Experimental investigation of ethylene hydroformylation to propanal on Rh and Co based catalysts

Article abstract of DOI:10.1016/j.apcata.2013.10.019

Intrinsic hydroformylation kinetics have been measured in a high-throughput kinetic test setup at temperatures varying from 448 to 498 K, with the total pressure ranging from 1 to 3 MPa. A gaseous feed containing CO, C ₂H₄ and H₂ was used with space times varying from 2.7 kg_cat s/mol_C2H₄,in to 149 kg_cat s/mol_C2H₄,in. Three catalysts have been investigated, i.e., 5%Rh on Al₂O₃, 1%Co on Al₂O₃ and 0.5%Co-0.5%Rh on Al₂O₃. The main products observed were ethane, propanal and propanol. The Rh catalyst showed the highest hydroformylation and hydrogenation site time conversions in the investigated range of operating conditions. Moreover it was found on all investigated catalysts that the hydrogenation activation energy was about 15-20 kJ mol^-1 higher than that for hydroformylation. On the Rh catalyst, higher ethylene feed concentrations have a more pronounced effect on CO conversion and production of propanal and propanol compared with an increase in the inlet concentration of the other reactants.

Full text of DOI:10.1016/j.apcata.2013.10.019

Applied Catalysis A: General 469 (2014) 357–366
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Experimental investigation of ethylene hydroformylation to propanal
on Rh and Co based catalysts
N. Navidi, J.W. Thybaut∗, G.B. Marin
Laboratory for Chemical Technology, Chemical Engineering Department, Ghent University, Technologiepark 914, Ghent 9052, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 25 June 2013
Received in revised form 9 September 2013
Accepted 8 October 2013
Available online 18 October 2013
Intrinsic hydroformylation kinetics have been measured in a high-throughput kinetic test setup at
temperatures varying from 448 to 498 K, with the total pressure ranging from 1 to 3 MPa. A gaseous
feed containing CO, C2H4 and H2 was used with space times varying from 2.7 kgcat s/molC
to
H
,in
4
149 kgcat s/molC
. Three catalysts have been investigated, i.e., 5%Rh on Al2O3, 1%Co on Al22O3 and
H
,in
4
0.5%Co–0.5%Rh 2on Al2O3. The main products observed were ethane, propanal and propanol. The Rh cat-
alyst showed the highest hydroformylation and hydrogenation site time conversions in the investigated
range of operating conditions. Moreover it was found on all investigated catalysts that the hydrogenation
activation energy was about 15–20 kJ mol−1 higher than that for hydroformylation. On the Rh catalyst,
higher ethylene feed concentrations have a more pronounced effect on CO conversion and production of
propanal and propanol compared with an increase in the inlet concentration of the other reactants.
© 2013 Elsevier B.V. All rights reserved.
Keywords:
Heterogeneous catalysis
Ethylene
Hydroformylation
Hydrogenation
Rhodium
Cobalt
1. Introduction
Due to the high activation energies involved, the reaction rate
remains rather low, however [4].
The ever increasing crude oil price has led to the exploration
of alternative feeds such as natural gas and, correspondingly, new
processes for the production of hydrocarbon chemicals. Hydro-
formylation, or oxo synthesis, that was discovered by the German
scientist Otto Roelen in 1938, is an important process for produc-
tion of aldehydes. The latter may play a key role in the upgrading
and/or separation of alkenes produced via alternative routes such as
oxidative coupling of methane or methanol to olefins. Hydroformy-
lation catalysts generally also exhibit double bound hydrogenation
which results in reactant alkene conversion into the correspond-
ing alkane as well as product aldehyde conversion into alcohols
[1]. Typical hydroformylation products are in the carbon number
range from 3 to 19, the reactant alkene being selected based on the
desired product aldehyde.
In principle, all the transition metals capable of forming car-
bonyls, such as Rh, Co, Ir, Ru,. are potential hydroformylation
catalysts. The order of activity from the most to the least active
metal, is as follows: Rh > Co > Ir, Ru > Os > Pt > Pd > Fe > Ni [5]. Cobalt
complexes had been the main industrial hydroformylation cata-
lysts until the early 1970s, i.e., prior to the commercialization of
rhodium based catalysts. Since then, cobalt catalysts for hydro-
formylation have been replaced in almost all major plants by the
more advantageous rhodium catalysts despite the higher price of
the noble metal. This substitution became feasible because the high
price of rhodium was offset by cheaper equipment, increased cat-
alyst activity, and a generally higher selectivity and efficiency.
Commercially, homogeneous Co or Rh complexes are typically
applied at temperatures ranging from as low as 360 K up to 573 K
[6,7]. The homogeneous character of industrial process configura-
tions leads to inherent operational problems such as difficulties in
catalyst separation from products, expensive metal losses and cor-
rosivity of catalytic solutions. The successful implementation of an
active and stable heterogeneous hydroformylation catalyst would
allow avoiding these drawbacks. Although the research into hetero-
geneously catalysed hydroformylation is in an early stage, a wide
range of research activities has been performed on experimental
heterogeneous hydroformylation with various reactants, catalysts
and operation conditions [8–17].
Aldehydes are useful intermediates in the production of valu-
able products such as alcohols, carboxylic acids, amines, and diols.
More recently, hydroformylation is also widely applied in the fine
chemicals and pharmaceutical industry for the production of drugs,
vitamins, herbicides and perfumes [2,3].
At 373 K, ethylene hydroformylation is a spontaneous, exother-
mic reaction with ꢀH◦ = −129.0 kJ mol−1 and ꢀG◦ = −56.9 kJ mol−1
.
Reaction mechanism elucidation is an important subject in
the reported heterogeneous hydroformylation studies. The most
∗
Corresponding author. Tel.: +32 9 331 17 52.
E-mail address: Joris.Thybaut@UGent.be (J.W. Thybaut).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.10.019

358
N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
H
CH2CH3
H
H
CH2
+C2H4
-CO
L
L
Rh
L
Rh
L
Rh
CO
CO
L
Rh
L
L
CH2
L
CO
CO
CO
+H2
+CO
O
-
propanal
ethane
CH2CH3
Rh CO
CO
COCH2CH3
COCH2CH3
L
L
H
+H2
Rh
L
Rh
L
L
L
H
CO
CO
Fig. 1. Wilkinson mechanism for heterogeneous hydroformylation [18].
referenced hydroformylation reaction mechanism was that pro-
posed by Wilkinson and his co-authors [18]. In essence, this
mechanism comprises 8 elementary steps, see Fig. 1, i.e., 1: carbon
monoxide desorption, 2: alkene chemisorption, 3: hydride addition
(towards a metal alkyl), 4: hydrogenation to ethane (side reaction),
5: carbon monoxide molecular chemisorption, 6: carbon monoxide
insertion into the metal alkyl bond, 7: hydrogen chemisorption and
8: aldehyde elimination.
The present work concentrates on heterogeneous ethylene
hydroformylation on a series of Rh and Co based catalysts. Gas
phase experimentation was performed in a high-throughput kinet-
ics setup to compare the Rh and Co catalysts and to investigate the
effect of the operation conditions on the main and by product for-
mation rates on these catalysts. We aim at a fundamental analysis
both in terms of operating conditions as well as in terms of cata-
lyst properties and how they could be potentially enhanced. Even if
the catalysts investigated give only moderate oxygenate yields, the
mechanistic investigation provides key insight into the correspond-
ing reaction mechanism that may be exploited in further catalyst
design.
reduction (TPR), temperature programmed desorption (TPD) nor
from X-ray diffraction (XRD) measurements. As a result, transmis-
sion electron microscopy (TEM) was performed to determine metal
particle sizes and achieve information about the distribution and
configuration of these metal particles on the surface of the three
investigated catalysts. TEM samples were prepared by applying
simple immersion of a carbon-support film on a nickel grid into
the catalyst powder. The excess powder was shaken off carefully
afterwards. Catalyst particles that had adhered to the carbon film
were investigated by different TEM modes: conventional and high
resolution (HR) TEM, electron diffraction, STEM, EDX and electron
energy loss spectroscopy (EELS). A JEOL, JEM2200FS-Cs-corrected
microscope was used. This microscope was operated at 200 kV
and equipped with Schottky-type FEG, EDX JEOL JED-2300D and
JEOL in-column omega filter. Secondary X-ray fluorescence from
the analytical sample holder was eliminated by means of a beryl-
lium retainer. However, Ni-peaks were detected in the EDX spectra
due to secondary fluorescence from the Ni support grid. EEL spec-
tra were obtained in STEM spot mode using an electron probe of
1.0 nm and objective aperture angular size of 6 mrad.
2. Experimental
2.3. Intrinsic kinetics measurements
2.1. Catalysts preparation
Gas phase hydroformylation measurements have been per-
formed in a high-throughput kinetics test setup, see Fig. 2. This
setup, which was constructed by Zeton B.V., consists essentially
of three separate sections: a feed section, a reactor section and an
analysis section. The setup consists of eight parallel reactors that are
positioned per pair in a split type bronze furnace, i.e., reactor tem-
perature and pressure are controlled per reactor pair. Each reactor
has three gas feed lines and one liquid feed line. Each of these lines
has an individual mass flow controller. The reactor tubes are made
of AISI 316 cold worked steel with 890 mm height and 11 mm inter-
nal diameter and are capable of operating up to 600 ◦C and 15 MPa,
i.e., far beyond the operating range as required for hydroformy-
lation. A three-zone reactor temperature control is established by
electrical heating within the bronze furnace, the catalyst being sit-
uated at the bottom of the second zone and the top of third zone.
The reactants were mixed before entering the reactor from the top
in the central axis. The pressure in the reactor was generated by
means of back pressure. The operation of the setup was controlled
by means of Labview software.
In order to provide a good comparison of the Co and Rh cata-
lysts which are used most frequently in industrial homogeneous
hydroformylation, 5%Rh deposited on Al2O3, 1%Co deposited on
Al2O3 and 0.5%Rh–0.5%Co deposited on Al2O3 catalysts, provided
by Johnson Matthey, have been investigated in this work. They were
received as powders, pressed into flakes and subsequently crushed
and sieved to retain a fraction with a diameter between 0.3 and
0.5 mm. The catalysts were reduced in situ at 673 K in order to
convert them into their active form.
2.2. Catalyst characterisation
The total surface area (BET) of each catalyst was measured using
a Micromeritics Gemini equipment. The crystallinity of the samples
was investigated via X-ray diffraction (XRD) measurements with a
Siemens Diffractometer Kristalloflex D5000, Cu K␣ radiation, in the
2ꢁ range 4–90◦ step.
The metal loading and corresponding particle size were too low
for retrieving useful information from temperature programme
Between 1 and 10 g of pelletised catalyst was diluted with an
identical mass of inert material prior to loading into the reactor

N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
359
Fig. 2. Schematic overview of the high-throughput kinetics test setup.
Table 1
Operating conditions and catalysts used in the experimental investigation of the ethylene hydroformylation kinetics.
Catalyst
Temperature (K)
Pressure (MPa)
Space time range (kgcat s/molC
)
,in
CO/C2H4/H2
H
2
4
Catalyst performance
5%Rh/Al2O3
1%Co/Al2O3
0.5%Rh/0.5%Co/Al2O3
1/1/1
1/1/1
473
2
5.4–149
Temperature and pressure effect
5%Rh/Al2O3
1%Co/Al2O3
448–498
473
1–3
2
2.7–149
2.7–80
Inlet composition effect
1/1/1; 2/1/1
1/2/1; 1/1/2
5%Rh/Al2O3

360
N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
tubes of the high-throughput setup. Table 1 presents the operating
conditions applied during the corresponding experiments, aiming
to compare the performance of the mentioned catalysts and to
investigate temperature, pressure and inlet composition effects on
the hydroformylation reaction behaviour.
hydrido-metal species in solution. They reported that the transfor-
mation of the acyl-metal species to the aldehyde, see steps 6 and 7
in Fig. 1, proceeds through reaction with a second catalyst species
in homogeneous media while in heterogeneous media the oxida-
tive addition of molecular hydrogen to an acyl-metal species is the
only way for aldehyde formation [7,23].
It is accepted today that Wilkinson’s mechanism is the most
likely one in hydroformylation [3,22,24]. In this mechanism, it is
assumed that carbon monoxide does not dissociate on the catalyst
surface. The absence of a CO dissociation step in ethylene hydro-
formylation on silica supported rhodium catalysts is also reported
by Hanaoka et al. [14]. These authors observed that the CO inser-
tion to an alkyl species to form an acyl species is the key step for
hydroformylation that controls the selectivity towards oxygenated
products.
Considering the above discussion, a reaction mechanism for
ethylene heterogeneous hydroformylation is proposed in what fol-
lows, see Fig. 3. It includes two parallel reaction pathways, starting
from a metal bound alkyl that is obtained after ethylene chemisorp-
tion on the metal surface. The first, and in hydroformylation desired
pathway, involves CO insertion and leads to aldehyde or alcohol
formation. The second of the parallel pathways is ethylene hydro-
genation into ethane. In this mechanism, the same hydrogen and
ethylene surface species are involved in the formation of both
ethane and propanal.
The potential presence of propene and higher carbon number
alkenes, alkanes, aldehydes and alcohols can be explained by CO
bond dissociation in the produced aldehydes or by CO bond dis-
sociation and subsequent C hydrogenation to methylene prior to
insertion in a metal alkyl species, as it may occur in Fischer–Tropsch
synthesis [25–27]. Not only historically, but also mechanistically,
heterogeneously catalysed hydroformylation is very closely related
to Fischer–Tropsch synthesis. In fact, the reaction of ethylene with
syngas has already been used as a probe reaction to study the activ-
ity and selectivity in the Fischer–Tropsch process on supported
transition metals [9].
Online effluent analysis was performed with a four channel
Agilent 3000 micro gas chromatograph equipped with thermal con-
ductivity detectors (TCDs). Hydrogen, and carbon monoxide were
analysed on a molar sieve 5A PLOT column of 10 m length and
0.32 inner diameter with a film thickness of 12 ␮m, while ethyl-
ene and ethane are detected on a PLOT U column of 8 m length
and 0.32 inner diameter with a film thickness of 30 ␮m. Heavy
hydrocarbons can be detected in an Alumina PLOT column of 10 m
length and 0.32 mm inner diameter with a film thickness of 8 ␮m.
Propanal, propanol and other oxygenates are only detected on the
OV-1 column of 10 m length and 0.15 mm inner diameter with a
film thickness of 2 ␮m. Argon was added as an internal standard
for the purpose of mass balance verification. The maximum error
of mass balance was less than 10% while most of the error values
were below 5%. The elemental balances were also verified and the
deviations were in line with those on the total mass balance. The
conversions and product molar flow rates were calculated using
calibration factors as reported by Dietz [19].
Long-term catalyst stability was assessed by repeating a refer-
ence experiment at regular time intervals, i.e., one every few days.
Typically the catalyst activity could be maintained for 3–4 weeks,
after which the catalyst was replaced by a fresh sample. The selec-
tivity towards a product P, Sp, was defined as the number of moles
transformed into that product, Fp, compared to the total number of
ethylene moles converted:
Fp
Sp
=
(1)
FC ,0 − FC
H
H
4
2
4
2
In order to compare the performance of the mentioned cata-
lysts under identical experimental conditions the conversions were
reported versus the so-called site time, which is obtained from the
space time by accounting for the metal loading and the fraction of
exposed metal atoms, as well as for atomic mass.
Because this comprehensive mechanism allows describing the
formation of all hydroformylation products observed in this work,
i.e., aldehydes and alcohols, as well as ethane, it is further used in
the assessment of the kinetic data.
space time × metal loading
site time =
× FE
(2)
MM
4. Results and discussion
The fraction exposed was obtained from the relation between
the particle size and dispersion as shown in Eqs. (3) and (4) [20].
4.1. Catalyst characterisation
dVS
dat
5.01
FE
drel(VS)
=
=
for FE < 0.2
(3)
4.1.1. BET Surface area
The BET surface areas for the investigated catalysts are reported
in Table 2. Only minor differences are observed between the cata-
lysts considered, indicating that the Al2O3 support is governing the
observed surface areas. Nevertheless, as may have been expected,
the catalyst with highest metal loading exhibits the lowest surface
area.
dVS
dat
3.32
FE1.23
drel(VS)
=
=
for 0.2 ≤ FE ≤ 0.92
(4)
It was verified by using the proper correlations, that at the
selected operating conditions so-called intrinsic kinetics are mea-
sured, i.e., the observations are not affected by mass or heat
transfer limitations [21]. Of course, upon extrapolation towards the
industrial scale, potential mass and heat transport effects must be
accounted for.
4.1.2. TEM analysis
The presence of Rh and Co on the investigated catalyst samples
was confirmed first by scanning transmission electron microscopy
(STEM) coupled with energy dispersive X-ray (EDX). The corre-
sponding pictures are given in Fig. 4a for 5%Rh on Al2O3, Fig. 4 for
3. Heterogeneous ethylene hydroformylation reaction
network
Table 2
The most generally reported products for ethylene hydro-
formylation, in descending order of importance, are ethane,
propanal, propanol and small amount of C3+ hydrocarbon prod-
ucts [7,10,12,13,16,22]. According to Henrici-Olivé and Olivé [23]
the key difference between the homogeneous and the heteroge-
neous reactions is the presence of a free, mobile and very reactive
BET surface area for the Rh and Co based catalysts used in this work.
Catalyst
BET surface area (m2/g)
5%Rh/Al2O3
0.5%Rh–0.5%Co/Al2O3
1%Co/Al2O3
120
147
143

N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
361
Fig. 3. Proposed reaction mechanism for ethylene hydroformylation to propanol [22].
1%Co on Al2O3 and in Fig. 4 for 0.5%Rh/0.5%Co on Al2O3. 5%Rh on
Al2O3, see Fig. 4a, has well dispersed particles with a small size, i.e.,
2 nm, resulting in a high fraction of exposed Rh.
The images show a, still small yet significantly larger particle
size, i.e., 11 nm, for the 1%Co catalyst while the 0.5%Rh/0.5%Co
contained large Rh clusters, i.e., 20–50 nm, but small Co particles,
i.e., 4.3 nm. In other words, while Rh was finely dispersed on the
5%Rh catalyst and Co was somewhat more poorly dispersed on
the 1%Co catalyst, the Rh dispersion on the 0.5%Rh/0.5%Co catalyst
was extremely poor, while the Co dispersion was again better on
this latter catalyst. The measured particle sizes and corresponding
fractions exposed are reported in Table 3.
Fig. 4. TEM images for: (a) 5%Rh on Al2O3 catalyst, (b) 1%Co on Al2O3 catalyst, (c and d) 0.5%Rh/0.5%Co on Al2O3 catalyst.

362
N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
Table 3
Particle size, fraction of exposed metal and accessible metal atoms for the Rh and Co based catalysts used in this work.
Catalyst
Particle size (nm)
TEM
Fraction of exposed metal (%)
Accessible metal atoms (mol/gcat.)
5%Rh/Al2O3
1%Co/Al2O3
2
11
4.3 for Co
35 for Rh
51.1
12.4
28.1 for Co
3.8 for Rh
2.4757 × 10−4
0.2098 × 10−4
0.2575 × 10−4
0.5%Rh–0.5%Co/Al2O3
Fig. 5. (a) CO conversion versus molar site time and (b) C2H4 conversion versus molar site time. Reaction conditions: temperature 473 K; pressure, 2.0 MPa; gas inlet
composition, C2H4:CO:H2:Ar = 30:30:30:10), 2 MPa.
4.2. Catalytic performance
The product distribution on these catalysts are reported in
Table 5. The 5%Rh on Al2O3 catalyst exhibited the highest selec-
tivity for oxygenates among these catalysts. The 0.5%Rh/0.5%Co on
Al2O3 catalyst also exhibits a higher selectivity for oxygenated com-
pounds compared to the Co catalyst, which can, again, be related to
the differences in Co particle size on the catalyst surface. The previ-
ously reported particle size effects [8,14] particularly concern the
CO insertion activity rather than the hydrogenation activity and,
hence, result in an increase of the oxo-selectivity.
Fig. 5 shows the observed CO and C2H4 conversion as a function
of the site time at 473 K and 2 MPa. The CO and C2H4 site time
conversions that were calculated based on the results presented in
these figures are reported in Table 4.
5%Rh on Al2O3 exhibited the highest site time conversions in
the entire range of operating conditions, while the lowest site time
conversions were obtained on 1%Co on Al2O3.
Based on the site time conversions obtained with the
monometallic catalysts and the knowledge of the fractions
exposed on the bimetallic catalyst, see Table 3, an expected
site time conversion on this bimetallic catalyst amounting to
0.00429 molCO, converted/molcat. atoms/s was calculated, which is
lower than the experimentally observed site time conversion. Giv-
ing the large Rh particle size on the 0.5%Rh/0.5%Co on Al2O3
catalyst, the higher than expected site time conversions observed
on the latter catalyst are attributed to the finely dispersed Co.
Indeed, particle size effects, also denoted as structure sensitivity,
have already been reported in hydroformylation [8,14] as well as
in Fischer–Tropsch synthesis [28].
A minor presence of methane in the product stream can be
the result of CO chemisorption, dissociation and hydrogenation
to methane. The selectivity of methane never exceeded 2% in
this work. Upon insertion of an intermediate methylene species
into a metal alkyl species, the latter will grow, as it occurs in
the Fischer–Tropsch synthesis reaction mechanism. Of course, CO
bond dissociation after a prior insertion in a metal alkyl species
may also explain the formation of heavier hydrocarbons [29]. The
amount of methane and C3+ hydrocarbon products on the tested
catalysts barely exceeded 5%, which is sufficiently low to assume
that practically all of the converted CO ended up in hydroformy-
lation products, i.e., propanal or propanol. Hence, the selected
Table 4
Site time conversions obtained on the Rh and Co based catalysts used in the present work at 473 K, 2.0 MPa and equimolar C2H4/CO/H2 gas inlet composition.
Catalyst
CO site time conversion
C2H4 site time conversion
(molCO, converted/molcat. atoms/s)
(molC ,converted/molcat. atoms/s)
H
2
4
5%Rh/Al2O3
1%Co/Al2O3
0.5%Rh–0.5%Co/Al2O3
0.01356
0.00359
0.00746
0.04071
0.01486
0.02626
Table 5
Product distribution obtained on the Rh and Co based catalysts used in the present work at 478 K, 2.0 MPa and equimolar C2H4/CO/H2 gas inlet composition.
Catalyst
Product selectivities (%)
CH4
C2H6
C3H6
C4H8 + C4H10
C3H5OH
C3H7OH
5%Rh/Al2O3
1%Co/Al2O3
0.5%Rh–0.5%Co/Al2O3
0.4
1.1
0.7
62.4
71.7
66.8
0.6
1.3
1
1.6
1.9
2.2
28
18
19.6
7
6
9.7

N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
363
Fig. 6. (a) CO conversion versus site time on 5%Rh on Al2O3, (b) C2H4 conversion versus site time on 5%Rh on Al2O3, (c) CO conversion versus site time on 1%Co on
Al2O3, and (d) C2H4 conversion versus site time on 1%Co on Al2O3. Reaction conditions: temperature, 448, 473 and 498 K; pressure, 2.0 MPa, gas inlet composition,
C2H4:CO:H2:Ar = 30:30:30:10.
mechanism for ethylene hydroformylation, see Fig. 3, can appro-
priately describe the reaction.
4.2.1. Temperature effect
The temperature and pressure effect on the hydroformylation
rate were investigated on the 5%Rh on Al2O3 and 1%Co on Al2O3
catalysts. The temperature effect on CO and C2H4 conversion and,
hence, the ethylene hydroformylation and hydrogenation rates on
the Rh and Co catalyst at a fixed pressure of 2 MPa is presented in
Fig. 6.
The correspondingly calculated site time conversions are rep-
resented in Fig. 7 as a function of the inverse of the temperature.
A temperature increase leads to a more pronounced increase of
the C2H4 site time conversion than of the CO site time conversion
on both tested catalysts. Correspondingly, the apparent activation
energy for ethane formation as determined from the Arrhenius
diagram presented in Fig. 7 exceeds that for propanal formation
on both tested catalysts by 15–20 kJ mol−1. Hence, as it can be
seen in Fig. 8, a temperature increase has a negative effect on the
selectivity towards oxygenated compounds on both catalysts. For
example by increasing the temperature from 448 K to 498 K on
5%Rh on Al2O3 at 2 MPa, the oxygenates selectivity decreased from
0.45 to 0.26. Similar observations are encountered in the literature
[9,12], where increased ethane selectivities with the temperature
were reported. It was observed by Balakos and Chuang [9], that
in heterogeneous ethylene hydroformylation on 4%Rh/SiO2 with
Fig. 7. Relation between space-time yield and temperature for hydrogenation rate
on Rh (ꢀ) and Co (ꢁ) and hydroformylation rate on Rh (ꢂ) and Co (ꢀ).
increasing the temperature from 483 K to 573 K the hydroformy-
see Fig. 3: with increasing temperature, adsorbed C2H4 molecules
apparently exhibit a higher affinity towards reductive elimination
into ethane, step 4, than reacting with adsorbed CO molecule, step
6, on the catalyst surface. This can be related to the expected evo-
lution in the hydrogen and CO surface concentrations. Because the
chemisorption heat of CO is typically about the double of that of
hydrogen [30], the CO concentration will decrease much faster
with increasing temperature than the hydrogen concentration. As a
lation selectivity, defined as TOFpropanal/TOFC 6 , decreased from
H
2
0.136 to 0.039. Hence, a good hydroformylation catalyst should
exhibit a high activity at lower operating temperatures to avoid
a too significant loss in product selectivity.
The more pronounced temperature effect on hydrogenation
than on hydroformylation can be assessed using the previously
discussed heterogeneous ethylene hydroformylation mechanism,

364
N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
Fig. 8. Product selectivities at 2.0 MPa and equimolar C2H4/CO/H2 gas inlet composition (a) on the 5%Rh/Al2O3 catalyst and (b) the 1%Co/Al2O3 catalyst.
result, the relative importance of hydrogenation in the overall eth-
ylene conversion will increase at the expense of hydroformylation.
selectivity indicates that in particular the CO surface concentration
has increased.
4.2.3. Inlet composition effect
4.2.2. Pressure effect
The effect of variations in C2H4, CO and H2 inlet composition on
the CO and C2H4 conversion, was investigated on 5%Rh on Al2O3
and is reported in Fig. 10. As it can be observed from Fig. 10a,
the CO conversion and, hence, the hydroformylation rate, increases
with the ethylene inlet concentration. Also the C2H4 conversion by
hydrogenation is enhanced by increasing its inlet concentration,
see Fig. 10b. This can be understood when C2H4 concentrations on
the catalyst surface are rather low and, hence, when an increasing
C2H4 surface concentration has a negligible impact on the other
surface concentrations. Increasing the inlet H2 concentration leads
The pressure effect on the CO and ethylene conversion as a
function of the site time at 473 on the Rh and Co catalyst is
shown in Fig. 9. All conversions and consequently hydroformyla-
tion and hydrogenation rates increase with the total pressure to
a similar extent. A slightly more pronounced enhancement of the
hydroformylation rate compared to the hydrogenation rate may
be discerned, in agreement with the literature [13]. At a higher
total pressure also the surface concentrations on the metal parti-
cles are correspondingly higher. The slightly increased oxygenates
Fig. 9. (a) CO conversion versus site time on 5%Rh on Al2O3, (b) C2H4 conversion versus site time on 5%Rh on Al2O3, (c) CO conversion versus site time on 1%Co on Al2O3 and (d)
C2H4 conversion versus site time on 1%Co on Al2O3 at different total pressures. Reaction conditions: temperature, 473 K; gas inlet composition, C2H4:CO:H2:Ar = 30:30:30:10.

N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
365
Fig. 10. CO and C2H4 conversions versus space time. Reaction conditions: catalyst 5%Rh on Al2O3, temperature, 473 K; pressure, 2.0 MPa.
to a significantly higher ethylene conversion but only to a slightly
higher CO conversion.
EDX
FA,0
Fp
energy dispersive X-ray
molar inlet flow rate of A (mol/s)
molar outlet flow rate of P (mol/s) fraction of exposed
metal
atoms (%)
Increasing the inlet CO concentration as it is depicted in Fig. 10a
and b has a considerably negative effect on both the CO as the ethyl-
ene conversion. This can be understood if CO is the most abundant
surface species and, hence; if the higher CO surface concentration
results in a more pronounced decrease of the H2 and C2H4 concen-
trations on the catalyst surface, such that both hydroformylation
and hydrogenation rates decrease.
These results were in line with what Balakos et al. reported for
dependence of the reaction rates on the partial pressures of reac-
tants [9]. They observed both ethane and propanal formation rates
are negative order in CO while positive order in both hydrogen and
ethylene in heterogeneous hydroformylation on Rh/SiO2 at 0.1 MPa
and 513 K.
FE
in
M
MM
Sp
TCD
TEM
TPD
TPR
XRD
inlet
metal atoms
standard atomic weight (kg/mol)
selectivity towards product P (%)
thermal conductivity detector
transmission electron microscopy
temperature programmed desorption
temperature programmed reduction
X-ray diffraction
Acknowledgments
5. Conclusions
The research leading to these results has received funding from
the European Community’s Sixth Framework Programme (contract
nr. 515792-2) and the Long Term Structural Methusalem Funding
by the Flemish Government.
The authors would like to thanks Hilde Poelman, Vladimir
Galvita and Vitaliy Bliznuk for assisting in XRD and TEM analysis.
Gas phase hydroformylation on 5%Rh on Al2O3, 1%Co on Al2O3
and 0.5%Co–0.5%Rh on Al2O3 practically exclusively resulted in
ethane, propanal and propanol formation. Finely dispersed Rh par-
ticles were the most active and selective (up to 45%) in ethylene
hydroformylation to propanal and propanol. Structure sensitivity
in ethylene hydroformylation has been explicitly observed for Co
based catalysts: finer Co particles appear to be more active and
selective than larger ones.
References
[1] B. Cornils, M. Rasch, Angew. Chem. Int. Ed. 33 (1994) 2144–2163.
[2] P.W. Van Leeuwen, Rhodium Catalyzed Hydroformylation, Springer, New York,
2002.
Irrespective of the investigated catalyst, the temperature effect
is more pronounced on ethylene hydrogenation than on hydro-
formylation. The 15–20 kJ mol−1 higher apparent activation energy
for ethane than for propanal formation is a result of the higher heat
of chemisorption for CO than for hydrogen. CO was found to be
the most abundant surface intermediate, resulting in an inhibi-
ting effect on both hydroformylation and hydrogenation. Surface
concentrations of hydrogen and ethylene were lower and did not
result in inhibition effects within the investigated range of oper-
ating conditions. An ideal catalyst for ethylene hydroformylation
must consist of a finely dispersed metal such as Rh with a maximum
activity at temperatures below 483 K.
[3] B. Breit, W. Seiche, Synthesis (2001) 1–36.
[4] K. Otsuka, T. Ando, I. Yamanaka, J. Catal. 165 (1997) 221–230.
[5] B. Cornils, Applied Homogeneous Catalysis with Organometallic Compounds,
VCH, Weinheim, 1996.
[6] T.A. Zeelie, A. Root, A.O.I. Krause, Appl. Catal., A 285 (2005) 96–109.
[7] Ullmann’s Encyclopedia of Industrial Chemistry, sixth ed., John Wiley & Sons,
2002.
[8] T.A. Zeelie, Rhodium and Cobalt Catalysts in the Heterogeneous Hydroformy-
lation of Ethane, Propene and 1-Hexene, 2007, Helsinki (PhD thesis).
[9] M.W. Balakos, S.S.C. Chuang, J. Catal. 151 (1994) 266–278.
[10] U.J. Jauregui-Haza, O. Diaz-Abin, A.M. Wilhelm, H. Delmas, Ind. Eng. Chem. Res.
44 (2005) 9636–9641.
[11] K. Tomishige, I. Furikado, T. Yamagishi, S. Ito, K. Kunimori, Catal. Lett. 103 (2005)
15–21.
[12] M.A. Brundage, M.W. Balakos, S.S.C. Chaung, J. Catal. 173 (1998) 122–133.
[13] V.I. Zapirtan, B.L. Mojet, J.G.V. Ommen, J. Spitzer, L. Lefferts, Catal. Lett. 101
(1–2) (2005) 43–47.
List of symbols
[14] T. Hanaoka, H. Arakawa, T. Matsuzaki, Y. Sugi, K. Kanno, Y. Abe, Catal. Today 58
(2000) 271–280.
Cat.
Conv.
dat
catalyst
conversion (%)
atomic diameter of given metal (m)
[15] F.S. Xiao, M. Ichikawa, J. Catal. 147 (1994) 578–593.
[16] C.U. Pittman Jr., G.M. Wilemon, J. Org. Chem. 46 (1981) 1901–1905.
[17] W. Junfan, S. Juntan, L. Hong, H. Binglin, React. Polym. 12 (1990) 177–186.
[18] D. Evans, J.A. Osborn, G. Wilkinson, J. Chem. Soc. 12 (1968) 3133–3142.
[19] W.A. Dietz, J. Gas Chromatogr. 5 (1967) 68–71.
drel(VS) the mean relative size of metallic crystallites
dVS
Ea
the mean size of crystallite (m)
[20] A. Borodzinski, M. Bonarowska, Langmuir 13 (1997) 5613–5620.
[21] R.J. Berger, E.H. Stitt, G.B. Marin, F. Kapteijn, J. Moulijn, CATTECH 5 (2001) 30–60.
activation energy (kJ mol−1
)

366
N. Navidi et al. / Applied Catalysis A: General 469 (2014) 357–366
[22] M.A. Brundage, S.S.C. Chaung, J. Catal. 164 (1996) 94–108.
[23] G. Henrici-Olive, S. Olive, The Chemistry of Catalyzed Hydrogenation of Carbon
Monoxide, Springer-Verlag, Heidelberg, 1980.
[24] C.D. Frohning, C.W. Kohlpainter, in: B. Cornils, W.A. Herrmann (Eds.), Applied
Homogeneous Catalysts with Organometallic Compounds, VCH, Weinheim,
1996 (Chapter 2.1).
[27] G. Lozano-Blanco, J.W. Thybaut, K. Surla, P. Galtier, G.B. Marin, Oil Gas Sci.
Technol. Rev. IFP 61 (2006) 489–496.
[28] G.L. Bezemer, J.H. Bitter, H.P.C.E. Kuipers, H. Oosterbeek, J.E. Holewijn, X.
Xu, F. Kapteijn, A.J. van Dillen, K.P. de Jong, J. Am. Chem. Soc. 128 (2006)
3956–3964.
[29] H. Pichler, H. Schulz, Chem. Ing. Tech. 42 (1970) 1162–1174.
[30] G. Lozano-Blanco, J.W. Thybaut, K. Surla, P. Galtier, G.B. Marin, Ind. Eng. Chem.
Res. 47 (2008) 5879–6589.
[25] J.P. Hindermann, G.J. Hutchings, A. Kiennemann, Catal. Rev. 35 (1993) 1–127.
[26] P. Biloen, W.M.H. Sachtler, Adv. Catal. 30 (1981) 165–216.