Supported ionic liquid phase (SILP) catalyzed hydroformylation of 1-butene in a gradient-free loop reactor

Article abstract of DOI:10.1016/j.jcat.2009.02.024

The supported ionic liquid phase (SILP) catalysis technology was applied to gas-phase hydroformylation of 1-butene using sulfoxantphos 1 modified rhodium complexes. Kinetic experiments were performed in a fixed bed reactor and compared to a gradient-free gas-phase loop reactor (Berty type). The influence of substrate concentration, temperature and syngas pressure was determined. Data from fixed bed and Berty reactor were found to be in good agreement with respect to activation energy and reaction order. Ex-situ NMR studies of fresh and used SILP catalysts confirmed that the ligand remained intact after prolonged time on stream.

Full text of DOI:10.1016/j.jcat.2009.02.024

Journal of Catalysis 263 (2009) 321–327
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Supported ionic liquid phase (SILP) catalyzed hydroformylation of 1-butene
in a gradient-free loop reactor
∗
Marco Haumann , Michael Jakuttis, Sebastian Werner, Peter Wasserscheid
Lehrstuhl für Chemische Reaktionstechnik, Universität Erlangen-Nürnberg, Egerlandstr. 3, D-91058 Erlangen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The supported ionic liquid phase (SILP) catalysis technology was applied to gas-phase hydroformylation
of 1-butene using sulfoxantphos 1 modified rhodium complexes. Kinetic experiments were performed in
a fixed bed reactor and compared to a gradient-free gas-phase loop reactor (Berty type). The influence of
substrate concentration, temperature and syngas pressure was determined. Data from fixed bed and Berty
reactor were found to be in good agreement with respect to activation energy and reaction order. Ex-situ
NMR studies of fresh and used SILP catalysts confirmed that the ligand remained intact after prolonged
time on stream.
Received 21 November 2008
Revised 24 January 2009
Accepted 20 February 2009
Available online 19 March 2009
Keywords:
Hydroformylation
Ionic liquid
Immobilization
Support
© 2009 Elsevier Inc. All rights reserved.
Kinetics
1. Introduction
The immobilization of precious homogeneous catalysts is a field
of intense research and several concepts have been developed dur-
ing the last two decades [1]. Basically two strategies can be found
in literature; the immobilization of the homogeneous catalyst in a
second phase (aqueous, fluorous, micellar, ionic liquid, supercriti-
cal CO2) or the heterogenization of the homogeneous complex by
means of attachment to a support [2]. The novel supported ionic
liquid phase (SILP) technology makes use of both strategies by dis-
solving a homogeneous catalyst in an ionic liquid which itself is
highly dispersed on the internal surface of a porous material as de-
picted in Fig. 1 [3]. The resulting ionic liquid catalyst film is only a
few nanometers thick and allows complete utilization of ionic liq-
uid and catalyst since the mass transport resistance from the gas
into the liquid phase is minimized [4]. The very low volatility of
the common ionic liquids allows these catalyst systems to be used
in continuous gas phase processes and time on stream can exceed
700 h without deactivation of the catalyst [5].
However, from a reaction engineering perspective the kinetic
data determined in a fixed bed reactor design at high conversions
are falsified due to the concentration gradient of the substrate
along the catalyst bed. In order to determine kinetic data from
such an integral reactor either elaborate data analysis and pa-
rameter estimation or sampling along the reactor axis, normally
accompanied by elaborate sampling procedures, are necessary. In
contrast, in a gradient-free reactor the kinetic data can be de-
Fig. 1. Schematic representation of the heterogenization of a transition metal com-
plex via supported ionic liquid phase (SILP) technology.
termined directly from inflow and outflow values. For gas phase
reactions, a reactor design can be applied as shown in Fig. 2 that
has been introduced by Berty in 1974 [6].
A turbine on top of the reactor vessel ensures circulation of the
inlet flow through the catalyst bed, placed in a holder in the mid-
dle of the vessel. The circulating flow around the catalyst bed (loop
flow) can be orders of magnitude higher than the flow through the
Corresponding author. Fax: +49 9131 8527421.
*
E-mail address: marco.haumann@crt.cbi.uni-erlangen.de (M. Haumann).
0
021-9517/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.02.024

322
M. Haumann et al. / Journal of Catalysis 263 (2009) 321–327
Fig. 2. Left: cross-section of a Berty reactor designed and manufactured in the workshop at CRT, Erlangen. 1: Reactor inlet; 2: reactor vessel; 3: catalyst basket; 4: turbine;
5: reactor outlet. Right: schematic representation of the flows inside the Berty reactor and resulting rate equation.
The ionic liquids were purchased from Solvent Innovation GmbH
now part of Merck KGaA). The water content was measured with
(
Karl Fischer titration and the ionic liquids were dried under high
◦
vacuum at 50 C overnight.
2.2. SILP catalyst preparation
0.0512 g (0.2 mmol) of Rh(CO)2(acac) was dissolved in 20 ml
dried MeOH and stirred for 10 min. 1.57 g (2 mmol) of ligand
sulfoxantphos was added and the orange solution was stirred for
another 10 min. Afterward, 1 ml (1.06 g) of ionic liquid was added
to the solution. After stirring for 30 min, 10 g calcined silica was
added and the solution was stirred for 60 min. The MeOH was
removed in vacuo and a pale red powder was obtained. The sup-
ported ionic liquid phase catalyst was stored under argon until
further use.
Scheme 1. Hydroformylation of 1-butene with sulfoxantphos 1 modified Rh-SILP
catalysts.
catalyst bed, thus resulting in CSTR-like gradient-free concentration
throughout the reactor. In such a system, the outlet concentration
is representative for the concentration inside the catalyst bed and
the kinetic data is not falsified by concentration gradients as given
in Fig. 2.
We studied the hydroformylation of propene [3–5] and 1-
butene [7] using rhodium based SILP catalysts in a fixed bed
reactor. Interestingly, the catalyst activity with 1-butene as sub-
strate was twice as high compared to propene under identical
conditions [7]. This was attributed to the higher solubility of 1-
butene, resulting in twice the substrate concentration in the ionic
liquid film.
2.3. Kinetic experiments
The continuous reactor setup (PFTR) for the SILP catalyzed hy-
droformylation has been described in detail previously [6]. The
Berty reactor was in-house build and consisted of a 250 ml stain-
less steel vessel with a feed inlet at the bottom and an outlet at
the side of the vessel. The catalyst basket is centered via three
bolts into the middle of the reactor vessel. The turbine is fixed
to the reactor lid and can thus be removed completely from the
vessel for catalyst replacement. The supported ionic liquid phase
(SILP) catalyst was filled into the reactor and the complete rig was
evacuated at room temperature. The rig was pressurized with 50
bar helium and left under pressure for 30 min while monitoring
the pressure. If no pressure drop was observed, the reactor was
heated to reaction temperature under helium pressure. The com-
plete setup was evacuated and flushed with helium three times
before syngas and 1-butene were fed into the system. 1-butene
was taken out of a reservoir in the liquid state and fed into a
heated evaporator via a HPLC pump (Knauer WellChrom K-120,
In order to investigate this higher activity in more detail, we
studied the 1-butene hydroformylation with sulfoxantphos 1 mod-
ified rhodium complexes (Scheme 1) in a gradient-free loop reactor
(Berty reactor) designed and manufactured in the Erlangen work-
shop.
2. Materials and methods
−1
1
0 ml min ) to control the molar flow of the substrate. Both car-
2
.1. Chemicals
bon monoxide and hydrogen flows were adjusted by means of
mass-flow controllers (MFC, 5850 S series, Brooks Instruments).
The preheated gases were combined with the 1-butene in the
mixing unit which was filled with glass beads in order to en-
sure proper mixing and isothermal conditions. The gas mixture
could then either enter the reactor or exit the system via a bypass.
A back pressure regulator valve (Samson Type 3277) was used to
maintain the desired reaction pressure and outlet gas flow. After
the regulator valve, the gas stream was split and a minor flow
was passed through a 134 μl sampling loop mounted on a gas
All syntheses were carried out using standard Schlenk tech-
niques under argon (99.999%). Rh(CO2)(acac), xanthene and MeOH
HPLC grade) were purchased from Aldrich and used without fur-
(
ther purification. The synthesis of sulfoxantphos ligand 1 was car-
ried out according to literature procedures [8]. Silica gel 100 (63
to 200 μm) was purchased from Merck KGaA and was thermally
pre-treated at 450 C for 24 h. Carbon monoxide (99.97%), hydro-
gen (99.99%) and 1-butene (99.5%) were purchased from Linde AG.
◦

M. Haumann et al. / Journal of Catalysis 263 (2009) 321–327
323
◦
Fig. 3. Left: cumulative residence time function F(Θ) as a function of turbine speed. Right: dimensionless Bodenstein number as a function of turbine speed. 25 C, atmo-
−1
spheric pressure, 175 ml min
helium flow, 250 μl pulse injection of propane.
chromatograph (HP 5890 II plus). Samples were taken at regular
intervals by injecting the volume of the sampling loop via a 6-port
valve (Vici AT60 series) into the GC.
some dead volume in the inlet and outlet tubing. With increasing
turbine speed the behavior resembled an ideal back-mixed sys-
tem of a CSTR. The dimensionless Bodenstein number was used
as an indication of the degree of back-mixing in the reactor with
Bo = u L/D where u0 is the flow velocity in the empty reactor,
2
.4. Gas chromatography
0
ax
L is the reactor length and Dax is the axial dispersion coefficient.
According to literature procedure, the Bo-number was calculated
from the slope of the graphs at the Θ = 1 position [9]. Bodenstein
numbers as a function of turbine speed are shown in the inlet
The conversion of 1-butene as a function of process condi-
tions was measured using on-line GC technique. A HP 5890 GC
equipped with a Pona column (50 m, 0.2 mm diameter, 0.25 μm
coating) and a flame ionization detector (FID) pre-calibrated for
−1
of Fig. 3. Between 0 and 6000 min
a constant decrease of the
1
-butene, n-pentanal and iso-pentanal (allowing the peak areas
Bodenstein number is observed, indicating a better degree of dis-
−1
to be transferred into 1-butene conversion) was applied: injec-
tor temperature 200 C, split ratio 43:1, helium carrier gas flow
persion in the reactor. Between 6000 and 10000 min
a constant
◦
Bodenstein number of 1 was obtained. This low value is indicative
of quasi-ideally back-mixed systems with no concentration gradi-
ent inside the reactor. All further experiments were conducted at
−1
◦
2.4 ml min , detector temperature 250 C. To detect possible high
boiling by-products (heavies), the following temperature program
was used: initial temperature 50 C, initial time 5 min, heating
ramp of 50 C min , final temperature 150 C, final time 3 min,
cooling ramp 50 C min , final temperature 50 C.
◦
−1
6000 min . For an identical Berty setup the ideally back-mixed
◦
−1
◦
−1
regime has been determined to start around 6000 min
hydrogenation of cyclohexene [10].
in the
◦
−1
◦
2
.5. NMR spectroscopy
3.2. Residence time variation
◦
Hydroformylation of 1-butene was studied at 100 C and 10 bar
The spectra of the fresh ligand solution and the solutions ob-
−1
tained by washing the SILP catalysts after reaction were analyzed
in d6-DMSO using a Jeol ECX 400 MHz spectrometer at room tem-
perature. 31P spectra were recorded at 161.8 MHz.
total syngas pressure (CO/H2 ratio 1:1). At 6000 min
the resi-
dence time inside the reactor was varied by changing the volume
flow at the reactor inlet. For seven different values between 30 and
70 s the results are compiled in Fig. 4.
3
. Results and discussion
The conversion increased linear up to 20% with increasing res-
idence time. At such low conversions this linear behavior can be
expected. With higher conversion the selectivity decreased only
slightly from 97.3% to 96.4% n-pentanal.
3
.1. Residence time distribution
In a first set of experiments, the Berty reactor was character-
ized with respect to the residence time distribution as a function
of turbine speed and thus loop flow. The dimensionless residence
time distribution function E(Θ) was calculated using a pulse func-
tion of propane at room temperature and under ambient pressure.
The propane concentration in the exit flow was measured by IR
3.3. Temperature variation
For SILP catalysts the influence of reaction temperature was
studied at 10 bar syngas pressure between 80 and 110 C. The ef-
fective reaction rate reff increased with increasing temperature in
an Arrhenius-type fashion. From a plot of lnreff versus T
apparent activation energy was calculated to be 64.1 kJ mol
shown in Fig. 5.
This value matches the activation energy of 63.8 kJ mol for
1-butene hydroformylation determined in a PFTR setup under sim-
ilar conditions [7]. This good agreement is not surprising, since the
◦
−
1
(BINOS, Emerson Process Management GmbH). Transformation into
the
as
−
1
the dimensionless cumulative RTD function F(Θ) was achieved by
numerical integration. In Fig. 3 the results for turbine speeds be-
−1
−1
tween 0 and 10000 min
are shown.
the Berty reactor behaved significantly dif-
ferent from an ideal plug flow reactor (PFTR), probably due to
−
1
Already at 0 min

324
M. Haumann et al. / Journal of Catalysis 263 (2009) 321–327
Fig. 4. Conversion and n-pentanal selectivity in SILP catalyzed hydroformylation
of 1-butene as a function of residence time. 1.90 g SILP catalyst, mRh = 0.2 wt%,
Fig. 6. Conversion and n-pentanal selectivity in SILP catalyzed hydroformylation of
1
-butene as a function of syngas composition. 1.90 g SILP catalyst, mRh = 0.2 wt%,
◦
˙
−1
,
[
BMIM][C8H17 OSO3], α = 0.1, p = 10 bar, T = 100 C, VCO = 10.40 N ml min
◦
−1
.
[
BMIM][C8H17 OSO3], α = 0.1, p = 10 bar, T = 100 C, V˙ 1-butene,liq = 0.1 ml min
V˙H = 10.40 N ml min
−1
,
˙
VHe = 20 N ml min
−1
,
V1-butene,liq = 0.1 ml min . Error
˙
−1
2
Error bars calculated from standard deviation of measured data points.
bars calculated from standard deviation of measured data points.
3
.5. Ex-situ NMR studies
The fate of the phosphine ligand immobilized on the SILP cat-
alyst was investigated before and after the hydroformylation reac-
tion by ex-situ NMR studies. Possible degradation can occur from
oxidation according to Scheme 2. The intense signal for the fresh
bisphosphine at −17.6 ppm (ca. 96%) indicated that the ligand was
intact before any contact with the carrier. Only traces of mono ox-
idized ligand 2 were present, indicative by signals at −21.1 ppm
and 25.7 ppm as shown in Fig. 7a.
The fresh ligand 1 was deliberately oxidized by treatment of a
fresh solution of 1 with H2O2. This resulted in complete disappear-
ance of the peak at −17.7 ppm while a strong signal at 30.7 ppm
was obtained as shown in Fig. 7b. The intact ligand 1 was then
used to prepare the SILP catalyst. After removal of the solvent, part
of the prepared dry SILP material was washed with small amounts
of ethanol in order to remove the ionic liquid film. Ethanol was re-
moved from the solution in vacuo and the light red solution was
analyzed in the NMR. The results for the 31P NMR are shown in
Fig. 7c.
−
1
Fig. 5. Arrhenius plot of lnreff versus T
for the SILP catalyzed hydroformylation
of 1-butene. 1.90 g SILP catalyst, mRh = 0.2 wt%, [BMIM][C8H17 OSO3], α = 0.1, p =
−1
˙
−1
˙
The 31P signal at −17.6 ppm was reduced in intensity to ca.
◦
˙
1
0 bar, T = 80 to 110 C, VCO = 10.40 N ml min , VH = 10.40 N ml min , VHe =
2
−1
˙
−1
2
0 N ml min , V1-butene,liq = 0.1 ml min
.
85% while the mono oxidized species 2 increased to more than
10%. A small doublet signal at 20.4 ppm with a coupling con-
stant of JRh-P = 99 Hz was attributed to the dimeric rhodium
species [Rh(μ-CO)(CO)(1)]2 4, indicating that under these condi-
tions enough ligand is left in the ionic liquid film to stabilize the
rhodium precursor. Previous experiments have indicated that part
of the ligand excess is irreversibly bound to the silica surface and
that a ligand/rhodium ratio of 10 is required to compensate this
loss [3d]. Beside the signals of free ligands 1 and 2 a small peak
at 30.7 ppm was observed that indicated the double oxidized lig-
and 3. This oxidation could either stem from oxygen impurity in
the ethanol washing solution or from the SILP preparation itself.
Since the same SILP catalyst batch showed excellent catalytic per-
formance in the hydroformylation over 120 h time on stream, it is
believed that the oxidation was rather due to the ethanol washing
procedure. After 120 h time on stream the catalyst was removed
from the Berty reactor and the ionic film was washed off with
ethanol in a similar procedure as with the fresh catalyst but us-
ing more ethanol to allow all ionic liquid to be washed from the
support. The same 31P signals were obtained as shown in Fig. 7d
except for the signal of the Rh-complex that are not visible any
more due to the much higher dilution. Interestingly, in Fig. 7d
the intensities has inversed compared to Fig. 7c with the mono
conversion was below 10% in the PFTR experiments, a prerequisite
for the differential analysis of kinetic data.
3
.4. Syngas composition variation
◦
At 100 C and 10 bar total pressure the syngas composition was
varied. Fig. 6 shows the conversion and selectivity as a function of
H2:CO ratio.
The conversion increased from 11.6% at a 1:1 ratio to 13.3% at
a 3:1 ratio. This behavior is well known from homogeneous cat-
alysts and stems from the Wilkinson mechanism [11]. Formation
of the active species HRh(CO)(P2) (P2 = bisphosphine ligand) from
HRh(CO)2(P2) requires the loss of one CO, thus this step is ham-
pered by high partial pressure of CO. On the other hand, reduc-
tive elimination of the aldehyde by H2 from the 18e acyl species
H2Rh(CO)(P2)C(O)R is favored at high partial pressure of hydrogen.
Upon further increase of the H2:CO ratio the conversion decreased
slightly to 12.5%. The selectivity was affected by the syngas com-
position and decreased linearly to a minor extend from 95.5% (1:1)
to 94.4% (5:1). No hydrogenated by-products (pentanols) were de-
tected under these conditions.

M. Haumann et al. / Journal of Catalysis 263 (2009) 321–327
325
Scheme 2. Oxidation of the sulfoxantphos ligand 1 during the course of the SILP hydroformylation.
Fig. 7. 31 P NMR spectrum of fresh sulfoxantphos (a), fresh SILP catalyst after washing with EtOH (c) and used SILP catalyst after washing with a high amount of ethanol (d)
in DMSO-d6. Ionic liquid film and catalyst were washed from the SILP material with ethanol. Ethanol was removed in vacuo prior to NMR analysis. (b) shows a sample of the
fresh ligand that was deliberately oxidized with H2O2 for a better comparison of the peak positions.
dp
oxidized ligand 2 signal at −21.1 (42%) now being significantly
larger than the one for the intact ligand at −17.6 ppm (18%). How-
ever, neither the partial oxidation of the sulfoxantphos ligand 1
nor the double oxidized ligand 3 at 30.4 ppm did influence the
performance of the SILP catalyst, since the activity and selectivity
remained high throughout the experiment. Experiments to verify
the source of ligand oxidation are currently being conducted in
our laboratory.
reff · ρcat · 2 · n
CMears =
ꢀ 0.15.
(1)
c
· k
bulk
film
The criterion includes the effective, mass related reaction rate r
eff
−
−3
mol kg s], the catalyst density ρcat [kg m ] and catalyst particle
cat
1
[
size dp [m], the reaction order n, the bulk substrate concentra-
−
3
−1
tion cbulk [mol m ] and the diffusion rate constant k
[m s ]
film
◦
through the film. At 130 C and 10 bar the reaction rate r was
calculated from the molar substrate flow, the conversion and the
catalyst mass according to Eq. (2).
eff
3
.6. Influence of film or pore diffusion
n˙ · X
mol
3
reff =
= 1.15 × 10−
.
(2)
Several criteria have been developed in order to estimate if the
mcat
kgcat s
overall reaction is influenced by mass transport processes. By us-
ing the Mears criterion CMears one can account for film diffusion
effects according to Eq. (1) [12].
The Reynolds number Re was calculated to be 0.03 under the
flow conditions applied (see Table 1).

326
M. Haumann et al. / Journal of Catalysis 263 (2009) 321–327
ꢂ
Table 1
4 · D · cb
eff
dp,max ꢀ
.
(8)
Parameters used for film and pore diffusion estimations.
r
eff
Parameter
Symbol
Unit
Value
Based on Eq. (8) a maximum particle diameter of 2.5 mm was
calculated. This value is large enough for several forms of SILP cat-
alysts to be applicable in a large scale hydroformylation process,
e.g. as cylinders, pellets or extrudates.
Based on the small Weisz–Prater modulus, the effectiveness fac-
tor was assumed to be close to 1. The correlation between the
Weisz–Prater modulus and the Thiele modulus for spherical par-
ticles is given in Eq. (9).
◦
bar
m
m
m
m s
Temperature
Pressure
T
p
dp
ν¯
C
130
10
1.32 × 10
1.45 × 10
0.01
−
5
Particle diameter
Viscosity of gas mixturea
Inner tube diameter
Gas velocity
Reynolds number
Molecular diffusion coefficient
2
−1
s
−5
dR
u0
Re
D12
DK
Deff
−1
−3
3.28 × 10
0.03
b
m2
m
m
−1
−1
−1
−6
−8
−9
s
s
s
2.48 × 10
1.27 × 10
8.97 × 10
0.9
c
2
Knudsen diffusion coefficient
2
Effective diffusion coefficient
Porosity
Tortuosityd
2
φ = η · ΦWP.
(9)
ε
τ
P
1.27
−
m s 1
mol s
mol m
−4
−5
Solving Eq. (8) for the Thiele modulus results in a value of
φ = 0.11.
Film diffusion rate constant
kfilm
n˙
1.20 × 10
1.43 × 10
55.64
0.16
56.11
−
1
1-Butene molar flow
1-Butene concentration
1-Butene conversion
1-Butene molar mass
−
3
cb
X
4
. Conclusion
kg kmol 1
kg
mol kg
mol m
kg m
−
M1
mcat
reff
−
3
Mass SILP catalyst
Mass related effective reaction rate
1.93 × 10
−
1
−3
In the present work we have applied a gradient-free gas phase
s
1.15 × 10
cat
reactor for the SILP catalyzed hydroformylation of 1-butene. The
use of such a Berty type reactor allowed the determination of
kinetic parameters without falsification from concentration or
temperature gradients. At low conversions around 10% the ac-
−
3
3
Volume related effective reaction rate
Density of catalyst bed
Mears criterion
Weisz–Prater modulus
Thiele modulus
reff
ρ
s
0.32
−
cat
430
−
3
CM
ΦWP
φ
4.75 × 10
−
2
1.11 × 10
0.11
◦
tivation energy has been determined between 80 and 110 C to
a
−1
Average viscosity calculated from the viscosities of pure compounds (1-butene,
H2, CO, He) according to the gas composition.
64.1 kJ mol , a value matching the result from fixed bed reactor
studies under similar conditions. Furthermore, this value is large
enough to account for absence of mass transport limitations from
the gas into the ionic liquid phase. In order to test for limitation
by film or pore diffusion, established criteria by Mears and Weisz–
Prater have been applied. In both cases it was found that the SILP
catalyst operates under no mass transport diffusion under the con-
ditions applied. This is an important outcome of our studies with
respect to possible application of the SILP technology for gas-phase
hydroformylation reactions.
The stability of the SILP catalyst material was investigated by
means of NMR in order to detect possible degradation products.
No ligand degradation products could be found in the NMR spec-
tra to a significant amount after 120 h time on stream. However,
partial and complete oxidation of the ligand was observed in the
NMR. Neither the partially nor the fully oxidized ligand seemed to
lower the catalyst performance in terms of activity or selectivity
significantly, thus it can be concluded that the oxidation occurred
during NMR sample preparation.
b
Calculated from a correlation by Hirschfelder et al. [15].
c
Calculated from Eq. (7).
d
Calculated according to a correlation by Dumanski [16].
The film diffusion rate constant kfilm was calculated from the
mass flow density J12 using correlations (3) and (4) [13,14].
J12 = 0.84 · Re−
0.51
= 5.04,
(3)
(4)
0
.67
Sh
k
· ν
film
J12 =
=
.
0.33
0.67
Re · Sc
u0 · D
1
2
From the data given in Table 1 a film diffusion rate constant
−4
−2
kfilm = 1.20 × 10
m s
and a Mears criterion of CMears = 0.0048
were calculated. The CMears-value is more than 30 times smaller
than the required Mears criterion limit. Under the conditions ap-
plied, film diffusion therefore has no influence on the SILP hydro-
formylation. The absence of film diffusion was further confirmed
by the large value for the activation energy.
We anticipate that the combination of SILP technology and
Berty reactor design will provide fast and reliable insight into var-
ious homogeneously catalyzed reactions in the near future.
Pore diffusion influence was tested with the help of the Weisz–
Prater criterion for spherical catalyst particles of diameter dp, given
by Eq. (5) for a first order reaction [17].
Acknowledgments
2
d
reff
p
ΦWP =
·
ꢀ 1.
(5)
4
Deff · cb
The authors would like to thank Michael Schmacks and Achim
Mannke for help with the Berty setup. Dr. Peter Schulz is gratefully
acknowledged for assistance with the NMR measurements. The au-
thors would like to thank Dr. Robert Franke from Evonik Oxeno
GmbH for fruitful discussions regarding the Mears and Weisz–
Prater criteria.
The effective diffusion coefficient for 1-butene was calculated
from Eq. (6) assuming an interplay between molecular and Knud-
sen diffusion (Eq. (7)) inside the catalyst particle.
ꢀ
ꢁ
1
τ
1
1
=
·
+
,
(6)
(7)
Deff
ε
P
D12
DK
References
ꢂ
T
DK = 48.5 · dp ·
.
[1] B. Cornils, W.A. Hermann, I.T. Horvath, W. Leitner, S. Mecking, H. Olivier-
Bourbigou, D. Vogt (Eds.), Multiphase Homogeneous Catalysis, Wiley–VCH,
Weinheim, 2005.
M1
From Eq. (5) a value for the Weisz–Prater modulus of ΦWP =
.0111 was calculated. This extremely low value indicates the ab-
sence of pore diffusion limitation on the overall reaction rate. By
rearranging Eq. (5) the maximum particle diameter was deter-
mined, when pore diffusion would begin to influence the overall
reaction rate and thus the effectiveness factor of the SILP catalyst.
[2] D.J. Cole-Hamilton, Science 299 (2003) 1702.
[3] (a) C.P. Mehnert, R.A. Cook, N.C. Dispenziere, M. Afeworki, J. Am. Chem.
Soc. 124 (2002) 12932;
0
(
(
(
b) A. Riisager, K.M. Eriksen, P. Wasserscheid, R. Fehrmann, Catal. Lett. 90
2003) 149;
c) A. Riisager, P. Wasserscheid, R. van Hal, R. Fehrmann, J. Catal. 219 (2003)
252;

M. Haumann et al. / Journal of Catalysis 263 (2009) 321–327
327
(
d) A. Riisager, R. Fehrmann, S. Flicker, R. van Hal, M. Haumann, P. Wasser-
[11] (a) J.F. Young, J.A. Osborn, F.A. Jardine, G. Wilkinson, J. Chem. Soc. Chem. Com-
mun. (1965) 131;
scheid, Angew. Chem. Int. Ed. 44 (2005) 185.
[
4] A. Riisager, R. Fehrmann, M. Haumann, P. Wasserscheid, Eur. J. Inorg. Chem.
(b) D. Evans, J.A. Osborn, G. Wilkinson, J. Chem. Soc. A (1968) 3133;
(c) D. Evans, G. Yagupsky, G. Wilkinson, J. Chem. Soc. A (1968) 2660.
(2006) 695.
[
5] A. Riisager, R. Fehrmann, M. Haumann, B.S.K. Gorle, P. Wasserscheid, Ind. Eng.
Chem. Res. 44 (2005) 9853.
[
[
12] D.E. Mears, Ind. Eng. Chem. Process Des. Dev. 10 (1971) 541.
13] O.A. Hougen, Ind. Eng. Chem. 53 (1961) 509.
14] M. Baerns, H. Hofmann, in: M. Baerns, et al. (Eds.), Technische Chemie, Wiley–
VCH, Weinheim, 2006.
15] J.O. Hirschfelder, C.F. Curtiss, R.B. Bird, Molecular Theory of Gases, John Wiley
[6] J. Berty, Chem. Eng. Prog. 70 (1974) 78.
[
[
[
[
7] M. Haumann, K. Dentler, J. Joni, A. Riisager, P. Wasserscheid, Adv. Synth.
Catal. 349 (2007) 425.
8] M.S. Goedheijt, P.C.J. Kamer, P.W.N.M. van Leeuwen, J. Mol. Catal. A Chem. 134
[
[
&
Sons, New York, 1967.
16] A. von Dumanski, Kolloid Z. 3 (1908) 210.
17] P.B. Weisz, Z. Phys. Chem. 11 (1957) 1.
(1998) 243.
9] U. Pallaske, Chem. Ing. Tech. 56 (1984) 46.
[
[
10] B. Etzold, A. Jess, University of Bayreuth, private communication.