Ruthenium-catalyzed isomerization of allylic alcohols: Oxidation state determines resistance against diene inhibition

Article abstract of DOI:10.1002/1099-0682(200208)2002:8<2147::AID-EJIC2147>3.0.CO;2-L

The novel complex mer-[RuCl₃(dmso)(phen)] (1) has been prepared and characterized by X-ray diffraction. The ruthenium center is in a distorted octahedral environment with the three chloride ions coordinated in a mer-fashion and an S-bonded dmso ligand trans to one of the phen nitrogen atoms. Electrochemical experiments show two reversible waves in acetonitrile solution, corresponding to the couples Ru^III/II (0.11 V) and Ru^IV/III (1.73 V). The analogous Ru^II complex cis, cis-[RuCl₂(dmso)₂(phen)] (2) shows one reversible wave at 1.08 V, corresponding to the Ru^III/II couple. Both complexes exhibit high catalytic activity in the isomerization of 3-buten-2-ol to butanone. The initial turnover frequency (TOF) for 1 in diglyme/water at 130 °C is 295 h^-1 with a firstorder k_ini of 0.6 h^-1, while 2 reaches an initial TOF of 260 h^-1 and a k_ini of 0.5 h^-1. A cumulative turnover number (TON) of 1025 has been obtained with 1 as a catalyst precursor. The activity of 1 and 2 has been compared to that of in situ mixtures with both Ru^III and Ru^II precursors. All Ru^II complexes are deactivated before 100% conversion has been attained. RuCl₃ is a good catalyst precursor with an initial TOF of 500 h^-1. If RuCl₃ is mixed in situ with one equivalent of phen, an induction period of 40 min results and the activity thereafter is much lower than without the addition of phen. With 2 equiv. of phen, no reaction is observed in the first 6 h and only a very low activity is obtained after that. An important difference between 1 and 2 becomes apparent when a conjugated diene is added to the reaction mixture; only 1 remains active. The results demonstrate that only Ru^III complexes with one phen ligand are active catalyst precursors for the isomerization of allylic alcohols in the presence of conjugated dienes. The consequence of this observation is further demonstrated by applying both catalyst precursors in the direct one-pot conversion of 1,3-butadiene to butanone. In this reaction, complex 1 (TON = 2050, t = 6 h) forms a much more active precursor than complex 2 (TON = 1300, t = 7 h). Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002.

Full text of DOI:10.1002/1099-0682(200208)2002:8<2147::AID-EJIC2147>3.0.CO;2-L

FULL PAPER
Ruthenium-Catalyzed Isomerization of Allylic Alcohols: Oxidation State
Determines Resistance Against Diene Inhibition
Robert C. van der Drift,[a] Jeroen W. Sprengers,[a] Elisabeth Bouwman,*[a]
Wilhelmus P. Mul,[b] Huub Kooijman,[c] Anthony L. Spek,[c] and Eite Drent[a]
Keywords: Isomerization / Allylic alcohols / Homogeneous catalysis / Ruthenium / Diene inhibition
The novel complex mer-[RuCl3(dmso)(phen)] (1) has been
prepared and characterized by X-ray diffraction. The ruth-
enium center is in a distorted octahedral environment with
the three chloride ions coordinated in a mer-fashion and an
S-bonded dmso ligand trans to one of the phen nitrogen
atoms. Electrochemical experiments show two reversible
waves in acetonitrile solution, corresponding to the couples
RuCl3 is a good catalyst precursor with an initial TOF of
500 h−1. If RuCl3 is mixed in situ with one equivalent of phen,
an induction period of 40 min results and the activity there-
after is much lower than without the addition of phen. With
2 equiv. of phen, no reaction is observed in the first 6 h and
only a very low activity is obtained after that. An important
difference between 1 and 2 becomes apparent when a con-
RuIII/II (0.11 V) and RuIV/III (1.73 V). The analogous RuII com- jugated diene is added to the reaction mixture; only 1 re-
plex cis,cis-[RuCl2(dmso)2(phen)] (2) shows one reversible
mains active. The results demonstrate that only RuIII com-
wave at 1.08 V, corresponding to the RuIII/II couple. Both com- plexes with one phen ligand are active catalyst precursors
plexes exhibit high catalytic activity in the isomerization of
3-buten-2-ol to butanone. The initial turnover frequency
for the isomerization of allylic alcohols in the presence of con-
jugated dienes. The consequence of this observation is fur-
(TOF) for 1 in diglyme/water at 130 °C is 295 h−1 with a first- ther demonstrated by applying both catalyst precursors in
order kini of 0.6 h−1, while 2 reaches an initial TOF of 260 h−1
and a kini of 0.5 h−1. A cumulative turnover number (TON) of In this reaction, complex 1 (TON = 2050, t = 6 h) forms a
the direct one-pot conversion of 1,3-butadiene to butanone.
1025 has been obtained with 1 as a catalyst precursor. The
activity of 1 and 2 has been compared to that of in situ mix-
much more active precursor than complex 2 (TON = 1300,
t = 7 h).
tures with both RuIII and RuII precursors. All RuII complexes ( Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany,
are deactivated before 100% conversion has been attained.
2002)
Introduction
use of stoichiometric amounts of concentrated sulfuric acid
as shown in Scheme 1.
Homogeneous catalysis opens new pathways to environ-
mentally benign industrial processes. Many processes in use
today produce stoichiometric amounts of inorganic salts.
Transition metal catalyzed reactions on the other hand,
may offer complete atom-economical routes to important
industrial products. For example, butanone (methyl ethyl
ketone, MEK) is a megaton-per-year scale solvent tradi-
tionally produced from butenes.[1] This process requires the
Scheme 1. Present-day synthesis of MEK from a mixture of butenes
1,3-Butadiene is an attractive alternative starting material
for the preparation of MEK. A viable route would require
water as the only other stoichiometric reagent. After acid-
catalyzed hydration of butadiene [see (a) in Scheme 2], the
1,2-addition product 3-buten-2-ol can be isomerized to
MEK [see (b) in Scheme 2].[2] The equilibrium of Scheme 2
(a) is 97% on the left-hand side, making a one-pot synthesis
preferable in order to circumvent an extensive butadiene re-
cycle. For the selective formation of MEK, this calls for
catalysts that catalyze the isomerization of the secondary
allylic alcohol 3-buten-2-ol but not the primary allylic alco-
hol 2-buten-1-ol, in the presence of dienes. Despite consid-
[a]
Gorlaeus Laboratories, Leiden Institute of Chemistry, Leiden
University,
P. O. Box 9502, 2300 RA Leiden, The Netherlands
Fax: (internat.) ϩ 31-71/527-4451
E-mail: bouwman@chem.leidenuniv.nl
[b]
Shell International Chemicals BV, Shell Research and
Technology Center Amsterdam,
Badhuisweg 3, 1031 CM Amsterdam, The Netherlands
[c]
Bijvoet Center for Biomolecular Research, Crystal and
Structural Chemistry, Utrecht University,
Padualaan 8, 3584 CH Utrecht, The Netherlands
Supporting information for this article is available on the
WWW under http://www.eurjic.com or from the author.
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155  WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002 1434Ϫ1948/02/0808Ϫ2147 $ 20.00ϩ.50/0
2147

E. Bouwman et al.
FULL PAPER
erable efforts this type of catalysts remain scarce.[3] Altern- compared in the absence and the presence of isoprene (2-
atively, other reagents such as alcohols and amines could methyl-1,3-butadiene) to reveal the role of the ruthenium
be used instead of water to obtain higher conversions, and oxidation state in the isomerization of allylic alcohols.
after isomerization and hydrolysis of the vinyl ether or en-
amine the carbonyl compound can be obtained.[4,5] In this
case an extra, costly purification step will be required to
separate MEK from butanal (or to separate the interme-
Results and Discussion
diates).
Synthesis and Structure of mer-[RuCl3(dmso)(phen)] (1) and
cis,cis-[RuCl2(dmso)2(phen)] (2)
Synthesis of 1 and 2
To obtain a clear view of the influence of the ruthenium
oxidation state in the catalytic process, two complexes are
required with ligand environments as identical as possible.
It was deduced from earlier studies[7,8] that the catalytically
active species should contain only one phen ligand. Fur-
thermore, the complexes should contain innocent ligands
that are easily replaced during the catalytic cycle. Numerous
ruthenium complexes with two or three phen ligands are
known,[9,10] but examples of mono(phen) complexes remain
scarce.[9Ϫ13] It is especially difficult to prevent ligand redis-
tribution during synthesis.[12] The synthesis of mono(phen)
Scheme 2. (a) Acid-catalyzed hydration of 1,3-butadiene (b) fol-
lowed by isomerization yields MEK in a complete atom-econom-
ical process
The feasibility of the one-pot synthesis of MEK from
1,3-butadiene was shown by Drent and co-workers.[6,7]
A
complexes was attempted in our laboratory starting from
[14]
[Ru(H2O)6](OTs)2, [Ru(MeCN)6](OTf)2
and RuCl3. In
catalytically active species obtained by in situ mixing of
[Ru(acac)3] (Hacac ϭ 2,4-pentanedione) and between 1 and
2 equiv. of 2,2Ј-bipyridine (bpy) or 1,10-phenanthroline
(phen) catalyzes the direct conversion of 1,3-butadiene to
MEK (in over 95% selectivity) with a maximum of 1200
turnovers in 32 h. Yet, this turnover number (TON) is not
sufficiently high to make this route economically successful.
An important catalyst deactivation route proved to be li-
gand redistribution that results in inactive ruthenium com-
plexes with two and three didentate nitrogen-donor ligands
(Scheme 3).[7] Indeed, it has been demonstrated recently
that reducing the extent of ligand redistribution may in-
crease the initial activity by a factor of almost ten.[8]
all cases, it proved impractical to obtain the ruthenium
atom in both the 2ϩ and 3ϩ oxidation states. The syntheses
of [Ru(acac)2(phen)] and [Ru(acac)2(phen)]PF6 were suc-
cessful,[15] but the acac ligand could not be replaced under
the reaction conditions.
A literature method to prepare the RuII complex 2[12]
from cis-[RuCl2(dmso)4] and phen in CHCl3 gave only low
yields in our hands, but changing the solvent to toluene[16]
resulted in direct precipitation of the pure product from the
reaction mixture [see (c) in Scheme 4]. An additional and
vital advantage of this method is the complete absence of
bis- and tris(phen)ruthenium complexes. Upon slow con-
centration of a CHCl3/toluene solution of 2, red crystals of
1 were obtained; a lead to the synthesis of the analogous
RuIII complex 1 was found. CHCl3 is known to contain
traces of HCl and recently the syntheses of several com-
plexes with the general formula [RuCl3(dmso)L], where L is
triazole, thiadiazole or diamine, by oxidation with HCl were
reported.[17,18] Indeed, it proved feasible to prepare 1 by
oxidation of cis-[RuCl2(dmso)4] with HCl in the presence of
phen [see (a) and (b) in Scheme 4]. The formation of an
RuIII species from an RuII precursor can be envisioned to
involve simultaneous reduction of the dmso ligand to di-
Scheme 3. Ligand redistribution of mono(phen)Ru complexes; S ϭ
vacant site or labile ligand; NʜN ϭ phen; 2x ϩ 3y ϭ 1, n ϭ 2, 3
Although the catalyst precursor consists of an RuIII com- methyl sulfide as proposed by Cingi et al. [see Equa-
plex, both RuIII and RuII complexes can be identified with tion (1)].[17] A general path to the RuII species that does not
mass spectrometry (MS) in a used catalyst. This observa- invoke ligand participation is discussed below. The room-
tion raises the question as to what the oxidation state of the temperature route [see (a) in Scheme 4] results directly in
ruthenium atom in the catalytically active species is. In this analytically pure 1, but the overall yield with the high-tem-
paper the synthesis and characterization of the RuIII com- perature route [see (b) in Scheme 4] is higher. It should be
pound mer-[RuCl3(dmso)(phen)] (1) and the analogous RuII noted, however, that at higher reaction temperatures with
complex cis,cis-[RuCl2(dmso)2(phen)] (2) are described. The longer reaction times and using concentrated HCl, other
activities of both complexes in the key step of the catalytic products may precipitate, which are probably oligomeric
process, the isomerization of 3-buten-2-ol to MEK, are analogs of 1.
2148
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155

Ruthenium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
ruthenium center is in a distorted octahedral environment
with coordination of one didentate phen ligand, three chlor-
ide ions and one S-bonded dmso ligand trans to one of the
phen nitrogen atoms. The SϪO bond is in the same plane
as the phen ligand. Both rutheniumϪnitrogen distances are
in the range found for other ruthenium complexes with aro-
matic nitrogen donor ligands.[13,17] The Ru(1)ϪS(21) bond
is longer than in analogous RuII complexes,[17] which re-
flects the smaller degree of π-back donation in this RuIII
Scheme 4. Synthesis of 1 (a) (b) and 2 (c) from cis-[RuCl2(dmso)4]
˚
complex. The S(21)ϪO(22) bond [1.478(2) A] is slightly
[19]
˚
shorter than in free dmso [1.492(1) A] indicating greater
SϪO double-bond character in the S-bonded dmso. The
bonds within the phen ligand are similar to those in other
(phen)ruthenium complexes.[13]
The Cl(1)ϪRu(1)ϪCl(2) angle is significantly smaller
than 180° due to steric repulsion of the dmso methyl
groups; the N(1)ϪRu1ϪN(10) bond angle of 78.99(8)° is
contracted compared to the ideal octahedral value by the
small phen bite angle, which is somewhat smaller than in
the complex where dmso is replaced by CO.[13] In this com-
plex the phen ligand itself is boat-shaped to some extent,
with angles between pairs of least-squares planes (through
each of the three rings) ranging from 4.97(13) to 10.30(12)°.
In the crystal packing a high degree of stacking is observed.
Only two rings of each phen ligand are stacked with two
rings of a phen ligand on an adjacent complex molecule
(generated by symmetry operation 1 Ϫ x, Ϫy, 1 Ϫ z),
avoiding steric repulsion between neighboring dmso groups.
Approximately perpendicular to this, the cocrystallized
toluene molecules form dimers, totally enclosed by molec-
ules 1, through stacking interactions.
(1)
Crystal Structure Determination of 1
An ORTEP drawing of 1 is shown in Figure 1. Some se-
lected bond lengths and angles are given in Table 1. The
Cyclic Voltammetry Studies on 1 and 2
The redox properties of both complexes have been invest-
igated with cyclic voltammetry. In an acetonitrile solution,
1 exhibits two reversible one-electron waves (Figure 2). The
first peak (E1/2 ϭ 0.11 V, ∆Ep ϭ 60 mV) corresponds to the
RuIII/II couple, whereas the second peak (E1/2 ϭ 1.73 V,
∆Ep ϭ 54 mV) corresponds to the RuIV/III couple. The na-
ture of both couples has been corroborated by linear sweep
voltammetry that shows the expected signs of the currents.
The current for both peaks is linearly dependent on v1/2 for
the scan rates studied (0.05Ϫ1 V sϪ1) indicating reversible
diffusion-controlled processes.[20] The potential of the first
wave is evidence for the ease of reduction of 1, which plays
an important role in isomerization catalysis (vide infra).
Figure 1. ORTEP plot and numbering scheme of the major com-
ponent of complex 1; hydrogen atoms and toluene solvent molec-
ules are omitted for clarity; thermal ellipsoids are drawn at the 50%
probability level
Table 1. Selected bond lengths and angles for 1·toluene
˚
Bond lengths [A]
Ru(1)ϪN(1)
Ru(1)ϪN(10)
Ru(1)ϪCl(1)
Ru(1)ϪCl(2)
2.099(2)
2.087(2)
2.3267(8)
2.3444(8)
Ru(1)ϪCl(3)
Ru(1)ϪS(21)
S(21)ϪO(22)
2.3477(8)
2.2971(8)
1.478(2)
Angles [°]
N(1)ϪRu(1)ϪN(10)
N(1)ϪRu(1)ϪCl(1)
N(1)ϪRu(1)ϪCl(2)
N(1)ϪRu(1)ϪCl(3)
N(1)ϪRu(1)ϪS(21)
N(10)ϪRu(1)ϪCl(1)
N(10)ϪRu(1)ϪCl(2)
N(10)ϪRu(1)ϪCl(3)
78.99(8)
86.27(6)
87.08(6)
172.34(6)
100.60(6)
90.42(6)
87.93(7)
93.35(6)
N(10)ϪRu(1)ϪS(21)
S(21)ϪRu(1)ϪCl(1)
S(21)ϪRu(1)ϪCl(2)
S(21)ϪRu(1)ϪCl(3)
Cl(1)ϪRu(1)ϪCl(2)
Cl(1)ϪRu(1)ϪCl(3)
Cl(2)ϪRu(1)ϪCl(3)
178.75(7)
90.73(2)
90.87(2)
87.06(2)
173.34(3)
93.49(2)
93.05(2)
Figure 2. Cyclic voltammogram of 1 in acetonitrile at room temper-
ature; scan rate: 0.1 V sϪ1
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155
2149

E. Bouwman et al.
FULL PAPER
The reversibility of the second peak is remarkable, demon- volatility (b.p. Ϫ4 °C) it is difficult to handle in standard
strating a robust RuIV complex, this was also observed in glassware, so it was decided instead to use isoprene in this
the analogous compound mer-[RuCl3(dmso)(tmen)] study. Isoprene is converted analogously to MIPK [methyl
(tmen ϭ N,N,NЈ,NЈ-tetramethylethylenediamine).[18] From isopropyl ketone, Equation (2)] and can thus function as
this peak it can be concluded that dmso in 1 is primarily a a reasonable substitute for 1,3-butadiene.[7] The results of
σ-donor and remains S-bonded even with the ruthenium isomerization of 3-buten-2-ol to MEK catalyzed by several
atom in the 4ϩ oxidation state.
RuIII (Entries 1Ϫ5) and RuII catalyst precursors (Entries
Complex 2 shows one reversible one-electron wave 6Ϫ10) both in the absence and presence of isoprene are
(E1/2 ϭ 1.08 V, ∆Ep ϭ 56 mV), corresponding to the RuIII/ collected in Table 2.
II
couple (Figure 3). In this case, the current for this peak
is also linearly dependent on v1/2 for the scan rates studied
(0.05Ϫ1 V sϪ1) and the oxidation state of 2 has been proven
(2)
with linear sweep voltammetry. In the reductive part of the
voltammogram (at potentials Ͻ Ϫ1.0 V) an irreversible
wave is observed that can be attributed to a ligand-based
reduction or reduction of the ruthenium center to Ru0.
It is clear that both RuIII and RuII complexes are capable
of catalyzing the isomerization of 3-buten-2-ol to MEK, but
RuIII precursors consistently show higher conversion. The
result of Entry 10 demonstrates that exclusion of air is not
essential to obtaining good activity. Comparison of Entries
1 and 6 with 2 and 7, respectively, shows that the phen
ligand as such is not a necessary requirement for activity.
Both in situ mixing of the components and the use of pre-
formed catalyst precursors gave comparable results. If a
higher substrate/1 ratio is applied, a TON of 1025 in 12 h
can be obtained with an average TOF of 85 hϪ1. The nega-
tive influence of an increased ligand-redistribution rate is
demonstrated by successive removal of chloride ions by AgI
salts from either 1 or 2, which results in the same or only
slightly lower catalytic activity (Entries 5 and 9). Presum-
ably, the anticipated higher activity, as a consequence of
more vacant coordination sites, is counterbalanced by an
increased rate of ligand redistribution. Addition of a four-
fold excess of AgI gives a color change of the reaction mix-
ture to dark blue and almost complete loss of activity.
While a large number of catalysts are able to catalyze the
isomerization of allylic alcohols to carbonyl compounds,[2]
almost all of these fail to do so in the presence of dienes.
The crucial next step therefore is to compare the catalytic
systems in the presence of isoprene. From Table 2, it be-
comes apparent that isoprene indeed has a dramatic effect
on the catalysis. Two factors are proven to play a role in
resistance against isoprene inhibition. First, although phen
is not required for activity in the absence of isoprene, the
significant difference in activity of the catalyst precursors
in Entries 1 and 2 shows that phen is an essential compon-
ent for catalytic activity in the presence of isoprene. Second,
all RuII catalyst precursors (Entries 6Ϫ9) loose 50% or
more of their activity with isoprene in the reaction mixture.
The presence of phen here (Entry 7) does not lead to resist-
ance against diene inhibition. Thus, both of the require-
ments, a 3ϩ oxidation state and one phen ligand, have to
be met. Complex 1 remains active, even if more isoprene is
added (Entry 4, Ͼ 1300 mol-equiv. to ruthenium). These
results unambiguously show, for the first time, the crucial
influence of the oxidation state of the ruthenium center in
allylic alcohol isomerization in the presence of a diene.
Comparison of several catalytic precursors lacking or con-
Figure 3. Part of the cyclic voltammogram of 2 in acetonitrile at
room temperature showing the RuIII/II wave; scan rate: 0.1 V sϪ1
Isomerization of 3-Buten-2-ol to MEK
The Influence of a Conjugated Diene
The key step in the direct conversion of 1,3-butadiene to
MEK has been identified as the isomerization of 3-buten-
2-ol to MEK in the presence of 1,3-butadiene.[3,7] Since 1,3-
butadiene is a suspected carcinogenic and due to its high
Table 2. Isomerization of 3-buten-2-ol to MEK catalyzed by RuIII
and RuII complexes
Entry[a]
Catalyst precursor
TON
TON
with isoprene[b]
1
2
[RuCl3·xH2O]
[RuCl3·xH2O]
ϩ 1 equiv. phen
1
285
390
95
300
3
4
5
6
7
465
n.d.
410
325
275
440
475
n.d.
110
75
1[c]
1 ϩ 3 equiv. AgOTs
cis-[RuCl2(dmso)4]
cis-[RuCl2(dmso)4]
ϩ 1 equiv. phen
2
8
9
10
290
295
270
150
n.d.
n.d.
2 ϩ 2 equiv. AgOTs
2[d]
[a]
Reactions were performed in a 25-mL sealed glass vessel at
130°C; substrate: 3-buten-2-ol (5.8 mmol); catalyst: ruthenium
complex (0.011 mmol) plus ligand as appropriate; substrate/cata-
lyst: 530; solvent: water/diglyme (1:3 mL); reaction time: 6 h; n.d. ϭ
[b]
not determined.
10 mmol of isoprene was added prior to addi-
tion of the substrate. [c] 15 mmol of isoprene. [d] In Ar with standard
Schlenk techniques.
2150
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155

Ruthenium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
taining various amounts of dmso ligands (cf. Table 2, Ent-
The results of a time-dependent investigation using cata-
ries 2 and 3, 7 and 8), especially in the presence of isoprene, lysts formed from [RuCl3·xH2O] and varying amounts of
shows that the influence of the ligands on the difference in phen at 100 °C are shown in Figure 5. In the absence of
activity is marginal.
phen, the isomerization is first order in 3-buten-2-ol with
an initial TOF of 500 hϪ1 after a negligible induction
period. When 1 equiv. of phen is added, the induction time
increases to 40 min. Moreover, the initial TOF drops to
105 hϪ1. In the original catalytic system with Ru(acac)3, ad-
dition of either 1 or 2 equiv. of phen resulted in equal react-
ivity.[7] In sharp contrast to this, the activity of RuCl3 with
2 equiv. of phen is virtually zero. After more than 16 h of
reaction time, a small amount of MEK (60 turnovers) is
obtained, but this must have been formed after an induction
period of at least 5 h (Figure 5). The variation in lengths of
the induction periods most likely originates from a kinetic
difference in displacement of the various ligands involved.
If isomerization is to occur with [RuCl3·xH2O] as the cata-
lyst precursor, water has to be replaced by the substrate,
which appears to be reasonably fast. Upon addition of 1
equiv. of phen, [RuCl3(H2O)(phen)] is formed. In this case,
dissociation of one or more chloride ions has to take place
prior to catalysis, which is considerably slower.[22] An induc-
tion period is not often observed in transition metal cata-
lyzed isomerization of allylic alcohols and it seems to point
at a cooperative effect; dissociation of the second and sub-
sequent chloride ions is faster after dissociation of the first
chloride ion. A similar effect has also been observed in iso-
merization of 3-buten-2-ol to MEK catalyzed by [Ru-
(MeCN)6].[14] Increasing the polarity of the reaction mix-
ture, by doubling the water content, should facilitate disso-
ciation. Indeed, the induction period is reduced to 20 min,
while the initial TOF is raised to 215 hϪ1. At the higher
reaction temperatures used with complexes 1 and 2 (Fig-
ure 4), chloride dissociation is much faster and only a small
induction period is observed. If 2 equiv. of phen are added
to RuCl3, [Ru(phen)2Cl2]Cl is formed. In an analogous iso-
merization reaction catalyzed by RuCl3 with 2 equiv. of
2,2Ј-bipyridine, only [RuCl2(bpy)2]Cl and related dinuclear
species are observed with electrospray MS,[15] while com-
plexes with only one bpy ligand are not found under these
Isomerization Profile in Time
The difference in isomerization activity of RuIII complex
1 and RuII complex 2 in the absence of isoprene is smaller
than suggested by the overall results given in Table 2. The
results plotted in Figure 4 clearly indicate that the reaction
does not obey clean first-order kinetics. Therefore, initial
TOFs and initial first-order rate constants (kini) have been
calculated for both complexes.[21] The values obtained for 1
(TOF ϭ 295 hϪ1; kini ϭ 0.6 hϪ1) are similar to those found
for 2 (TOF ϭ 260 hϪ1; kini ϭ 0.5 hϪ1). Although initial ac-
tivity is comparable, 2 is deactivated much faster under the
reaction conditions. The influence of isoprene is nicely illus-
trated (Figure 4). In the initial stages of the reaction, com-
plex 2 still catalyzes the formation of MEK, but both initial
TOF (75 hϪ1) and kini (0.1 hϪ1) are significantly lower than
they are in the absence of isoprene. This remarkable finding
may also explain the relatively low conversion reached by
RuII complexes in the absence of isoprene: under the reac-
tion conditions, some reversion (dehydration) occurs [see
(a) in Scheme 2] that yields 1,3-butadiene as evidenced by
GLC analysis of the gas cap. The increasing concentration
of 1,3-butadiene in time causes a decrease of the number of
active catalyst centers until finally catalysis is stopped com-
pletely.
Figure 4. Isomerization of 3-buten-2-ol to MEK catalyzed by com-
plexes 1 (squares) and 2 (triangles) at 130 °C in water/diglyme (1:3
mL) in the absence (filled symbols) and presence (open symbols)
of 10 mmol of isoprene; substrate/catalyst precursor ϭ 530
The initial rate of isomerization in the presence of iso-
prene is also lower with 1 as the catalyst precursor (TOF ϭ
140 hϪ1, kini ϭ 0.3 hϪ1), but deactivation in time does not
occur and a high conversion can be attained in 6 h. Previ-
ous studies with Ru(acac)3 and phen or bpy likewise showed
a decrease in the reaction rate in the presence of 1,3-butadi-
ene, but this system also remained active.[7]
Figure 5. Isomerization of 3-buten-2-ol to MEK with RuCl3 (᭜)
and RuCl3 plus 1 (᭿) and 2 (᭡) equiv. of phen at 100 °C in water/
diglyme (1:3)
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155
2151

E. Bouwman et al.
FULL PAPER
conditions. Given the slow dissociation of chloride ions
Analysis of reaction mixtures containing either catalyst
precursors 1 or 2 with electrospray MS clearly shows the
presence of RuII and RuIII species in both cases. Since both
catalyst precursors exhibit a significantly different activity,
apparently an equilibrium has not been reached during the
early stages of the reaction. The formation of RuII com-
plexes, even to a significant extent in the presence of (oxid-
izing) acid, is unfortunate and should receive further atten-
tion in attempts to maximize catalytic activity. However,
simply increasing the acid concentration to keep the ruth-
enium centers in the 3ϩ oxidation state results in an in-
creased amount of side reactions such as 1,3-butadiene
polymerization.
The direct conversion of 1,3-butadiene was previously
shown to also be catalyzed by an in situ mixture of RuCl3
and one phen ligand.[8] In fact, the performance of this sys-
tem with an initial TOF of 960 hϪ1 and a cumulative TON
of 2700 in 5 h is significantly better than the Ru(acac)3/phen
system.[7] In connection with the present study it is interest-
ing to recall that electrospray MS spectra showed the pres-
ence of species with ruthenium in the oxidation state 3ϩ or
2ϩ.[8] It can now be concluded that the latter species are
inactive under the actual 1,3-butadiene hydration/iso-
merization reaction conditions used.
from RuIII [22]
catalytic activity is now difficult to achieve.
,
Direct Conversion of 1,3-Butadiene to MEK
In the presence of p-toluenesulfonic acid, the preformed
complex 1 is active in the direct conversion of 1,3-butadiene
to MEK (Scheme 2). A TON of 2050 is reached in 6 h with
a selectivity for MEK of over 95% in water/diglyme at 145
°C. Remarkably, in view of the isomerization results above,
2 also catalyzes the formation of MEK from 1,3-butadiene.
The TON of 1300 in 7 h, however, is considerably lower
than the TON reached with 1 as the catalyst precursor. The
activity of 2 may be ascribed to the in situ oxidation of RuII
to RuIII by the acid present in the reaction mixture for the
hydration of 1,3-butadiene. As schematically shown in
Equation (3), oxidative addition of a proton initially leads
to an RuIV hydride species. After protonation of the hydride
yielding dihydrogen, the resulting RuIV species may com-
proportionate with a second RuII complex to give two RuIII
complexes. On the other hand, complex 1 can be reduced,
under these reaction conditions, to the inactive RuII species
according to Equation (4). Coordination of a substrate al-
koxide followed by β-hydrogen elimination gives an RuIII
hydride and an aldehyde. After reductive elimination of 1
equiv. of HCl, the resulting RuI species may compropor-
tionate with a starting RuIII complex to give two RuII spe-
cies.
Mechanistic Rationale of Diene Inhibition
Two general mechanisms have been proposed for iso-
merization of allylic alcohols: the intramolecular π-al-
lylmetal hydride mechanism and the intermolecular metal
hydride addition/elimination mechanism.[2] Neither mech-
anism assigns a specific role to the oxygen moiety during
the catalytic cycle. However, complexes 1 and 2 are inactive
in the isomerization of unsubstituted alkenes such as 1-oc-
tene. So, a third, fundamentally different mechanism that
invokes coordination of the oxygen moiety was proposed
that is shown in Scheme 5.[3,23]
(3)
(4)
Scheme 5. Mechanism of isomerization of 3-buten-2-ol invoking oxygen coordination; indicated with dotted arrows is a possible deactiva-
tion/regeneration route involving a π-allyl species; S ϭ labile ligand or 2eϪ vacant site; NʜN ϭ phen, n ϭ 2, 3
2152
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155

Ruthenium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
The first step is the coordination of 3-buten-2-olate to and various explanations may be offered. Perhaps ruth-
the ruthenium center after deprotonation of the substrate. enium(III) hydrides are less easily formed, as dissociation
This coordination can either be directly didentate as shown of MVK, coordinated through both the oxygen moiety and
in Scheme 5, or occur in two consecutive steps with initial the double bond, from the higher charged RuIII center may
coordination of the double bond.[3] β-Hydrogen abstraction be relatively slower than from RuII. Alternatively, the re-
yields the (enone)ruthenium hydride species 3. After readdi- moval of a π-allyl species by protonation might be faster
tion of the hydride ion to give an (oxaallyl)ruthenium com- with (allyl)RuIII species. We currently favor the former ex-
plex, protonation affords MEK and regenerates the start- planation as more plausible, but molecular modeling aided
ing complex.
by well-designed experiments should further clarify the pe-
Two possibilities of catalyst deactivation by conjugated culiar difference in reactivity between RuIII and RuII com-
dienes can be envisioned. The simplest mode of action is plexes.
preferential coordination; if isoprene coordinates much
stronger to the ruthenium center than 3-buten-2-ol, no iso-
Conclusion
merization can take place. Although at this point we cannot
completely exclude formation of a stable (isoprene)ruthen-
ium complex with either 1 or 2, NMR studies with ruth-
enium(II) complexes in our laboratory indicated a stronger
initial coordination of 3-buten-2-ol.[15]
In conclusion, the novel ruthenium(III) complex 1 has
been synthesized and characterized by X-ray diffraction.
The comparison with its RuII analog 2 revealed an import-
ant difference in catalytic activity in isomerization of allylic
alcohols. Whereas both complexes show high activity in the
isomerization of 3-buten-2-ol to MEK, only 1 remains act-
ive in the presence of the conjugated diene isoprene. Com-
parison with other catalyst precursors demonstrated that
not only an RuIII oxidation state is required for activity in
the presence of isoprene, but also that one phen ligand has
to be present in the complex. This is the first time that the
role of the oxidation state has been pinpointed, implying
that a second major deactivation route is discovered in the
ruthenium-catalyzed direct conversion of 1,3-butadiene to
MEK, next to ligand redistribution. In fact, while 1 gives
high turnovers of MEK (up to 2050), 2 is less active
(TON ϭ 1300). The activity of 2 is most likely caused by
in situ oxidation to 1 by the acid present in the reaction
mixture. The reduction potential of 1 (0.11 V, RuIII/II) and
the oxidation potential of 2 (1.08 V, RuIII/II) indicate that
under the applied reaction conditions a delicate balance be-
tween both RuIII and RuII complexes will be present in solu-
tion, which is confirmed by MS. Further studies are aimed
at complexes with higher redox potentials with ligands that
Intermediate 3 forms a second possible handle for diene
inhibition as dienes are known to form stable π-allyl com-
plexes after reaction with ruthenium hydrides.[24] Through-
out a normal catalytic cycle with extremely fast intramolec-
ular hydrogen transfer no long-lived (‘‘free’’) hydride is pre-
sent. However, if methyl vinyl ketone (MVK) dissociates
from 3, the resulting ruthenium hydride may react with iso-
prene to form a (presumably inactive) (methylallyl)ruth-
enium complex. Whereas pure 1-octene is not isomerized in
the presence of 1 or 2, trans-2-octene (ca. 10 turnovers with
either 1 or 2 as catalyst precursor) can be detected by GLC
during the isomerization of 3-buten-2-ol in the presence of
1-octene, a strong support for the presence of free hydrides.
Hydrides resulting from direct oxidative addition of acids
to the ruthenium atom in 1 or 2 (vide supra) appear to be
short-lived as 1-octene isomerization is not observed upon
addition of HOTs in the absence of allylic alcohols.
The behavior of both catalyst precursors 1 and 2 is sim-
ilar with the GLC detection of approximately 30 turnovers
of MVK with respect to ruthenium. It should be noted that
in both cases a mixture of RuIII and RuII species is pro-
duced and it cannot be determined which of the two is re-
sponsible for MVK formation. The higher selectivity to
MEK with catalyst precursor 1 suggests a lower rate of
MVK dissociation in this case. The amount of MVK
formed is more than stoichiometric and there must there-
fore be a way to regenerate the original catalytically active
species from a ruthenium hydride or a (π-allyl)ruthenium
species. A tentative route for a ruthenium hydride involves
may provide stabilization of RuIII
.
Experimental Section
General Remarks: cis-[RuCl2(dmso)4][12] was prepared according to
a literature procedure. [RuCl3·xH2O] (Aldrich), 3-buten-2-ol (Ald-
rich), isoprene (Acros) and other reagents and solvents were com-
mercially available and used as received, except 1-octene, from
reaction of the hydride with the protic allylic alcohol to which peroxides were removed by flash chromatography on alu-
mina immediately prior to use. Quantitative gas liquid chromato-
graphy (GLC) analyses were carried out with a Chrompack appar-
atus equipped with a CP wax 58 (FFAP) CB column (25 m ϫ
1.2 µm) with toluene as internal standard. Melting points were
measured with a Büchi apparatus and are uncorrected. Matrix As-
sisted Laser Desorption Ionization (MALDI) Time of Flight
(TOF) mass spectra were recorded with a Finnigan MAT Vision
2000 spectrometer. The analytes were mixed with a 2,5-dihydroxy-
benzoic acid (DHB) or a α-cyano-4-hydroxycinnamic acid (α-
generate dihydrogen gas. Alternatively, protonation of an
allylruthenium species will yield alkenes. A GLC gas-cap
analysis of RuCl3/phen-catalyzed isomerization of 3-buten-
2-ol in the presence of isoprene demonstrates the presence
of a mixture of C5 alkenes, which is nicely explained by the
protonation of π-allyl species.
Thus, the reaction of a diene with a free ruthenium hy-
dride may explain the deactivation of the catalyst, but at
this point no conclusion can be drawn as to why RuIII com-
CHCA) matrix. Elemental analyses were performed with
a
plexes are much more resistant to this deactivation route PerkinϪElmer series II 2400 CHNS/O analyzer. IR spectra were
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155 2153

E. Bouwman et al.
FULL PAPER
obtained with a PerkinϪElmer Paragon 1000 FT-IR spectrophoto-
meter equipped with a golden gate ATR device, using the re-
flectance technique (4000Ϫ300 cmϪ1). 1H NMR spectra
(300.1 MHz) were measured with a Bruker 300 DPX. Chemical
shifts (δ) are reported in ppm. Proton chemical shifts are relative
to TMS.
to room temperature, the mixture was analyzed by GLC. Several
time-dependent measurements were performed at 100 °C under ar-
gon by using standard Schlenk techniques (see Figure captions). In
isoprene inhibition experiments the amount of isoprene [typically
1.0 mL (10 mmol)] was added prior to the addition of the substrate.
All reported values for TONs and TOFs are the arithmetic means
of two or more reproducible experiments. Experiments are consid-
ered reproducible when the deviation of the obtained results are
less than 15% of the arithmetic mean.
mer-[RuCl3(dmso)(phen)] (1):
A
bright yellow solution of
[RuCl2(dmso)4] (1.0 g, 2.1 mmol) and phen (0.39 g, 2.2 mmol) in 6
 HCl was heated at 80 °C for 2.5 h, after which time the resulting
orange-red mixture was filtered. The filtrate was concentrated in a
rotary evaporator. Flash chromatography of the red-brown solid
on Al2O3 with CH2Cl2/MeOH (10:1) as eluent gave 1 as a dark red
solid. Yield: 0.71 g, 74%. Alternatively, 1 can be prepared in situ
from cis-[RuCl2(dmso)4] and phen at room temperature in MeOH
with 1 equiv. of 0.1  HCl (yield: 30%). Crystals suitable for X-ray
analysis can be obtained from slow concentration of an MeOH
solution at room temperature. M.p. 182 °C (decomp.). Paramag-
Direct Conversion of 1,3-Butadiene to MEK: In a typical experi-
ment, a high-pressure autoclave was filled with 0.09 mmol of 1 or
2 and 3.5 mmol of p-toluenesulfonic acid (for the acid-catalyzed
hydration of 1,3-butadiene). After addition of 100 mL of diglyme/
water (70/30) solvent mixture, the autoclave was closed and purged
three times with dinitrogen. Next, 1,3-butadiene (10 mL) was added
as a liquid using an ISCO high-pressure pump and the autoclave
was heated to 145 °C. After 10 h, the autoclave was cooled to room
temperature and the contents were analyzed with GLC.
1
netic H NMR (CDCl3): δ ϭ Ϫ17.3 (br. s, 6 H, dmso), Ϫ8.9 (br.
s, 1 H, phen), 2.2 (br. s, 1 H, phen), 8.2 (br. s, 1 H, phen), 8.5 (br.
s, 1 H, phen), 11.5 (br. s, 1 H, phen) ppm. Other phen peaks were
too broad to be observed. MALDI-TOF MS: m/z ϭ 497 [M Ϫ 2 Cl
Ϫ dmso ϩ (α-CHCAϪH)]ϩ. IR: ν˜ ϭ 3055 (νCϪH), 2923 (νCϪH),
1429Ϫ1411 (δCϪH), 1342Ϫ1309 (δCϪH), 1106 (νSϪO), 1094
(νSϪO), 328 (νRuϪCl). C14H14Cl3N2ORuS (465.5): calcd. C 36.1,
H 3.0, N 6.0, S 6.9; found C 36.9, H 2.9, N 6.0, S 7.0.
X-ray Crystallographic Study of 1: Crystals of 1 suitable for X-ray
analysis were obtained by slow concentration of a CHCl3/toluene
solution of
2 at room temperature. Pertinent data for 1:
C14H14Cl3N2ORuS·C7H8, Mr ϭ 557.90, red, block-shaped crystal
¯
(0.2 ϫ 0.3 ϫ 0.3 mm), triclinic, space group P1 with a ϭ
˚
7.7503(10), b ϭ 12.046(2), c ϭ 12.872(3) A, α ϭ 73.972(10), β ϭ
3
˚
80.949(10), γ ϭ 79.691(10)°, V ϭ 1128.9(4) A , Z ϭ 2, Dc ϭ 1.641 g
cmϪ3, µ(Mo-Kα) ϭ 1.157 mmϪ1. 20526 Reflections were measured
(5133 independent, Rint ϭ 0.0442, 1.6° Ͻ θ Ͻ 27.48°, T ϭ 150 K,
cis,cis-[RuCl2(dmso)2(phen)] (2): This complex was prepared ana-
logous to the method described in ref.[16] from cis-[RuCl2(dmso)4]
(0.20 g, 0.41 mmol) and phen (0.07 g, 0.4 mmol). Yield: 0.18 g,
86%; m.p. Ͼ 300 °C (decomp.). 1H NMR (CDCl3): δ ϭ 2.44 (s,
dmso), 3.12 (s, dmso), 3.59 (s, dmso), 3.62 (s, dmso), 7.76 (dd, 1
˚
Mo-Kα radiation, graphite monochromator, λ ϭ 0.71073 A) with
a Nonius Kappa CCD diffractometer with a rotating anode; no
absorption correction was applied. The structure was solved by
automated direct methods.[25] The structure displays relatively high
3
3
H, ArH3 or ArH8, J ϭ 5.1, J ϭ 8.1 Hz), 7.87 (dd, 1 H, ArH3 or
residual density peaks in the area around Ru (peak height up to
ArH8, J ϭ 5.1, J ϭ 8.1 Hz), 7.92 (d, 1 H, ArH5 or ArH6, J ϭ
3
3
3
Ϫ3
˚
˚
2.7 e A at 0.78 A from Ru). The peaks appear to be related to
the heavy atoms positions (Ru, Cl and S) by a noncrystallographic
twofold rotation, approximately parallel to the local twofold rota-
tion axis of the phenanthroline ligand. The rotation images of some
heavy atoms coincide with existing atom sites. Similar peaks were
observed in data sets collected on two other crystals. There are no
signs of twinning. Rough models, describing these peaks as the re-
sult of orientational disorder of the Ru complex, indicated a dis-
order fraction of approximately 3%. In view of the low occupation
factor of the minor component, the disorder model was aban-
doned. Hydrogen atoms were introduced on calculated positions
and included riding on their carrier atoms. Non-hydrogen atoms
were described with anisotropic displacement parameters. The iso-
tropic displacement parameters of the hydrogen atoms were
coupled to the equivalent isotropic displacement parameters of
their carrier atoms. Full-matrix least-squares refinement[26] of 265
parameters on F2 resulted in a final R1 value of 0.0306 [for 4914
reflections with I Ͼ σ(I)], wR2 ϭ 0.0729, GoF ϭ 1.044. The final
9.0 Hz), 7.98 (d, 1 H, ArH5 or ArH6, J ϭ 9.0 Hz), 8.37 (d, 1 H,
3
ArH4 or ArH7, J ϭ 8.1 Hz), 8.47 (d, 1 H, ArH4 or ArH7, 3J ϭ
3
8.1 Hz), 9.98 (d, 1 H, ArH2 or ArH9, J ϭ 5.1 Hz), 10.07 (d, 1 H,
3
ArH2 or ArH9, J ϭ 5.1 Hz) ppm. MALDI-TOF MS: m/z ϭ 435
3
[M Ϫ 2Cl ϩ (DHBϪH)]ϩ. IR (cmϪ1): ν˜ ϭ 3049 (νCϪH), 2924
(νCϪH), 1418 (δCϪH), 1302 (δCϪH), 1082 (νSϪO), 930 (νSϪO),
328 (νRuϪCl). C16H20Cl2N2O2RuS2·0.1toluene (517.3): calcd. C
38.7, H 4.0, N 5.4, S 12.4; found C 38.4, H 3.9, N 5.7, S 11.5.
Cyclic Voltammetry Experiments: The electrochemistry measure-
ments were performed with an Autolab PGstat 10 potentiostat con-
trolled by GPES4 software. A three-electrode system was used, con-
sisting of a platinum (Pt) working electrode, a platinum (Pt) auxili-
ary electrode and an Ag/AgCl reference electrode. The experiments
were carried out in acetonitrile at room temperature under argon
with tetrabutylammonium hexafluorophosphate as electrolyte (0.1
). Under these conditions the ferrocenium/ferrocene couple was
located at ϩ0.436 V with a peak separation of 0.099 V. All poten-
tials are reported relative to Ag/AgCl. Linear voltammograms were
obtained with a rotating (500 rpm) platinum disc as working elec-
residual density was in the range of Ϫ1.09 to 2.66 e AϪ3. Geometric
˚
calculations and molecular graphics were performed with the PLA-
TON package.[27] CCDC-176535 contains the supplementary crys-
tallographic data for this paper. These data can be obtained free
of charge at www.ccdc.cam.ac.uk/conts/retrieving.html or from the
Cambridge Crystallographic Data Centre, 12, Union Road, Cam-
bridge CB2 1EZ, UK [Fax: (internat.) ϩ44-1223/336-033; E-mail:
deposit@ccdc.cam.ac.uk].
trode at a scan rate of 0.005 V sϪ1
.
Isomerization Experiments: Catalytic reactions were performed in
a closed glass vessel under air. The reaction vessel was charged with
ruthenium precursor (0.011 mmol), ligand as appropriate
(0.011 mmol, see text) and substrate (5.8 mmol). In some experi-
ments an AgI salt (silver tosylate and silver triflate gave identical
results) was added in the required stoichiometric amount. After
addition of the solvent mixture water/diglyme (1:3 mL) and internal
standard (toluene, 0.3 mL), the vessel was closed and the mixture
was stirred in a pre-heated oil bath at 130 °C for 6 h. After cooling
Acknowledgments
This research was supported by the Technology Foundation STW,
applied science division of NWO and the technology program of
2154
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155

Ruthenium-Catalyzed Isomerization of Allylic Alcohols
FULL PAPER
[14]
[15]
[16]
[17]
R. C. Van der Drift, M. Van Meurs, W. P. Mul, E. Bouwman,
E. Drent, manuscript in preparation.
R. C. Van der Drift, E. Bouwman, E. Drent, unpublished re-
sults.
H. A. Hudali, J. V. Kingston, H. A. Tayim, Inorg. Chem. 1979,
18, 1391Ϫ1394.
M. B. Cingi, M. Lanfranchi, M. A. Pellinghelli, M. Tegoni,
Eur. J. Inorg. Chem. 2000, 703Ϫ711.
J. J. Rack, H. B. Gray, Inorg. Chem. 1999, 38, 2Ϫ3.
M. Calligaris, O. Carugo, Coord. Chem. Rev. 1996, 153,
83Ϫ154.
D. H. Evans, K. M. OЈConnell, R. A. Petersen, M. J. Kelly, J.
Chem. Educ. 1983, 60, 290Ϫ293.
It is assumed that all reactions are first order in 3-buten-2-ol.
In view of any deactivation, initial k values have been deter-
mined by fitting straight lines through ln(1 Ϫ TON) vs. time
curves over the first 20 min of the reactions rather than fitting
power curves through the complete turnover vs. time plots. Ini-
tial TOF values have been determined by using kini and the
final TON after 6 h.
R. G. Wilkins in Kinetics and Mechanism of Reactions of Trans-
ition Metal Complexes, 2nd ed., VCH, Weinheim, 1991, p. 400.
B. M. Trost, R. J. Kulawiec, J. Am. Chem. Soc. 1993, 115,
2027Ϫ2036.
M. A. Bennett, M. I. Bruce, T. W. Matheson in Comprehensive
Organometallic Chemistry, 1st ed. (Ed.: G. Wilkinson), Perga-
mon Press, Oxford, 1982, vol. 4, p. 691.
G. M. Sheldrick, SHELXS-86 Program for crystal structure de-
termination, University of Göttingen, Germany, 1986.
G. M. Sheldrick SHELXL-97 Program for crystal structure re-
finement, University of Göttingen, Germany, 1997.
A. L. Spek, PLATON A multi-purpose crystallographic tool,
Utrecht University, The Netherlands, 2000; internet: http//
www.cryst.chem.uu.nl/platon
the Ministry of Economic Affairs. This work was supported in part
(A. L. S.) by the Council for Chemical Sciences of the Netherlands
Organization for Scientific Research (CW-NWO). The authors are
indebted to Mr. W. W. Jager (SRTCA, The Netherlands) for his
skilful technical assistance and to Dr. W. J. L. Genuit (SRTCA,
The Netherlands) for recording the mass spectra. Dr. J. G. de Vries
(DSM, The Netherlands) and Mr. W. G. Reman (SRTCA, The
Netherlands) are thanked for stimulating discussions.
[18]
[19]
[20]
[21]
[1]
Ullmann’s Encyclopedia of Industrial Chemistry, 6th electronic
release ed., Wiley-VCH, Weinheim, 2000.
R. C. Van der Drift, E. Bouwman, E. Drent, J. Organomet.
[2]
Chem. 2002, 650, 1Ϫ24.
R. C. Van der Drift, M. Vailati, E. Bouwman, E. Drent, J.
[3]
Mol. Catal. A: Chem. 2000, 159, 163Ϫ177.
J. Kanand, R. Paciello, M. Roper, U.S. Pat. No. 6166265, 2000.
J. Kanand, M. Roper, R. Paciello, A. Thome, U.S. Pat. No.
[4]
[5]
5892125, 1999.
E. Drent, Eur. Pat. No. 457387, 1991.
F. Stunnenberg, F. G. M. Niele, E. Drent, Inorg. Chim. Acta
[22]
[23]
[24]
[6]
[7]
1994, 222, 225Ϫ233.
R. C. Van der Drift, W. P. Mul, E. Bouwman, E. Drent, Chem.
[8]
Commun. 2001, 2746Ϫ2747.
E. A. Seddon, K. R. Seddon, The Chemistry of Ruthenium,
[9]
[25]
[26]
[27]
Elsevier, Amsterdam, 1984.
X. Meddin in Comprehensive Coordination Chemistry (Ed.: G.
[10]
Wilkinson), Pergamon Press, Oxford, 1987, vol. 4, p. 277.
F. P. Dwyer, H. A. Goodwin, E. C. Gyarfas, Aust. J. Chem.
[11]
1963, 16, 42Ϫ50.
I. P. Evans, A. Spencer, G. Wilkinson, J. Chem. Soc., Dalton
[12]
Trans. 1973, 204Ϫ209.
F. P. Pruchnik, E. Galdecka, Z. Galdecki, A. Kowalski, Poly-
[13]
Received January 7, 2002
hedron 1999, 18, 2091Ϫ2097.
[I02004]
Eur. J. Inorg. Chem. 2002, 2147Ϫ2155
2155