A.F. Borowski, S. Sabo-Etienne, B. Chaudret, J. Mol. Catal. A
174 (2001) 69.
€
[19] L. Plasseraud, G. Suss-Fink, J. Organomet. Chem. 539 (1997) 163.
[37] In order to check the possibility of nanocluster catalysis in this
case, we started a collaboration with Professor Richard G. Finke
(Colorado State University) for a detailed kinetic study of this
reaction under biphasic conditions. The results will be published
in due time.
[20] P.J. Dyson, D.J. Ellis, D.G. Parker, T. Welton, Chem. Commun.
(1999) 25.
€
[21] E. Garcia Fidalgo, L. Plasseraud, G. Suss-Fink, J. Mol. Catal. A
132 (1998) 5.