P.J. Baricelli et al. / Applied Catalysis A: General 490 (2015) 163–169
165
Qualitative analysis of the known products was made by GC
coupled with mass spectrometry in a Shimadzu GC2010/QP2010-
plus instrument fitted with a Restek Rtx-5 MS capillary column
(30 m × 0.25 mm × 0.25 m), operating at 70 eV.
homogeneous Rh/TPP system under similar conditions (Table 1,
cf. run 3 and 5; P/Rh = 10). Both homogeneous and biphasic reac-
5000 h−1, respectively. At P/Rh = 20 (Table 1, run 4), the reaction in
the biphasic system was even faster (TOF = 8000 h−1).
The total pressure of the equimolar gas mixture had a strong
positive effect on the rate of the hydroformylation of eugenol
(Table 1, runs 3, 6 and 9). Typically, the increase in the total pres-
sure does not affect significantly the hydroformylation rate because
either the rate-determining step is prior to the hydrogen addition
or CO has a negative order, competing with hydrogen for the metal
sites and thus resulting in a compensative effect if both gases have
the same increase in their partial pressures.
3. Results and discussion
The hydroformylation of eugenol (1a) as a model for natu-
rally occurring allylbenzene family has been studied in aqueous
biphasic systems in the absence of additional organic solvents using
[Rh(COD)(-OMe)]2 as the catalyst precursor and various water
soluble ligands (TPPTS, BDPPETS, BDPPPTS and BISBIS) to keep
the catalyst dissolved in the aqueous phase. For comparison, the
reaction was also performed in a conventional homogeneous sys-
tem with triphenylphosphine (TPP) as the auxiliary ligand. The
reactions with estragole (1b) and safrole (1c) were performed in
biphasic systems under optimized conditions.
All three substrates gave two major products derived from the
responding propenylbenzenes 2a–c, formed due to the substrate
isomerization, were detected as minor products along with trace
amounts of the products of substrate hydrogenation (not shown in
Scheme 1). The hydroformylation of propenylbenzenes 2a–c would
give branched aldehydes 4a–c and 5a–c with the formyl group in
-and ␣- positions, respectively; however, ␣-isomers 5a-c were
formed only in small amounts because of the lower reactivity of
internal olefins in hydroformylation.
The comparison of relative hydroformylation rates in various
runs was made by measuring the pressure drop of syngas inside
the reactor along the reaction time. The internal pressure was mea-
sured through a pressure transducer and automatically logged by
a field logger connected to a computer. Kinetic curves were nearly
straight lines up to ca. 90% conversions in most of the runs, what
indicates that the reaction rate is independent on the substrate con-
centration. The reported rates correspond to the slope of the curves
in their linear part (stationary period of the reaction).
In addition to the total pressure effect, we examined the effects
of different partial pressures (different CO/H2 ratios) on the hydro-
formylation rate (Table 1, runs 3, 7 and 8). It is interesting to observe
that the rate is roughly first order with respect to the hydrogen
pressure, but quite insensitive (nearly zero order) to the CO pres-
sure. For example, in run 8, the hydrogen pressure was 1.8 times
higher than that in run 7, and the rate was also 1.8 times higher in
spite of the difference in the CO pressure. In run 9, the hydrogen
pressure was doubled as compared to run 3, which resulted in two
times increase in the reaction rate. On the other hand, the rate of
the reaction at 10 atm of the equimolar gas mixture was dispropor-
tionably lower (four times lower) than that at 20 atm (Table 1, runs
6 and 3). Thus, the existence of mass transference limitations could
not be ruled out in the biphasic systems at low gas pressure.
The selectivity for the aldehydes was higher than 93% in all the
runs presented in Table 1, except for run 1 and run 6, which were
performed at either low ligand concentration (P/Rh = 3 in run 1) or
low gas pressure (10 atm in run 10). On the other hand, the highest
regioselectivity of 81% for linear aldehyde 3a was obtained in run 6.
The results obtained can be interpreted in the light of the puta-
tive mechanism for the rhodium-catalyzed hydroformylation of
terminal olefins. After initiation, i.e., the formation of the square-
planar [RhH(CO)(L)(L’)] species (L, L’ = CO or P(III) ligand), the
expected organometallic steps are: (i) substrate coordination; (ii)
hydride migration to form linear or branched metal-alkyl species;
(iii) CO coordination; (iv) migration of the alkyl moiety to the
coordinated CO to form metal acyl species; (v) oxidative addi-
tion of hydrogen to metal acyl species; vi) reductive elimination
of the aldehyde. The latter step (vi) is considered to be essen-
tially irreversible, while the others may be reversible, depending
on substrates and reaction conditions. From the branched metal-
alkyl species formed during step (ii), a -hydrogen elimination,
instead of step (iii), may lead to isomeric olefins. Thus, if in any
step of the hydroformylation the path after the hydride migration
(i) is hindered, the tendency is to increase the importance of the
isomerization path.
3.1. Hydroformylation of eugenol in the absence of the surfactant
formylation of eugenol occurred smoothly under mild conditions
to give aldehydes 3a, 4a and 5a with a 78% combined selectiv-
ity (Table 1, run 1). The isomeric propenylbenzene, isoeugenol
2a, was responsible for almost all the rest of the mass balance.
The effect of the ligand concentration is illustrated by runs 1–4
in Table 1. It is remarkable that the increase in the P/Rh atomic
ratio from 3 to 20 has resulted in a strong enhancement of the
reaction rate as well as improved the chemoselectivity for the alde-
hydes to 97%. Although the results suggested that the TPPTS/Rh
ratio higher than 20 could be beneficial for the system, we have
decided to carry the studies with P/Rh = 10 to spare the ligand
during the screening of the conditions. The regioselectivity was
nearly 70% for linear aldehyde 3a within the whole range of ligand
In rhodium systems promoted by monodentate phosphorus(III)
ligands, the active species shown in Eqn. 1 are formed. Consid-
ering the geometric isomers of the square–planar [RhH(CO)(L)2]
complex, four active species can be formed, with their concen-
trations being depending on both the CO pressure and ligand L
concentration. The activity and selectivity of the system will be
the combination of the activities and selectivity of the individual
species taking into account their concentrations.
(1)
concentrations used, with ␣-aldehyde 5a being formed in very
small amounts.
It is important to note that the performance of the Rh/TPPTS
biphasic system was comparable to that of the conventional
At low ligand concentrations, the equilibrium in Eq. (1) tends
to the left, and in average the rhodium species tend to have less
electron density. Thus, the hydrogen oxidative addition will be
more difficult and, as a consequence, the hydroformylation rate will