Isomerisation of ω-hydroxyalkenes under hydroxycarbonylation conditions in palladium catalysed aqueous phase systems

Article abstract of DOI:10.1016/j.jorganchem.2006.05.038

The ω-hydroxyolefins 3-buten-1-ol, 3-buten-1-methyl-1-ol and 4-penten-1-ol were subjected to hydroxycarbonylation conditions in water in the presence of PdCl₂(PhCN)₂ and 4-8 equiv. of water soluble tris(3-sodiumsulfonatophenyl)phosphine (TPPTS), or N-bis(N′,N′-diethyl-2-aminoethyl)-4-aminomethylphenyl-diphenylphosphine (N3P). Under conditions of high conversion, the olefins primarily undergo isomerisation through a chain walking mechanism with selectivities for aldehyde ranging from 65% to 98%, with the lower values for longer chain alcohols. The lactones formed as the minor product are almost exclusively branched, indicating that in the first step 2,1-insertion is strongly favoured over 1,2-insertion. In the N3P system also linear lactone is produced at lower conversion. Running the reaction in D₂O produces multiple deuterium incorporation in all positions of the carbon chain. A mechanism is discussed.

Full text of DOI:10.1016/j.jorganchem.2006.05.038

Journal of Organometallic Chemistry 691 (2006) 3806–3815
www.elsevier.com/locate/jorganchem
Isomerisation of x-hydroxyalkenes under
hydroxycarbonylation conditions in palladium catalysed
aqueous phase systems
Adriana Ionescu, Markus Ruppel, Ola F. Wendt *
Organic Chemistry, Department of Chemistry, Lund University, P.O. Box 124, S-221 00 Lund, Sweden
Received 21 April 2006; received in revised form 16 May 2006; accepted 16 May 2006
Available online 3 June 2006
Abstract
The x-hydroxyolefins 3-buten-1-ol, 3-buten-1-methyl-1-ol and 4-penten-1-ol were subjected to hydroxycarbonylation conditions in
water in the presence of PdCl (PhCN) and 4–8 equiv. of water soluble tris(3-sodiumsulfonatophenyl)phosphine (TPPTS), or N-
2
2
0
0
bis(N ,N -diethyl-2-aminoethyl)-4-aminomethylphenyl-diphenylphosphine (N3P). Under conditions of high conversion, the olefins pri-
marily undergo isomerisation through a chain walking mechanism with selectivities for aldehyde ranging from 65% to 98%, with the
lower values for longer chain alcohols. The lactones formed as the minor product are almost exclusively branched, indicating that in
the first step 2,1-insertion is strongly favoured over 1,2-insertion. In the N3P system also linear lactone is produced at lower conversion.
Running the reaction in D O produces multiple deuterium incorporation in all positions of the carbon chain. A mechanism is discussed.
2
Ó 2006 Elsevier B.V. All rights reserved.
Keywords: Palladium; Biphasic reactions; Hydroxycarbonylation; Isomerisation; Homogeneous catalysis
1
. Introduction
have been shown to undergo isomerisation reactions. Stille
reported that while simple olefins undergo methoxycarb-
onylation in the presence of PdCl , CuCl and sodium
Hydroxycarbonylation of olefins to carboxylic acids
2
2
(
Scheme 1) in aqueous biphasic systems was simulta-
butyrate to provide succinates, functionalised olefins like
3-butenol gave mixtures of products, due to isomerisation
of the double bond [3]. On the other hand, later synthetic
and mechanistic studies of palladium catalysed alkoxycarb-
onylation reactions of 3-butenols and acetylenic alcohols to
give lactones, report no signs of isomerisation [4]. Isomeri-
sation of higher olefins (1-hexene, 1-octene, 1-decene, and
4-pentenoic acid) under hydroxycarbonylation conditions
has also been reported [1a,5].
neously discovered by three research groups [1]. Since then,
there has been an extensive study of this reaction, mostly
using the water-soluble phosphine tris(3-sodiumsulfona-
tophenyl)phosphine (TPPTS) as ligand and styrene and
styrene derivatives as model substrates [2]. As expected
for substrates with moderate to low water solubility [2a],
the rate determining step is mass transfer between aqueous
and organic layers. One way to avoid this problem and to
allow for mechanistic investigations is to use water-soluble
olefins, such as 3-buten-1-ol (1). However, substrates like 1
Here, we report on the Pd catalysed reactions of x-
hydroxyolefins under hydroxycarbonylation conditions,
0
using two water soluble phosphine ligands: N-bis(N ,
0
N -diethyl-2-aminoethyl)-4-aminomethylphenyl-diphenyl-
*
phosphine (N3P) and the bulkier, but electron poorer
TPPTS (Chart 1). The main products of these reactions
Corresponding author. Tel.: +46 46 2228153.
E-mail address: ola.wendt@organic.lu.se (O.F. Wendt).
0
022-328X/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2006.05.038

A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
3807
COOH
Pd/TPPTS
+
R
COOH
R
+ CO + H O
R
2
+
H /H O
2
(n)
(i)
Scheme 1. The hydroxycarbonylation reaction.
O
O
O
O
O
OH
H
1
2
3
4
O
O
O
OH
O
O
5
6
7
8
O
OH
O
O
H
9
10
11
N
N
N
SO Na
3
NaO3S
P
P
SO Na
3
N3P
TPPTS
Chart 1.
were not the lactones from the intramolecular alkoxycarb-
onylation, but the corresponding aldehyde, as a result of an
isomerisation reaction. To gain a deeper insight into the
mechanism of isomerisation versus hydroxycarbonylation
we also report on HP NMR spectroscopy studies of reac-
tion intermediates and a deuterium labelling study.
2. Experimental
2.1. General procedures
All reactions and manipulations were conducted using
standard Schlenk techniques. All starting materials were

3808
A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
purchased and used without further purification. All sol-
vents were deoxygenated with a nitrogen or argon purge
prior to use. H and P NMR spectra were recorded at
ether. The combined organic phases were transferred into a
freshly prepared solution of 200 mg 2,4-dinitrophenylhydr-
azine, 2 ml conc. H SO (96%), 3 ml H O and 8 ml metha-
1
31
2
4
2
3
3
00 and 121 MHz, respectively, on a VARIAN Unity
00 MHz spectrometer. H and C NMR spectra were
nol. The resulting orange solution was cooled to 5 °C over
night, and this provided thin, shiny, orange needles. They
were filtered and then recrystallised from a hot solution
1
13
recorded at 500 and 125 MHz, respectively, using a BRU-
KER ARX 500 MHz spectrometer. The high pressure
NMR experiments were performed using a sapphire tube
assembly [6]. H NMR and C{ H} NMR chemical shift
are reported in ppm downfield from tetramethylsilane. H
of 4 ml methanol and 1 ml H O by cooling the solution
2
to ꢀ18 °C. Yield: 0.052 g (20%).
1
13
1
1
3
H NMR (500 MHz, CDCl ): d 1.02 (t, J
7.41 Hz,
HH
3
1
3
3H, CH ), 1.66 (t, JHH 7.41 Hz, 2H, CH ), 2.41 (dt,
3
2
3
3
3
NMR chemical shifts were referenced to the residual
hydrogen signal of the deuterated solvents. In the
JHH 7.41 Hz, JHH 5.47 Hz, 2H, CH ), 7.53 (t, JHH
2
5.47 Hz, 1H, C(O)H), 7.92 (d, J
3
9.59 Hz, 1H, H6arom.),
HH
1
3
1
3
4
C{ H} NMR measurements, the signal of CD Cl
8.29 (dd, J
9.59 Hz, J
2.62 Hz, 1H, H5arom.), 9.11
HH
2
2
HH
3
1
4
(
55.8 ppm) was used as reference. P NMR chemical shifts
(d, J
2.62 Hz, 1H, H3arom.), 11.01 (b, 1H, NH).
HH
C NMR/DEPT (125 MHz, CDCl ): d 13.70 (CH ),
1
3
are reported in ppm downfield from an external 85% solu-
tion of orthophosphoric acid. Fast atom bombardment
3
3
6
19.74 (CH ), 34.44 (CH ), 116.51 (C H ), 123.52
arom.
2
2
5
3
1
(
FAB) mass spectroscopic data were obtained on a JOEL
(C Harom.), 129.96 (C Harom.), 137:82 ðC
Þ; 145:16
arom:
2
4
SX-102 spectrometer applying 3-nitrobenzyl alcohol as
matrix and CsI as calibrant. The hydroxycarbonylations
were performed in a 50 ml AUTOCLAVE ENGINEERS
HASTALOYä autoclave equipped with a sampling valve
and temperature and stirring speed controllers. Gas chro-
matographic analyses were carried out on a VARIAN
ðC
Þ; 151:21 ðC
Þ, 152.39 (C(O)H). EI-MS: m/z =
arom:
arom:
+
252 [M] .
2.4. Synthesis of butyraldehyde-d 2,4-
3
dinitrophenylhydrazone (12b)
3
300 instrument equipped with a 25 m CP-SIL 19 fused sil-
ica capillary column.
Using the same procedure as in Section 3.3, but applying
pure D O as the solvent gave shiny, orange needles of 12b.
2
2
.2. Aqueous phase hydroxycarbonylations
Yield: 0.044 g (17%).
1
H NMR (500 MHz, CDCl ): d 0.99–1.09(m), 1.63–
3
3
In a typical experiment, [PdCl (PhCN) ] (0.021 mmol)
1.75(m), 2.35–2.44(m), 7.53 (d, J
3.9 Hz, 1H, C(O)H),
HH
2
2
3
3
and the specified amount of ligand (P/Pd ratio = 4 or 8)
were placed in a flask in the glove box. The flask was
charged with 12 ml of water, previously acidified with p-
toluenesulfonic acid (acid/olefin ratio: 0.7), and the mixture
was vigorously stirred until the Pd precursor was com-
pletely dissolved. The resulting aqueous solution was
placed in the autoclave previously loaded with the sub-
strate (substrate/Pd ratio = 50). The flask, which had con-
tained the catalytic solution, was rinsed with 3 ml of water
and the resulting solution was transferred into the auto-
clave. After five pressurising–depressurising cycles with
CO to remove traces of nitrogen and air, the aqueous solu-
tion was heated to the selected temperature and then pres-
surised with CO under vigorous stirring. After the reaction,
the autoclave was cooled to room temperature, slowly
depressurised, and the reaction mixture was transferred
into a Schlenk tube under nitrogen. The hydroxycarbony-
lation mixture was extracted with ether (6.5 ml) for
7.93 (d, JHH 9.89 Hz, 1H, H6arom.), 8.30 (dd, JHH
4
4
9.89 Hz, JHH 2.34 Hz, 1H, H5arom.), 9.12 (d, JHH
2.40 Hz, 1H, H3arom.), 11.02 (b, 1H, NH).
1
3
C NMR/DEPT (125 MHz, CDCl ): see Table 3 for the
3
6
shifts of the alkyl chain, 116.38 (C H
), 123.37
arom.
1
5
3
(C Harom.), 129.83 (C Harom.), 137:71 ðC
Þ; 145:04
arom:
2
4
ðC
Þ; 151:08ðC
Þ, 152.25 (C(O)H). EI-MS: m/z =
arom:
arom:
+
252–256 [M] .
2.5. NMR studies of reaction intermediates
A D O solution of PdCl (6 mM) and TPPTS (P/Pd =
2
2
6) was subjected to CO at 1.7 bar for 15 min. Trifluoro-
acetic acid was added to give a 60% (v/v) solution, which
was stirred for 2 h. The yellow solution was subjected to
three freeze–pump–thaw cycles to remove dissolved CO
and, under argon, 1.5 equiv. of 3-butene-1-ol was added.
An NMR sample was taken and subjected to 1.7 bar CO
for 1.5 h and then 50 bar CO for 3 h. The solution was
afterwards concentrated under vacuum and 1,4-dioxane
was added, furnishing a yellow precipitate, which was
analysed by NMR and IR. Half of the precipitate was
hydrolysed at 80 °C for 1 h under argon and the other
half was hydrolysed under 50 bar CO at 80 °C for 1 h.
During the whole procedure, NMR samples were taken
at regular intervals. The products of hydrolysis were
identified by GC.
1
0 min. A sample of 3 ml of the ether layer was separated,
dried over MgSO and analysed by GC and GC–MS.
4
2
(
.3. Synthesis of butyraldehyde 2,4-dinitrophenylhydrazone
12a)
Following the general procedure in 2.2. (T = 65 °C,
PCO = 15 bar) and using 3-buten-1-ol (1.043 mmol,
0
.0892 ml) as substrate gave a yellow aqueous product
solution, which was extracted with 5 ml and 3 ml of diethyl

A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
3809
3
. Results and discussion
matically to 11% at low temperature and low pressure (run
1). Only at 95 °C and a P/Pd ratio of 8 the conversion was
quantitative. Thus, a high P/Pd ratio not only provides sta-
bilisation but also decreases the conversion due to coordi-
native saturation.
2
3
.1. The hydroxycarbonylation of x-hydroxyalkenes
The aqueous phase hydroxycarbonylation of 1 was car-
ried out in the presence of [PdCl (PhCN) ] and the phos-
Also the product distribution is affected by the P/Pd
ratio and the CO pressure. At a TPPTS/Pd ratio of 8 the
selectivity for the aldehyde was always higher than 90%.
However, at a TPPTS/Pd ratio of 4 the 2/3 ratio was con-
siderably lower, in the range of ca. 1–10:1. It cannot be
excluded that there is a contribution to the catalysis from
the Pd black formed in entries 1–5. A similar aldehyde/lac-
tone ratio is observed in the N3P experiments where palla-
dium black is formed. Optimal isomerisation requires a
TPPTS/Pd = 8 (run 7) but a N3P/Pd = 4 (run 13). For
N3P/Pd = 8, the aldehyde/lactones ratio is under 10 at
65 °C (runs 21, 22) but much higher at 95 °C (runs 19,
20), showing that the isomerisation is favoured at higher
temperatures with this ligand whereas the opposite holds
true for the case where TPPTS/Pd = 8 (vide infra).
Moreover, at a given P/Pd ratio, there was also a varia-
tion in the product distribution as the partial pressure of
CO was varied, but this variation is not continuous; instead
the aldehyde/lactone ratio goes through a minimum at an
intermediate CO pressure of 50 bar.
2
2
phine ligand. The reactions were run under various
temperatures and carbon monoxide pressures, in the pres-
ence of p-toluenesulfonic acid. Beside 3-buten-1-ol (1),
two heavier hydroxy-substituted olefins, 3-buten-1-
methyl-1-ol (5) and 4-penten-1-ol (9) were used as sub-
strates (entries 11 and 12, Table 1). The products are
known compounds and were identified by GC–MS. All
results are summarised in Table 1 and all structures are
given in Chart 1.
Under the present conditions, the stability of the cata-
lyst system is dependent on the P/Pd ratio. The formation
of metallic palladium was observed in experiments operat-
ing at a TPPTS/Pd molar ratio of 4 and this decomposition
of the catalyst probably explains the relatively low conver-
sion in entries 1–5 (see Table 1) compared to entries 6–12 in
which a TPPTS/Pd molar ratio of 8 was applied. In these
latter experiments, no formation of metallic palladium
metal was observed and the conversion of the substrate
was quantitative. N3P provided better stabilisation of the
catalyst and at a ratio of 4, palladium black was observed
only at high temperature (95 °C). At a ratio of 8, no palla-
dium black was observed, but the conversion dropped dra-
With TPPTS only the 5-ring 3-methyltetrahydro-2-fura-
none (3) is formed, but with N3P both 3 and the 6-ring tet-
rahydro-2H-pyran-2-one (4) are formed. Their distribution
Table 1
Results of the aqueous phase hydrocarboxylation of hydroxyalkenes under various conditions
Run
Substrate (ligand)
P/Pd
T (°C)
P
CO (bar)
Conversion (%)
Selectivity (%)
Aldehyde
Aldehyde/lactone ratio and (n/i)
Lactone
a,c
1
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
1 (TPPTS)
5 (TPPTS)
9 (TPPTS)
1 (N3P)
1 (N3P)
1 (N3P)
1 (N3P)
1 (N3P)
1 (N3P)
1 (N3P)
1 (N3P)
1 (N3P)
4
4
4
4
4
8
8
8
8
8
8
8
4
4
4
4
4
4
8
8
8
8
65
65
65
65
75
75
65
65
65
65
65
65
65
65
65
65
95
95
95
95
65
65
0
15
50
75
15
15
15
15
50
75
15
15
15
15
50
75
15
50
50
15
15
50
8.2
63.5
48.1
68.0
62.1
100.0
100.0
100.0
100.0
100.0
100.0
100.0
54.1
100.0 (2)
89.3 (2)
82.2 (2)
90.7 (2)
41.5 (2)
90.9 (2)
98.2 (2)
92.5 (2)
96.0 (2)
97.3 (2)
68.0 (6)
65.0 (10)
99.0 (2)
78.5 (2)
82.0 (2)
97.4 (2)
83.6 (2)
85.1 (2)
96.0 (2)
98.0 (2)
67.4 (2)
66.7 (2)
–
–
8.3
4.6
9.7
0.7
10
54.6
12.3
24.0
36.0
2.1; (1)
1.9
99.0
3.6; (5.4)
4.6; (0.3)
37.5
5.1; (0.9)
5.7
c
2
10.7 (3)
17.8 (3)
9.3 (3)
58.5 (3)
9.1 (3)
1.8 (3)
7.5 (3)
4.0 (3)
c
3
c
4
c
5
6
7
8
b
9
1
1
1
1
1
1
1
1
1
1
2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
2.7 (3)
16.0 (7); 16.0 (8)
35 (11)
1.0 (4)
b
32.7
83.5
60.6
62.0
3.4 (3); 18.1 (4)
14.0 (3); 4.0 (4)
2.6 (3)
8.6 (3); 7.8 (4)
14.9 (3)
4.0 (3)
2.0 (3)
28.3 (3); 4.3 (4)
33.3 (3)
c
c
69.6
100.0
100.0
11.1
24
49
2.1; (0.1)
2
1 (N3P)
68.8
General reaction conditions. Precursor: PdCl
2
2 2
(PhCN) (1,4 mM); olefin/Pd ratio: 50; acid/olefin ratio: 0.7; acid: p-toluenesulfonic acid; solvent: H O; total
volume: 15 ml; 3 h.
a
This reaction was carried out under an argon atmosphere as a blank reaction.
b
c
D
2
O as solvent; the given values are averages over 3 independent experiments under described conditions.
Formation of metallic palladium.

3810
A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
varies from pure 4 at low P/Pd ratio, temperature and pres-
sure (run 13) to pure 3 at very high pressure (run 16), or
when at least two of the above three reaction conditions
are increased (runs 18, 19, 20, 22). The isomerisation is
the main reaction also for substrates 5 and 9. However,
the extent of isomerisation is lower than for 1, for which
3.2. Mechanistic considerations
The proposed mechanism for the hydroxycarbonylation
of 3-buten-1-ol is shown in Scheme 2. As outlined by Shel-
don, the complex [PdCl(TPPTS) ] of type I is obtained by
dissolving PdCl2 or the bis-benzonitrile complex in an
+
3
9
7
8% aldehyde was formed under identical conditions (run
). Instead, 32–35% of the reacting alkenes end up as the
aqueous TPPTS solution. We confirmed the structure of I
by P NMR analysis, showing a triplet at 33.9 ppm and
a doublet at 30.7 ppm [7a]. This solution was subjected to
3
1
lactones 7 and 8 in the case of substrate 5 (run 11), or
the lactone 11 in the case of substrate 9 (run 12). The rela-
tive rate of isomerisation is normally lower for higher ole-
fins and the methyl group in olefin 5 can account for a
degree of steric hindrance, hampering the formation of 6.
1.7 bar CO providing the zerovalent complex Pd(TPPTS)
3
3
1
II, which exhibited a broad P NMR resonance at
22.9 ppm [7]. When the solution was acidified, three new,
rather broad signals appeared under 1.7 bar CO, at
0
2
Pd(PhCN) Cl2
+ 3P
[PdClP ]+
3
+CO; H O
2
Pd P
3
or
PdCl2
I
-CO ; 2HCl
2
II
+
H+ [HPdP3] + [Pd H(CO)P ]
+
+
2
4
III
IV
2
,1-insertion
(P = TPPTS, N3P)
+ 3-buten-1-ol
1
,2-insertion
(
P = N3P)
n+
X
repeated olefin insertion/
Pd
β−H elimination
P
temp.
O
VI
H
+
n+
X
OH n+
P
P
Pd
Pd
CO
P
H
O
O
H
VII
V
n+
P
OH
+P; fast
+ CO
Pd
X
P
H
VIII
2
O
+
+
[HPdP3]
O
n+
O
P
Pd
O
X
+
+
[HPdP3]
O
H
IX
4
RDS
+2P
O
O
+
+
[HPdP3]
3
ꢀ
2
Scheme 2. The proposed mechanism of hydroxycarbonylation of 3-buten-1-ol (P = TPPTS or N3P; X = Cl (n = 0); X = CO, D O (n = 1)).

A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
3811
2
2
7.5 ppm, 21.6 ppm and 20.5 ppm. Two of them, at
as a ligand and that the signal at 19.9 ppm is assigned to
a complex of type VI.
7.5 ppm and 20.5 ppm correspond to the unresolved dou-
+
blet and triplet of the Pd-hydride complex [PdH(TPPTS)3]
III as reported by Sheldon [8]. We also observed this mono-
meric hydride in H NMR spectroscopy, showing a double
It is generally accepted that monodentate phosphines
give more branched acids (n/i < 1) [12] and this is also
the case for the TPPTS system [1]. The formation of the
branched lactones 3, 7 and 11 through 2,1-insertion may
reflect this general preference, which is then the result of
the selectivity of the first insertion step. At low to moderate
CO pressures the chelate binding obviously precludes the
cis coordination of CO, which is a prerequisite for CO
insertion into the Pd-C bond. We cannot fully exclude that
complex V is also formed (in fast equilibrium with VI, since
1
2
2
triplet at ꢀ7.9 ppm with J
= 175 Hz and J
=
Pcis–H
Ptrans–H
3
1
1
2
5 Hz. The third signal in the P NMR spectrum at
1.6 ppm, can be attributed to the dimeric hydride-bridged
+
complex [Pd (l-H)(l-CO)(TPPTS) ] IV. This complex
was not observed earlier under 1 bar CO [8], but by increas-
ing the carbon monoxide pressure it was easily formed. The
2
4
1
dimeric hydride gives a pentet at ꢀ7.4 ppm in the H NMR
2
31
spectrum with
J
= 40 Hz, due to coupling to four
we see only one P NMR signal), but it seems unreason-
P–H
equivalent phosphorous atoms. We obtained further con-
able that it should be less reactive than VI and if it forms
we would expect to see also the six-membered lactone
product. No hydroxycarbonylation products were formed
by hydrolysis (80 °C, no CO present) of the precipitate con-
taining complex VI, only butanal. Upon heating under
hydroxycarbonylation conditions (80 °C, 50 bar CO), com-
plex VI gives merely butanal and traces of lactone 3. As in
the catalysis experiments, the 6-ring d-valerolactone (4)
formed by hydroxycarbonylation of the linear isomer V is
not observed, again corroborating that the alkyl complex
be formulated as VI. Similarly, substrate 9 exclusively gives
the six-membered lactone 11, originating from an initial
2,1-insertion. The situation is different for substrate 5,
where the methyl group a to the hydroxyl obviously desta-
bilises the 2,1-insertion product to give equal amounts of
lactones 7 and 8.
For N3P the situation is different. Here the 1,2-insertion
is a viable alternative giving both 5- and 6-ring lactones 3
and 4. The latter is favoured only at low P/Pd ratio, low
temperature and low pressure (runs 13, 14 in Table 1). It
can be noted that conditions giving high conversion also
give exclusive formation of 3. Contrary to expectations
based on steric requirements it therefore seems that the
1,2-insertion is favoured by coordinative unsaturation
and the use of a less bulky phosphine. It has been observed
earlier in the context of hydroformylation that the use of
bulky phosphines give more branched product, and this
was explained by bulkier phosphines giving a lower phos-
phine coordination number [13].
The last step in the lactone formation is a nucleophilic
attack by the hydroxyl group at the carbonyl carbon atom.
Obviously only at high CO pressures can this process com-
pete with chain walking to give butanal; the formation of
lactone increases as the partial pressure of CO increases
as seen when comparing runs 2 and 3, 7 and 9 or 13 and
15. Surprisingly, lactone formation does not continue to
increase as the CO pressure is further increased (runs 4,
10, 16). A high CO concentration should promote the acyl-
ation reaction and inhibit the chain-walking reaction, by
coordinating to the open coordination site. However, in
order to obtain the lactone, the acyl complex has to be
quenched by an intramolecular nucleophilic attack of the
hydroxyl group at the carbonyl carbon atom, which pre-
sumably is an irreversible, rate determining step. There
1
3
firmation by a high pressure C NMR spectrum at
1
3
5
2
0 bar CO, where the dimer gives a double pentet at
2
2
13.6 ppm with J
= 30 Hz and J
= 10 Hz. The fast
C–H
C–P
ligand exchange, which accounts for the equivalence of all
four phosphorous atoms, could not be frozen even at low
temperatures, as observed earlier for corresponding
diphosphine complexes [9].
Using N3P, we have earlier isolated two reaction inter-
mediates of the type V and VIII [5b], the alkyl complex
being the most stable one, due to its chelated form.
Attempts to obtain similar complexes in the TPPTS system
were therefore undertaken. Thus, the solution containing
the hydride-complexes of type III and IV (with TPPTS)
was carefully degassed and after the addition of the olefin
under argon, a new signal at 19.9 ppm appeared at the
expense of the hydride signals. This signal can be assigned
to an alkyl complex since it is formed in the absence of car-
bon monoxide. The peak remained unchanged also when
the solution was left under 50 bar CO for several hours.
The same peak was obtained when the addition of the ole-
fin was made under CO atmosphere. For characterisation,
the solution was concentrated and diluted with 1,4-dioxane
3
1
to obtain a precipitate which was first analyzed by
P
NMR spectroscopy in D O. This proved to be a mixture
2
of TPPTS oxide, complex II and the alkyl complex at
1
9.9 ppm. IR analysis of the precipitate revealed, beside
the stretching frequencies of complex II [10], several bands
of stretching vibration of saturated CH groups between
ꢀ
1
2
968 and 2858 cm , the band of deformational vibration
ꢀ
1
of saturated CH groups (1465 cm ) and an intense band
in the region of low frequencies, which can be assigned to
the stretching vibration of a Pd–O bond (537 cm ) [11].
ꢀ
1
1
3
A
C NMR spectrum showed no evidence for the presence
of any Pd-acyl species. It seems that the a-insertion and
formation of a Pd-acyl complex is hampered also at high
carbon monoxide pressures, indicating that the equilibrium
for CO migratory insertion lies to the left. The IR spectrum
and the high stability of the alkyl complex even under CO
pressure are indications of a cyclic alkyl complex. This
could either be the result of a 1,2- or 2,1-insertion, com-
plexes V or VI. Based on the exclusive formation of lactone
3
and the deuterium labelling experiments (vide infra), we
propose that only 2,1-insertion takes place with TPPTS

3812
A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
are two different possible mechanisms for this intramolecu-
lar attack – it can occur with or without coordination of
the hydroxyl group to the metal centre (from intermediate
VIII). The observation that the formation of the lactone
increases as the CO-pressure increases from 15 to 50 bar
but then decreases as the CO pressure is increased even fur-
ther to 75 bar can possibly be explained by an initial coor-
dination of the alkoxy group under formation of the cyclic
derivative IX [14]. The hydroxyl group which is a relatively
weak ligand can probably not effectively compete with CO
as the concentration of CO in the reaction solution exceeds
a certain level. Lactone formation as a function of the par-
tial pressure of CO therefore goes through a maximum at
an intermediate CO pressure.
D-atoms in the carbon chain could not be established by
the MS analysis. Instead, the aldehyde was converted to
its corresponding 2,4-dinitrophenylhydrazone (12b), which
was characterised by MS and NMR ( H, C, DEPT) anal-
ysis. Comparison between the data obtained for 12b and
data for the unlabeled hydrazone (12a) showed the pres-
ence of several deuterium atoms, distributed on different
carbons of the carbon chain (Scheme 3).
1
13
The deuterium isotope effect and the labelling experi-
ments show that butanal is formed via alternating olefin
insertion and b-hydrogen elimination reactions. Deuterium
is incorporated upon olefin insertion into a Pd–D bond and
if bond rotation takes place prior to b-hydrogen elimina-
tion, the olefin is left with an incorporated D-atom
+
+
At higher temperature, however, the situation changes
in the TPPTS system and the esterification process, which
is rather slow at low temperature, becomes relatively fast
at 75 °C (runs 5 and 6) and the yield of lactone formed thus
becomes larger than that obtained at 65 °C. At a first
glance, one should expect the contrary, an increase of alde-
hyde/lactone ratio, due to a faster b-H elimination at
higher temperature, like in the N3P system. However, both
rate determining steps are first-order intra-molecular pro-
cesses and in view of this it is not surprising that a change
of ancillary ligand can shift the selectivity.
(Scheme 4). If there is a D /H exchange on the palladium
hydride, repeated olefin insertion/b-hydrogen elimination
can result in the incorporation of multiple D-atoms per
molecule [16]; this is also what we observe. The presence
of a palladium hydride species explains why we see such
a high degree of isomerisation. Earlier studies, where there
is no or little isomerisation, have postulated non-hydridic
pathways [17]. Palladium hydride species are known to be
effective catalysts for olefin isomerisation [18] and late
metal catalysts often exhibit facile metal migration along
the chain (the so-called ‘‘chain walking mechanism’’) [19].
Quantitative EI-MS spectrum of the deuterium labelled
hydrazones (compared with the MS of the unlabelled one)
shows up to four deuterium atoms per molecule of hydra-
zone in the TPPTS system and up to five deuterium atoms
3
.3. Deuterium labelling experiments
2
1
In order to gain further information about the mecha-
nism behind the catalytic transformation of 3-buten-1-ol
to butanal some deuterium labelling experiments, using
in the N3P system. The H NMR and H NMR spectra
show the deuterium content in different positions (Table
2
2). The results were obtained from quantitative H NMR
D O as the solvent, were carried out. The conditions cho-
sen were the same as those previously shown to give a high
yield of butanal (entries 7 and 13, Table 1). In general,
spectra by integration of the signals of the groups with deu-
2
1
terium incorporation and comparison with H NMR spec-
1
3
tra. C NMR DEPT (135°) spectra show complex patterns
2
3
compared to the experiments carried out in H O, the corre-
due to various amounts of deuterium in positions C , C
2
4
sponding reactions in D O exhibited lower butanal/lactone
and C . However, six predominant species with zero to
two deuterons can be distinguished in the TPPTS system
(see Table 3 and Fig. 1). At least seven species with zero
to three deuterons are identifiable in the N3P system (see
Table 3 and Fig. 2). Groups with two protons come up
as anti-phase singlets (–CH –) or triplets (CH D–). Groups
2
ratios (entries 7, 8 and 13, 14 in Table 1). This difference
reflects the kinetic isotope effect and implies that more
C–H bond breaking is involved in the reaction to give but-
anal [15].
A mass spectroscopic analysis of the deuterium labelled
products indicated the presence of one to five deuterium
atoms in the deuterated butanal but the position of these
2
2
with three and one proton come up as in-phase singlets
(CH –) and triplets (–CHD–), respective. Any CD groups
3
2
O2N
H
2
,4-dinitrophenyl-
N
NO2
O
hydrazine
N
X2
C
C1
H O, EtOH
2
3
X2
C
C1
4
2
3
C X3
C
H
H SO4
4
2
X2
2
C X3
C
H
X2
2
12
Scheme 3. The formation of butyr-hydrazone (X = H or D).

A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
3813
OD
OD
CDH2
+
Pd-D
3
+ Pd-D/D+
OD - Pd-H/H+
Pd-C rotation
HC
H
Pd
2
,1-insertion
H
H
Pd
D
H
CDH2
CDH2
OD
HDC
CH
2
β-insertion
Pd-C rotation
OD
H
CDH2
Pd
Pd
Pd
D
OD
H
D
H
CDH2
HDC
O
+
DH
C
CH D O
3
tautomerisation
H DC
2
C
H
-
Pd-H
DH
O
D O
2
D
Scheme 4. The deuterium incorporation into 3-buten-1-ol via the chain walking mechanism and enol tautomerisation (some hydrogens are left out for the
sake of simplicity).
Table 2
Deuterium content in the labeled hydrazone 12b
Ligand
Deuterium content via MS analysis (%)
Deuterium incorporation at each carbon
via H and H NMR
2
1
b
4
3
2
b
0
D
1D
2D
3D
4D
5D
0
NDM
C
C
C
NDM
TPPTS
N3P
a
7.62
39.19
36.39
14.93
1.87
1.64
0.68
0.36
0.50
1.54
a
a
a
(0.79)
5.00
(0.1)
23.73
(0.01)
0.48
61.26
8.43
1.1
2.75
0.60
0.42
1.60
2.62
a
(
0.05)
The numbers in brackets show the natural occurrence (%) of the respective deuterated species.
Average number of deuterium atoms per molecule.
b
Table 3
Carbon shifts (ppm) in C NMR DEPT for different species of the deuterated hydrazone 12b (only the aldehyde chain)
13
4
3
2
Aldehyde chain
C
C
C
TPPTS
N3P
TPPTS
N3P
TPPTS
N3P
CH
CH
CH
CH
CH
CH
CH
3
3
3
3
2
2
2
2
ACH
2
ACH
ACHDACH@
ACH@
ACHDACHDACH@
2
ACH@
13.57
13.54
13.46
13.43
13.28
13.25
13.72
13.61
13.46
13.43
13.31
19.59
19.44
19.24
19.16
19.51
19.36
19.53
19.38
19.19
19.11
19.46
34.31
33.94
34.22
33.86
34.28
33.92
34.34
33.95
34.24
33.89
34.31
ACH
2
ACHDACH
2
DACH
DACH
DACH
2
2
2
ACH
ACHDACH@
ACD ACH@
ACH@
2
ACH@
2
13.28
13.25
18.95
18.68
33.59
33.53
CH
DACHDACD
2
1
JCD (Hz)
19.2
19.6
19.6
1
3
4
disappear. It is hard to identify CD signals in the C spec-
The higher deuterium incorporation at C compared to
2
3
trum, because they are in a lower quantity and appear as
septets, therefore could being lost in the noise.
C in both ligands systems corroborates the earlier conclu-
sion that 2,1- insertion is favoured over 1,2-insertion. It can

3814
A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
13
Fig. 1. C NMR DEPT (135°) spectrum of the deuterium labelled hydrazone (TPPTS as ligand).
13
Fig. 2. C NMR DEPT (135°) spectrum of the deuterium labelled hydrazone (N3P as ligand).
be noted that for the N3P system there is a less pronounced
preference for deuterium incorporation in the 4-position
and here we also see formation of the linear 6-ring lactone.
Chain walking occurs along the carbon chain and in those
cases where Pd–C rotation and H /D exchange take
place, deuterium can be incorporated in all positions. It
is clear that bond rotation is often slower than elimina-
tion/insertion resulting in an average of 1.5 (TPPTS) or
tautomerisation reaction as indicated in Scheme 4 would
lead to a high D-incorporation in the 2-position. In the
TPPTS system the tautomerisation probably then takes
place through metal catalysis, as proposed earlier for the
Wacker chemistry [21].
+
+
4. Conclusions
2
.6 (N3P) deuterium atoms/molecule. This is in line with
In conclusion, x-hydroxyolefins primarily undergo
isomerisation through chain walking under hydroxycarb-
onylation conditions. For the bulky TPPTS system the lac-
tones formed are almost exclusively branched, indicating
that in the first step 2,1-insertion is strongly favoured over
1,2-insertion. In the N3P system this preference is less pro-
nounced. The rate determining step in lactone formation is
measured reaction barriers [20]. The very high deuterium
incorporation on C in the N3P system is an indication that
a different mechanism is operating. We suggest that in the
N3P case with higher substitution rates, the enol tautomer
is easily displaced from the metal centre and the tautomer-
isation is not metal catalysed. A classical acid catalysed
2

A. Ionescu et al. / Journal of Organometallic Chemistry 691 (2006) 3806–3815
3815
probably an intramolecular attack by a chelated alkoxy
function on the carbonyl. Multiple deuterium incorpora-
[7] (a) G. Papadogianakis, J.A. Peters, L. Maat, R.A. Sheldon, J. Chem.
Soc., Chem. Commun. (1995) 1105;
(
b) C. Binkowski, J. Cabou, H. Bricout, F. Hapiot, E. Monflier, J.
tion in D O points to Pd–H/Pd–D exchange being on the
2
Mol. Catal. 215 (2004) 23.
same time scale as chain walking.
[8] G. Verspui, I.I. Moiseev, R.A. Sheldon, J. Organomet. Chem. 586
(
1999) 196.
[9] (a) M. Portnoy, D. Milstein, Organometallics 13 (1994) 600;
b) J.C. Calabrese, E.E. Bunel, Organometallics 20 (2001) 337.
10] G. Papadogianakis, L. Maat, R.A. Sheldon, Inorg. Synth. 32 (1998)
5.
11] N.V. Kramareva, A.E. Koklin, E.D. Finashina, N.S. Telegina,
A.Yu. Stakheev, L.M. Kustov, Kinet. Catal. 45 (2004) 43.
12] I. del Rio, C. Claver, P.W.N.M. van Leeuwen, Eur. J. Inorg. Chem.
Acknowledgements
(
[
[
[
[
We thank the Swedish Foundation for Strategic Re-
search for financial support. O.F.W. thanks the Swedish
Research Council for generous support.
2
References
2
719 (2001), and references therein.
13] B. Cornils, W.A. Herrmann (Eds.), Applied Homogeneous Catalysis
[
1] (a) F. Bertoux, S. Tilloy, E. Monflier, Y. Castanet, A. Mortreux,
New J. Chem. 21 (1997) 529;
with Organometallic Compounds, vol. 1, VCH, Weinheim, 1996, p.
58.
(
b) G. Papadogianakis, G. Verspui, L. Maat, R.A. Sheldon, Catal.
Lett. 47 (1997) 43;
c) B. Xie, Y. Kou, Y. Ying, Fenzi Cuiha 11 (1997) 81.
[14] Y.-S. Lin, A. Yamamoto, Organometallics 17 (1998) 3466.
[15] J.H. Espenson, Chemical Kinetics and Reaction Mechanisms, second
ed., McGraw-Hill, Inc., New York, 1995, p. 214ff.
[16] (a) C. Benedek, G. Szalontai, A. G o¨ m o¨ ry, Z. T o¨ r o¨ s, B. Heil, J.
Organomet. Chem. 579 (1999) 147;
(
[
2] (a) F. Bertoux, S. Tilloy, E. Monflier, Y. Castanet, A. Mortreux, J.
Mol. Catal. 138 (1999) 53;
(
b) S. Jayasree, A. SeaYad, Bibhas R. Sarkar, R.V. Chaudari, J.
(b) K.H. Shaughnessy, R.M. Waymouth, Organometallics 16 (1997)
1001.
Mol. Catal. 181 (2002) 221.
[
[
3] J.K. Stille, R.J. Divakaruni, J. Org. Chem. 44 (1979) 3474.
4] (a) H. Alper, D. Leonard, Tetrahedron Lett. 26 (1985) 5639;
[17] (a) T.R. Murray, J.R. Norton, J. Am. Chem. Soc. 101 (1979) 4107;
(b) E.G. Samsel, J.R. Norton, J. Am. Chem. Soc. 106 (1984) 5505.
[18] X. Wang, L.K. Woo, J. Mol. Catal. 130 (1998) 171, and references
therein.
(b) T.F. Murray, E.G. Samsel, V. Varma, J.R. Norton, J. Am.
Chem. Soc. 103 (1981) 1981;
(
(
c) Y. Tamaru, M. Hojo, Z. Yoshida, J. Org. Chem. 56 (1991) 1099;
d) S. Toda, M. Miyamoto, H. Kinoshita, K. Inomata, Bull. Chem.
[19] (a) L.K. Johnson, C.M. Killian, M. Brookhart, J. Am. Chem. Soc.
117 (1995) 6414;
Soc. Jpn. 64 (1991) 3600.
5] (a) G. Verspui, J. Feiken, G. Papadogianakis, R.A. Sheldon, J. Mol.
Catal. 146 (1999) 299;
(b) D.J. Tempel, L.K. Johnson, R.L. Huff, P.S. White, M. Brookhart,
J. Am. Chem. Soc. 122 (2000) 6686.
[20] L.H. Shultz, D.J. Tempel, M. Brookhart, J. Am. Chem. Soc. 123
(2001) 11539.
[
[
(
b) E. Monflier, S. Tilloy, F. Bertoux, Y. Castanet, A. Mortreux,
New. J. Chem. 21 (1997) 857;
c) M. Karlsson, A. Ionescu, C. Andersson, J. Mol. Catal. A,
˚
[21] J.E. B a¨ ckvall, B. Akermark, S.O. Ljunggren, J. Am. Chem. Soc.
(
101 (1979) 2411, This mechanism was recently challenged in
favour of a chloride mediated reductive elimination: Keith, J.A.;
Oxgaard, J., Goddard, W.A. III, J. Am. Chem. Soc. 128 (2006)
3132.
accepted.
6] A. Cussanelli, U. Frey, D.T. Richens, A.E. Merbach, J. Am. Chem.
Soc. 118 (1996) 5265.