M.A. Klingshirn et al. / Journal of Organometallic Chemistry 690 (2005) 3620–3626
3623
sible that small amounts of residual starting material
could have been lost in the distillation process. 1-Octene
gave a 4.3:1 ratio of linear to branched product which is
somewhat higher than that typically seen for this cata-
lyst system in alcohol solvents (2–3:1) [17].
A side reaction that occurred in these reactions was
the acid-catalyzed addition of methanol to the alkene
(Eq. (2)). Methanol addition was a minor side product
for styrene and 4-chlorostyrene (<5%), but became a
major competitive pathway with 4-t-butylstyrene. Due
to the increased electron density, the alkene in 4-t-butyl-
styrene would be more susceptible to protonation than
styrene. In the case of 1-octene, no methanol addition
was observed.
ane extraction. The IL solution was returned to the reac-
tor with fresh reagents. Additional triphenylphosphine
was added in each cycle to account for leaching losses.
Polystyrene production, which occurred when recycling
studies were carried out at 90 °C, was nearly totally sup-
pressed at 70 °C. Ester yields remained constant through
the first two runs with near complete conversion of sty-
rene followed by a gradual decline in product yields.
Yields remained greater than 50% until the sixth cycle.
The catalyst solution remained yellow throughout the
recycling, suggesting the catalyst remains intact.
Mechanical loss during the extraction likely accounts
for some of the loss of activity, as the IL volume de-
creased from 2 to 1 mL during the recycling study. Inter-
estingly, the l:b ratio increased from 4.3 to 9.8 from the
first to sixth run. This increase may reflect an increased
amount of phosphine in the system.
OCH3
p-TsOH
ð2Þ
CH3OH
R
R
Separation of the products from the catalyst-IL-
4. Discussion
methanol phase was easily accomplished by hexane
extraction, with the catalyst being retained within the
IL-methanol phase. The hexane extracts were colorless
and showed no catalyst by 31P NMR. Approximately
one third of the added phosphine was extracted (24%
as PPh3 and 8% as PPh3O), however. Upon concentra-
tion of the IL-MeOH solution remaining after hexane
extraction, a yellow crystalline precipitate formed,
which was identified as trans-(Ph3P)2PdCl2. Isolation
of this solid from the residual IL solution showed recov-
ery of ꢁ55% of the original catalyst. In contrast, exten-
sive catalyst leaching was observed when pure methanol
was used as the solvent and the IL is omitted.
The ease of separation and ability to recover un-
changed a large fraction of the catalyst precursor
encouraged us to explore the recyclability of the IL/cat-
alyst solution. Recycling experiments were carried out at
70 °C (Table 4) rather than 90 °C, since catalyst decom-
position was more prevalent at higher temperatures.
After each run, the ester products were removed by hex-
ILs are effective solvents for the palladium-catalyzed
hydroesterification of alkenes. The PPh3/Pd/TsOH sys-
tem used by us and Dupont [19] is known to give modest
linear selectivities (1.5:1 l:b) in methanol under similar
reaction conditions to this work. In IL solvents, the
PPh3/Pd/TsOH catalyst system gave high linear to
branched ratios (3–5) that were relatively independent
of reaction conditions. Typically, the regioselectivity of
hydroesterification catalysts with weakly coordinating
anions is strongly dependent on the reaction conditions.
For example, the [(Ph3P)2Pd(CH3CN)2]2+ catalyst sys-
tem gave a l:b ratio of 2.7 with 75 psi CO at 80 °C, while
the l:b ratio was 0.11 using 300 psi CO at 50 °C [17]. In
contrast, our results showed little change in the l:b ratio
(3–5) when both the temperature and CO pressure were
varied, although product yields were significantly af-
fected by these changes. Thus, it appears that under
our conditions, the IL induces a linear preference that
is insensitive to the reaction conditions.
Claver and co-workers [36] have proposed that cata-
lysts with high branched selectivity proceed through a
neutral catalytic cycle, while those with a linear prefer-
ence follow a cationic catalytic cycle (Scheme 1). In
the neutral catalytic cycle, styrene displaces a phosphine
to give a neutral Pd-styrene complex. Since the (mono-
phosphine)Pd-styrene complex is relatively unhindered,
migratory insertion preferentially occurs to give the elec-
tronically favored benzylic-Pd species. CO insertion and
methanolysis gives the branched product. In the cationic
pathway, styrene displaces chloride to give a cationic
Pd-styrene complex. Because the (diphosphine)Pd center
is more sterically crowded than the monophosphine
complex of the neutral pathway, insertion to give the
primary Pd-alkyl is preferred leading to the linear ester.
The preferred pathway can be influenced by choice of
Table 4
Recycling of IL-catalyst solution used in the hydroesterification of
styrenea
Run
Yieldb (%)
l:bc
1
2
3
4
5
6
a
85
83
64
50
60
30
4.3
5.0
5.6
7.2
6.8
9.8
4 mL [C4mim][NTf2] IL, 4 mL methanol, 70 °C, 200 psi CO,
0.25 mL styrene. Runs 2–6 used IL/catalyst solution recovered from
the previous run.
b
Determined by GC.
l:b = linear to branched product ratio determined by GC.
c