Chemical Engineering Science 65 (2010) 273 -- 281

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: w w w . e l s e v i e r . c o m / l o c a t e / c e s

Hydroformylation of 1-octene using rhodium–phosphite catalyst in a thermomorphic

solvent system
Maizatul S. Shaharun, Binay K. Dutta ∗ , Hilmi Mukhtar, Saikat Maitra
Chemical Engineering Department, Universiti Teknologi PETRONAS, Malaysia

A R T I C L E I N F O A B S T R A C T

Article history: The use of a liquid–liquid biphasic thermomorphic or temperature-dependent multicomponent solvent
Received 29 June 2008 (TMS) system, in which the catalyst accumulates in one of the liquid phases and the product goes prefer-
Received in revised form 26 June 2009 ably to the other liquid phase, can be an enabling strategy of commercial hydroformylation processes
Accepted 30 June 2009
with high selectivity, efficiency and ease of product separation and catalyst recovery. This paper describes
Available online 8 July 2009
the synthesis of n-nonanal, a commercially important fine chemical, by the hydroformylation reaction
Keywords:
of 1-octene using a homogeneous catalyst consisting of HRh(PPh3 )3 (CO) and P(OPh)3 in a TMS-system
Hydroformylation consisting of propylene carbonate (PC), dodecane and 1,4-dioxane. At a reaction temperature of 363 K,
Thermomorphic solvent systems syngas pressure of 1.5 MPa and 0.68 mM concentration of the catalyst, HRh(CO)(PPh3 )3 , the conversion of
Kinetics 1-octene and the yield of total aldehyde were 97% and 95%, respectively. With a reaction time of 2 h and
1-octene a selectivity of 89.3%, this catalytic system can be considered as highly reactive and selective compared
Rhodium catalyst to conventional ones. The resulting total turnover number was 600, while the turnover frequency was
400 h−1 . The effects of increasing the concentration of 1-octene, catalyst loading, partial pressure of CO and
H2 and temperature on the rate of reaction have been studied at 353, 363 and 373 K. The rate was found
to be first order with respect to concentrations of the catalyst and 1-octene, and the partial pressure of
H2 . The dependence of the reaction rate on the partial pressure of CO showed typical substrate inhibited
kinetics. The kinetic behavior differs significantly from the kinetics of conventional systems employing
HRh(CO)(PPh3 )3 in organic solvents. Most notable are the lack of olefin inhibition and the absence of a
critical catalyst concentration. A mechanistic rate equation has been proposed and the kinetic parameters
evaluated with an average error of 5.5%. The activation energy was found to be 69.8 kJ/mol.
© 2009 Elsevier Ltd. All rights reserved.

1. Introduction The advantages of the rhodium-catalyst system are mild reaction

One of the widely studied homogeneous catalyzed reactions, the
hydroformylation reaction, provides a versatile route for the synthe-
sis of a vast array of bulk and specialty chemicals, such as indus-
trial solvents, biodegradable detergents, surfactants, lubricants and
plasticizers (Bohnen and Cornils, 2003). The overall reaction can be
represented by Eq. (1). A cobalt- or rhodium-based catalyst is often
used.
(1)
∗ Corresponding author. Present address: Chemical Engineering Department, The
Petroleum Institute, Abu Dhabi, UAE. Tel.: +971 2 6075246; fax: +971 2 6075200.
E-mail addresses: bdutta@pi.ac.ae, binaykdutta@yahoo.com (B.K. Dutta).
conditions, higher n:i (normal to iso) ratio of the products and higher
activity. In general, a linear aldehyde is the desired product. How-
ever, separation of the products and recovery of the precious catalyst
remain as challenging problems. Molecular catalysts immobilized
on different types of support have been explored but low catalytic
activity and catalyst leach-out are difficult to overcome (Arhancet
et al., 1991). Aqueous biphasic hydroformylation (for example,
propene to butanal) by Ruhrchemie/Rhone–Poulenc process was
applied successfully at an industrial scale. Water soluble rhodium
catalyst could be easily and almost completely separated from wa-
ter insoluble products. But this technique is less efficient for higher
olefins. Since the catalytic reaction occurs in aqueous phase, the
success of the technique is limited to the solubility of the olefins in
water phase (Beller et al., 1995). Other strategies such as the fluorous
biphasic hydroformylation developed by Horváth and Rábai (1994)
did not prove to be commercially successful (Kolhofer and Plenio,
2003; Liu et al., 2003). A novel solvent system that itself reversibly
changes from biphasic to monophasic at an elevated temperature,
known as a thermomorphic biphasic or temperature-dependent

0009-2509/$ - see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ces.2009.06.071
274 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281
multicomponent solvent (TMS) system, has gained interest as the
reaction medium. The TMS-system provides easy separation of the
expensive catalyst from the products on cooling of the reaction mix-
ture (Behr et al., 2005; Behr and Roll, 2005; Behr and Fangewisch,
2003). The new concept was also applied to rhodium-catalyzed co-
oligomerization of sunflower fatty acid methyl ester (SFAME) and
ethylene with PC/SFAME/dioxane solvents system and in the absence
of any `tagged ligand' (Behr and Miao, 2004). Similarly, the thermo-
morphic biphasic system was used in the hydroaminomethylation of
1-octene using PC/dodecane/morpholine solvent system (Behr and
Roll, 2005). The isomerizing hydroformylation of trans-4-octene in
the TMS-system of PC/dodecane/p-xylene has been carried out pro-
ducing a very high conversion (about 99%) of the trans-4-octene and
offering very attractive selectivities of n-nonanal ranging from 79%
to 90% (Behr et al., 2005), but a strong rhodium leaching as high as
47% was reported. On the other hand, Tijani and Ali (2006) has de-
veloped a thermomorphic biphasic rhodium-catalyzed system using
an inexpensive and conventional ligand such as P(OPh)3 to catalyze
the hydroformylation of higher olefins ( > C6). However, neither ex-
perimental data nor any empirical model on the kinetics of the hy-
droformylation of higher olefins in the TMS-system are available in
the literature.
The effects of process conditions on the reaction rate and
the kinetics of hydroformylation of 1-octene were studied in the
temperature-dependent multicomponent solvent systems consist-
ing of propylene carbonate (PC), dodecane and 1,4-dioxane. 1,4-
Dioxane was chosen as the solvent mediator because it has a greater
polarity compared to p-xylene and we anticipated that it might
reduce the effect of the catalyst leaching. In the above context, the
specific objectives of this work are to: (a) experimentally deter-
mine the effects of process parameters of the hydroformylation of
1-octene with HRhCO(PPh3 )3 /P(OPh)3 catalyst in the TMS-system of
PC/dodecane/1,4-dioxane over a temperature range of 353–373 K,
(b) study the kinetics of the hydroformylation of 1-octene with
the above catalyst and reaction medium, and (c) develop a kinetic
model of the reaction and evaluate the rate parameters by fitting
experimental data.
2. Experimental
2.1. Materials
1-Octene, triphenylphosphite, propylene carbonate, dodecane
and 1,4-dioxane, purchased from Merck, were of purity 98–99%. The
chemicals were used without further purification. GC analysis did
not show any peaks of impurities. Nitrogen and syngas (1:1 CO/H2 )
were supplied by MOX, with a purity of 99.99%.
2.2. Experimental setup and procedure
The catalyst was prepared in situ by mixing a rhodium catalyst
precursor, HRh(CO)(PPh3 )3 (ABCR, Germany) and triphenylphos-
phite P(OPh)3 . Experimental runs were conducted at four different
temperatures (353, 363, 373 and 383 K), three total gas pressures
(1.5, 2.0 and 2.5 MPa) and catalyst concentrations varying from
8.67×10−5 to 6.8×10−4 kmol/m3 (0.08 to 0.63 g/L). The molar ra-
tio of HRhCO(PPh3 )3 catalyst to P(OPh)3 ligand was 1:12 to en-
sure the optimal stability of the catalytic complex (van Leeuwen
et al., 2000). Hydroformylation of 1-octene was carried out in a
stirred high pressure reactor (model: Parr 4843). A schematic of
experimental setup is shown in Fig. 1. The reactor was equipped
with an automatic temperature control system, which consisted
of an external electric heating jacket and an internal cooling
loop. A pressure transducer-monitor system with high precision
was also connected to the reactor for on-line measurement of
6
5 Vent
P
T
3
PT
TI PI
1 2
4
Fig. 1. Schematic of the experimental setup for kinetic studies: (1) nitrogen; (2)
syngas (CO/H2 ); (3) cooling water in; (4) high pressure reactor; (5) stirrer; (6)
sampling valve; (T) thermocouple; (P) pressure gauge.
reactor pressure in the course of semi-batch hydroformylation
reaction.
The feed mixture (solvents, 1-octene and catalyst; volume =
220 mL) was taken in the preheated reactor. The reactor was purged
with nitrogen and syngas successively, and then adjusted to the de-
sired temperature. It was pressurized with an equimolar mixture of
CO and H2 to a desired total pressure which was maintained con-
stant during the whole run. The stirrer was set at 450 rpm while the
reaction continued. Liquid samples were withdrawn and analyzed
at regular time intervals to follow the progress of the reaction. The
analysis of reactants and products was carried out on a Shimadzu
GC-2010 chromatograph equipped with a flame ionization detector
and a 5% phenylmethyl siloxane capillary column (30 m×0.32 mm ID,
3.0 m). Initial experiments were repeated three times to check for
reproducibility. Measurements are, in general, reproducible within
a maximum of 10% but mostly within a few percent.
3. Results and discussions
3.1. Composition of the temperature-dependent multicomponent
solvent system
A TMS-system consists of a polar and an essentially non-polar
solvent. The third solvent which is semi-polar acts as a mediator
between the polar and the non-polar component. In this work, PC
was the polar solvent, dodecane, the non-polar liquid component,
and 1,4-dioxane, the semi-polar one. To determine the appropri-
ate composition of the solvent mixture for the reaction, the phase
diagram of the TMS-system was determined by cloud titration at
different temperatures from 298 to 373 K following the procedure
described elsewhere (Shaharun et al., 2008). The phase diagram with
the corresponding working points is presented in Fig. 2. The hydro-
formylation of 1-octene has been carried out in PC/dodecane/1,4-
dioxane with varying compositions denoted by A, B, C and D and
listed in Table 1. At a reaction temperature of 363 K, a syngas pres-
sure of 1.5 MPa and 0.17 mM catalyst concentration, the conversion
of 1-octene and the yield of total aldehyde were 53% and 47%, re-
spectively. Much higher conversion was achieved at increased cata-
lyst concentrations. With a reaction time of 2 h and a selectivity of
approximately 90%, this catalytic system can be considered as more
reactive than the aqueous biphasic catalytic system (Suárez et al.,
2006). The total turnover number was 600 while the turnover fre-
quency was 400 h−1 . Gas chromatographic analysis of the two liquid
phases obtained after cooling the mixture at the end of the reac-
tion showed that about 67% of the product preferentially remained
 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281 275

1,4-Dioxane 100
1 0

0.2 80
0.8

Conversion of 1-octene (%)

0.4
0.6 60

0.6
0.4
40 Catalyst concentration

0.8 0.0866 mM
0.2
0.18 mM
20
0.35 mM
1
0 0.68 mM
PC 1 0.8 0.6 0.4 0.2 0 Dodecane
0
Fig. 2. Phase diagram of the solvent system PC/dodecane/1,4-dioxane. () T = 298 K;
0 20 40 60 80 100 120 140
() T = 353 K; (%) T = 373 K and () operating points (A, B, C, D) of this work.
Reaction time (min)

Fig. 3. Time evolution of the yield of total aldehyde as at several concentra-

Table 1
tions of HRh(CO)(PPh3 )3 catalyst. Reaction conditions: P = 1.5 MPa, yH2 /yCO = 1/1,
Effect of the composition of TMS-systems.
T = 363 K, 1-octene = 1.9 kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, PC/dodecane/
Legend PC/dodecane/ Yield of n/i TON TOF (h−1 ) Rh 1,4-dioxane = 0.30/0.10/0.60.
1,4-dioxane total loss
(wt%) aldehyde (%)
(%) the addition of excess ligand (PPh3 :Rh = 5:1 molar ratio) system-
A 30.0/10.0/60.0 47.5 10.5:1 600.5 404.0 2.8 atically led to high rates of hydroformylation and high selectivity
B 25.5/13.5/61.0 47.0 9.7:1 574.9 383.3 3.0 with reduced rates of isomerization and hydrogenation of the olefin
C 21.8/15.8/62.4 47.9 9.5:1 463.2 308.8 6.3 (Huang et al., 2004). As shown in Fig. 3, the initial rate and conver-
D 17.8/18.2/64.0 50.6 8.5:1 618.3 412.2 10.4
sion increased with the concentration of the HRhCO(PPh3 )3 /P(OPh)3
Reaction conditions: P = 1.5 MPa, yH2 /yCO =1/1, HRh(CO)(PPh3 )3 = 1.7×10−4 kmol/m3 , catalyst. After 2 h, conversion of 1-octene increased from 50.0%
1-octene = 1.9 kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, temperature = 363 K, to 95.3% for an eight fold increase in catalyst concentration. The
reaction-time = 2 h.
n-product was favoured under all conditions studied even though
the regioselectivity to n-nonanal decreased with time. With 0.68 mM
in the less polar phase. The catalyst remained in the propylene car- HRh(CO)(PPh3 )3 , the conversion of 1-octene and the yield of total
bonate phase giving it a light yellow color. As shown in Table 1, the aldehyde were 97% and 95%, respectively. After 2 h of reaction time,
n-selectivity increases at a higher PC concentration in the TMS- the n/i aldehyde ratio was 8.4 and formation of the 2-octene isomer
system. Similar results on the effect of the fraction of PC in the was only 1.6%. A comparison of conversion and selectivity of hydro-
solvent mixture were reported by Behr et al. (2005) for the hydro- formylation reported in the literature for a few systems with that
formylation of trans-1-octene using Rh(acac)(CO)2 /BIPHEPHOS cat- achieved in the present work is presented in Table 2.
alyst. The solvent composition according to the operation point A
(PC/dodecane/1,4-dioxane: 0.30/0.10/0.60) given in Table 1, was se- 3.3. Effect of temperature
lected as a practical one for further study. The catalyst loss could be
correlated to the polarity and the solubility of the solvent mediator The reaction temperature can significantly influence the distribu-
in the product phase. 1,4-Dioxane proved to be a better solvent me- tion of products due to enhanced isomerization of the double bond
diator with low catalyst loss of 3% only compared to p-xylene which of the substrate olefin and the dissociation of ligand from the cat-
caused a higher rhodium leaching of about 47% (Behr et al., 2005). alytic complexes at higher temperatures (van Leeuwen, 2004). The
total conversion of 1-octene at reaction temperatures of 363 K and
3.2. Effect of catalyst concentration 383 K do not differ significantly. However, the total yields of alde-
hyde were higher at temperature of 363 K compared to 383 K. After
Industrial hydroformylation processes carried out in a single 2 h, the total yield of aldehyde was as high as 95% and the ratio of n/i
phase solvent using rhodium catalyst generally require high catalyst aldehydes was 8.4 (Fig. 4). Thus, a higher temperature increases the
loading up to 1 mM HRh(CO)(PPh3 )3 . In this study, a significantly rate of the hydroformylation but decreases the n/i ratio. The isomer-
less amount of catalyst was found to give satisfactory conversion ization of 1-octene into other internal octenes are favorable at higher
and selectivity. The high activity of the system may be due to the temperatures. The resulting higher concentrations of the branched
presence of an excess of the triphenylphosphite, P(OPh)3 ligand. octenes adversely affect the n/i ratio.
According to Beller et al. (1995), by using phosphite-modified cat-
alyst, even less reactive olefins such as 1-octene, 2,3-dihydrofuran 3.4. Effect of the total pressure of syngas
and 2,5-dihydrofuran, 1-butene and 2-butene are hydroformylated
at much higher rates compared to those achieved with phosphine- The effect of total pressure (yH2 /yCO = 1/1) on the total yield of
modified catalysts. Similar observation was reported before in the aldehyde is presented in Fig. 5. In general, an increase in the total
case of hydroformylation of 1-octene at 10 bar and 353 K, where pressure increased the rate and conversion of 1-octene but the total
276 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281

Table 2
A summary of the effects of reaction conditions on the conversion and selectivity involving Rh-catalyzed hydroformylation reaction.

Reference Catalyst Substrate [Catalyst] P (bar) T (K) t (h) Solvent Olefin con- Yield of n/iso
precursor (mol/m3 ) version (%) aldehyde
(%)

Klein et al. (2001) Rh(acac)(CO)2 NAPHOS; 1-pentene 0.080 10 393 16 Aanisole – 76.0 99.0
NAPHOS: Rh = 5

Huang et al. (2004) Rh4 (CO)12 1-octene 11.6 10 353 1.5 THF 96.7 6.5 1.4
Rh4 (CO)12 +PPh3 ; P:Rh = 5 97.7 93.1 2.4
Rh4 (CO)12 /MCM-41(NH2 ) 25.2 5.9 1.7
Rh4 (CO)12 /MCM-41(NH2 )+PPh3 ; 98.9 95.9 2.7
P:Rh = 5
RhCl(PPh3 )3 20.1 21.4 3.0
RhCl(PPh3 )/MCM-41(NH2 ) 73.6 61.0 2.9
RhCl(PPh3 )/MCM- 92.9 94.5 2.7
41(NH2 )+PPh3 ; P:Rh = 5

van Rooy et al. (1995) Rh(CO)2 (acac)tris(2- 1-octene 0.1 10 353 – Toluene – 44.0 1.9
tertbutyl-4-methylphenyl)
phosphate; P:Rh = 50

Behr et al. (2005) Rh(acac)(CO)2 biphephos; 4-octene 3.33 10 398 4 PC/dodecane/p-xylene 99.0 9.0
biphephos:Rh = 5

Tijani and Ali (2006) HRh(CO)(PPh3 )3 P(OPh)3 ; 1-octene 0.83 14 363 1.5 PC/heptane – 84.0 8.1
P:Rh = 12

Suárez et al. (2006) RhCl(CO)(TPPMS)2 1-hexene 0.19 68 373 1.5 biphasic toluene/H2 O 95.0 73.2 1.1
[TPPMS = P(C6 H5 )2 (C6 H4 SO3 )]
RhCl(CO)(TPPDS)2 96.0 73.9 1.3
[TPPDS = P(C6 H5 )(C6 H4 SO3 )2 ]
RhCl(CO)(TPPTS)2 94.0 74.3 1.1
[TPPTS = P(C6 H4 SO3 )3 ]

Present work HRh(CO)(PPh3 )3 P(OPh)3 ; 1-octene 0.17 15 363 1.5 PC/dodecane/ 1,4-dioxane 53.0 47.0 9.0
P:Rh = 12

yield of aldehydes were in the range of 88–90% and did not differ
significantly. However, with an increase in the total pressure, a de-
crease in selectivity of the linear aldehyde was observed. The n/i ra-
tio decreased from 7.7 at total pressure of 1.5 MPa to 3.8 at 2.5 MPa.
Typical experimental results on conversion and selectivity are given
in Table 3.
3.5. Kinetics
The reaction proceeds through catalytic addition of hydrogen and
the formyl (CHO) group across the double bond of the olefin to give
aldehydes. The main steps in the catalytic cycle follow the Heck
and Breslow mechanism developed for the cobalt-catalyzed oxo re-
action (van Leeuwen, 2004). Based on the observations reported
in literature, it can be stated that the rate of hydroformylation is
influenced positively by increasing the concentrations of the cat-
alyst and hydrogen, whereas increased carbon monoxide pressure
exerts a negative effect. However, when vinyl acetate or 1-hexene
was the substrate, Chaudhari noted a critical catalyst concentration
(C*) of 0.2–0.4×10−3 kmol/m3 below which no reaction occurs. Be-
yond C*, the dependence on catalyst concentration was first order
(Deshpande and Chaudhari, 1988, 1989). The inactivity at low cata-
lyst concentrations was attributed to a high substrate/catalyst ratio,
leading to catalytically inactive dimer formation. In addition, sev-
eral previous researchers noted distinct substrate inhibition, which
is attributed to the inactive dimeric rhodium species that form at
high substrate/catalyst ratios (Bhanage et al., 1997; Deshpande and
Chaudari, 1988). The structures of these inactive species have not
been confirmed. For the conventional system, the substrate/catalyst
ratio at which inhibition took place had values between 400 and
1000, depending on temperature and substrate. However, this ratio
is obtained from the catalytic system using HRh(CO)(PPh3 )3 catalyst
and without the presence of free PPh3 ligand.
The initial rates of hydroformylation were determined in the
range of conditions as shown in Table 4. Since the reaction occurs
in the liquid phase and two of the reactants are supplied as gas,
mass transfer limitation may influence the rate of reaction. This was
tested by conducting the reaction at different stirrer speeds under
otherwise unchanged experimental conditions. Increase of the stir-
rer speed above 400 rpm had no effect on the rate of reaction and ex-
istence of the kinetic regime was confirmed (Fig. 6). This conclusion
was further supported by a comparison of the reaction rates with
the maximum rates of mass transfer (R0 /(kL aCH ∗ ), R /(k aC ∗ )) of CO
2
0 L CO
and H2 at the respective conditions. Here, R0 represents the initial
rate of hydroformylation, in kmol/(m3 s); kL a is the volumetric mass
∗ and C ∗ are their interfacial con-
transfer coefficient, in s−1 ; and CH
2 CO
centrations (i.e., equilibrium solubilities) of H2 and CO in kmol/m3 ,
respectively. We measured the maximum rates of mass transfer of
both H2 and CO independently. The requisite volume of the solvent
only was taken in the reactor which was then quickly pressurized.
From the recorded time rate of change of pressure of the gas, the
initial rate of physical absorption could be estimated. The calculated
ratio of initial reaction rate to the maximum mass transfer rate of
gas-to-liquid was found to be less than 0.10 at 450 rpm. This con-
firms that the reaction is kinetically controlled under the conditions
used in this study.
 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281 277

100 Table 3
Typical results on conversion of 1-octene, selectivity and yields.a

Tempera- [catalyst] [1-octene] P (atm) Conversion Yield of n/i

ture (K) (mol/m3 ) (kmol/m3 ) of 1-octene aldehyde
(%) (%)
80
353 0.68 1.90 15.0 95.3 60.0 10.5
363 0.68 1.90 15.0 97.2 95.0 8.4
Yield of total aldehyde (%)

373 0.68 1.90 15.0 98.6 83.0 7.5

383 0.68 1.90 15.0 99.1 68.0 7.0
60 363 0.086 1.90 15.0 50.0 48.2 3.0
363 0.18 1.90 15.0 53.0 47.0 5.0
363 0.35 1.90 15.0 93.0 89.7 7.7
363 0.35 1.90 20.0 95.3 89.0 5.5
363 0.35 1.90 25.0 98.8 88.2 3.8
40 a
Reaction conditions: H2 /CO = 1, reaction time = 2 h, catalyst: P(OPh)3 = 5:1
molar ratio.
Temperature

T at 353 K
20 Table 4
T at 363 K
Range of process conditions for kinetic study of the hydroformylation of 1-octene
T at 373 K in TMS-system of PC/dodecane/1,4-dioxane using HRh(CO)(PPh3 )3 /P(OPh)3 catalyst.
T at 383 K
Conc. of catalyst (kmol/m3 ) 8.66×10−5 –6.78×10−4
0 Initial conc. of 1-olefin (kmol/m3 ) 0.21–4.2
0 20 40 60 80 100 120 140 Pt (MPa) 1.5–2.5
T (K) 353–373
Reaction time (min) Reaction volume (mL) 240
Solvent PC/dodecane/1,4-dioxane (0.30/0.10/0.60)
Fig. 4. Time evolution of the yield of total aldehyde at four reaction tempera-
tures. Reaction conditions: P = 1.5 MPa, yH2 /yCO = 1/1, HRh(CO)(PPh3 )3 = 6.8×10−4
kmol/m3 , 1-octene = 1.9 kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, PC/dodecane/
1,4-dioxane = 0.30/0.10/0.60. 3

2.5
100
R x105 (kmolm-3s-1)

80
1.5
Yield of total aldehyde (%)

60 1

0.5
40

Total pressure 0
350 400 450 500 550 600
1.5 MPa rpm
20 2.0 MPa
Fig. 6. Effect of agitation speed on the rate of reaction in hydroformylation of
2.5 MPa
1-octene. Reaction conditions: P = 1.5 MPa, yH2 /yCO =1/1, HRh(CO)(PPh3 )3 = 6.8×10−4
kmol/m3 , 1-octene = 1.9 kmol/m3 , P(OPh)3 /HRh (CO)(PPh3 )3 = 12, PC/dodecane/
1,4-dioxane = 0.30/0.10/0.60, temperature = 363 K.
0
0 20 40 60 80 100 120 140
Reaction time (min)
(gas pressure = 2.0–2.5 MPa), and the initial light color of the re-
Fig. 5. Yield of total aldehyde as a function of reaction time at different total syngas action mixture sustained. Deshpande et al. (1992) reported that a
pressure. Reaction conditions: yH2 /yCO = 1/1, HRh(CO)(PPh3 )3 = 3.5×10−4 kmol/m3 , higher partial pressure of CO could prevent the formation of the
1-octene = 1.9 kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, PC/dodecane/ 1, 4-dioxane inhibiting species that led to catalyst deactivation. Wilkinson (see
=0.30/0.10/0.60, temperature = 363 K.
van Leeuwen et al., 2000) also reported the formation of the orange
dirhodium species, Rh2 (CO)5 L3 at a low partial pressure of H2 and
high Rh catalyst concentration. Since the formation of dirhodium
The kinetics of the reaction has been investigated at three tem- species is a reversible process, at a high partial pressure of H2 , the
peratures −353, 363 and 373 K. At the higher temperature (373 K), active rhodium hydrides (HRh(CO)2 L2 ) are regenerated and might be
a little color develops in the reaction mixture perhaps due to the another cause of the positive influence of the total pressure of the
decomposition of the triphenylphosphite ligand. This affects the ac- syngas on the reaction rate.
tivity and stability of catalyst. However, the catalyst was found to The initial reaction rates are plotted against the catalyst concen-
be more stable at higher concentrations of dissolved CO and H2 tration (Fig. 7 ), the concentration of 1-octene (Fig. 8) and the partial
278 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281

3 1.4
353 353
363 1.2 363
2.5 373
373
modeled (Eq. 9)
Rate, Rx104 (kmol/m3.s)

Rate, Rx104 (kmol/m3.s)

2

0.8
1.5
0.6
1
0.4

0.5
0.2

0 0
0 0.2 0.4 0.6 0.8 0 2 4 6 8 10 12
Catalyst concentration, Cx103 (kmol/m3) PH2 (atm)

Fig. 7. Effect of catalyst concentration on the rate of reaction. Reaction conditions: Fig. 9. Effect of partial pressure of H2 on the rate of reaction. Reaction conditions:
P = 1.5 MPa, yH2 /yCO = 1/1, 1-octene = 1.9 kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, P = 1.5 MPa, yH2 /yCO = 1/1, 1-octene = 1.9 kmol/m3 , HRh(CO)(PPh3 )3 = 1.73×10−4
PC/dodecane/1,4-dioxane = 0.30/0.10/0.60. kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, PC/dodecane/1,4-dioxane = 0.30/0.10/0.60.

1.8 1.4
353
1.6
363 1.2
373
1.4
1
Rate, Rx104 (kmol/m3.s)
Rate, Rx104 (kmol/m3.s)

1.2

1 0.8

0.8 0.6

0.6 353 K
0.4 363 K
0.4 373 K
0.2 ⎯ Modeled (Eq. 9)
0.2

0 0
0 1 2 3 4 5 0 2 4 6 8 10 12
PCO (atm)
Concentration of 1-octene (kmol/m3)

Fig. 8. Effect of 1-octene concentration on the rate of reaction. Reaction con- Fig. 10. Effect of partial pressure of CO on the rate of reaction. Reaction conditions:
ditions: P = 1.5 MPa, yH2 /yCO = 1/1, HRh(CO)(PPh3 )3 = 1.73×10−4 kmol/m3 , P(OPh)3 / P = 1.5 MPa, yH2 /yCO = 1/1, 1-octene = 1.9 kmol/m3 , HRh(CO) (PPh3 )3 = 1.73×10−4
HRh(CO)(PPh3 )3 = 12, PC/dodecane/1,4-dioxane = 0.30/0.10/0.60. kmol/m3 , P(OPh)3 /HRh(CO)(PPh3 )3 = 12, PC/dodecane/ 1,4-dioxane = 0.30/0.10/0.60.

pressure of H2 (Fig. 9) at all the three temperatures. The plots are have developed a rate equation based on the following simplified
linear with a high correlation coefficient ( > 0.995). Thus the reaction reaction steps.
rate is first order with respect to each of the above two concentra-
k1
tions (catalyst and 1-octene) and the partial pressure of H2 as well. HRh(CO)L2 + CO  HRh(CO)2 L2 (2)
However, the rate was found to be inversely dependent on the CO k−1

partial pressure in the range of 4.0–11.5 atm (Fig. 10). In the present
k2
study, substrate inhibition was absent over the range of substrate and HRh(CO)2 L2 + RCH&CH2  Rh(CO)L2 (RCH2 CH2 · CO) (3)
catalyst concentrations selected, where the substrate/catalyst ratio k−2
varied from 580 to 2300. This may be due to several factors—solvent
k
effects, the phosphite ligands, and the increased H2 and CO con- Rh(CO)L2 (RCH2 CH2 · CO) + H2 → HRh(CO)L2 + RCH2 CH2 · CHO (4)
centrations relative to conventional systems (Shaharun et al., 2008).
The parametric effects on the reaction kinetics was found to be sim- Reactions (1) and (2) are assumed to be in `equilibrium' while re-
ilar to that observed for a homogeneous hydroformylation system action (3)—the oxidative addition of hydrogen—is rate controlling. If
under industrial operating conditions (van Leeuwen et al., 2000). K1 and K2 are the equilibrium constants of the first two reactions
Although the reaction reportedly involves a number of steps, we that stand for addition of a mole of CO to the catalyst followed by
 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281
addition of a molecule of the olefin with simultaneous acylation
reaction, then,
[HRh(CO)2 L2 ] = K1 [HRh(CO)L2 ][CO]
and
[Rh(CO)L2 (RCH2 CH2 · CO)] = K2 [Rh(CO)2 L2 ][RCH2 CH2 ]
Here a square bracket denotes concentration of a species. The total
concentration of the rhodium catalyst may be expressed as
[Catalyst] = [HRh(CO)2 L2 ] + [HRh(CO)L2 ] + [Rh(CO)L2 (RCH2 CH2 .CO)]
Using Eqs. (5) and (6),
[Catalyst]
[Rh(CO)L2 (RCH2 CH2 · CO] =
1 + K1 [CO] + K2 [CO][RCH&CH2 ]
The rate of reaction can be expressed as
Rate = k[Rh(CO)L2 (RCH2 CH2 · CO)][H2 ]
k [Catalyst][CO][RCH&CH2 ][H2 ]
= (7)
1 + K1 [CO] + K1 K2 [CO][RCH&CH2 ]
van der Veen et al. (2000) reported that the association/dissociation
of CO Eq. (2) can be treated as a pre-equilibrium step, since the rate
of this reaction is two orders of magnitude faster that the overall hy-
droformylation rate and the HRh(CO)L2 complex is not identifiable
under hydroformylation conditions or in the CO dissociation stud-
ies. However, the relative concentrations of HRh(CO)2 L2 intermedi-
ate is arguably controlled by the CO concentration (van Leeuwen
et al., 2000). Theoretical studies indicate that a transition state in-
volving migratory insertion of ethene into the Rh-hydride species,
HRh(CO)2 L2 , exists (Decker and Cundari, 2002). The rotation of the
alkene from the in-plane coordination to a perpendicular coordina-
tion mode contributes to the barrier. Using the ONIOM molecular
mechanics approach, the calculated activation energy was estimated
as 49 kJ/mol. A transition state involving intramolecular CO insertion
to form acyl complexes was also found which require an activation
energy of 84 kJ/mol at the MP2 level (Matsubara et al., 1997). Since
the olefin insertion reaction given by Eq. (3) is likely to be an al-
ternative rate determining step (van Leeuwen, 2004), it is expected
that the equilibrium constant, K2 is an order of magnitude smaller
than K1 . Thus, at a reasonably high pressure of CO, the equilibrium
constant of the reaction 1 is presumably much larger than that for
reaction 2, i.e., K1  K2 . The rate Eq. (7) then reduces to
k[Catalyst][CO][RCH&CH2 ][H2 ]
Rate =
1 + K1 [CO]
The aldehyde concentration versus time data for each experiment
were fitted to a polynomial, which was then differentiated to deter-
mine the slope (experimental reaction rate in kmol/(m3 s) at each
data point. This rate versus concentration data were then used to
determine the parameters by fitting the experimental rate. We ob-
served that the rate data fitted much better if the denominator is
slightly modified as in Eq. (9) below. The reason may be the simpli-
fications made in the reaction steps given by Eqs. (2)–(4).
k[octene][catalyst]PCO PH2
Rate =
(1 + KCO PCO )m
Estimation of the kinetic parameters by non-linear least square re-
gression method needs appropriate initial guess values in order to
obtain optimized values which otherwise might converge to local
minima (Wojciechowski and Rice, 2003). The semi-empirical rate
Eq. (9) fits the observed rate data within the limit of experimental
279
Table 5
Estimated kinetic model parameters with 95% confidence limit.
T (K) k m3 /(kmol atm2 s) Kco (atm−1 ) m min (×10−11 ) SEE (%)
(5)
353 0.55 0.52 1.72 1.54 5.97
363 1.11 1.18 1.60 5.00 6.35
373 1.97 2.24 1.36 3.79 4.26
(6)
error. Activation energy was estimated bus sing the Arrhenius
equation with `temperature centering'.
  
−E 1 1
k = k0 exp − (10)
R T Tm
The technique of `temperature centering' admits of a robust param-
eter estimation of the Arrhenius equation (Pant and Kunzru, 1997;
Wojciechowski and Rice, 2003; Patel and Pant, 2007). In the present
study Tm = 363 K was used for temperature centering and parameter
estimation using Eq. (9).
The non-linear least square regression based on the criterion
of minimization of the statistical parameter mean residual sum of
squares () was performed to determine the kinetic parameters,
N
− Rexpt )2
i=1 (Rcalc
= (11)
Nexpt − Nparam
where Nexpt is the number of experimental data, Nparam is number
of model parameters, Rcalc and Rexpt represent calculated and exper-
imental rates, respectively.
It will be pertinent to compare the above rate equation, Eq. (9),
with those reported by other workers for hydroformylation of 1-
octene in different solvent media—van Rooy et al. (1995) who stud-
ied the kinetics using toluene as the solvent, and Palo and Erkey
(1999) who did it in supercritical carbon dioxide medium. van Rooy
et al. (1995) used Rh(CO)2 (acac) as the catalyst and a bulky tris
(2-tert-butyl-4-methyl phenyl phosphate) as the ligand. The rate
equation they proposed did not show any dependence on the olefin
concentration but had an inverse dependence on CO concentration.
In contrast, the rate of reaction for the present thermomorphic sys-
tem, Eq. (9), has a first order dependence on the olefin concentration
and essentially linear dependence on CO concentration if it is low.
The rate equation derived by Palo and Erkey (1999), which is con-
siderably different from that of van Rooy et al. (1995), shows half-
order dependence of the rate on concentrations of both olefin and
hydrogen. The dependence on the catalyst concentration is nearly
linear (0.84, to be more precise). A stronger inverse dependence on
CO concentration was suggested.
The semi-empirical model is able to describe the experimental
(8) data quite satisfactorily in the entire range of concentrations of re-
actants and of the catalyst. Table 5 summarizes the values of the
rate and equilibrium constants and the average standard error of
estimation (SEE) and  values of the non-linear regression analysis
for the above rate equation at 353, 363 and 373 K and Fig. 11 shows
the comparison of experimental and predicted values of the rate of
hydroformylation. The average deviation between experimental and
calculated rate of reaction was found to be in the range of ± 5.5%
and the activation energy was found to be 69.8 kJ/mol, which lies
in the range of activation energy values (66–75 kJ/mol) reported by
other workers for the hydroformylation of 1-octene with different
Rh-complexes by homogeneous, biphasic and supported aqueous
(9)
phase catalysis, SAPC (Jáuregui-Haza et al., 2001; Deshpande et al.,
1996; Herrmann et al., 1992; Arhancet et al., 1991).
4. Conclusions
The hydroformylation of 1-octene using HRh(CO)(PPh3 )3 /P(OPh)
catalyst can be conveniently performed in the thermomorphic
280 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281

2 Acknowledgments
353 K
1.8 363 K Financial assistance from Ministry of Science, Technology and
373 K Innovation (MOSTI), Government of Malaysia, under the Research
1.6 Project no. 03-02-02-SF0019: Development of a green process for the
104 x calculated reaction rate (kmolm-3s-1)

production of higher aldehydes from olefins by hydroformylation is

1.4 gratefully acknowledged.

References
1.2
Arhancet, J.P., Davis, M.E., Hanson, B.E., 1991. Supported aqueous-phase, rhodium
1 hydroformylation catalyst I. New methods of preparation. Journal of Catalysis
121, 94–99.
Behr, A., Fangewisch, C., 2003. Rhodium-catalysed synthesis of branched fatty
0.8 compounds in temperature-dependent solvent systems. Journal of Molecular
Catalysis A: Chemical 197, 115–126.
Behr, A., Miao, Q., 2004. A new temperature-dependent solvent system based on
0.6 polyethylene glycol 1000 and its use in rhodium catalyzed cooligomerization.
Journal of Molecular Catalysis A: Chemical 222, 127–132.
0.4 Behr, A., Obst, D., Turkowski, B., 2005. Isomerizing hydroformylation of trans-4-
octene to n-nonanal in multiphase systems: acceleration effect of propylene
carbonate. Journal of Molecular Catalysis A: Chemical 226, 215–219.
0.2 Behr, A., Roll, R., 2005. Hydroaminomethylation in thermomorphic solvent systems.
Journal of Molecular Catalysis A: Chemical 239, 180–184.
Beller, M., Cornils, B., Frohning, C.D., Kohlpaintner, C.W., 1995. Progress in
0 hydroformylation and carbonylation. Journal of Molecular Catalysis A: Chemical
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 104, 17–85.
Bhanage, B.M., Divekar, S.S., Deshpande, R.M., Chaudari, R.V., 1997. Kinetics of
104 x experimental reaction rate (kmolm-3s-1) hydroformylation of l-dodecene using homogeneous HRh(CO)(PPh3 )3 catalyst.
Journal of Molecular Catalysis A: Chemical 115, 247–257.
Fig. 11. Parity plot of the experimental and calculated reaction rates. Bohnen, H., Cornils, B., 2003. Hydroformylation of alkenes: an industrial view of the
status and importance. ChemInform 34, 17–25.
Decker, S.A., Cundari, T.R., 2002. A quantum mechanics/molecular mechanics study
solvent systems consisting of propylene carbonate, dodecane and of the steric influence of the PR3 spectator ligands on the energetics of ethylene
insertion into the Rh–H bond of HRh(PR3 )2 (CO)(2 -CH2 = CH2 ). New Journal of
1,4-dioxane. High conversions of 1-octene up to 97% in 2 h at 363 K
Chemistry 26, 129–135.
and good selectivities to the n-aldehyde up to 89% can be achieved. Deshpande, R.M., Chaudari, R.V., 1988. Kinetics of 1-hexene using homogeneous
After cooling down of the reaction mixture at the end, the catalyst HRh(CO)(PPh3 )3 complex catalyst. Industrial and Engineering Chemistry Research
can be easily recovered by a simple phase separation. The rhodium 27, 1996–2002.
Deshpande, R.M., Chaudhari, R.V., 1989. Hydroformylation of vinyl acetate using
leaching to the product phase is very low (3%). The kinetics has homogeneous HRh(CO)(PPh3 )3 catalyst: a kinetic study. Journal of Molecular
been investigated in a temperature range of 353–373 K. The rate Catalysis 57, 177–189.
versus partial pressure of CO shows typical substrate inhibited Deshpande, R.M., Purwanto, R.P., Delmas, H., Chaudhari, R.V., 1996. Kinetics of
hydroformylation of 1-octene using [Rh(COD)Cl]2 -TPPTS complex catalyst in
kinetics. The rate was found to be first order with respect to the a two-phase system in the presence of a cosolvent. Industrial Engineering
concentrations of 1-octene and the catalyst and the partial pressure Chemistry Research 35, 3927–3936.
of hydrogen. The activation energy was found to be 69.8 kJ/mol. The Deshpande, R.M., Divekar, S.S., Bhanage, B.M., Chaudhari, R.V., 1992. Effect of solvent
on the kinetics of hydroformylation of 1-hexene using HRh(CO)(PPh3 )3 catalyst.
observed kinetics differ significantly from those previously obtained Journal of Molecular Catalysis 77, L13–L17.
for HRh(CO)(PPh)3 in other media (van Rooy et al., 1995; Palo and Herrmann, W.A., Kohlpaintner, C.W., Bahrmann, H., Konkol, W., 1992. Water-soluble
Erkey, 1999). The most significant observation is the lack of olefin metal complexes and catalysts part 6. A new, efficient water-soluble catalyst
for two-phase hydroformylation of olefins. Journal of Molecular Catalysis 73,
inhibition, and the absence of a critical catalyst concentration. This 191–201.
may be due to solvent effects, the phosphite ligand and the in- Horváth, I.T., Rábai, J., 1994. Facile catalyst separation without water: fluorous
creased H2 and CO concentrations relative to conventional systems. biphase hydroformylation of olefins. Science 266, 72–75.
Huang, L., He, Y., Kawi, S., 2004. Catalytic studies of aminated MCM-41-tethered
These results establish the potential of the TMS-system for practical
rhodium complexes for hydroformylation of 1-octene and styrene. Journal of
application in industrial hydroformylation of higher alkenes. Molecular Catalysis A: Chemical 213, 241–249.
Jáuregui-Haza, U.J., Dessoudeix, M., Kalck, Ph., Wilhelm, A.M., Delmas, H., 2001.
Multifactorial analysis in the study of hydroformyaltion of oct-1-ene using
Notation supported aqueous phase catalysis. Catalysis Today 66, 297–302.
Klein, H., Jackstell, R., Wiese, K-D., Borgmann, C., Beller, M., 2001. Highly selective
catalysts systems for the hydroformylation of internal olefins to linear aldehydes.
[catalyst] concentration of catalyst HRh(CO)(PPh3 )3 /P(OPh)3 Angew Chemica International Edition 40, 3408–3411.
in the reaction mixture, kmol/m3 Kolhofer, A., Plenio, H., 2003. Homogeneous catalysts supported on soluble polymers:
E activation energy for rate constant, kJ/mol biphasic Sonogashira coupling of aryl halides and acetylenes using MeOPEG-
bound phosphine-palladium catalysts for efficient catalyst recycling. Chemistry
k reaction rate constant, m3 /(kmol atm2 s)
— European Journal 9, 1416–1422.
k0 pre exponential rate constant, m2 /(s kmol) Liu, C., Jiang, J., Wang, Y., Cheng, F., Jin, Z., 2003. Thermoregulated phase transfer
KCO constant in Eqs. (8) and (9), atm−1 ligands and catalysis XVIII: synthesis of N,N-dipolyoxyethylene-substituted-2-
(diphenylphosphino)phenylamine (PEO-DPPPA) and the catalytic activity of its
[octene] concentration of 1-octene in the reaction mixture,
rhodium complex in the aqueous-organic biphasic hydroformylation of 1-decene.
kmol/m3 Journal of Molecular Catalysis A: Chemical 198, 23–27.
PCO partial pressure of CO, atm Matsubara, T., Koga, N., Ding, Y., Musaev, D.G., Morokuma, K., 1997. Ab initio MO
PH2 partial pressure of H2 , atm study of the full cycle of olefin hydroformylationc catalyzed by a rhodium
complex, RhH(CO)2 (PH3 )2 . Organometallics 16, 1065–1078.
R universal gas constant, 8.314 J/(mol K) Palo, D.R., Erkey, C., 1999. Kinetics of the homogeneousand. Engineering Chemistry
T temperature, K Research 38, 3786–3792.
Pant, K.K., Kunzru, D., 1997. Catalytic pyrolysis of n-heptane: kinetic and modeling.
Greek letter Industrial and Engineering Chemistry Research 36, 2059–2065.
Patel, S., Pant, K.K., 2007. Experimental study and mechanistic kinetic modeling for
selective production of hydrogen via catalytic steam reforming of methanol.
 objective function
Chemical Engineering Science 62, 5425–5435.
 M.S. Shaharun et al. / Chemical Engineering Science 65 (2010) 273 -- 281
Shaharun, M.S., Mukhtar, H., Dutta, B.K., 2008. Solubility of carbon monoxide
and hydrogen in propylene carbonate and thermomorphic multicomponent
hydroformylation solvent. Chemical Engineering Science 63, 3024–3035.
Suárez, T., Fontal, B., Gustavo León, Reyes, M., Bellandi, F., Contreras, R.R., Cancines,
P., 2006. Aqueous biphasic olefin hydroformylation catalyzed by water-soluble
rhodium complexes. Transition Metal Chemistry 31, 974–976.
Tijani, J., Ali, B.E., 2006. Selective thermomorphic biphasic hydroformylation of higher
olefins catalyzed by HRhCO(PPh3 )3 /P(OPh)3 . Applied Catalysis A: General 303,
158–165.
van Leeuwen, P.W.N.M., 2004. Homogeneous Catalysis Understanding the Art. Kluwer
Academic Publishers, Dordrecht, The Netherlands.
281
van Leeuwen, P.W.N.M., Claver, M., Carman, N., 2000. Rhodium Catalyzed
Hydroformylation. Kluwer Academic Publishers, Dordrecht, The Netherlands.
van der Veen, L.A., Keeven, P.H., Schoemaker, G.C., Reek, J.N.H., Kramer, P.C.J., van
Leeuwen, P.W.N.M., Lutz, M., Spek, A.L., 2000. Origin of the bite angle effect on
rhodium diphosphine catalyzed hydroformylation. Organometallics 19, 872–883.
van Rooy, A., Orij, E.N., Kamer, P.C.J., van Leeuwen, P.W.N.M., 1995. Hydroformylation
with a rhodium/bulky phosphite modified catalyst catalyst comparison for oct-
1-ene, cyclohexene and styrene. Organometallics 14, 34–43.
Wojciechowski, B.W., Rice, N.M., 2003. Experimental Methods in Kinetic Studies.
Elsevier Science BV, Amsterdam.