Two reactions of allylic alcohols catalysed by ruthenium cyclopentadienyl complexes with didentate phosphine ligands: Isomerisation and ether formation

Article abstract of DOI:10.1016/S1381-1169(00)00221-1

The activity of chloro(η⁵-cyclopentadienyl)(didentate phosphine)ruthenium(II) complexes in the catalysis of two types of reaction with allylic alcohols is described. The isomerisation of 3-buten-2-ol to butanone (MEK) and allyl alcohol to propanal proceeds at high rates. A trend in catalytic activity is observed upon variation of the carbon-chain length in the didentate phosphine ligand: dppm < dppe < dppp < dppb. Complexes with rigid didentate phosphine ligands like cis-dppv and dppph show no activity. The second type of reaction constitutes the first example of a ruthenium-catalysed ether formation directly from allylic alcohols. Homo-coupled ethers like di-allyl ether (DAE) are easily formed as well as ethers from heterocoupling of allyl alcohol with aromatic and aliphatic alcohols. In fact, the ruthenium complexes achieve much higher turnover frequencies and turnover numbers than reported before in palladium-catalysed ether formation. Complexes with dppe and dppp in the presence of a conjugated diene switch from isomerisation to ether formation, but the new complex [RuClCp(o-MeO-dppe)] (Cp = η⁵-cyclopentadienyl) (3) has proven to be very active in ether formation, even in the absence of a diene. The mechanisms of the reactions have been studied by using both deuterium-labelled substrates and ³¹P nuclear magnetic resonance (NMR). (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S1381-1169(00)00221-1

Ž
.
Journal of Molecular Catalysis A: Chemical 159 2000 163–177
www.elsevier.comrlocatermolcata
Two reactions of allylic alcohols catalysed by ruthenium
cyclopentadienyl complexes with didentate phosphine ligands:
isomerisation and ether formation
Robert C. van der Drift, Matteo Vailati, Elisabeth Bouwman), Eite Drent
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden UniÕersity, P.O. Box 9502, Leiden 2300 RA, Netherlands
Received 15 November 1999; accepted 27 March 2000
Abstract
5
Ž
.Ž
.
Ž .
The activity of chloro h -cyclopentadienyl didentate phosphine ruthenium II complexes in the catalysis of two types of
Ž
.
reaction with allylic alcohols is described. The isomerisation of 3-buten-2-ol to butanone MEK and allyl alcohol to
propanal proceeds at high rates. A trend in catalytic activity is observed upon variation of the carbon-chain length in the
didentate phosphine ligand: dppm-dppe-dppp-dppb. Complexes with rigid didentate phosphine ligands like cis-dppv
and dppph show no activity. The second type of reaction constitutes the first example of a ruthenium-catalysed ether
Ž
.
formation directly from allylic alcohols. Homo-coupled ethers like di-allyl ether DAE are easily formed as well as ethers
from heterocoupling of allyl alcohol with aromatic and aliphatic alcohols. In fact, the ruthenium complexes achieve much
higher turnover frequencies and turnover numbers than reported before in palladium-catalysed ether formation. Complexes
with dppe and dppp in the presence of a conjugated diene switch from isomerisation to ether formation, but the new complex
5
w
Ž
.x Ž
. Ž .
RuClCp o-MeO-dppe Cpsh -cyclopentadienyl 3 has proven to be very active in ether formation, even in the absence
of a diene. The mechanisms of the reactions have been studied by using both deuterium-labelled substrates and 31P nuclear
Ž
.
magnetic resonance NMR . q 2000 Elsevier Science B.V. All rights reserved.
Keywords: Ruthenium; Phosphine ligands; Isomerisation; Ether formation; Allylic alcohols
1. Introduction
For example, hydration produces allylic alco-
hols, which may serve as intermediates in ke-
tone and aldehyde synthesis. The isomerisation
of allylic alcohols to aldehydes and ketones has
been investigated extensively and many transi-
tion metal complexes are capable of catalysing
Conjugated dienes form interesting starting
w
x
materials for a whole range of products 1–4 .
Ž
.
AbbreÕiations: dppm, bis diphenylphosphino methane; dppe,
Ž
.
Ž
1,2-bis diphenylphosphino ethane; dppp, 1,3-bis diphenylphos-
w
x
the isomerisation reaction 5–9 . Thus far, only
one system has proven to be able to catalyse the
conversion of butadiene to butanone in a one-pot
.
Ž
.
phino propane; dppb, 1,4-bis diphenylphosphino butane; dppph,
Ž
.
Ž
1,2-bis diphenylphosphino benzene; cis-dppv, cis-1,2-bis diphen-
.
Ž Ž
ylphosphino ethene; o-MeO-dppe, 1,2-bis di ortho-metho-
.
.
Ž
xyphenyl phosphino ethane; dcpe, 1,2-bis dicyclohexylphos-
w
x
synthesis 10 . This catalytic system consists of
.
phino ethane
Ž
.
)
a strong Brønsted acid, a ruthenium III salt and
Corresponding author. Fax: q31-71-5274451.
Ž
.
2,2X-bipyridine or 1,10-phenanthroline. One of
E-mail address: bouwman@chem.leideuniv.nl E. Bouwman .
1381-1169r00r$19.00 q 2000 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S1381-1169 00 00221-1

(
)
164
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
the key features of this system is that it can
selectively catalyse the isomerisation of allylic
alcohols to ketones in the presence of alkenes
and conjugated dienes, which usually are strong
merisation to form carbonyl compounds, or the
formation of allyl ethers with other alcohols.
Both a study with deuterated substrates and a
31
Ž
.
P nuclear magnetic resonance NMR study
were used to get more insight in the reaction
mechanisms of these two interesting reactions.
poisons in isomerisation catalysis.
5
Ž
.
.
Ž
Chloro- h -cyclopentadienyl bis triphenyl-
Ž . Ž .
phosphine ruthenium II
1 was reported by
x
w
Trost and Kulawiec 11,12 to isomerise allylic
alcohols to ketones and aldehydes without iso-
merising a simple isolated alkene moiety in the
same molecule. This selective isomerisation
prompted us to investigate the isomerisation of
2. Experimental
2.1. General
Ž
.
3-buten-2-ol to butanone MEK catalysed by 1
and analogous complexes with didentate phos-
phine ligands in the presence or absence of a
conjugated diene. Although most complexes
showed a very high activity towards isomerisa-
tion, in the presence of a conjugated diene, no
isomerisation was observed at all. However,
some complexes showed an unexpected, but
highly interesting activity. The presence of a
diene worked as a switch: isomerisation was
switched off and ethers were formed instead.
Allylation of nucleophiles is a versatile reac-
tion in organic chemistry to form new carbon–
Substrate alcohols, silver para-toluene-
Ž
.
sulfonate AgOTs , isoprene, phosphines and
other reagents were commercially available and
used without further purification. Solvents were
reagent grade and used as received except for
THF, which was distilled over CaH2 prior
Ž
Ž
.
to use. 1,2-Bis di ortho-methoxyphenyl phos-
.
Ž
.
phino ethane o-MeO-dppe was kindly sup-
plied by Shell International Chemicals
Ž .
w
x
Ž
Ž
Both 1 25 and chloro- 1,2-bis diphenyl-
5
.
.Ž
.
Ž .
phosphino -ethane h -indenyl -ruthenium II
26 were prepared according to literature proce-
dures. The other chloro- h -cyclopentadienyl -
didentate phosphine -ruthenium II complexes
w
x
5
Ž
.
w
x
Ž
.
Ž .
carbon or carbon–heteroatom bonds 13–16 ,
e.g. in the protection of alcohols. Allylic alco-
hols are usually much cheaper than allylic ac-
etates and allylic halides and in fact most other
were prepared from 1 analogous to a published
w
x
procedure for a substituted dppe complex 27 ,
except for the dcpe complex, which was pre-
w
x
w x
allyl sources are synthesised from them 17 .
Efficient ether formation with allyl alcohol could
lead to a halogen-free route to epoxy resins.
But, despite increasing interest in the use of
pared by a modified procedure 28 . 3-Buten-2-
ol-2-d was prepared by reduction of methyl
vinyl ketone with lithium aluminium deuteride
w
x
29 .
All syntheses and catalytic reactions were
w
x
allylic alcohols as an ‘allyl feed stock’ 17–22 ,
catalysed ether formation from allyl alcohol
sources remains largely unsuccessful. By far,
most reports use a palladium salt as catalyst and
have only moderate to low turnover numbers
performed in an argon atmosphere using stan-
dard Schlenk techniques. Reactions at elevated
temperatures and pressures were performed in a
100 ml Hastelloy C22 Premex HPM-T high-
pressure autoclave.
Ž
.
varying from 20 up to 300 . Ruthenium has, to
the best of our knowledge, only once success-
Quantitative gas liquid chromatography anal-
fully been employed to form allylic ethers with
yses were carried out on a Chrompack apparatus
Ž
equipped with a CP Sil 5-CB column 50 m=
w
x
glycols from homo-allylic alcohols 23,24 .
In this report, it is shown that several ruthe-
nium cyclopentadienyl complexes with diden-
tate phosphine ligands have the possibility of
catalysing two reactions of allylic alcohols: iso-
.
0.4 mm with toluene or anisole as internal
Ž
.
standard. Melting points m.p. were measured
1
on a Buchi apparatus and are uncorrected. H
¨
13
Ž
.
Ž
NMR spectra 199.5 MHz and C NMR 49.88

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
165
.
MHz were performed on a Jeol JNM 200 FT-
mixture was stirred at 1808C for 24 h and then
cooled to room temperature. The resulting white
solid was collected by filtration and washed
with pentane. A second crop of product was
obtained by reducing the volume of the filtrate
to about one third. The combined products were
Ž .
NMR spectrometer. Proton chemical shifts d
are reported in parts per million relative to
Me4Si ds0 . Carbon chemical shifts d are
reported in parts per million referenced to inter-
nal solvent carbon atoms. P H NMR spectra
121.5 MHz were recorded on a Bruker DPX
300 spectrometer. Chemical shifts d are re-
ported in parts per million relative to external
85% aqueous H3PO4 ds0 . H NMR spectra
46.07 MHz were recorded on a Bruker DPX
300 spectrometer. Chemical shifts d are re-
ported in parts per million referenced to internal
residual solvent deuterons. Mass spectra were
recorded on a Finnigan MAT TSQ-70 equipped
Ž
.
Ž .
31
1
Ä
4
Ž
.
Ž
dried at reduced pressure to yield 8.1 g 65%
Ž .
. Ž .
Žw
based on anthracene . Mass ESI : 249 M–
1
x . Ž w x .
H – . m.p. 1838C lit. 30 1878C . H NMR
2
Ž
.
Ž
. Ž . Ž
CDCl3 11.0 br s, 1H, COOH , 7.32–7.22 m,
Ž
.
. Ž . Ž
4H, ArH , 7.14–7.06 m, 4H, ArH , 4.66 d,
Ž .
. Ž .
1H, ArCHCHCOOH , 4.33 t, 1H, ArCHCH2 ,
Ž
.
Ž
2.92 m, 1H, CH2CHCOOH , 2.05 m, 2H,
.
CHCH2CHCOOH .
Ž
.
(
with a custom made electrospray interface ESI .
2.3.2. 9,10-Dihydroxy-9,10-ethano-11- hydroxy-
)
methyl-d2 anthracene
5
(
)(
2.2. Synthesis of chloro h -cyclopentadienyl o-
To a stirred suspension of lithium aluminium
)
( ) ( )
Ž
.
Ž
.
MeO-dppe ruthenium II
3
deuteride 0.51 g, 12.1 mmol in THF 60 ml ,
9,10-dihydro-9,10-ethano-anthracene-11-
Ž
.
In a two-neck flask equipped with a reflux
carboxylic acid 3.0 g, 12.0 mmol was slowly
added as a solid. After complete addition, the
mixture was heated to reflux for 24 h. The
resulting mixture was evaporated to dryness and
5
Ž
.
condenser, chloro- h -cyclopentadienyl -bis-
Ž
.
Ž . Ž
triphenylphosphine -ruthenium II 0.18 g, 0.25
mmol and o-MeO-dppe 0.15 g, 0.29 mmol
.
Ž
.
Ž
.
Ž .
were dissolved in toluene 20 ml . After 8 h
heating to reflux, the resulting orange-brown
solution was concentrated to about 1 ml on a
rotary evaporator. Precipitation with hexane af-
then taken up in diethyl ether 20 ml . The ether
Ž
.
solution was quenched with water 5 ml and
then separated. The ether layer was dried with
MgSO4, followed by evaporation on a rotary
Ž
.
Ž
.
forded 0.14 g of a yellow powder 78% . Mass
evaporator. A white solid 2.5 g, 88% resulted.
Residual proton content at the methanol carbon
Ž
Ž
.
.
Ž
w
x
.
ESI : mrz 685 M–Cl q . m.p. 200–2028C
dec. . P H NMR acetone-d6 70.3 ppm.
31
1
1
Ä
4
Ž
.
was less than 2% as measured by H NMR.
m.p. 1088C lit. 30 1108C . H NMR CD-
1
Ž
w
x
.
Ž
. Ž . Ž
2.3. Synthesis of allyl-1,1-d2 alcohol
Cl3 7.31–7.22 m, 4H, ArH , 7.12–7.07 m,
.
Ž
4H, ArH , 4.43 d, 1H, J s 2.6 Hz,
. Ž
ArCHCHCD2OH , 4.28 t, 1H, Js2.6 Hz,
Allyl-1,1-d2 alcohol was prepared in a three-
step synthesis from acrylic acid analogous to
published procedures that start with ethyl acryl-
ate 5,30,31 . Protection of the double bond by a
Diels–Alder reaction with anthracene was nec-
essary to prevent extensive polymerisation.
.
Ž
.
ArCHCH2 , 2.14 br m, 1H, CHCHCD2OH ,
Ž
1.92 dt, 1H, Js2.6 Hz, Js12.0 Hz, CH-
HCHCD2OH , 1.10 br s, 1H, OH , 1.08 dm,
w
x
.
Ž
.
.
Ž
1H, CHHCHCD2OH .
2.3.3. Allyl-1,1-d2 alcohol
2.3.1. 9,10-Dihydro-9,10-ethano-anthracene-
11-carboxylic acid
A round-bottom one-neck flask was con-
nected to a round-bottom two-neck flask by a
U-shaped tube. The two-neck flask was con-
nected to the Schlenk line. 9,10-Dihydroxy-
An autoclave was charged with anthracene
Ž
.
Ž
8.94 g, 50.2 mmol and acrylic acid 5.4 g,
.
Ž
.
Ž
.
74.9 mmol dissolved in p-xylene 50 ml . The
9,10-ethano-11- hydroxymethyl-d2 anthracene

(
)
166
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
Ž
.
Ž
.
2.5 g, 10.5 mmol was heated in the one-neck
other alcohols, the other alcohol 5 mmol was
added before heating. In cases where isoprene
flask to 300–4008C while the two-neck flask
was simultaneously cooled in liquid nitrogen.
When no more product distilled over, the two-
neck flask was warmed to room temperature
Ž
.
was used, 2 ml 20 mmol was added to the
reaction mixture prior to heating.
and the product was obtained as a colourless
1
Ž
.
Ž
.
Ž
liquid 0.5 g, 79% . H NMR CDCl3 5.93 dd,
3. Results and discussion
.
1H, Js10.3 Hz, Js17.7 Hz, CH2 5C H ,
Ž
5.21 dd, 1H, J s 1.7 Hz, J s 17.7 Hz,
3.1. Isomerisation of 3-buten-2-ol
.
Ž
C H H5CH , 5.08 dd, 1H, Js1.7 Hz, Js10.3
2
.
Ž .
Ž
.
Hz, CH H5CH , 3.12 s, 1H, OH . H NMR
13
Ž
.
Ž
.
Ž
In view of our interest in the formation of
CDCl3 4.0 s . C NMR CDCl3 137.1 s,
Ž
.
.
Ž
.
Ž
butanone methyl ethyl ketone, MEK from bu-
CH2 5CH , 115.0 s, CH2 5CH , 62.8 quin,
Ž
. w x
.
tadiene vide supra 10 , we started our investi-
gations with the isomerisation of 3-buten-2-ol
CD2 .
Ž
.
w
x
Eq. 1 . Trost and Kulawiec 11,12 showed that
2.4. General procedure for isomerisation of al-
lylic alcohols
1 selectively isomerises allylic alcohols in the
presence of unfunctionalised alkene moieties.
The reported yields are high, but a catalyst
concentration of 1 mol% is used. However, with
Ž
In a typical experiment, 3-buten-2-ol 46
.
Ž
.
.
mmol , a catalyst complex 0.008 mmol ,
AgOTs 0.010 mmol and anisole 3.7 mmol
Ž
.
Ž
3-buten-2-ol, 1 is capable of reaching a turnover
frequency TOF of over 200,000 h 1, which is
higher than reported before by three orders of
magnitude. This dramatic increase in rate may
be attributed to three factors. First of all, 3-
buten-2-ol is much smaller than the substrates
used by Trost. Secondly, the use of silver tosy-
y
Ž
.
were brought into a two-neck round-bottom flask
equipped with a reflux condenser and a septum.
Next, the flask was lowered into a pre-heated
oil bath of 1008C. After 15 min, a sample was
taken with an air-tight syringe, and the sample
1
was analysed by H NMR andror gas–liquid
Ž
.
Ž
.
late this manuscript instead of Et3NHPF6
chromatography GLC . In some cases, several
Ž
.
Ž
Trost is more effective in the removal of the
samples were taken at certain time intervals see
text . In cases where isoprene was used, 2 ml
20 mmol was added to the reaction mixture
.
chloride ion, which in our case leads quantita-
tively to the formation of the cationic ruthenium
species. Finally, the use of a potentially co-
ordinating solvent like dioxane may possibly
retard the reaction. In our case, no additional
solvent is used.
Ž
.
prior to heating.
2.5. General procedure for ether formation
Analogous to the isomerisation experiments,
a two-neck flask was charged with allyl alcohol
Ž .
1
Ž
.
Ž
.
Ž
45 mmol , toluene 4 mmol and catalyst 0.005
.
mmol . In all cases, a slight excess of AgOTs
was added to remove the chloride ion from the
ruthenium complex. The reaction mixture was
lowered into a pre-heated oil bath at the speci-
To modify the reactivity and perhaps gain
selectivity, a series of ruthenium cyclopentadi-
enyl complexes with didentate phosphine lig-
ands has been prepared. The results of the iso-
merisation reaction of 3-buten-2-ol to MEK
catalysed with these complexes are reported in
Ž
.
fied temperature usually 1008C and samples
Ž
.
were taken at certain time intervals see text .
All samples were analysed with GLC. In cases
Ž
.
of coupling of allyl alcohol 12–15 mmol with

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
167
Table 1
ethene bridge, show no catalytic activity at all
w
x
Isomerisation of 3-buten-2-ol to MEK catalysed by RuClCpL
Ž
.
entries 8 and 9 . The complex with the ligand
complexesa
dppm is active in the early stages of the reac-
tion, but it almost completely loses activity
y1
Ž
TOF h
.
Entry
L
Ž .b
1
2
3
4
5
6
7
8
9
2PPh3
dppmc
dppm
dppe 2
Ž .
dppp 5
dppb 6
dcpe
dppph
cis-dppv
1
)200,000
2000
530
Ž
.
within 15 min cf. entries 2 and 3 . This may be
due to the fact that the ruthenium dppm com-
plex has a considerable ring strain energy in the
Ž .
1675
5500
18,000
20,000
0
w
x
order of 13 kcalrmol 32 and the ligand may
Ž .
therefore dissociate completely.
0
a
3.2. Formation of di-allylic ethers from 3-buten-
2-ol and allyl alcohol
Ž
Reactions were carried out at 1008C in neat 3-buten-2-ol 45
.
mmol with anisole as IS and 0.008 mmol of catalyst. In all cases,
a slight excess of AgOTs was added to remove the chloride ion.
Activities were determined after 158 min by 1H NMR, shown as
moles of MEK formed per mole catalyst per hour. No products
other than MEK were observed.
In the following experiments, the isomerisa-
tion of 3-buten-2-ol in the presence of isoprene
was examined to mimic the reaction conditions
bActivity after 2 min.
cActivity after 5 min.
w
x
for the target reaction 10 . Unfortunately, most
of the complexes showed no catalytic activity
under these conditions. Although no isomerisa-
tion was observed, complex 2 showed an inter-
Table 1. In all cases, only the selective forma-
Ž
.
tion of MEK )95% was observed. All com-
plexes with didentate phosphine ligands are
clearly less active than 1. Upon increasing the
carbon-chain length between the phosphorus-
donor atoms, the reactivity of the ruthenium
complexes increases. The mechanism of iso-
merisation of allylic alcohols by 1, as proposed
Ž
.
esting reaction Eq. 2 , which we explored in
further detail. The formation of an ether from
allylic alcohols under such relative mild condi-
w
x
tions is rare, but has been reported 17–22 . All
reported catalysts, however, are less active than
1
w
x
by Trost and Kulawiec 11 , involves the disso-
ciation of one triphenylphosphine ligand. The
increased reactivity with increasing chain length
in the phosphine ligand reflects the greater ease
of ring opening of larger chelate rings. Indeed,
complexes with didentate phosphine ligands that
are extremely rigid, such as with a phenylene or
complex 2, which reaches a TOF of 1050 hy
Ž
.
Table 2 . A total of 3300 turnovers can be
reached after a few hours, which is, to our
knowledge, the highest activity reported so far.
Previous results with ruthenium bipyridine-
catalysed isomerisation of 3-buten-2-ol 10
showed that conjugated dienes have an in-
w
x
Table 2
Formation of BMAE and MEK from 3-buten-2-ola
y1
b
y1
Ž
.
Ž
.
MEK
Entry
Complex
TOF h
BMAE
TOF h
w
Ž
.x Ž .
1
2
3
4
5
6
RuClCp dppe
2
.
0
1050
200
0
150
110
1675
25
340
0
400
0
Ž
2qisoprene 2 ml
w
Ž
.x Ž .
3
RuClCp o-MeO-dppe
Ž
.
3qisoprene 2 ml
w
Ž
.Ž
.x Ž .c
RuCl Ind dppe
4
Ž
.
4qisoprene 2 ml
a
Ž
.
Reactions were carried out at 1008C in neat 3-buten-2-ol 45 mmol with toluene as IS and 0.005 mmol of catalyst. In all cases, a slight
excess of AgOTs was added to remove the chloride ion. Activities were determined after 2 h by GLC, shown as moles of product formed
per mole catalyst per hour.
b BMAE is formed in a 1:1 mixture of DL-BMAE and rac-BMAE as observed by GLC analysis.
c Ind: h5-indenyl.

(
)
168
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
hibitory effect on the MEK formation rate. In
the present study with phosphine ligands, iso-
prene completely stops isomerisation and ether
formation is started instead. This makes iso-
prene an ideal ‘on–off switch’ between both
reactions. The effect is best observed with 2.
BMAE is now already observed in the absence
of isoprene, but the steric effect is demonstrated
by the fact that the overall TOF is lower than
that observed with complex 2.
Complex 3 turned out to be a more active
catalyst than complex 2 when we turned our
attention to allyl alcohol itself. The results are
reported in Table 3. Although the TOF is some-
what lower than with the ether formation of
3-buten-2-ol, selectivity is virtually complete.
Ž
Without isoprene, only MEK is formed entry
.
1 . When isoprene is added, the same catalyst
Ž
.
Ž
.
switches to bis methylallyl ether BMAE for-
Ž
.
mation with only traces of MEK entry 2 .
Ž
.
Complex 3 produces di-allyl ether DAE with a
TOF of 615 hy1. Addition of isoprene leads in
all cases only to a slower reaction. As with the
isomerisation of 3-buten-2-ol, a trend is visible
with carbon-chain length between the phospho-
rous donor atoms. This time, however, increas-
ing the chain length of the diphosphine bridge
decreases activity. Apparently, in the mecha-
nism for ether formation, ring opening is not
required in or before the rate-determining step.
Ring opening seems to be necessary in some
stage of the reaction, because the complex with
Ž .
2
For the isomerisation of allylic alcohols, Trost
w
x
and Kulawiec 11 reported an increase in cat-
alytic activity when the cyclopentadienyl ligand
is replaced by the indenyl ligand. In our case
Ž
.
entry 5, Table 2 , complex 4 gives a lower
Ž
activity for isomerisation and clearly gives lower
selectivity than 2. The formation of BMAE is
only slightly slower in the presence of isoprene
with 4, but again, the formation of MEK has
cis-dppv does not show ether formation nor
.
isomerisation, Table 1, entry 9 . If the ring is
opened too easily, isomerisation becomes pre-
dominant again as can be seen with complex 6.
The homocoupling of alcohols is limited to
allylic alcohols with a terminal double bond.
Both an allylic alcohol with an internal double
bond, 2-buten-1-ol and a homo-allylic alcohol,
3-buten-1-ol, remain unchanged even after pro-
longed heating.
Ž
.
completely stopped cf. entries 5 and 6 .
To probe the steric requirements of the reac-
tion, o-MeO-dppe, a more sterically demanding
ligand successfully employed in our laboratory
w
x
in hydrogenation reactions 33 , was tested us-
ing complex 3. The formation of the ether
Table 3
Formation of DAE and propanal from allyl alcohola
y1
y1
Ž
.
Ž
.
propanal
Entry
Complex
TOF h
DAE
TOF h
w
Ž
.x Ž .
1
2
3
4
5
6
7
8
RuClCp dppe
2
.
150
0
94
trace
615
225
0
0
0
0
815
0
0
Ž
2qisoprene 2 ml
w
w
w
Ž
Ž
Ž
.x Ž .
RuClCp dppp
RuClCp dppb
RuClCp o-MeO-dppe
5
.x Ž .
6
.x Ž .
3
Ž
.
3qisoprene 2 ml
w
Ž
.Ž .x Ž .
RuCl Ind dppe
4
150
0
Ž
.
4qisoprene 2 ml
0
a
Ž
.
Reactions were carried out at 1008C in neat allyl alcohol 45 mmol with toluene as IS and 0.005 mmol of catalyst. In all cases, a slight
excess of AgOTs was added to remove the chloride ion. Activities were determined after 2 h by GLC, shown as moles of product formed
per mole catalyst per hour.

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
169
3.3. Coupling of allyl alcohol with other alco-
hols
than the formation of the cross-coupling product
AMPE. Apparently, the amount of DAE de-
creases with time, but this is most likely due to
losses during sampling. No reaction is observed
when DAE is heated with p-cresol in the pres-
ence of 3. Reaction of p-cresol does not start
until around 40 min. In the period until 1 h,
AMPE is selectively formed. Heating for a
longer period does not increase the yield of
AMPE, but increases the amount of ring ally-
lated products. A direct ring attack to form
o-allyl cresol and other ring allylated products
cannot be excluded. However, since both toluene
and anisole fail to react with allyl alcohol under
the reaction conditions, this seems less likely.
The best catalyst for coupling of allyl alcohol
to DAE proved to be 3. Therefore, 3 was used
to extend this reaction to other nucleophiles.
Thus, when p-cresol is added to the reaction
mixture, not only DAE is formed, but a consid-
Ž
.
erable amount of allyl p-methylphenyl ether
Ž
.
AMPE is also formed. Under the reaction
conditions, at 1008C, AMPE partially undergoes
a Claisen rearrangement to form a mixture of
Ž
.
several ring allylated cresols and ethers Eq. 3 .
In Fig. 1, it is shown that formation of DAE in
the initial stages of the reaction is much faster
Ž .
3
Other aromatic and aliphatic alcohols can be
tion conditions required for heterocoupling are
more severe than for homocoupling. Whereas
Ž
.
coupled with allyl alcohol Table 4 . The reac-

(
)
170
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
Aliphatic alcohols can be coupled to allyl
Ž
alcohol as well, although much slower entries
.
13 and 14 . If the reaction is carried out with
methanol as a solvent, a complete conversion to
methyl allyl ether takes place, with no forma-
Ž
tion of DAE. Coupling with 3-buten-1-ol entry
.
11 fails because it co-ordinates strongly to the
ruthenium centre and thereby inhibits further
reaction. In this case also, no formation of DAE
is observed.
Fig. 1. Coupling of allyl alcohol with p-cresol catalysed by 3.
Ž
Ž
.
Ž
.
p-Cresol 3.5 mmol was reacted with allyl alcohol 15 mmol at
.
Ž
.
1008C with toluene as IS and 0.014 mmol of 3 .
'
Allyl
Ž
.
Ž
.
Ž
. Ž . Ž
.
.
p-methylphenyl ether AMPE .
B p-Cresol. ` DAE. X
3.4. Mechanistic considerations
Ž
. Ž . Ž
.
Ž
o-Allyl p-cresol . l o-Allyl p-methylphenyl allyl ether. q
2,6-Di-allyl cresol.
From the results mentioned above, it is clear
that ruthenium cyclopentadienyl complexes with
didentate phosphine ligands are, in principle,
capable of catalysing both isomerisation and
ether formation of allylic alcohols. Which reac-
tion is catalysed predominantly depends both on
the phosphine ligand and on the substrate. Trost
the reaction proceeds smoothly at 1008C, no
Ž
.
reaction is observed at 808C or less entry 5 .
DAE is formed even at 508C, albeit at a much
lower rate. Aromatic alcohols are good nucle-
ophiles in this reaction and substituents are
tolerated. Both p-methoxy phenol and 2,4,6-tri-
1
methyl phenol react to give a TOF of 700 hy
and Kulawiec 11 have proposed an intramolec-
ular mechanism for the isomerisation of allylic
alcohols by 1. We performed a cross-over ex-
periment to prove the proposed intramolecular-
ity. Deuterated 3-buten-2-ol was prepared by
w
x
1
to the mono allyl aryl ether and 55 hy
,
Ž
.
respectively. The latter only gives one product,
since Claisen rearrangements are not possible in
this case.
Table 4
w
Ž
.x Ž .a
Coupling of allyl alcohol with nucleophiles catalysed by RuClCp o-MeO-dppe
3
y1
Ž
.
Entry
Nucleophile
Product
TOF h
b
1
2
3
4
–
DAE
19
0
0
144
0
toluene
anisole
allyl toluene
allyl anisole
allyl phenyl ether
allyl phenyl ether
phenol
5
phenolc
Ž
.
6
7
8
9
10
11
12
13
p-cresol
allyl p-methylphenyl ether
254
700
41
875
0
0
100
63
Ž .
p-MeO-phenol
2,4,6-trimethylphenol
2-naphthol
3-buten-1-ol
MeOH
allyl p-methoxyphenyl ether
Ž
.
allyl 2,4,6-trimethylphenyl ether
allyl 2-napthyl ether
allyl 3-butenyl ether
allyl methyl ether
Ž
.
MeOH 5 ml
allyl methyl ether
allyl ethyl ether
EtOH
aReactions were carried out at 1008C with 12–15 mmol allyl alcohol and 5 mmol nucleophile. Toluene was used as IS and 0.005 mmol
of catalyst was applied. In all cases, a slight excess of AgOTs was added to remove the chloride ion. Activities were determined after 45
min by GLC, shown as moles of product formed per mole catalyst per hour. In all cases, DAE was formed in considerable amounts as well.
Ž
The selectivity towards the cross-coupled ether is time-dependent. For an example, see Fig. 1. Selectivities for other nucleophiles are
.
comparable.
bTs508C.
cTs808C.

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
171
reduction of methyl vinyl ketone with lithium
expected to be formed as well, which originate
w
x
Ž
.
aluminium deuteride 29 . This was stirred to-
gether with allyl alcohol at room temperature in
the presence of 1. If the hydrogen shift during
the isomerisation is intramolecular, two prod-
ucts are expected to be formed: butanone-4-d
from the initial formation of a ruthenium IV
hydride. To avoid difficulty in observing alde-
hyde molecular ion peaks by mass spectrome-
try, the reaction mixture was analysed by NMR
techniques. 13C NMR proved to be best suitable,
1
Ž
.
and propanal Eq. 4 . An intermolecular hydro-
gen shift would give rise to four products. Next
to the already mentioned butanone-4-d and
propanal, also butanone and propanal-3-d would
be formed. In this case, some unsaturated car-
bonyl compounds, e.g. methyl vinyl ketone, are
because in H NMR, all peaks overlapped. Dur-
ing the reaction, two product peaks appeared,
Ž
.
assigned to butanone-C4 7.24 ppm and
Ž
.
propanal-C3 5.69 ppm after independent syn-
thesis. Only the peak at 7.24 ppm was a 1:1:1
triplet indicative of monodeuteration at that car-
Ž .
Ž
Scheme 1. Proposed mechanism for the isomerisation of allylic alcohols with ruthenium II Cp didentate phosphine complexes. Cp:
5
.
h -cyclopentadienyl .

(
)
172
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
bon atom. No deuteration was observed at any
other carbon atom. This proves the proposed
intramolecularity of the isomerisation of allylic
alcohols catalysed by 1.
31
w
Ž
.xq
Fig. 2. P NMR monitoring of RuCp dppe
in reaction with several alkene moieties. All spectra were taken at room temperature in
w
Ž
.xq
Ž . w
Ž
.x Ž . w
Ž
.xq
CDCl3 directly after mixing of the RuCp dppe
solution with the substrate. a RuClCp dppe . b RuCp dppe
after reaction with
.xq Ž .
Ž . w
.xq
Ž
w
Ž
allyl alcohol.
.xq
c
RuCp dppe
.xq
after reaction with 2-buten-1-ol. The peak at 77.5 ppm is from uncomplexed RuCp dppe .
after reaction with 1-octene. The peak at 77.5 ppm is from uncomplexed RuCp dppe
. d
w
Ž
w
Ž
RuCp dppe

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
173
Ž .
4
In analogy to the mechanism for catalysis by
Scheme 1. Removal of the chloride ion from
w
x
w
Ž
.
x
Ž .
1 11 , we propose a mechanism as depicted in
RuCpCl dppe by a silver I salt gives a vacant
5
Ž .
Ž
Scheme 2. Proposed mechanism for formation of DAE catalysed by ruthenium II Cp didentate phosphine complexes Cp: h -cyclopenta-
.
dienyl .

(
)
174
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
site on the ruthenium centre. The behaviour of
w
Ž
.
xq
the resulting RuCp dppe towards allyl alco-
31
Ž
.
hol was examined with P NMR Fig. 2 . The
appearance of two doublets at approximately the
Ž
.
same d-value Fig. 2b suggests an asymmetric
co-ordination of allyl alcohol, with both phos-
phorous atoms co-ordinated to the ruthenium.
Both allylic and homo-allylic alcohols give the
same pattern in the 31P NMR spectrum around
Ž
.
the same d-value, as does 1-octene Fig. 2c .
Allylic alcohols with an internal double bond
w
Ž
.
xq
react much slower. Reaction of RuCp dppe
with 2-buten-1-ol leaves most of the cations
Ž
.
uncomplexed Fig. 2d . These results show that
initial co-ordination of the allylic alcohol to the
ruthenium centre with its alkene moiety is most
likely.
Ž
The results in the isomerisation with diden-
tate phosphine ligands, as shown in Table 1,
further suggest that the chelate ring has to open
up in order to create a vacant site at the ruthe-
nium centre. If the ring cannot be opened easily,
as is the case with cis-dppv, no catalysis takes
place. The ring might be opened under the
influence of co-ordination of the oxygen moi-
allyl species. The second pathway shown in
.
Scheme 2 consists of exchange of the alcohol
with the hydroxide resulting in a ruthenium
alkoxide accompanied by release of a water
molecule. Reductive elimination then yields the
product ether. At present, no discrimination can
be made between both pathways. Some support
for the direct attack pathway has recently been
Ž
w
x
w x
ety. Contrary to a previous report 11 , unfunc-
tionalised alkenes are also isomerised by 1 and
complexes with didentate phosphine ligands, al-
though at a much lower rate. Still, chelating
co-ordination of both the oxygen and the alkene
moiety can explain the selectivity towards al-
lylic alcohols when both allylic alcohols and
published 34 . Oxidative addition of allyl bro-
w
Ž
.Ž
.
2 xq
mide on RuCp PPh3 MeCN
followed by
stoichiometric addition of triethylamine resulted
in RuBrCp PPh3 CH2 5CHCH2 NEt3 q: the
product resulting from direct attack on the allyl
moiety. In the reaction of 3-buten-2-ol catalysed
w
Ž
.Ž
.x
Ž
by 2, only the branched ether is formed in a 1:1
mixture of DL-BMAE and meso-BMAE . This
.
.
unfunctionalised alkenes are present. Co-
ordination of an allylic alcohol through both the
oxygen moiety and the alkene moiety is then
followed by hydrogen abstraction from C1. The
can be explained by assuming that rotation of
Ž
the methyl allyl moiety is not possible or rota-
.
tion is slower than reductive elimination . Di-
Scheme 3. Formation of two possible regio isomers in the ether formation of allyl-1,1-d2 alcohol.

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
175
rect attack could be sterically andror electroni-
cally directed towards the ipso position.
1H NMR and two doublets at 5.4 ppm in the 2 H
Ž
.
NMR Fig. 3 . The ratio between both products
is dependent on the nature of the catalyst. The
difference in ratio can be explained by both
To probe the nature of the reaction with allyl
alcohol, allyl-1,1-d2 alcohol was prepared. As
shown in Scheme 3, exclusive attack on the
ipso position would only lead to deuteration on
aliphatic carbons, whereas attack on both sides
would lead to a mixture of deuterated aliphatic
and vinylic carbons. Indeed, attack on both
sides takes place, which can be deduced from
the appearance of a doublet at 3.95 ppm in the
w
x
steric and electronic reasons 35 , although in
this case, steric reasons will probably dominate.
Again, two rationales can be given: via the
direct attack route or the reductive elimination
route. The sterically more crowded 3 gives pre-
Ž
dominant attack on the C3 position aliphatic to
.
vinylics1:3 , which is probably the most eas-
Fig. 3. Formation of DAE from allyl-1,1-d2-alcohol catalysed by 3. a 1H NMR spectrum in CDCl3. Note the appearance of a doublet at
Ž .
3.95 ppm due to attack at C3. b 2 H NMR spectrum in CDCl3. Note the appearance of two doublets at 5.4 ppm due to attack at C3.
Ž .

(
)
176
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
ily accessible for the attacking nucleophile. In
the less bulky catalysts 2 and 5, both sides
become more and more accessible and the ratio
drops to 1:2 and 1:1, respectively. The latter
ratio is expected when a symmetrical h3-allyl
ligand is well accessible from both sides. In
case the pathway, as shown in Scheme 2, is
operative, the co-ordinated allyl must be able to
rotate freely. Apparently, rotation is more re-
stricted in complex 3 than in 2 and 5; in the
latter complex apparently, the allyl moiety can
rotate freely. These results exclude a mecha-
Ether formation of allylic alcohols to di-al-
lylic ethers is much faster with this type of
Ž .
ruthenium II complex than reported before with
palladium and nickel catalysts. The new com-
w
Ž
.x
plex RuClCp o-MeO-dppe forms ethers even
in the absence of isoprene and is the most active
of the tested complexes in the coupling of allyl
alcohol to aromatic and aliphatic alcohols.
The mechanism of isomerisation is intra-
molecular and starts with the co-ordination of
allyl alcohol to the ruthenium centre through its
alkene moiety. It is proposed that the mecha-
nism of ether formation involves the intermedi-
acy of a ruthenium allyl species. Attack of a
nucleophile on that allyl species is probably
sterically directed, as ipso attack predominates
with o-MeO-dppe as a ligand, whereas a 1:1
mixture of products resulting from both ipso
attack and C3 attack is observed with the steri-
cally less hindered dppp ligand. If the allyl
group is substituted on C1 as is the case in the
reaction of 3-buten-2-ol, attack is sterically
andror electronically directed towards the ipso
position.
nism in which direct SN attack on the alcohol
2
takes place with the ruthenium complexes
merely acting as a Lewis acid.
The role of isoprene is not yet fully under-
stood. Co-ordination of isoprene to the ruthe-
nium centre may prevent the didentate co-
Ž
ordination of the allylic alcohol by both the
.
alkene and oxygen moieties; see Scheme 1 , and
thereby inhibit the isomerisation reaction. How-
ever, oxidative addition may still be possible
from the monodentate co-ordination with the
Ž
.
alkene moiety of the allylic alcohol Scheme 2 .
Both mechanisms and the influence of conju-
gated dienes on switching from one reaction to
the other are currently under investigation.
Acknowledgements
This research was supported by the Technol-
ogy Foundation STW, Applied Science Division
of NWO and the technology programme of the
Ministry of Economic Affairs. M.V. was in-
volved in the project through the Socrates Stu-
dent Exchange Programme. Dr. J.G. de Vries
and Mr. W.G. Reman are thanked for stimulat-
ing discussions.
4. Conclusions
Ž .
One type of ruthenium II complex, viz.
chloro cyclopentadienyl didentate phosphine -
Ž
.
Ž .
ruthenium II , can catalyse two types of reac-
tion of allylic alcohols: isomerisation and ether
formation. Conjugated dienes can act as
w
a switch between both reactions. Chloro-
References
.
Ž
.
cyclopentadienyl -bis triphenylphosphine -ruthe-
w x
1
G.W. Parshall, S.D. Ittel, in: Homogeneous Catalysis,
Wiley-Interscience, New York, 1992, p. 51.
Ž .
x
ium II is most active in the isomerisation of
allylic alcohols to carbonyl compounds. Replac-
ing triphenylphosphine by didentate phosphine
ligands lowers the activity. The increased reac-
tivity of those complexes with increasing chain
length reflects the greater ease of ring opening
of larger chelate rings.
w x
2
F. Fringuelli, A. Taticchi, Dienes in the Diels–Alder Reac-
tion, Wiley-Interscience, New York, 1990.
w x
Ž
.
3
G. Wilke, A. Eckerle, in: B. Cornils, W.A. Herrmann Eds. ,
Applied Homogeneous Catalysis with Organometallic Com-
pounds, Applications Vol. 1, VCH, Weinheim, 1996, p. 358.
J. Kanand, R. Paciello, M. Pinkos, A. Roeper, A. Thome,
w x
4
Ž
.
German Patent No. 4400837 1995 .

(
)
R.C. Õan der Drift et al.rJournal of Molecular Catalysis A: Chemical 159 2000 163–177
177
w x
Ž .
D.V. McGrath, R.H. Grubbs, Organometallics 13 1994 224.
w
w
x
Ž
.
5
w x
6
21 S.-C. Yang, C.-W. Hung, J. Org. Chem. 64 1999 5000.
Ž
.
x
J.-E. Backvall, U. Andreasson, Tetrahedron Lett. 34 1993
22 H. Bricout, J.-F. Carpentier, A. Mortreux, J. Mol. Catal. A:
Ž
.
5459.
Chem. 136 1998 243.
w x
Ž
.
w
w
x
Ž
.
7
S.H. Bergens, B. Bosnich, J. Am. Chem. Soc. 113 1991
958.
23 D. Wang, C.-J. Li, Synth. Commun. 28 1998 507.
x
24 D. Wang, D. Chen, J.X. Haberman, C.-J. Li, Tetrahedron 54
w x
8
Ž
.
C. Bianchini, E. Farnetti, M. Graziani, G. Nardin, A. Vacca,
1998 5129.
Ž
.
w
w
w
w
w
x
F. Zanobini, J. Am. Chem. Soc. 112 1990 9190.
25 M.I. Bruce, C. Hameister, A.G. Swincer, R.C. Wallis, Inorg.
w x
Ž
.
Ž
.
9
R.W. Goetz, M. Orchin, J. Am. Chem. Soc. 85 1963 1549.
Synth. 21 1982 78.
w
w
w
w
x
x
10 F. Stunnenberg, F.G.M. Niele, E. Drent, Inorg. Chim. Acta
26 L.A. Oro, M.A. Ciriano, M. Campo, C. Foces-Foces, F.H.
Ž
.
Ž
.
222 1994 225.
Cano, J. Organomet. Chem. 289 1985 117.
x
Ž
.
.
x
11 B.M. Trost, R.J. Kulawiec, J. Am. Chem. Soc. 115 1993
27 F. Morandini, G. Consiglio, B. Straub, G. Ciani, A. Sironi, J.
Ž
.
2027.
Chem. Soc., Dalton Trans. 1983 2293.
x
Ž
x
12 B.M. Trost, R.J. Kulawiec, Tetrahedron Lett. 32 1991
28 F.L. Joslin, M.P. Johnson, J.T. Mague, D.M. Roundhill,
Ž
.
3039.
Organometallics 10 1991 2781.
x
x
13 B.M. Trost, T.R. Verhoeven, in: G. Wilkinson, F.G.A. Stone,
29 B.S. Furniss, A.J. Hannaford, P.W.G. Smith, A.R. Tatchell,
in: Vogel’s Textbook of Practical Organic Chemistry, Long-
man, London, 1989, p. 519.
Ž
.
E.W. Abel Eds. , Comprehensive Organometallic Chemistry
Vol. 8, Pergamon, Oxford, 1982, p. 799.
w
x
w
x
Ž
.
14 J.A. Davis, in: E.W. Abel, F.G.A. Stone, G. Wilkinson
30 P.D. Bartlett, F.A. Tate, J. Am. Chem. Soc. 75 1953 91.
Ž
.
w
x
Eds. , Comprehensive Organometallic Chemistry II Vol. 9,
31 W.T. Hendrix, F.G. Cowherd, J.L. v. Rosenberg, J. Chem.
Ž
.
Pergamon, Oxford, 1995, p. 291.
Soc., Chem. Commun. 1968 97.
w
w
w
x
Ž
.
w
w
w
x
15 J. Tsuji, Tetrahedron 42 1986 4361.
x
32 C. Li, M.E. Cucullu, A. McIntyre, E.D. Stevens, S.P. Nolan,
Ž
.
Ž
.
16 B.M. Trost, D.L. Van Vranken, Chem. Rev. 96 1996 395.
Organometallics 13 1994 3621.
x
x
17 M. Sakamoto, I. Shimizu, A. Yamamoto, Bull. Chem. Soc.
33 I.M. Angulo, A.M. Kluwer, E. Bouwman, Chem. Commun.
Ž
.
Ž
.
Jpn. 69 1996 1065.
1998 2689.
w
w
w
x
x
18 T. Satoh, M. Ikeda, M. Miura, M. Nomura, J. Org. Chem. 62
34 E. Ruba, W. Simanko, K. Mauthner, K.M. Soldouzi, C.
¨
Ž
.
1997 4877.
Slugove, K. Mereiter, R. Schmid, K. Kirchner, Orga-
x
Ž
.
19 D.E. Bergbreiter, D.A. Weatherford, J. Chem. Soc., Chem.
nometallics 18 1999 3843.
Ž
.
w
x
Ž
.
Commun. 1989 883.
35 B.M. Trost, T.R. Verhoeven, J. Am. Chem. Soc. 102 1980
4730.
x
20 T. Yamamoto, M. Akimoto, O. Saito, A. Yamamoto,
Ž
.
Organometallics 5 1986 1559.