The kinetics of gas-phase propene hydroformylation over a supported ionic liquid-phase (SILP) rhodium catalyst

Article abstract of DOI:10.1016/j.jcat.2012.05.011

An investigation of the kinetics of propene hydroformylation in the gas phase has been conducted over a silica-supported Rh-sulfoxantphos complex stabilized by the ionic liquid [bmim][OctSO₄]. The reaction temperature was found to have a strong effect on the kinetics of n- and iso-butanal formation. For both products, it was observed that increasing the temperature decreased the apparent activation energy, altered the reaction orders with respect to reactants, and decreased the molar ratio of n- to iso-butanal. The observed changes in the kinetics are discussed in terms of the generally accepted mechanism for olefin hydroformylation and are attributed to a change in the rate-determining step (RDS). It is concluded that at low temperature, the RDS is alkene insertion into an Rh-H bond but becomes the oxidative addition of H₂ at high temperature. The change in the RDS is rationalized in terms of a change in the elementary step with the largest Gibbs free energy of activation (ΔG^?). A greater loss in entropy for the oxidative addition of H₂ over alkene insertion causes the ΔG^? of the oxidative addition to be greater than the ΔG^? of alkene insertion at high temperature.

Full text of DOI:10.1016/j.jcat.2012.05.011

Journal of Catalysis 292 (2012) 166–172
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
The kinetics of gas-phase propene hydroformylation over a supported ionic
liquid-phase (SILP) rhodium catalyst
⇑
David G. Hanna, Sankaranarayanapilla Shylesh, Sebastian Werner, Alexis T. Bell
Department of Chemical and Biomolecular Engineering, University of California, Berkeley, Berkeley, CA 94720, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
An investigation of the kinetics of propene hydroformylation in the gas phase has been conducted over a
Received 7 March 2012
Revised 11 May 2012
Accepted 12 May 2012
Available online 25 June 2012
4
silica-supported Rh–sulfoxantphos complex stabilized by the ionic liquid [bmim][OctSO ]. The reaction
temperature was found to have a strong effect on the kinetics of n- and iso-butanal formation. For both
products, it was observed that increasing the temperature decreased the apparent activation energy,
altered the reaction orders with respect to reactants, and decreased the molar ratio of n- to iso-butanal.
The observed changes in the kinetics are discussed in terms of the generally accepted mechanism for ole-
fin hydroformylation and are attributed to a change in the rate-determining step (RDS). It is concluded
that at low temperature, the RDS is alkene insertion into an Rh–H bond but becomes the oxidative addi-
Keywords:
Propene
Hydroformylation
Butanal
Ionic liquid
Rhodium
SILP
tion of H
2
at high temperature. The change in the RDS is rationalized in terms of a change in the elemen-
à
tary step with the largest Gibbs free energy of activation (
addition of H
D
G ). A greater loss in entropy for the oxidative
à
à
2
over alkene insertion causes the
D
G of the oxidative addition to be greater than the
DG of
alkene insertion at high temperature.
Kinetics
Ó 2012 Elsevier Inc. All rights reserved.
1
. Introduction
silica [8]. Subsequent work by Davis and coworkers led to the
development of supported aqueous-phase (SAP) catalysts in which
an aqueous solution of a Rh complex is supported on microporous
glass [9]. Such catalysts proved effective for the hydroformylation
of water-insoluble alkenes in nonpolar solvents [10]. However,
subsequent work revealed that SAPs are not suitable for gas-phase
hydroformylation due to evaporation of the aqueous phase [11].
Since ionic liquids have negligible vapor pressure, Mehnert and
coworkers developed supported ionic liquid catalysts (SILCs) by
The liquid-phase hydroformylation of alkenes to aldehydes, cat-
alyzed by Rh complexes, is used extensively to produce intermedi-
ates for a number of commodity chemicals (e.g., plasticizers,
detergents, etc.) [1]. Two of the ongoing challenges with the use
of homogeneous Rh complexes are separation of the organic prod-
ucts from the catalyst and loss of Rh due to leaching [2]. Biphasic
systems such as the Ruhrchemie/Rhône-Poulenc process are cur-
rently used to facilitate product separation and reduce Rh leaching
for propene hydroformylation [1]. The reaction is carried out in an
aqueous phase containing the solubilized Rh complex, whereas an
organic phase is used to extract the product aldehyde. Although
satisfactory for propene hydroformylation, this process is not suit-
able for longer chain olefins due to decreased alkene solubility in
the aqueous phase [2,3]. Another disadvantage of such systems is
the cost for recovering the alkene from the organic phase [3].
To circumvent the issues associated with using biphasic sys-
tems, a number of authors have examined the possibility of dis-
persing the active Rh complex in a liquid contained within the
pores of a high surface area solid or grafting the complex to the
surface of a support material [4–7]. In the early 1980s, Gerritsen
and coworkers showed that molten triphenlyphosphine could be
4 6
impregnating a solution of Rh-tppts in [bmim][BF ] or [bmim][PF ]
into a 1-n-butyl-3-methyl imidazolium modified silica gel support
[12]. In later studies conducted by Wasserscheid and coworkers,
analogous supported ionic liquid-phase (SILP) catalysts were found
to be exceptionally stable and active under continuous gas-phase
reaction conditions [13,14]. The stability, high activity, and facile
separation of the products from the Rh complex makes SILP cata-
lysts particularly promising for the commercial application of pro-
pene hydroformylation.
While previous studies of SILP catalysts have suggested that the
Rh complex is homogeneously dispersed in an IL film [12–15], our
3
1
recent study using P MAS NMR and in situ FT-IR suggests that
interactions of the IL and metal complex with the support are re-
quired to yield a stable catalysts [16]. As shown in Scheme 1, the
phosphine ligand, sulfoxantphos (SX), interacts with the slightly
acidic silanol groups via the lone pair of electrons on phosphorus
3
used as a solvent for HRh(CO)(PPh ) supported on c-alumina and
(
conformation A) or in an ion–dipole interaction with the charged
⇑
Corresponding author. Fax: +1 510 642 4778.
E-mail address: bell@cchem.berkeley.edu (A.T. Bell).
sulfonate group (conformation B). Similarly, the IL ([bmim][Oct-
0
021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.05.011

D.G. Hanna et al. / Journal of Catalysis 292 (2012) 166–172
167
4
Scheme 1. Interaction of SX (conformations A and B) and [bmim][OctSO ] (conformation C) with surface silanol groups.
SO
4
]) interacts with the silanol groups via the sulfonate group of
]) (conformation C). Conformation B is desired
sured with a K-type thermocouple sheathed in a quartz capillary
placed in direct contact with the catalyst bed.
the anion ([OctSO
4
to yield active catalysts because the phosphine groups on SX are
free to coordinate to the Rh cations.
In this work, we report an investigation of the kinetics for the
gas-phase hydroformylation of propene over a SILP catalyst. Previ-
Prior to reaction, the 0.3 g of catalyst was heated to the reaction
ꢀ1
temperature at a rate of 2 K min in pure He (Praxair, 99.999%)
3
ꢀ1
flowing at 20 cm min at STP. The feed to the reactor consisted
of propene (Praxair, 99.9%), CO (Praxair, 99.99%), H (Praxair,
99.999%). A reactant ratio of C :CO:H of 1:1:1 was used unless
2
ous studies have suggested that the oxidative addition of H
2
is the
H
3 6
2
RDS [7,17,18] despite arguments that the RDS occurs earlier in the
hydroformylation cycle [19,20]. Other authors have claimed that
the RDS is dependent upon the partial pressures of propene and
hydrogen [21]. We show that the kinetics for forming n- and iso-
butanal are functions of the reaction temperature suggesting that
the rate-determining step (RDS) is also dependent upon the reac-
tion temperature. Rate equations were developed based upon the
Wilkinson mechanism for olefin hydroformylation [22], and an
explanation was given to rationalize the temperature effects on
the RDS.
specified otherwise. Experiments were carried out at total gas
pressures of 2 atm by throttling a needle valve located down-
stream from the reactor. Total gas flow rate was typically
120 cm min at STP. Using these conditions, the conversion of
propene was always less than 1%. Reaction products were analyzed
using an Agilent 6890N gas chromatograph containing a bonded
and crosslinked (5%-Phenyl)-methylpolysiloxane capillary column
(Agilent, HP-1) connected to a flame ionization detector.
3
ꢀ1
3
. Results
.1. Catalytic activity of SX–Rh/SiO
To avoid artifacts due to catalyst deactivation, catalyst stability
2
. Material and methods
3
2
SILP catalysts
2
.1. Synthesis of SX–Rh/SiO SILP catalysts
2
with time on stream was examined. Fig. 1 shows a representative
example of the results, demonstrating that the catalyst is stable for
SILP catalysts were prepared following the procedure described
in Ref. [15]. In brief, 5 mg of Rh(CO) (acac) (Fluka, purum) was dis-
2
solved in 10 mL of anhydrous methanol (Aldrich, 99.8%), and then,
approximately 150 mg of sulfoxantphos (SX) [23] was added to
this solution to achieve a molar ratio of SX/Rh = 10. After 10 min
4
of stirring, 120 mg of [bmim][OctSO ] [24] was added to the mix-
ture. The volume of IL added to solution was 10% of the pore vol-
ume of the silica support. After another 1 h of stirring, 1 g of
2
ꢀ1
silica (Silicycle, 500 m g , average pore diameter 60 Å) stored un-
der vacuum at 353 K was added to solution and was stirred for an
additional 1 h. Methanol was then slowly removed under vacuum
in a rotary evaporator. The resulting catalyst is a red colored, free
ꢀ1
flowing powder containing 0.002 g Rh (g SiO
2
)
.
2.2. Gas-phase hydroformylation of propene
Measurements of reaction rates were performed in a 6.35 mm
OD quartz tube with an expanded section (ꢁ12.7 mm OD,
20 mm length). The reactor was packed with quartz wool below
the catalyst bed to hold the powder in place. The reactor was
placed inside a resistively heated ceramic furnace with external
temperature control, and the catalyst bed temperature was mea-
ꢁ
Fig. 1. Time on stream study with SX–Rh/SiO
Total = 2 atm,
silica pretreated at 373 K, catalyst mass = 0.3 g, total gas flow rate = 120 cm min
2
SILP catalysts. T = 393 K,
], = 10%, SX/Rh = 10,
P
P
H2 ¼ PCO ¼ PC3H6 ¼ 1 : 1 : 1, [bmim][OctSO
4
a
3
ꢀ1
at STP.

1
68
D.G. Hanna et al. / Journal of Catalysis 292 (2012) 166–172
4
0 h. The only products observed under any of the conditions
Table 1
Activity of different sized particles of SX–Rh/SiO
2
SILP catalyst.
investigated were n- and iso-butanal.
Rate (h 1)a
ꢀ
The activity and the molar n/i ratio reported in literature for an
Particle size (
lm)
ꢀ1
analogous SX–Rh SILP catalyst are 381 h and 14 when measured
at 10 atm and 393 K, respectively [14]. In order to compare with
the results of the present study, the measured activity was ad-
justed from 2 atm to 10 atm using the partial pressure dependen-
cies on each of the reactants measured at 393 K (see the
Supporting information for details). Under the same reaction con-
363 K
413 K
<
150
50
32
35
151
155
>
a
ꢀ1
ꢀ1
mol n-butanal (mol Rh)
(h) .
ꢀ1
ditions, the catalyst discussed here is 572 h , which is 50% higher
than that of the catalyst described in Ref. [14]. A molar n/i ratio of
An Arrhenius plot for propene hydroformylation in the temper-
ature interval 353–413 K is shown in Fig. 3. The apparent activa-
1
1 was observed at 2 atm; however, as discussed below, the molar
ꢀ
1
tion energy for n-butanal formation is 60 kJ mol measured over
n/i ratio is weakly dependent on pressure.
ꢀ
1
the interval 353–373 K and 24 kJ mol measured over the interval
83–413 K. The value measured at low temperature for n-butanal
Fig. 2 shows the effects of reaction temperature on the catalytic
activity and the molar n/i ratio. The rates of n- and iso-butanal for-
mation increase as the reaction temperature increases. At 403 K,
the rate of n-butanal formation begins to plateau while the rate
of iso-butanal formation continues to rise. Thus, the n/i ratio de-
creases from 24 at 343 K to 6 at 413 K, demonstrating that temper-
ature can be used as a parameter to control the n/i isomer ratio.
Above 413 K, a loss in catalytic activity was observed due to ligand
decomposition, as identified by differential thermal gravimetric
analysis (see Supporting information).
3
is in agreement with previous reports for SILP catalysts [14]. The
apparent activation energy for iso-butanal formation is greater
than that for n-butanal and decreases from 72 kJ mol measured
over the interval 353–373 K to 55 kJ mol
interval 383–413 K. To determine whether the decrease in activa-
tion energy was due to internal mass transfer limitations, the cat-
ꢀ1
ꢀ1
measured over the
alyst (>150
lm) was crushed and sieved into small particles
(
<50 m). Table 1 shows that the activities at low and high temper-
l
ature are independent of particle size, leading to the conclusion
that the catalytic activity remains kinetically controlled at high
temperature. As discussed below, the observed decrease in the
apparent activation at high temperature can be ascribed to a
change in the rate-determining step (RDS) in the mechanism of
propene hydroformylation.
3.2. Kinetics of propene hydroformylation
The dependences of the rate of n- and iso-butanal formation on
the partial pressures of propene, CO, and H were measured by
2
holding the partial pressures of two components at 0.67 atm while
the partial pressures of the third reactant and He were varied. The
total gas flow rate and pressure were kept constant at
3
ꢀ1
3
ꢀ1
6
0 cm min and 2 atm (120 cm min at STP).
Fig. 4a–f illustrates the effect of varying the partial pressures of
3 6 2
C H , CO, and H partial pressures on the catalytic activity at 363 K
and 413 K. The molar n/i ratio was also found to be a weak function
of the reactant partial pressures, changing by less than ꢁ10% over
the pressure range investigated (see Supporting information). The
Fig. 2. Effect of temperature on rate of butanal synthesis and regioselectivity. (d,
s) Rate, (N) n/i ratio,
P
Total = 2 atm,
P
H2 ¼ PCO ¼ PC3H6 ¼ 1 : 1 : 1, catalyst
3
ꢀ1
mass = 0.3 g, total gas flow rate = 120 cm min at STP. [bmim][OctSO
SX/Rh = 10, silica pretreated at 353 K.
4
],
a
= 10%,
order of reaction with respect to H
and iso-butanal formation are similar and follow the same trend.
As shown in Fig. 4a–b, the partial pressure dependence of H on
the rate of n- and iso-butanal formation is nearly zero order at
63 K but is nearly first order at 413 K. At an intermediate temper-
ature of 393 K, the partial pressure dependence of H is 0.3 (see
2 3 6
, CO, and C H for n-butanal
2
3
2
Supporting information). The partial pressure dependence on CO
shown in Fig. 4c–d is roughly ꢀ0.5 at 363 K but increases to nearly
3 6
zero at 413 K. As shown in Fig. 4e–f, the order with respect to C H
decreases slightly from 0.9 to 0.7 when measured at 363 K and
4
13 K, respectively.
4
. Discussion
4.1. Rate expressions for butanal synthesis
The kinetics of butanal formation can be interpreted using the
mechanism proposed by Heck and Breslow [25] adapted by Wil-
kinson [22]. As shown in Scheme 2, the hydroformylation of pro-
pene is initiated by the loss of a CO ligand from HRh(CO)
2
SX
Fig. 3. Arrhenius plot showing the apparent activation energies of (d) n- and (s)
(
species A) to form the 16-electron species B via Reaction 1. Pro-
iso-butanal formation.
P
Total = 2 atm,
P
H2 ¼ PCO ¼ PC3H6 ¼ 1 : 1 : 1, catalyst
2
3
ꢀ1
pene then coordinates to species B to form a g –Rh complex (spe-
mass = 0.3 g, total gas flow rate = 120 cm min at STP. [bmim][OctSO
silica pretreated at 353 K. Rate is plotted on logarithmic scale.
4
], a = 10%,
cies C). Migratory insertion into the Rh–H bond forms a propyl-Rh

D.G. Hanna et al. / Journal of Catalysis 292 (2012) 166–172
169
Fig. 4. Log–log plot of the effect of reactant partial pressures on the rates of (d, N) n- and (s,
D
) iso-butanal synthesis. (d, s) T = 413 K, (N,
D
) T = 363 K. PTotal = 2 atm, (a and
b) PCO ¼ PC3H6 ¼ 0:67 atm, balance He, (c and d) PC3H6 ¼ PH2 ¼ 0:67 atm, balance He, (e and f) PCO ¼ PH2 ¼ 0:67 atm, balance He. Catalyst mass = 0.3 g, total gas flow
3
ꢀ1
4
rate = 120 cm min at STP. [bmim][OctSO ], a = 10%, silica pretreated at 353 K.
complex (species D) via Reaction 3. CO then adds to species D to
form an 18-electron intermediate species E in Reaction 4. Migra-
tory insertion of the propyl to a carbonyl group leads to the forma-
tion of an acyl intermediate (species F) via Reaction 5. Oxidative
coworkers using homogeneous Rh–xantphos complexes [26].
1
2
12 13
While isotopic experiments ( H/ H and
C/ C) were unable
to determine the RDS irrefutably, their hybrid quantum
mechanical/molecular mechanical (QM/MM) calculations sug-
gested that alkene insertion is the RDS. In agreement with
van Leeuwen’s conclusion that olefin chemisorption is a bar-
rier-less process, other theoretical studies were unable to locate
2
addition of H (Reaction 6) forms the 18-electron Rh(III) species
G, followed by the reductive elimination of butanal (Reaction 7).
Isobutanal is formed in a catalytic cycle by an analogous mecha-
nism, in which species D–G involve a secondary, rather than a pri-
mary, alkyl species.
The observed reaction orders at 363 K suggest that the RDS
is either alkene coordination (Reaction 2) or alkene insertion
3 2
a transition state for olefin coordination to HRh(PH ) CO com-
plexes [27,28]. For these reasons, alkene insertion into the
Rh–H bond is taken as the RDS. If Reaction 1 and 2 are as-
sumed to be quasi-equilibrated and Reaction 3 is the irrevers-
ible, rate-limiting step, then the rate of butanal formation is
given by the following relation:
(
Reaction 3) into the Rh–H bond. Support for Reaction 3 being
the RDS can be drawn from the work of van Leeuwen and

1
70
D.G. Hanna et al. / Journal of Catalysis 292 (2012) 166–172
Scheme 2. Proposed mechanism for the hydroformylation of propene.
K
1
K
2
k
3
P
þ K
C3H6 ½Rhꢂ
K
1
K
2
K
3
K
4
K
5
k
þ K
6
P
K
C3H6
P
H2 ½Rhꢂ
r
butanal ¼ P
ð1Þ
r
butanal
¼
ð2Þ
CO þ K
1
1
K
2
P
C
3H6
1 þ ðK
1
K
2
K
3
K
4
1
2
K
3
K
4
K
5
ÞPC3H6
Terms in the denominator inversely proportional to the partial pres-
sure of CO must be negligible in order to be consistent with the
experimental results. The less than first order dependence observed
for propene suggests that the terms in the denominator propor-
tional to the partial pressure of propene are significant in magni-
tude relative to unity. Further evidence for a change in the RDS is
provided by the observation of a partial pressure dependence on
where k
i
is the rate coefficient for Reaction i, K
i
is the equilibrium
j
constant for Reaction i, P is the partial pressure of species j,
and [Rh] is the total moles of Rh. Implicit in this rate equation
is the assumption that the concentrations of species D–G are
negligible.
Eq. (1) is consistent with the observed kinetics at 363 K. Similar
kinetics have been reported for homogenous hydroformylation Rh
complexes [29–31]. The rate of butanal synthesis is zero order in
H
2
of 0.3 at 393 K. At this intermediate temperature, Eqs. (1) and
(2) contribute to the overall rate, resulting in an increase in the or-
der with respect to H found at 363 K.
the partial pressure of H
pressure of C . As the partial pressure of C
C3H6 term in the denominator becomes increasingly signifi-
2
and less than first order in the partial
2
H
3 6
H
3 6
increases, the
MATLAB was used to estimate the parameters appearing in Eqs.
(1) and (2). Values for the parameters or groups of parameters are
listed in Table 2 for n- and iso-butanal formation. As seen in
Fig. 5a–d, the derived rate expressions (Eqs. (1) and (2)) adequately
describe the data acquired at 363 K and 413 K. The coefficient of
1 2
K K P
cant, thereby reducing the effective order from unity. The experi-
mentally observed, inhibitory effect of CO is also represented
effectively by Eq. (1). Since the order of CO is ꢀ0.4, the remaining
terms in the denominator are significant in magnitude compared
2
determination, R , was above 90% for all cases considered. Values
2
to PCO
.
of R and the root-mean-square error (RMSE) are presented in
At 413 K, the partial pressure dependence on H
2
and CO be-
the Supporting information.
comes nearly first order and zero order, respectively, suggesting
a change in the RDS from olefin insertion (Reaction 3) to oxidative
4.2. Regioselectivity
2
addition of H (Reaction 6). If Reaction 6 is considered to be the
irreversible, rate-limiting step and Reactions 1–5 are considered
quasi-equilibrated, then the rate of butanal formation will be given
by the following relation:
The decrease in n/i molar ratio with increasing temperature
shown in Fig. 2 can be rationalized in terms of the RDS in each tem-
perature regime and the corresponding rate expressions for n- and

D.G. Hanna et al. / Journal of Catalysis 292 (2012) 166–172
171
Table 2
Nonlinear regression results for Eqs. (1) and (2).
3
63 K
413 K
K
COþK
1
K
2
k
3
P
C
H
½Rhꢂ
a
P
C
H
P
H
½Rhꢂ
2
3
6
3
6
r
butanal ¼ P
r
butanal
¼
1þbð1þK
1
þK
1
K
2
P
C
3
H
6
5
ÞP
C H
3 6
Parameter
Value
Unit
atm
Parameter
Value
Unit
n-Butanal
iso-Butanal
n-Butanal
iso-Butanal
ꢀ
2
ꢀ1
K
K
1
2
0.16
2.7
134
0.05
3.9
10.7
a = K
b = K
1
K
K
2
K
K
3
K
K
4
K
k
5 6
589
85.6
0.22
1.8
atm
atm
–
h
ꢀ
1
ꢀ1
atm
1
2
3
4
0.89
0.20
3287
ꢀ
1
k
3
h
K
5
ꢀ
1
ꢀ1 ꢀ1
k
6
5
= a(bK )
210
atm
h
Fig. 5. Rate of (a and b) n- and (c and d) iso-butanal synthesis vs. partial pressure of (d) CO, (
D
) C
3
H
6
, (j) H
2
. Experimental points are indicated by data markers. Spline curves
], = 10%, silica pretreated at 353 K.
indicate predicted values. PTotal = 2 atm, catalyst mass = 0.3 g, total gas flow rate = 120 cm3 min at STP. [bmim][OctSO
ꢀ1
a
4
iso-butanal formation. The regioselectivity is determined by the al-
kene insertion into the Rh–H bond to form either a primary or sec-
ondary alkyls. At low temperature, alkene insertion is the RDS, and
the apparent activation energies measured for n- and iso-butanal
consistent with previous deutroformylation studies that found
that Rh-alkyl formation is irreversible at low temperature but is
reversible at higher temperatures [34,35].
ꢀ1
formation are 60 and 72 kJ mol , respectively. Therefore, the
resulting n/i ratio is high because primary alkyls are favored kinet-
ically over secondary alkyls. Theoretical studies by Bo and cowork-
ers have demonstrated that the transition for the formation of
4.3. Entropic effects on the RDS
The shift in the RDS with increasing temperature can be as-
cribed to a change in the elementary step with the largest Gibbs
ꢀ1
à
à
à
à
secondary alkyls is approximately 9 kJ mol greater than primary
alkyls due to the steric effects from the phenyl groups of the phos-
phine ligand [32]. To a first approximation, the ratio of rate con-
stants for alkene insertion to form either a primary or secondary
free energy of activation (
D
G =
D
D
H ꢀT
DS ). At 363 K, the DG for
à
alkene insertion is greater than
G for oxidative addition due to
a larger enthalpic barrier. Theoretical analysis of alkene hydrofor-
mylation on HRh(PH (CO) has shown that the activation enthal-
py barrier for alkene insertion is 40 kJ mol
oxidative addition of H [27]. However, the contribution of the
becomes more significant with increasing
temperature. The intrinsic entropy of activation is expected to be
more negative for oxidative addition of H (Reaction 6) than for al-
kene insertion (Reaction 3) due to the loss in entropy resulting
3 2
)
ꢀ1
alkyl (k
3
) can effectively model the temperature effect on the n/i
larger than for
ratio. However, the model begins to fail above 383 K due to the
change in the RDS. The deviations arise due the influence of equi-
librium constants corresponding to Reactions 3–5. The decrease in
the n/i ratio at high temperature is due to a shift in the equilibrium
from the thermodynamically favorable primary alkyl complex [33]
to the less favorable secondary alkyl complex. Our results are
2
à
entropic term to
DG
2
à
2
from the binding of gas-phase H . The value of DS for oxidative

1
72
D.G. Hanna et al. / Journal of Catalysis 292 (2012) 166–172
ꢀ
1
ꢀ1
addition was estimated to be approximately ꢀ100 J mol
K
References
based upon the removal of translational and rotational degrees of
à
[1] C.W. Kolpainter, R.W. Fischer, B. Cornils, Appl. Catal. A 221 (2001) 219.
[2] D.J. Cole-Hamilton, Science 299 (2003) 1702.
freedom of H
2
[36]. This value of
D
S is sufficient to cause the mag-
for the oxidative addition step to become greater
than that of alkene insertion at 393 K given the difference in the
à
nitude of
DG
[
3] J. Herwig, R. Fischer, in: P.W.N.M. van Leeuwen, C. Claver (Eds.), Rhodium
Catalyzed Hydroformylation, Kluwer Academic Publishers, AA Dordrecht,
à
2000, p. 189.
enthalpic barriers and assuming that
ligible. Hence, the change in the RDS from alkene insertion to the
oxidative addition of H can be ascribed to a change in the entropic
contribution to the intrinsic rate parameter.
DS of alkene insertion is neg-
[
[
4] R.J. Davis, J.A. Rossin, M.E. Davis, J. Catal. 98 (1986) 477.
5] A.J. Sandee, L.A. van der Veen, J.N.H. Reek, P.C.J. Kamer, M. Lutz, A.L. Spek,
P.W.N.M. van Leeuwen, Angew. Chem. Int. Ed. 38 (1999) 3231.
2
[6] L. Yan, Y.J. Ding, L.W. Lin, H.J. Zhu, H.M. Yin, X.M. Li, Y. Lu, J. Mol. Catal. A: Chem.
00 (2009) 116.
3
[
[
[
7] T. Kim, F.E. Celik, D.G. Hanna, S. Shylesh, S. Werner, A.T. Bell, Top. Catal. 54
(2011) 299.
8] L.A. Gerritsen, A. van Meerkerk, M.H. Vreugdenhil, J.J.F. Scholten, J. Mol. Catal. 9
5
. Conclusion
(
1980) 139.
The kinetics of propene hydroformylation over a silica-sup-
9] J.P. Arhancet, M.E. Davis, B.E. Hanson, J. Catal. 129 (1990) 94.
ported Rh–sulfoxantphos complex has been investigated. Reaction
temperature was found to have profound effects on the regioselec-
tivity and the nature of the rate-determining step (RDS). Evidence
for this change was observed by performing kinetic experiments at
[10] C. Disser, C. Muennich, G. Luft, Appl. Catal. A 296 (2005) 201.
11] J.P. Arhancet, M.E. Davis, J.S. Merola, B.E. Hanson, Nature 339 (1989) 454.
12] C.P. Mehnert, R.A. Cook, N.C. Dispenziere, M.J. Afeworki, J. Am. Chem. Soc. 124
[
[
(
2002) 12932.
[13] A. Riisager, K.M. Eriksen, P. Wasserscheid, R. Fehrmann, Catal. Lett. 90 (2003)
149.
14] A. Riisager, R. Fehrmann, M. Haumann, B.S.K. Gorle, P. Wasserscheid, Ind. Eng.
Chem. Res. 44 (2005) 9853.
15] A. Riisager, R. Fehrmann, S. Flicker, R. van Hal, M. Haumann, P. Wasserscheid,
Angew. Chem. Int. Ed. 44 (2005) 815.
16] S. Shylesh, D.G. Hanna, S. Werner. A.T. Bell, ACS Catal. (Communicated).
17] R.M. Deshpande, R.V. Chaudhari, Ind. Eng. Chem. Res. 27 (1988) 1996.
18] D.Y. Murzin, A. Bernas, T.J. Salmi, J. Mol. Catal. A: Chem. 315 (2010) 148.
[19] P.W.N.M. van Leeuwen, C.P. Casey, G.T. Whiteker, Phosphines as ligands, in:
W.N.M. van Leeuwen, P. Claver, C. Rhodium (Eds.), Catalyzed
Hydroformylation, Kluwer Academic Publishers, AA Dordrecht, 2000.
20] L.A. van der Veen, P.H. Keeven, G.C. Shoemaker, J.N.H. Reek, P.C.J. Kamer,
P.W.N.M. van Leeuwan, M. Lutz, A.L. Spek, Organometallics 19 (2000) 872.
21] P.C.J. Kamer, A. van Rooy, G.C. Schoemaker, P.W.N.M. van Leeuwen, Coord.
Chem. Rev. 248 (2004) 2409.
22] C.K. Brown, G. Wilkinson, J. Chem. Soc. A (1970) 2753.
[23] W.P. Mul, K. Ramkisoensing, P.C.J. Kamer, J.N. H Reek, A.J. van der Linden, A.
Marson, P.W.N.M. van Leeuwen, Adv. Synth. Catal. 344 (2002) 293.
3
63 K and 413 K. At low temperature, the RDS is alkene insertion
into the Rh–H bond, but at high temperature, the oxidative addi-
tion of H becomes rate-limiting. Although the partial pressure
dependence of C was roughly unaltered by temperature, the
partial pressure dependences of H and CO were found be to sub-
stantially dependent on temperature. The partial pressure depen-
dence of H was nearly zero order at low temperature but is
[
2
[
3
H
6
[
[
[
2
2
nearly first order at high temperature. The partial dependence of
CO was ꢀ0.4 at low temperature but was zero order at high tem-
perature. Based upon the mechanism for hydroformylation, rate
expressions were derived in the high and low temperature regime.
The parameters appearing in these rate expressions were esti-
mated by nonlinear regression. Variation in the RDS can be ratio-
nalized in terms of the elementary step with largest Gibbs free
energy of activation. The enthalpic barrier of activation is greater
for alkene insertion than oxidative addition, which causes alkene
insertion to be the RDS at low temperature. However, the greater
loss in entropy for the oxidative addition reaction causes the oxida-
[
[
[
[
[
24] P. Wasserscheid, R. van Hal, A. Bosmann, Green Chem. 4 (2002) 400.
25] R.F. Heck, D.S. Breslow, J. Am. Chem. Soc. 84 (1962) 2499.
[26] E. Zuidema, L. Escorihuela, T. Eichelsheim, J.J. Carbo, C. Bo, P.C.J. Kamer,
P.W.N.M. van Leeuwen, Chem. Eur. J. 14 (2008) 1843.
27] S.A. Decker, T.R. Cundari, Organometallics 20 (2001) 2827.
28] T. Matsubara, N. Koga, Y. Ding, D.G. Musaev, K. Morokuma, Organometallics 16
[
[
tive addition of H
2
to become the RDS at high temperature.
(
1997) 1065.
[
[
29] G. Kiss, E.J. Mozeleski, K.C. Nadler, E. VanDriessche, C. DeRoover, J. Mol. Catal.
A: Chem. 138 (1999) 155.
30] R.M. Deshpande, B.M. Bhanage, S. Kanagasabapathy, R.V. Chaudhari, Ind. Eng.
Chem. Res. 37 (1998) 2391.
Acknowledgments
2
This work was supported by the XC program funded by BP. The
authors would like to thank the Wasserscheid Group in Erlangen,
Germany for providing the sulfoxantphos ligand and Eric Bloch
for his assistance in performing dTGA.
[31] S.S. Divekar, R.M. Desphande, R.V. Chaudhari, Catal. Lett. 21 (1993) 191.
[
[
[
32] J.J. Carbo, F. Maseras, C. Bo, P.W.N.M. van Leeuwen, J. Am. Chem. Soc. 123
2011) 7630.
(
33] R.H. Crabtree, The Organometallic Chemistry of the Transition Metals, fourth
ed., John Wiley & Sons, Inc., Hoboken, 2005. p. 194.
34] R. Lazzaroni, R. Settambolo, A. Raffaelli, S. Pucci, G. Vitulli, J. Organomet. Chem.
3
39 (1988) 357.
[35] C.P. Casey, L.M. Petrovich, J. Am. Chem. Soc. 117 (1995) 6007.
36] D.A. McQuarrie, Statistical Mechanics, University Science Books, Sausalito,
000. p. 108.
Appendix A. Supplementary material
[
2
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.05.011.