Hydroformylation of 1-octene in supercritical carbon dioxide and organic solvents using trifluoromethyl-substituted triphenylphosphine ligands

Article abstract of DOI:10.1002/adsc.200606150

Two different in situ prepared catalysts generated from Rh(CO) ₂acac and trifluoromethyl-substituted triphenylphosphine ligands have been evaluated for their activity and selectivity in the hydroformylation of 1-octene. The solvents used were supercritical carbon dioxide, hexane, toluene, and perfluoromethylcyclohexane. The highest value for the turnover frequency, 9820 mol_1-octene mol_Rh^-1 h^-1, has been obtained in supercritical carbon dioxide using ligand I, P[C ₆H₃-3,5-(CF₃)₂]₃. For both supercritical carbon dioxide and hexane employing ligand II, P(C ₆H₄-3-CF₃)₃, a selectivity towards the linear aldehyde product, nonanal, and an n:iso ratio of 79.3% and 4.6-4.8 have been obtained, respectively. These values are significantly higher than those obtained with triphenylphosphine as ligand (nonanal: 74-76%, n:iso: 3.1-3.3). An increase in trifluoromethyl substitution on the triphenyl ligand results in an increase in the 1-octene conversion rate, an increase in the n:iso ratio and a decrease in the overall selectivity towards aldehydes. In terms of turn-over frequency and selectivity the three ligands give comparable results in supercritical carbon dioxide and hexane. This leads to the conclusion that the properties of supercritical carbon dioxide as a solvent for hydroformylation can be better compared with those of hexane rather than with those of toluene.

Full text of DOI:10.1002/adsc.200606150

COMMUNICATIONS
DOI: 10.1002/adsc.200606150
Hydroformylation of 1-Octene in Supercritical Carbon Dioxide
and Organic Solvents using Trifluoromethyl-Substituted
Triphenylphosphine Ligands
a
b
a,
Ard C. J. Koeken, Michiel C. A. van Vliet, Leo J. P. van den Broeke, *
b
a
Berth-Jan Deelman, and Jos T. F. Keurentjes
Process Development Group, Eindhoven University of Technology, P. O. Box 513, 5600 MB Eindhoven, The Netherlands
a
Fax: (+31)-40-244-6104; e-mail: L.J.P.van.den.Broeke@tue.nl
Arkema Vlissingen B. V., P. O. Box 70, 4380 AB Vlissingen, The Netherlands
b
Fax: (+31)-113-612-984
Received: March 27, 2006; Accepted: June 2, 2006
Abstract: Two different in situ prepared catalysts
one of the most important examples of homogeneous
catalysis applied on an industrial scale.
generated from Rh(CO) acac and trifluoromethyl-
2
substituted triphenylphosphine ligands have been
evaluated for their activity and selectivity in the hy-
droformylation of 1-octene. The solvents used were
supercritical carbon dioxide, hexane, toluene, and
perfluoromethylcyclohexane. The highest value for
Commercial hydroformylation processes are com-
monly carried out in a gas-liquid system or in a gas-
liquid-liquid system. Rigorous mixing of the phases is
required to reduce or prevent mass transfer limita-
tions. The Ruhrchemie/Rhône–Poulenc (RCH/RP)
hydroformylation process, in which propene is con-
verted to butanal using a phosphine-modified rhodi-
um catalyst, is an example of homogeneous catalysis
in a gas-liquid-liquid system. In this process, the sul-
fonated phosphine renders the catalyst preferentially
soluble in the aqueous phase, facilitating easy catalyst
separation and recovery through liquid-liquid phase
separation. By this means one of the major disadvan-
tages of homogeneous catalysis, which is the trouble-
some recovery of the catalyst, is overcome.
À1 À1
the turnover frequency, 9820 mol1-octene molRh h ,
has been obtained in supercritical carbon dioxide
using ligand I, P[C H -3,5-(CF ) ] . For both super-
A
H
R
U
G
6
3
3 2 3
[
2]
critical carbon dioxide and hexane employing ligand
II, P(C H -3-CF ) , a selectivity towards the linear
6
4
3 3
aldehyde product, nonanal, and an n:iso ratio of
9.3% and 4.6–4.8 have been obtained, respectively.
7
These values are significantly higher than those ob-
tained with triphenylphosphine as ligand (nonanal:
74–76%, n:iso: 3.1–3.3). An increase in trifluoro-
methyl substitution on the triphenyl ligand results in
an increase in the 1-octene conversion rate, an in-
crease in the n:iso ratio and a decrease in the overall
selectivity towards aldehydes. In terms of turn-over
frequency and selectivity the three ligands give com-
parable results in supercritical carbon dioxide and
hexane. This leads to the conclusion that the proper-
ties of supercritical carbon dioxide as a solvent for
hydroformylation can be better compared with those
of hexane rather than with those of toluene.
Research into hydroformylation reactions has fo-
cused on ligand design to make the catalyst complex
more selective, more active, or better separable from
[
3]
the product. To facilitate catalyst separation phos-
phine ligands have been covalently attached to a sup-
port, such that the catalyst complexes form a separate
[
4–8]
[9]
solid
or liquid phase, which can be recovered by
filtration or decantation, respectively. Another strat-
[2]
egy is to design ligands that dissolve in an aqueous ,
[
10–13]
[14,15]
fluorous
, or ionic
solvent that is immiscible
with an organic phase containing the product.
Keywords: fluorinated ligands; homogeneous catal-
ysis; 1-octene hydroformylation; organic solvents;
supercritical carbon dioxide
The RCH/RP process is limited to the conversion
of short-chain alkenes, such as propene and 1-butene,
because the solubility of longer-chain alkenes in water
is too low to realize an acceptable reaction rate. Cur-
rently, industrial hydroformylation of long-chain al-
kenes, including 1-octene, is carried out using a cobalt
catalyst. Although the cobalt-based processes have
Supercritical carbon dioxide, scCO , has been estab- been optimized over the years, relatively harsh condi-
2
lished as an environmentally benign alternative to or- tions, typically 30MPa and 1758C for the Exxon pro-
[
1]
ganic solvents in the field of homogeneous catalysis.
cess, are required to keep the catalyst active and
In terms of production volume, hydroformylation is stable, while the activity and selectivity for linear al-
Adv. Synth. Catal. 2006, 348, 1553 – 1559
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1553

Ard C. J. Koeken et al.
COMMUNICATIONS
dehyde are lower than in rhodium-catalyzed hydrofor-
[
2]
mylation. Employing a rhodium catalyst and scCO2
as the reaction medium could be advantageous for
the hydroformylation of longer-chain alkenes when
combined with a facile recovery of the expensive rho-
dium catalyst. The use of scCO has the additional ad-
2
vantage that it reduces mass transfer limitations, re-
sulting in a more efficient use of reactor volume at
elevated process pressure. Moreover, the develop-
ment in the synthesis of bulky phosphites and diphos-
phines has led to improved activity and regio- or ste-
reoselectivity for rhodium-catalyzed hydroformylation
[
16]
reactions.
Furthermore, it has been shown that a
homogeneous catalyst can be recovered using a mi-
croporous ceramic membrane in combination with
[17,18]
scCO as the reaction medium.
2
A number of researchers have illustrated the poten-
tial of scCO as an alternative reaction medium for
2
Figure 1. Reaction scheme for the hydroformylation of 1-
octene (1a), with the two main products nonanal (2a) and 2-
methyloctanal (2b). The side products are (E,Z)-2-octene
the hydroformylation long-chain alkenes, such as 1-
[
19–21]
octene.
Besides the requirement of a good selec-
tivity and activity, it is important that the catalyst is
(
1b, 1c), (E,Z)-3-octene (1d, 1e), (E,Z)-4-octene, 2-ethylhep-
[
19]
also sufficiently soluble in the supercritical mixture.
tanal (2c), 2-propylhexanal (2d), and n-octane (3).
Generally, it is concluded that the hydroformylation
rates in scCO are comparable or even exceed the
2
rates observed in organic solvents. It has been demon-
strated that the attachment of perfluoroalkyl substitu-
ents on a ligand improves the solubility in scCO con-
2
[
19,22]
siderably.
Additionally, the attachment of per-
fluoroalkyl substituents on the aryl rings of triphenyl-
[
23]
phosphine has led to an improved catalytic activity.
In terms of activity and selectivity the comparison
between an organic solvent and scCO as a reaction
2
medium for hydroformylations has only been made to
a limited extent. In general, toluene is used to bench-
Figure 2. The three triphenylphosphine analogues used. I=
[21,24]
mark the performance in scCO2.
In this work the
tris[3,5-bis(trifluoromethylphenyl)]phosphine, II=tris(3-tri-
A
H
R
U
G
hydroformylation of 1-octene (1a), see Figure 1, is fluoromethylphenyl)phosphine, III=triphenylphosphine.
evaluated for four different solvents, including scCO2,
n-hexane, toluene, and perfluoromethylcyclohexane.
Also, the effect of ligand modification is addressed.
The batch reactors used did not allow for a visual
Beside triphenylphosphine, two different monoden- inspection of the mixture during the catalyst pre-for-
tate triarylphosphine analogues with trifluoromethyl mation or the catalytic reaction. Prior to the reaction
substituents on the meta positions were used for hy- in scCO , the catalyst complex was formed in situ
2
droformylation of 1a. The chemical structures of the from Rh(CO) acac (acac=acetylacetonate) and one
2
three phosphines used are shown in Figure 2.
of the phosphine analogues in the presence of CO, H2
First, the hydroformylation in supercritical carbon and CO (under the conditions as stated in Table 1).
2
dioxide is discussed. In all the experiments a high The reaction was started upon addition of 1a. We em-
À1
ligand to rhodium ratio was used, in the order of 50:1, ployed a CO2 density of 0.63 gmL , which was
which approaches conditions used on an industrial higher than the density used by Koch and Leitner,
[
21]
scale. A high ligand to rhodium ratio usually results who observed a one-phase mixture.
Furthermore,
in an increase in the selectivity towards the linear employing a higher density and a higher temperature
[
2]
product. Also, we chose a relatively high reaction at the same concentration of solvent and reactants in-
temperature of 708C to ensure a high catalytic activi- creases the likelihood that a one-phase mixture is
[
25]
ty. Most results reported for the hydroformylation of present during the progress of the reaction.
The
1
-octene in scCO2 were obtained at temperatures maximum initial total pressure obtained after addition
[20,21]
lower than 708C.
In Table 1 an overview is given of 1a was ca. 50MPa. This corresponds to a total reac-
À1
of the experimental reaction conditions for the case tion mixture density of approximately 0.78 gmL .
scCO or an organic solvent is used.
The total concentration (the sum of 1a and its reac-
2
1554
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1553 – 1559

Hydroformylation of 1-Octene
COMMUNICATIONS
Table 1. Overview of experimental conditions.
Symbol Unit
ScCO2 Organic sol-
vent
Reactor tempera-
ture
T
[8C]
70
70
[
a]
Reactor pressure
Reactor volume
Mode of operation[
Amount of solvent
P
V
A
H
R
U
G
1
[mL]
107.6 100
batch semi-batch
87
b]
[c]
[d]
[mL]
21
25.0
1.0
[b]
Amount of 1a
A
H
E
N
[
e]
Amount of syngas
PCO/H2
ACHTREUNG
Amount of
Rh
A
T
E
N
12.5
[b]
Rh(CO) acac
2
b]
[
Ligand
L
A
H
R
N
[mmol] 2.75
0.625
[
[
[
[
a]
b]
c]
d]
Maximum pressure reached after injection of 1a.
Average value.
At 208C and 6.0MPa.
20 mL solvent plus 1.07 mL (5.5 mmol) n-decane as in-
ternal standard.
Molar ratio CO:H =1:1.
At 208C. This pressure corresponds with 110 mmol of
[
[
e]
f]
2
syngas.
tion products, excluding CO and H , as measured
2
using off-line GC) became constant after approxi-
mately 20 min and then did not deviate more than
1
2% from the initial concentration based on the
amount of 1a added to the reactor vessel. Further-
more, the homogeneity of the reaction mixture was
confirmed by taking samples at different locations,
i.e., the top and bottom, of the reactor. Considering
Figure 3. (a): Normalized concentrations of the main reac-
tant and products and (b) normalized concentrations of iso-
mers as a function of time for the experiment with ligand I.
these arguments it is likely that during the reactions Experimental conditions: T=708C, [1a] =0.98 mol/L,
0
3
there is a one-phase reaction system.
[1a] :[Rh]=2.0·10 , [CO ]=14.5 mol/L, [CO] =[H ] =1
0
2
0
2 0
The normalized concentrations of the main reactant mol/L.
and products as a function of time for the experiment
conducted in scCO with I as the ligand are depicted
2
in Figure 3, panel a. In Figure 3, panel b the normal-
ized concentrations of the remaining octene isomers,
aldehyde isomers, and octane are given. The catalyst
derived from I gave rise to a significant amount of
isomers of 1a, resulting in a maximum concentration
of about 6% and 5.5% for 1b and 1c after approxi-
mately 0.5 h of reaction, respectively. It can be seen
that after almost complete conversion of 1a the con-
centration of 2a remains constant. However, as a
result of the hydroformylation of 1b and 1c the con-
centrations of 2b and 2c still increase, respectively.
After 2 h the hydroformylation product 2d also starts
to appear.
In Figure 4 the conversion of 1a is plotted as a func-
tion of time for the reaction in scCO for the differ-
2
Figure 4. Conversion of 1-octene as a function of time using
I, II, and III with an L:Rh ratio of 50:1. The markers indi-
cate the times at which samples were taken. Experimental
ent catalytic systems. No noticeable induction time
has been observed at the sampling frequency used.
Therefore, the slope of the conversion profile, the
conditions:
T=708C,
[1a] =0.974Æ0.005
mol/L,
0
3
straight lines in Figure 4, is a measure of the initial ac- [1a] :[Rh]=2.0·10 , [CO ]=14.3Æ0.2 mol/L, [CO] =[H ] =
0
2
0
2
0
tivity. Although III is significantly less soluble in 1 mol/L.
Adv. Synth. Catal. 2006, 348, 1553 – 1559
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1555

Ard C. J. Koeken et al.
COMMUNICATIONS
In Figure 5 the overall selectivity towards aldehydes
is given as a function of time. Davis and Erkey have
reported a minimum overall selectivity of 95% to-
wards aldehydes at a TOFof
ca. 1500 mol1a
À1 À1
molRh
h
by applying I, with L:Rh=3:1, at a tem-
[26]
perature of 508C at similar reactant concentrations.
We obtained an overall selectivity of 90.5% towards
À1 À1
aldehydes at a TOFof 9820 mol 1a molRh h . This
difference can be explained by the higher tempera-
ture applied in our experiments, which leads to more
[
2]
isomerization of 1a. The catalyst derived from I
shows a good activity for the hydroformylation of in-
ternal alkenes, 1b and 1c. This has also been observed
in a study on the hydroformylation of dienes where
[
24]
the same catalyst was used.
To be able to make a comparison with the results in
scCO , the three ligands have also been used to pre-
Figure 5. The overall selectivity for aldehydes as a function
of time using I, II, and III with an L:Rh ratio of 50:1. The
markers indicate the times at which samples were taken. Ex-
2
perimental conditions: T=708C, [1a] =0.974Æ0.005 mol/L, pare the catalysts, in situ, for the hydroformylation of
0
3
[
1a] :[Rh]=2.0·10 , [CO ]=14.3Æ0.2 mol/L, [CO] =[H ] = 1a in the organic solvents toluene, hexane, and the
0
2
0
2 0
1
mol/L.
fluorinated
pfmch). In this case the reactions were performed
under a constant syngas pressure (molar ratio
solvent
perfluoromethylcyclohexane
(
scCO than I or II, a reasonable activity is observed H :CO=1:1) of 1.0MPa, whereas during the reac-
2
2
À1 À1
for this ligand. A TOFof 470 mol 1a molRh
h
was tions in scCO2 no additional syngas was added
obtained. The number of trifluoromethyl groups per (4.9MPa initial syngas pressure, a near stoichiometric
ligand follows the order: I>II>III. Figure 4 shows amount). In Table 2a detailed comparison is made of
that the activity of the corresponding catalytic system the reaction rates and selectivities obtained for the
follows the same sequence. The highest activity, a three catalyst systems in the four solvents. Additional-
À1 À1
TOFof 9820 mol 1a molRh h , was observed for the ly, results are given for experiments in toluene where
catalyst derived from I.
an L:Rh ratio of 100 was used.
Table 2. Comparison of different ligands and solvents.
[
a]
[b]
[b,c]
[b,d]
Entry Solvent Ligand [1a]0
[1a] :[Rh]
L:Rh TO3F
X
n:iso
S
0
3
À1
(
mol/L) (10 )
(10 h )
(%)
2
a
2a–d
1b–g
3
(%)
(%)
(%)
(%)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
scCO2
scCO2 II
scCO2 III
toluene
toluene II
toluene III
toluene
toluene II
toluene III
hexane
hexane II
hexane III
I
0.98
0.97
0.97
0.96
0.96
0.96
1.0
1.0
1.0
1.0
1.0
2.0
1.9
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
50
50
51
50
50
9.82
1.33
0.47
5.90
3.66
1.15
5.90
1.70
0.90
8.96
1.80
1.12
0.55
1.20
<0.04
99.2
91.7
55.3
98.1
97.9
93.1
98
3.33
4.59
3.33
3.7
4.0
2.8
4.5
5.3
3.1
3.9
4.8
3.1
2.4
3.2
1.5
69.6
79.3
76.0
71.5
76.0
73.7
75
90.5
96.6
98.8
90.6
95.1
98.6
91
6.1
2.9
<1
7.5
4.2
<1
7
3.3
<1
<1
1.9
<1
<1
1
I
50
I
100
100
100
50
50
50
50
50
50
97
89
80
74
95
97
4
2
<1
<1
1.0
<1
<1
5.2
<1
0
1
2
3
4
5
I
98.7
98.9
90.6
91.5
99.3
9.2
74.4
79.3
74.0
28.9
72.2
9.5
93.7
95.7
98.0
41.1
94.7
15.7
5.3
4.1
1.5
53.6
4.6
84.3
1.0
1.0
1.0
1.0
[e]
pfmch
pfmch II
pfmch III
I
[f]
nd
[
a]
The initial turn-over frequency, TOF, is given by the product of the slope of the linear fit and the (initial) substrate to cat-
alyst ratio.
[
[
[
[
[
b]
Conversion (X), the n:iso ratio, and the selectivity (S) shown here are at approximately 3 h of reaction.
n:iso=[2a]/([2b]+[2c]+[2d]).
The overall selectivity for a product(s) at a certain time: S (t)=[i](t)/
Perfluoromethylcyclohexane.
The 3 signal overlaps with the 1a signal.
c]
d]
e]
f]
A
H
R
U
G
i
0
1556
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1553 – 1559

Hydroformylation of 1-Octene
COMMUNICATIONS
For the hydroformylation in scCO the parent phos- a slightly lower selectivity towards aldehydes at a con-
2
phine, triphenylphosphine III, offers a better total al- version of 39.6%. The observed differences in activity
dehyde selectivity than I or II, but this is at the cost and selectivity are probably a result of the higher
of a lower activity and, when compared to II, a lower temperature and the lower initial pressure used than
n:iso ratio resulting in less linear aldehyde. For I the in our experiments. This could have resulted in a
total aldehyde selectivity drops in the early phases of lower solubility for III and its corresponding catalyst
the reaction only to recover slowly over the course of complex. Kainz et al. have reported a conversion of
the reaction. This is caused by the formation of iso- 26% and an n:iso ratio of 3.5 after 19 h for the hydro-
mers of 1a and subsequent hydroformylation of these formylation of 1a with a concentration of 0.48 mol/L
4
isomers to aldehydes. The observed trend for the ac- at 608C using III and Rh
A
H
R
U
G
A
H
R
U
G
8
12
tivity as a function of the number of trifluoromethyl 6:1 ratio.[19] The density of the mixture applied there
substituents could be a result of differences in solubil- was approximately 0.75 gmL , which is similar to the
ity of the different ligands. It can be expected that the density applied in our case. The experimental condi-
solubility increases with the number of trifluorometh- tions we applied, a high density combined with a rela-
yl groups on the ligand. For III a solubility of 0.101 tively high temperature, can account for the higher
mol/L in pure carbon dioxide at 25.2MPa and 578C activity of III/Rh(CO) acac observed as compared to
2
[
22]
was reported.
However, the presence of carbon these previously reported values.
From Table 2 it follows that, despite the experimen-
monoxide and hydrogen could have an opposite anti-
solvent effect. The amount of phosphine applied cor- tal differences, the results obtained in scCO and the
2
responds to a concentration of 0.025 mol/L assuming organic solvents hexane and toluene are comparable,
total solubility. This value is four times lower than the while the fluorous solvent pfmch gives rise to inferior
reported solubility value; this implies that III is most results. Clearly, the three catalysts systems employed
likely completely dissolved under the conditions ap- are not fluorous enough to completely dissolve in
plied. The addition of 1a will probably improve the pfmch. For I/Rh(CO) acac the highest TOFs are ob-
2
solubility of the ligand and its corresponding complex served in scCO and hexane. For II/Rh(CO) acac and
2
2
due to co-solvent effects. The observation that there III/Rh(CO) acac the activity order scCO <hexane<
2
2
is no, or very little, induction time is another indica- toluene and scCO <hexane ꢀ toluene, respectively,
2
tion that the catalyst is completely dissolved before is observed. The trends in n:iso ratio as well as the
the addition of 1a. Therefore, it is concluded that the total aldehyde selectivity are essentially reproduced
observed reaction rates represent the real activity of in all 4 solvent systems. The highest n:iso selectivity
the complex. Furthermore, comparable activities with towards 2a is always obtained with II in scCO and
2
the same dependence on the ligand were found in hexane. It can be seen in Table 2 that in toluene this
hexane, supporting this conclusion.
high selectivity with II/Rh(CO) acac can only be
2
The dependence of the activity on the number of reached when the amount of ligand is increased to an
fluorine atoms in the phosphines corresponds with re- L:Rh ratio of 100. For the total selectivity towards
sults obtained by Palo and Erkey (508C, L:Rh=3:1, the aldehydes, 2a–d, III gives the best result. The for-
P=27.3MPa) but absolute activities were significantly mation of octene isomers and their successive hydro-
lower under these conditions (TOF=500–900 mol1a formylation has an unfavorable effect on the n:iso
À1 À1 [23]
molRh h ).
Also, they obtained relatively low ratio. For the different solvents, scCO , hexane, and
2
n:iso ratios in the range 3.0–3.3 for I, II, and P(C H - toluene, similar trends in activity and selectivity are
6
4
4
-CF ) while the concentration of rhodium was com- observed, indicating that the extent and location of
3 3
parable to that in our experiments. The optimum in trifluoromethylation, rather than difference in solubil-
n:iso ratio obtained for II, both in scCO (n:iso=4.6) ity are responsible for the variations observed.
2
and organic solvent (n:iso=4.0–4.8) can thus be con-
The hydroformylation in scCO2 was performed
sidered as quite remarkable. Moreover, this indicates batch-wise, while for the organic solvents the reaction
that the n:iso ratio becomes sensitive to variation in was performed semi-batch. As can be derived from
perfluoralkylation at higher L:Rh ratios, while the Table 2 the difference in the mode of operation has a
higher activity obtained with tris(trifluoromethylaryl)- small effect on the overall outcome of the reaction.
phosphines is essentially maintained.
The fact that the syngas pressure on the reaction rate
Few results have been reported on the use of III as is almost zero-order seems to be a plausible explana-
[
2]
a ligand for homogeneous hydroformylation catalysts tion for this phenomenon. Furthermore, the effects
in scCO because of its limited solubility and poor ac- of the partial pressure of hydrogen and carbon mon-
2
tivity. Sellins and co-workers have used III for the hy- oxide are opposite, and it is to be expected that for an
droformylation of 1a in scCO at 1008C using equimolar syngas mixture the partial pressure effects
2
[
27]
Rh
A
C
H
T
R
E
U
N
G
(OAc) as the metal precursor with L:Rh=3.3.
will cancel each other out.
2
4
They obtained a low activity, (TOF=104 mol
molRh h ) but a comparable n:iso ratio of 3.4 and cluded in terms of the effect of the solvent and the
From the different results the following can be con-
aldehyde
À1 À1
Adv. Synth. Catal. 2006, 348, 1553 – 1559
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1557

Ard C. J. Koeken et al.
COMMUNICATIONS
effect of the ligand. The activity increases up to one An hour in total, 30 min at 508C and 30 min at 708C, was
considered sufficient to allow the active catalyst complexes
order of magnitude with the number of trifluorometh-
yl substituents attached to the meta positions on the
to be formed in situ from the Rh(CO)2
(acac) precursor and
A
H
R
U
G
the different fluorous ligands. The 1a in the feed pump was
then compressed to a pressure just above the pressure pres-
ent in the reactor. The reaction was subsequently initiated
by opening the valve between pump and reactor, causing a
fast pressure equalization, and by consecutively pumping
the desired volume of 1a into the reactor, which as a rule
did not take more than 30 s. Initially, high pressure samples
aryl rings for scCO , hexane and toluene in the order
2
I>II>III. The highest activity was found for I, using
scCO . Similar ligand-dependent trends in activity
2
and selectivity were observed for the hydroformyla-
tion in scCO , toluene and hexane, with the results in
2
hexane matching closest those obtained in scCO2.
The catalyst II/Rh(CO) acac gives the best compro- were taken every 10–30 min. At higher conversions the sam-
2
mise between activity and selectivity towards the pling frequency was reduced to 1 per h. The content of the
linear aldehyde. Conditions were found, using hexane sample volume was carefully bubbled through a vial with a
solution of n-decane in toluene, and after depressurization
and scCO2 as
a reaction medium, where II/
the sample volume was rinsed with additional toluene solu-
tion to collect 1a and its reaction products quantitatively.
Subsequently the sample volume was dried by alternately
applying an argon flow and vacuum. A reaction time of 3 h
was observed.
Rh(CO) acac gives a 2.5-fold higher activity and a
2
5
% higher selectivity for the linear aldehyde than III/
Rh(CO) acac. A similar selectivity for 2a with II/
2
Rh(CO) acac can also be reached using the solvent
2
toluene. However, this required a much higher L:Rh
ratio of 100. The combined selectivity towards the al-
dehydes (linear and branched as well as other iso-
mers) in scCO hexane, and toluene follow the order: Organic Solvent Hydroformylation Procedure
2
,
III>II>I. The higher 1-alkene isomerization activity
of the catalyst systems resulting from ligands I and II Rh(CO)2(acac) (12.5 mmol; S/C=2000; 1 mL stock solution
is responsible for this effect. The difference in experi- in toluene) and ligand (0.625 mmol; 50 equivs.) were mixed
A
H
R
U
G
in 10 mL of degassed toluene and stirred for 0.5 h at 678C
under an atmosphere of 0.8 MPa syngas. Then 1a and n-
decane (25 + 5.5 mmol; 5.0 mL of a standard mixture) were
mental procedure for scCO and the organic solvents
2
has a small effect on the overall performance of the
hydroformylation of 1a. This leads to the general con-
added via a short column of alumina, followed by degassed
clusion that hexane, rather than toluene, best matches
toluene (10 mL). The mixture was pressurized with 10 bar of
:1 synthesis gas and stirred at 708C. Additional syngas was
the solvent properties of scCO in the context of hy-
2
1
droformylation catalysis.
added to maintain the pressure at 1Æ0.1 MPa. Samples
were taken at regular intervals. Conversions and selectivities
were determined after a reaction time of 3 h. The TOFwas
determined after 15 min.
For experiments with hexane or perfluoromethylcyclohex-
ane (pfmch) as solvent a slightly different procedure was
followed. Rh(CO)2(acac) (12.5 mmol; S/C=2000; 1 mL
stock solution in MTBE) and ligand (0.625 mmol; 50
equivs.) were mixed in 5 mL of degassed Et O and stirred
for 0.5 h at 708C under an atmosphere of 0.8 MPa syngas.
Volatiles were removed under vacuum and then 1a and n-
decane (25 + 5.5 mmol; 5.0 mL of a standard mixture) were
added via a short column of alumina, followed by degassed
hexane or pfmch (20 mL). The remainder of the experi-
ments was carried out following the procedure applied using
toluene as the solvent.
Experimental Section
Ligands I and II were supplied by Arkema (Vlissingen).
Ligand III was purchased from Aldrich. Rhodium(I) dicar-
A
H
R
U
G
bonylacetylacetonate [Rh(CO) acac] was obtained from
2
2
Fluka. Ligands I and II and Rh(CO) acac were stored under
2
argon; 1a was purchased from Aldrich, dried over molecular
sieves and stored under argon.
ScCO Hydroformylation Procedure
2
The general procedure for a hydroformylation experiment
started with conveying 54.8Æ0.6 mmol of Rh(CO) acac and
2
2
7.5Æ0.6 mmol of fluorinated phosphine ligand into the re-
Analysis
actor. After closing the reactor, the reactor volume was
carefully flushed with argon and evacuated three times.
Next, the stirrer was turned on with a revolution speed of
The samples were analyzed off-line using a Fisons Instru-
ments GF-FID equipped with a Restek Rtx-5 column (fused
silica, length 30 m, internal diameter 0.53 mm) where
helium was the carrier gas. Calibration was done for 1a, 3
and 2a. The sensitivity coefficients for the octene (1b–g) and
aldehyde isomers (2b–d) were taken to be equal to those of
7
00 rpm and the desired amount of carbon monoxide and
hydrogen gas were fed to the reactor at room temperature.
A total pressure of 4.9MPa was used to obtain a concentra-
tion of 1 mol/L for both CO and H . The reactor content
was then heated to a temperature of 508C and subsequently
CO was pumped into the reactor to a pressure of 26MPa.
2
1
a and 2a, respectively.
2
These conditions were maintained for at least half an hour
before heating further to the desired reaction temperature.
1558
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1553 – 1559

Hydroformylation of 1-Octene
COMMUNICATIONS
Acknowledgements
[13] A. Aghmiz, C. Claver, A. M. Masdeu-Bultó, D. Mail-
lard, D. Sinou, J. Mol. Catal. A 2004, 208, 97–101.
[14] P. B. Webb, M. F. Sellin, T. E. Kunene, S. Williamson,
A. M. Z. Slawin, D. J. Cole-Hamilton, J. Am. Chem.
Soc. 2003, 125, 15577–15588.
This work was supported by an EET grant from SenterNo-
vem (agency of the Dutch Ministry of Economic Affairs).
[
15] M. Solinas, A. Pfaltz, P. G. Cozzi, W. Leitner, J. Am.
References
Chem. Soc. 2004, 126, 16142–16147.
16] M. Beller, B. Cornils, C. D. Frohning, C. W. Kohlpaint-
er, J. Mol. Catal. A 1995, 104, 17–85.
[17] L. J. P. van den Broeke, E. L. V. Goetheer, A. W. Ver-
kerk, E. de Wolf, B.-J. Deelman, G. Van Koten, J. T. F.
Keurentjes, Angew. Chem. Int. Ed. 2001, 40, 4473–
4474.
[
[
1] P. G. Jessop, T. Ikariya, R. Noyori, Chem. Rev. 1999, 99,
75–493.
4
[2] C. D. Frohning, C. W. Kohlpainter, H.-W. Bohnen, in:
Applied Homogeneous Catalysis, 2nd edn., Vol. 1,
(Eds.: B. Cornils, W. A. Herrmann), Wiley-VCH, Wein-
heim, 2002, pp. 31–103.
[18] E. L. V. Goetheer, A. W. Verkerk, L. J. P. van den
Broeke, E. de Wolf, B.-J. Deelman, G. van Koten,
J. T. F. Keurentjes, J. Catal. 2003, 219, 126–133.
[19] S. Kainz, D. Koch, W. Baumann, W. Leitner, Angew.
Chem. 1997, 109, 1699–1701.
[
[
3] D. J. Cole-Hamilton, Science 2003, 299, 1702–1706.
4] N. J. Meehan, M. Poliakoff, A. J. Sandee, J. N. H. Reek,
P. C. J. Kamer, P. W. N. M. van Leeuwen, Chem.
Commun. 2000, 1497–1498.
[
[
[
[
[
5] A. J. Sandee, J. N. H. Reek, P. C. J. Kamer, P. W. N. M.
van Leeuwen, J. Am. Chem. Soc. 2001, 123, 8468–8476.
6] S. Bektesevic, T. Tack, M. R. Mason, M. A. Abraham,
Ind. Eng. Chem. Res. 2005, 44, 4973–4981.
7] I. Kani, R. Flores, J. P. Fackler, Jr., A. Akgerman, J. Su-
percritical Fluids 2004, 31, 287–294.
8] F. Shibahara, K. Nozaki, T. Hiyama, J. Am. Chem. Soc.
[
[
[
20] D. R. Palo, C. Erkey, Ind. Eng. Chem. Res. 1998, 37,
203–4206.
21] D. Koch, W. Leitner, J. Am. Chem. Soc. 1998, 120,
3398–13404.
22] K.-D. Wagner, N. Dahmen, E. Dinjus, J. Chem. Eng.
Data 2000, 45, 672–677.
4
1
[
[
23] D. R. Palo, C. Erkey, Organometallics 2000, 19, 81–86.
24] S.-I. Fujita, S Fujisawa, B. M. Bhanage, Y. Ikushima, M.
Arai, Eur. J. Org. Chem. 2004, 2881–2887.
2
003, 125, 8555–8560.
9] M. Solinas, J. Jiang, O. Stelzer, W. Leitner, Angew.
Chem. Int. Ed. 2005, 44, 2291–2295.
[
[
[
25] J. Ke, B. Han, M. W. George, H. Yan, M. Poliakoff, J.
[
10] B. Richter, A. L. Spek, G. van Koten, B.-J. Deelman, J.
Am. Chem. Soc. 2000, 122, 3945–3951.
11] D. J. Adams, J. A. Bennett, D. J. Cole-Hamilton, E. G.
Hope, J. Hopewell, J. Kight, P. Pogorzelec, A. M.
Stuart, Dalton Trans. 2005, 3862–3867.
12] D. F. Foster, D. Gudmunsen, D. J. Adams, A. M. Stuart,
E. G. Hope, D. J. Cole-Hamilton, G. P. Schwarz, P. Po-
gorzelec, Tetrahedron 2002, 58, 3901–3910.
Am. Chem. Soc. 2001, 123, 3661–3670.
26] T. Davis, C. Erkey, Ind. Eng. Chem. Res. 2000, 39,
[
3
671–3678.
27] M. F. Sellin, I. Bach, J. M. Webster, F. Montilla, V.
Rosa, T. Avilꢁs, M. Poliakoff, D. J. Cole-Hamilton, J.
Chem. Soc., Dalton Trans. 2002, 4569–4576.
[
Adv. Synth. Catal. 2006, 348, 1553 – 1559
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1559