Rhodium-catalyzed hydroformylation of C6 alkenes and alkene mixtures - A comparative study in homogeneous and aqueous-biphasic media using PPh 3, TPPTS and TPPMS ligands

Article abstract of DOI:10.1016/j.jorganchem.2004.06.051

The complexes RhH(CO)L₃, where L = PPh₃, P(m-C ₆H₄SO₃Na)₃ (TPPTS), and (C ₆H₅)₂P(m-C₆H₄SO ₃Na) (TPPMS) were used as catalyst precursors for a comparative study of the catalytic hydroformylation of several C6 alkenes and alkene mixtures under moderate reaction conditions in homogeneous (PPh₃) and aqueous-biphasic (TPPTS, TPPMS) media. The biphasic systems are efficient for the hydroformylation of hex-1-ene, 2,3-dimethyl-1-butene, styrene, cyclohexene, and mixtures thereof, in water/n-heptane at 80 °C. The main problem associated with these catalysts is their tendency to promote alkene isomerization if the effective syngas concentration in the liquid phases is low, but this side-reaction can be suppressed by using higher CO/H₂ pressures (54 atm). The selectivity of both water-soluble catalysts for linear products of hex-1-ene and for branched products of styrene is modest in comparison with the homogeneous system, which may limit their utility for classical oxo uses, but this is not a disadvantage for other interesting applications related to the hydroformylation of alkene mixtures and particularly to naphtha upgrading where linear and branched products are equally useful. The catalysts can be recycled without significant loss of activity and are resistant to the presence of benzothiophene in the mixture.

Full text of DOI:10.1016/j.jorganchem.2004.06.051

Journal of Organometallic Chemistry 689 (2004) 3782–3792
www.elsevier.com/locate/jorganchem
Rhodium-catalyzed hydroformylation of C6 alkenes and
alkene mixtures – a comparative study in homogeneous and
aqueous-biphasic media using PPh3, TPPTS and TPPMS ligands
a,
P.J. Baricelli , E. Lujano , M. Modron˜o , A.C. Marrero , Y.M. Garcıa ,
a
a
a
b
*
´
b
A. Fuentes , R.A. Sanchez-Delgado
b,
*
´
a
´ ´
Centro de Investigaciones Quımicas, Facultad de Ingenierıa, Universidad de Carabobo, Valencia, Venezuela
b
´
Centro de Quımica, Instituto Venezolano de Investigaciones Cientıficas (IVIC), Caracas, Venezuela
´
Received 17 May 2004; accepted 28 June 2004
Available online 5 August 2004
Abstract
The complexes RhH(CO)L3, where L = PPh3, P(m-C6H4SO3Na)3 (TPPTS), and (C6H5)2P(m-C6H4SO3Na) (TPPMS) were used
as catalyst precursors for a comparative study of the catalytic hydroformylation of several C6 alkenes and alkene mixtures under
moderate reaction conditions in homogeneous (PPh3) and aqueous-biphasic (TPPTS, TPPMS) media. The biphasic systems are effi-
cient for the hydroformylation of hex-1-ene, 2,3-dimethyl-1-butene, styrene, cyclohexene, and mixtures thereof, in water/n-heptane
at 80 ꢀC. The main problem associated with these catalysts is their tendency to promote alkene isomerization if the effective syngas
concentration in the liquid phases is low, but this side-reaction can be suppressed by using higher CO/H2 pressures (54 atm). The
selectivity of both water-soluble catalysts for linear products of hex-1-ene and for branched products of styrene is modest in com-
parison with the homogeneous system, which may limit their utility for classical oxo uses, but this is not a disadvantage for other
interesting applications related to the hydroformylation of alkene mixtures and particularly to naphtha upgrading where linear and
branched products are equally useful. The catalysts can be recycled without significant loss of activity and are resistant to the pres-
ence of benzothiophene in the mixture.
ꢁ 2004 Elsevier B.V. All rights reserved.
Keywords: Hydroformylation; Rhodium; Phosphines; Aqueous biphasic catalysis
1. Introduction
industrial applications of transition metal complex-cata-
lyzed reactions, with a current world production of over
6 million Tons of aldehydes and alcohols per annum [1,2].
Although industrial hydroformylation is carried our
predominantly in homogeneous phase, new technologies
have emerged in order to overcome the difficulties in the
separation of the products from the catalysts and in the
recycling of expensive noble metals at the end of the reac-
tion. A particularly important alternative is the aqueous-
biphasic version of the process in which the catalyst is
immobilized in an aqueous phase that is immiscible with
the phase containing the organic substrates and products
[3,4]. This allows an easy and efficient recycle of the
The rhodium-catalyzed hydroformylation of alkenes
continues to attract considerable attention because of
its fundamental and applied interests. The chemistry
associated with this reaction is rich and complex and de-
spite the fact that it has been extensively studied, many
fundamental challenges in catalyst design remain open.
Hydroformylation is no doubt one of the most important
*
Corresponding authors. Tel.: +58-212-504-1741; fax: +58-212-
504-1350.
´
E-mail address: rsanchez@ivic.ve (R.A. Sanchez-Delgado).
0022-328X/$ - see front matter ꢁ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.06.051

P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
3783
catalyst and it has been used with great success in the
Ruhrchemie/Rhone–Poulenc process, which is applied
mainly to the hydroformylation of propylene to butyral-
dehyde and, to a lower extent, of C4 alkene mixtures to
the corresponding aldehydes [1–4]. The Rh-derived cata-
lyst employed in that process is solubilized in water
through the use of triarylphosphine ligands modified by
introducing highly polar sulfonate groups in the phenyl
rings attached to phosphorus, specifically, the sodium
salt of tris[(meta-sulfonato)phenyl]phosphine P(m-
C6H4SO3Na)3, also known as triphenylphosphine tri-
sulfonate (TPPTS) [3,4]. For longer-chain alkenes
(C5+), Union Carbide has proposed a process involving
Rh together with the sodium salt of [(meta-sulfo-
nato)phenyl]diphenylphosphine Ph2P(m-C6H4SO3Na),
(triphenylphosphine monosulfonate, TPPMS), or better
still with diphenylphosphinebutylsulphonate DPBS, in
conjunction with N-methylpyrrolidone as a solubilizer.
This produces a homogeneous mixture that can be easily
split into two phases by addition of water or methanol at
the end of the reaction [1,5]. Aqueous biphasic technol-
ogies are generally considered very ‘‘green’’, as they
avoid the use of large amounts of polluting volatile
organic solvents.
extremely favorable atom economy with virtually no
waste.
In view of the large volumes of refinery cuts that
need to be treated, a classical homogeneous system
requiring costly or complicated product separation
from the catalyst would not be practical and therefore
we have turned our attention to liquid biphasic systems
(aqueous and ionic liquids). We have recently reported
the use of Ru(II) catalysts containing TPPMS in the
aqueous-biphasic hydroformylation of alkene mixtures
as models of naphtha components [7]. In the present
paper we describe a comparative study of the Rh-cata-
lyzed hydroformylation of C6 alkenes and alkene mix-
tures in homogeneous and aqueous-biphasic media
using PPh3, TPPTS, and TPPMS ligands, as a further
demonstration of the considerable potential of these
‘‘green’’ systems for applications related to fuel
upgrading issues.
2. Experimental
2.1. General procedure
A different set of very interesting new applications of
hydroformylation chemistry involves the treatment of
mixtures of valuable medium-chain alkenes -mainly in
the C4–C7 range- present in refinery cuts and sometimes
referred to as ‘‘orphan olefins’’, for which there are few
commercial uses. Examples of such applications have
been reported e.g. for the conversion of raffinate-2
(but-1-ene + but-2-ene) [1,2] and also for the production
of high-performance plasticizers like di-iso-nonylphta-
late [6]. Along these lines, we are interested in the fact
that refinery naphtha (the basis for gasoline) contains
up to 50% vol. of C5–C7 alkenes of varied structures,
while specifications for the final product allow no more
than about 12%. Currently, the excess of olefins in naph-
tha is lowered by hydrogenation over standard solid cat-
alysts, which is costly and results in a marked decrease
in the octane rating of the fuel. This is subsequently cor-
rected by addition of methylterbutylether (MTBE) but
this gasoline additive is highly polluting because of its
water solubility and it is therefore being rapidly phased
out, while less harmful oxygenated components are ex-
pected to replace it in the near future. Hydroformylation
could thus be an interesting novel method for lowering
the olefin content of naphtha through the in situ conver-
sion of some of the alkenes present into valuable C5–C8
oxygenates (aldehydes or alcohols), capable of improv-
ing the combustion properties of the fuel while at the
same time being less water-soluble and less harmful than
MTBE. For this type of application, the regioselectivity
of the catalysts (n/i ratios) is unimportant, since both lin-
ear and branched aldehydes or alcohols would be
acceptable in the final product, thus resulting in an
All manipulations were carried out under nitrogen
using standard Schlenck techniques. Organic solvents
and olefins were dried and purified by distillation over
standard agents under N2 prior to use. RhCl3 Æ3H2O
was purchased from Pressure Chemicals and alkenes
were obtained from Aldrich and distilled from Na under
nitrogen or argon. All other chemicals were commercial
products and were used without further purification.
All gases were of high purity (>99%) and were pur-
chased from AGA Gases. RhH(CO)(PPh3)3 [8], TPPTS,
TPPMS [9], RhH(CO)(TPPTS)3, and RhH(CO)
(TPPMS)3 [10] were synthesized according to published
methods. Infrared spectra were recorded in Perkin–El-
mer 1000 or Nicolet Magna 560 FTIR spectrometers
using pressed samples as KBr disks. GC analyses were
performed on a Hewlett Packard 5890 Plus Series II
chromatograph with a flame ionization detector and ul-
tra 2.5% phenyl methyl silicone, 25m, 320 lm column.
Quantification was achieved by using naphthalene as
the internal standard and all peaks were identified by
GC/MS on a HP 5890/5972 coupled system using a
Quadrex PONA 5% phenyl methyl silicone, 25 m,
320 lm column.
2.2. Catalytic hydroformylation
All the catalytic reactions were performed under con-
stant pressure in glass-lined stainless steel autoclaves
(Parr) fitted with internal mechanical stirring, a high-
pressure reservoir, temperature control unit, and a sam-
pling valve. For each catalyst, a hydroformylation run
was carried out in the presence of excess mercury

3784
P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
(1 ml, 7 · 10À2 mol) for HX1E and for the four-compo-
nent mixture, in order to ascertain the possible involve-
ment of metallic particles or colloids in the catalysis [11].
DM1B
Chart 1.
DM2B
HX1E
2.2.1. Homogeneous reactions
Typically the catalyst (10 mg, 10À5 mol), the alkene
(0.5 ml, 4.0 · 10À3 mol), and the solvent (40 ml) were
introduced into the reactor. The system was purged
three times with CO/H2 (1:1, 5 atm), charged to the de-
sired pressure and heated to the chosen reaction temper-
ature with constant stirring at 600 rpm (the time
required to attain thermal equilibrium was ca. 1 h, dur-
ing which some hydroformylation (610% conv.) was
observed). Samples were periodically withdrawn for
analysis and at the end of the run the reactor was cooled
in ice and vented before opening it for the analysis of the
final reaction mixture.
3. Results and discussion
3.1. Homogeneous hydroformylation of C6 alkenes and
alkene mixtures by use of RhH(CO)(PPh3)3
In order to establish a basis for comparison for the
aqueous biphasic systems of particular interest here,
we have carried out a series of experiments in the hydro-
formylation of hex-1-ene (HX1E), 2,3-dimethylbut-1-
ene (DM1B), and 2,3-dimethylbut-2-ene (DM2B), which
are representative of C6 mono-, di- and tetra-substituted
alkenes present in naphtha (Chart 1). HX1E has been
employed numerous times in hydroformylation studies
but the more hindered terminal alkene DM1B and its
tetra-substituted internal isomer DM2B have been less
frequently used in this reaction [1,2].
Fig. 1(a) shows the behavior of RhH(CO)(PPh3)3 in
the hydroformylation of HX1E. The apparent induction
period observed during the first hour of reaction corre-
sponds to the time required to attain thermal equilibra-
tion and in part also to form the active species
RhH(CO)n(PPh3)m(alkene) in appreciable concentra-
tions. Under the moderate reaction conditions chosen
as our standard (60 ꢀC, 15 atm CO/H2) the reaction
proceeds at reasonable rates to produce the two corre-
sponding isomeric aldehydes, n-heptanal and 2-methyl-
hexanal in a 3:1 ratio, as typically found for this
catalyst in the absence of excess phosphine. No isomeri-
zation or hydrogenation products were detected under
these conditions.
2.2.2. Biphasic reactions
In a typical experiment, an aqueous solution of the
catalyst precursor (10À5 mol) in 20 ml of water and the
alkene or alkene-mixture (6 · 10À3 mol) in 20 ml of n-
heptane were introduced into the reactor. The solution
was purged three times with CO/H2 (1:1, 5 atm), then
it was charged to the required pressure, heated to the
desired temperature and stirring at 600 rpm was com-
menced, which defines the zero time of the reactions.
Samples of the reaction mixture were periodically
extracted and the total pressure was adjusted via a
high pressure reservoir. The samples were cooled, the
phases separated and the organic phase analyzed by
GC-MS. At the end of the run, the reactor was cooled
in ice and vented before opening it for analysis of the
final reaction mixture. In the anaerobic recycling
experiments, after the end of each run the reactor
was cooled to room-temperature, the olefin was in-
jected into the reactor by means of a high-pressure
burette, the pressure was adjusted to the required value
and stirring was re-started immediately to initiate a
new run.
On raising the reaction conditions to 100 ꢀC and 20
atm H2/CO the hydroformylation of HX1E is achieved
with a rate slightly over twice that observed at the lower
100
80
100
80
60
40
20
0
60
%
%
40
20
0
0
60 120 180 240 300 360
t (min)
0
120 240 360 480
t (min)
(a)
(b)
Fig. 1. Hydroformylation by RhH(CO)(PPh3)3 in toluene. 60 ꢀC, 15 atm CO/H2 (1:1), [olefin]:[Rh] = 400. (a) (r) Hex-1-ene; (j) 1-heptanal; (m) 2-
Me-hexanal. (b) (r) 2,3-Dimethyl-1-butene; (j) 3,4-dimethylpentanal.

P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
3785
conditions employed. When DM1B is hydroformylated
under the milder conditions (60 ꢀC, 15 atm CO/H2),
the reaction profile shown in Fig. 1(b) is observed. In this
case, the induction period is of about 2h, indicating that
even after reaching the thermal equilibrium, the reaction
is very slow for another hour, after which a linear behav-
ior up to ca. 60% conversion is observed, albeit at a lower
rate than the one observed for HX1E. This implies that
the formation of the active species in sufficient concentra-
tion is slower for the more hindered alkene. As expected,
the sole reaction product is 3,4-dimethylpentanal, the
‘‘linear’’ isomer of the hydroformylation; it is known that
the ‘‘branched’’ product is not formed in this case be-
cause of the strong steric congestion that would be
caused by two methyl groups in the vicinity of the rho-
dium center in the corresponding alkyl intermediate.
Similarly to what was observed for HX1E, no isomeriza-
tion or hydrogenation products were detected under the
conditions employed. The turnover frequencies (TOF) of
these two reactions were calculated from the slopes of the
linear portion of each plot in Fig. 1, and the correspond-
ing values are collected in Table 1.
conversion of HX1E dominated the reaction, while the
relative activity for the conversion of DM1B was low-
ered in comparison with the case of the pure olefins by
about 30%, and DM2B did not react. This indicates that
there is a competition for the active Rh sites in favor of
the coordination of the less hindered alkene HX1E, and
the result of that is that the hydroformylation of DM1B
becomes significant only when more than 50% of HX1E
has been consumed. For practical purposes concerning
possible naphtha upgrading applications, it was interest-
ing to evaluate this catalyst in the hydroformylation of
the three-component mixture, under the more stringent
conditions necessary to effect the conversion of the least
reactive, most hindered alkene DM2B (120 ꢀC, 68 atm
CO/H2). As can be seen in Fig. 2, no induction period
is present in this case, as a much faster heating rate
can be employed to stabilize our reactor at 120 ꢀC. Also,
it is noted that the three substrates react in a consecutive
manner: HX1E is completely consumed within the first
30 min in this case. The formation of 3,4-dimethylpent-
enal starts when most of the HX1E has been converted
and it is fairly rapid at the early stages of the reaction
but after 50% conversion a slower reaction rate is
observed. These two kinetic regimes correspond to the
When the tetra-substituted alkene DM2B was sub-
jected to hydroformylation under analogous conditions,
only 4% conversion was observed after 24h reaction and
the only product detected was 3,4-dimethylpentanal,
again the ‘‘linear’’ isomer of the hydroformylation of
DM1B. This indicates that the catalysts is capable of
slowly isomerizing DM2B to DM1B under the condi-
tions employed and hydroformylation occurs only when
the C‚C bond of the former has shifted to the terminal
position. In this case the TOF was estimated from the
percentage conversion after 24 h and the corresponding
value is included in Table 1. From these data it can be
deduced that the rate of hydroformylation of HX1E is
about three times faster than that of DM1B and about
250 times faster than the rate of isomerization–hydro-
formylation of DM2B. On raising the temperature to
120 ꢀC and the pressure to 68 atm, the isomerization–
hydroformylation of the very un-reactive tetra-substi-
tuted olefin can be achieved with conversions of about
50% in 5h, corresponding to a reasonable TOF of ca.
100
80
60
40
20
0
0
200
400
600
t (min)
Fig. 2. Hydroformylation of a 1:1:1 mixture of HX-1E, 2,3-DM1B,
2,3-DM2B by RhH(CO)(PPh3)3 in toluene. 120 ꢀC, 68 atm CO/H2
(1:1), [olefin]tot:[Rh] = 400. (a) (m) n-Heptanal + 2-Me-hexanal; (j)
3,4-dimethylpentanal.
0.8 minÀ1
.
When a 1:1:1 mixture of the three alkenes was hydro-
formylated under our standard conditions for 5 h, the
Table 1
Hydroformylation of alkenes by RhH(CO)(PPh3)3 (conditions: H2/CO (1:1), [alkene]:[Rh] = 400)
Alkene
T (ꢀC)
P (atm)
t (h)
% conv.
TOF (minÀ1 a
)
Rel. Act.
HX1E
HX1E
DM1B
DM2B
DM2B
a
100
60
20
15
15
15
68
0.75
2
56
50
50
4
5.0
2.3
–
1.0
0.3
0.004
–
60
60
6
0.75
0.01b
0.8b
24
5
120
50
Values calculated from the maximum slopes in the plots shown in Fig. 1.
b
Average values estimated from (% conversion/t).

3786
P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
direct hydroformylation of DM1B (faster), followed by
the slower isomerization–hydroformylation of DM2B.
A last point of interest concerning reactions of refin-
ery cuts is that they usually contain small amounts of
organo-sulfur compounds that may act as poisons to-
ward the catalyst. In order to model such a situation,
the hydroformylation of a mixture containing the three
alkenes plus thiophene or benzothiophene in a 1:1:1:1
molar ratio was carried out under analogous conditions
to the ones reported in Fig. 2 (120 ꢀC, 68 atm CO/H2
(1:1), [olefin]tot:[Rh] = 400). No inhibition of the hydro-
formylation reaction was observed, but instead a slight
increase in the overall conversion could be measured
(about 10% for thiophene and about 20% for benzothio-
phene). Sulfur tolerance has been previously recognized
for Rh-phosphine catalysts [2] but it is important to note
that in our experiments RhH(CO)(PPh3)3 was not only
resistant to poisoning by organosulfur compounds at
concentrations much higher than the ones normally ex-
pected to be present in a naphtha, but a promotion effect
was actually observed. This point will be the object of
future studies.
that isomerization competes with hydroformylation at
the early stages of the reaction but after about 30 min
the production of aldehydes prevails. Representative
data indicating the total aldehyde production (lin-
ear + branched) with this aqueous system as a function
of temperature and pressure are shown in Fig. 3(b).
Thus, by adequately adjusting the operation parameters,
the biphasic system can be tuned to efficiently promote
alkene hydroformylation.
These observations are in contrast with the behavior
of the homogeneous system RhH(CO)(PPh3)3, which
does not promote C‚C bond isomerization. However,
they are consistent with the low solubility of hex-1-ene
and of syngas in water. At low pressures, the olefin
can come into contact with the catalyst, either in the
water phase or possibly in the liquid–liquid interface,
but the concentration of syngas in that region is very
low. Consequently, a classical metal-hydride promoted
C‚C bond migration mechanism becomes dominant
and most of the terminal alkene is converted rapidly
to the internal isomer that is more resistant to hydro-
formylation. As the gas pressure is increased the effective
concentration of CO and H2 at the catalytic site is suffi-
cient for hydroformylation to be faster than isomeriza-
tion. From the data in Fig. 3(b), optimum
hydroformylation conditions are chosen as 80 ꢀC and
55 atm CO/H2, under which the isomerization side-reac-
tion is almost completely suppressed and ca. 100% con-
version to aldehydes is achieved in 2.5 h. The n/i ratios in
the experiments at 80 ꢀC or 100 ꢀC were systematically
of the order of 1.5, where 2-Me-hexanal originates from
both hex-1-ene and the small amounts of hex-2-ene
formed. Also, a small proportion (<10%) of 2-Et-pent-
anal, the other product of hex-2-ene hydroformylation,
was consistently detected. Addition of free TPPTS in
concentrations up to 8 moles per mol of complex did
not result in a significant variation of the selectivity.
This indicates that under the relatively high CO pres-
sures required to obtain reasonable rates and a good
chemoselectivity of hydroformylation in the biphasic
3.2. Aqueous biphasic hydroformylation of C6 alkenes and
alkene mixtures by use of RhH(CO)(TPPTS)3
Under reaction conditions analogous to the ones se-
lected for the homogeneous reactions, the complex
RhH(CO)(TPPTS)3 is not a good catalyst for the two-
phase hydroformylation of HX1E in water/n-heptane
mixtures. The reaction is very slow at temperatures of
60 ꢀC or below but reasonable rates are achieved at 80
ꢀC or above. However, at low pressures of syngas
(<30 atm) HX1E is rapidly converted mainly to hex-2-
ene (ca. 60–70%) while the production of aldehydes is
only observed as a minor reaction. Nevertheless, at pres-
sures between 50 and 68 atm isomerization is strongly
disfavored (<20% of the total conversion) and hydro-
formylation dominates. A typical reaction profile under
optimized conditions is shown in Fig. 3(a), indicating
60
45
30
15
0
100
90
80
70
60
50
40
30
20
10
0
80˚ C
100˚C
0
40
80
t (min)
120 160
0
10 20 30 40 50 60 70
P(CO/H2) (atm)
(a)
(b)
Fig. 3. Biphasic hydroformylation of hex-1-ene in H2O/n-heptane by use of RhH(CO)(TPPTS)3. (a) Reaction profile at 80 ꢀC, 55 atm CO/H2,
[olefin]:[Rh] = 600; (j) hex-2-ene; (m) 1-heptanal; (r) 2-Me-hexanal; (·) 2-Et-pentanal. (b) Effect of T and P on the total production of aldehydes
after 2.5 h.

P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
3787
system, the monophosphine species a, which is the least
regioselective intermediate, predominates in the equilib-
rium shown in Scheme 1 and addition of free-phosphine
does not shift this equilibrium to an important degree
toward the more sterically hindered bis(phosphine)
intermediate b that would be more selective for the pro-
duction of linear products.
and is known to produce very high n/i ratios in the pres-
ence of excess PPh3 via bis(phosphine)intermediates like
b. Attempts to carry out the analysis of such mixtures by
31P NMR spectroscopy failed, because the samples
decomposed during the time required for their manipu-
lation from the high-pressure reactor into the NMR
tube, even under a syngas stream.
In agreement with this, the IR spectra of samples
withdrawn under a syngas stream from the thermally
equilibrated catalytic solutions at the working pressure,
with or without excess TPPTS but in the absence of the
olefin, were dominated by features corresponding to a
In view of the good performance of the aqueous cat-
alyst for hex-1-ene hydroformylation, and of the results
obtained for other alkenes with the analogous homoge-
neous system, it was decided to further evaluate the cat-
alytic properties of RhH(CO)(TPPTS)3 for the
hydroformylation of DM1B, styrene (STY) and cyclo-
hexene (CHX), which better represent the components
of a refinery cut. DM2B was no longer considered in this
part of the study, because of its highly un-reactive nat-
ure and due to the fact that real refinery naphtha con-
tains only very small amounts of tetra-substituted
olefins [12]. Fig. 4 depicts typical reaction profiles for
these three alkenes.
Under the reaction conditions optimized for HX1E
(80 ꢀC, 55 atm CO/H2, [substrate]:[Rh] = 600) DM1B
is hydroformylated with reasonable rates, achieving
100% conversion after about 3 h with a short induction
period (<30 min) (Fig. 4(a)). Styrene is also hydroform-
ylated efficiently (100% conv. in less than 5 h) to the two
corresponding aldehydes (Fig. 4(b)) with a selectivity for
dicarbonyl species, possibly a (mCO 2000, 2020 cmÀ1
;
mRh–H 2052 cmÀ1). This behavior is again in contrast
with the homogeneous system, which does not require
very high pressures to operate with reasonable rates
H
-P
-CO
CO
CO
P
P
CO
P
P
H
P
H
Rh
Rh
Rh
OC
(b)
(a)
CO
linear product
branched product
Scheme 1.
100
80
60
40
20
0
100
80
60
40
20
0
0
70 140 210 280 350 420
t (min)
0
70 140 210 280 350 420
t (min)
(b)
100
(a)
100
90
80
70
60
50
40
30
20
10
0
80
60
40
20
0
0
70 140 210 280 350 420
t (min)
0
70 140 210 280 350 420
t (min)
(c)
(d)
Fig. 4. Biphasic hydroformylation by RhH(CO)(TPPTS)3 in H2O/n-heptane; 80 ꢀC, 55 atm CO/H2, [subst.]:[Rh] = 600. (a) (r) 2,3-Dimethyl-1-
butene; (j) 3,4-dimethylpentanal. (b) (r) Styrene; (j) 2-phenyl-propanal; (m) 3-phenyl-propanal. (c) (r) Cyclohexene; (j) cyclohexanecarbox-
aldehyde. (d) Total conversions for (r) hex-1-ene; (j) 2,3-dimethyl-1-butene; (m) styrene; (·) cyclohexene.

3788
P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
the branched isomer typical of this reaction at tempera-
tures around 100 ꢀC (n/i ca. 0.3 at <50% conversion and
about 0.5 at high conversions) [1]. Cyclohexene, as ex-
pected, was the least reactive of the alkenes studies,
reaching 100% conversion after about 7 h and displaying
an induction period of ca. 30 min, presumably reflecting
the slower formation of RhH(CO)n(TPPTS)n(CHX) in
appreciable concentrations (Fig. 4(c)). For comparison,
Fig. 4(d) shows the total conversions for each alkene as
a function of time. Plots of the linear section of each
curve (in turnover number units, TON) allowed us to
calculate the TOF values collected in Table 2.
From these data it is seen that the hydroformylation
of HX1E is about three times faster than that of
DM1B, four times faster than that of STY and 7.5 faster
than that of CHX. All the alkenes tested can thus be effi-
ciently converted to aldehydes by RhH(CO)(TPPTS)3 in
a biphasic mixture under moderate reaction conditions.
Only HX1E produced small amounts of isomerization
by-products, thus making this a promising system for
naphtha upgrading applications.
When a 1:1:1:1 mixture of the four olefins was hydro-
formylated, the results displayed in Fig. 5(a) were ob-
tained. It is clear from these data that in the mixture,
the reactions are consecutive and not parallel, and under
these conditions STY is converted more rapidly than
DM1B, in contrast to what is observed for the individ-
ual substrates. Cyclohexene only reacts appreciably after
about 7 h, when most of the other alkenes have been
consumed. These observations indicate that when sev-
eral olefins are present simultaneously, a competition
for the metal centers takes place leading to a reactivity
order in this case HX1E > STY > DM1B > CHX. The
inversion in the relative reactivity of STY and DM1B
observed in the mixture with respect to the individual
components probably results from a combination of
the differences in their water-solubility and in their
intrinsic strengths of coordination to Rh.
The reaction profile in terms of total olefin conversion
depicted in Fig. 5(b) also shows the trend of a rapid ini-
tial period (up to about 90 min) with no induction peri-
od, corresponding to the complete hydroformylation of
HX1E plus the partial conversion of the two other termi-
nal olefins, after which a slower regime is observed essen-
tially for the hydroformylation of STY and DM1B, while
the reaction of CHX contributes little to the overall
conversion. Addition of 50 ppm of benzothiophene did
Table 2
Aqueous biphasic hydroformylation of alkenes by RhH(CO)(TPPTS)3
(conditions: 80 ꢀC, 54 atm H2/CO (1:1), [alkene]:[Rh] = 600)
Alkene
t (h)
% conv.
TOF (minÀ1 a
)
Rel. Act.
HX1E
DM1B
STY
CHX
a
1.5
3
100
100
100
100
13.5
5.0
3.4
1.8
1.0
0.4
4.5
7
0.25
0.13
Values calculated from the maximum slopes of the plots in Fig. 4.
100
80
60
40
20
0
150
120
90
60
30
0
0
100 200 300 400 500
t (min)
0
70 140 210 280 350 420
t (min)
(a)
(b)
100
90
80
70
60
50
40
30
20
10
0
300
250
200
150
100
50
0
0
70 140 210 280 350 420
t (min)
0
70 140 210 280 350 420
t (min)
(c)
(d)
Fig. 5. Hydroformylation of mixtures of hex-1-ene, styrene, 2,3-dimethyl-1-butene and cyclohexene by RhH(CO)(TPPTS)3 in H2O/n-heptane at
80 ꢀC, 55 atm CO/H2, [olefin]tot:[Rh] = 600. For a 1:1:1:1 molar ratio: (a) TON values: (r) hex-1-ene; (m) styrene; (j) 2,3-dimethyl-1-butene; (·)
cyclohexene. (b) Total olefin % conversion. For a HX1E : STY: CHX : DM1B molar ratio of 1:1:1.5:7: (c) Total olefin % conversion. (d) TON values:
(r) hex-1-ene; (j) 2,3-dimethyl-1-butene; (m) styrene; (·) cyclohexene.

P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
3789
not have any effect on the catalyst behavior under the
conditions of Fig. 5(b).
Finally, another mixture of the four olefins was pre-
regime operates up to about 20% conversion, followed
by a slower reaction reaching a total conversion of ca.
70% after 7 h. TON values for each individual olefin
are also presented in Fig. 5(d), showing again a behavior
of the whole mixture corresponding to a set of consecu-
tive reactions that correlate well with the initial compo-
sition of the mixture.
The terminal olefins (HX1E + DM1B + STY) are all
hydroformylated with reasonably fast rates that are
comparable to each other. Their combined conversion
corresponds to the faster regime in Fig. 5(a). Once
pared
containing
the
proportions
HX1E:S-
TY:CHX:DM1B = 1:1:1.5:7, which are analogous to
those found in a real naphtha for mono-substituted, cyc-
lic, and di-substituted alkenes [12]. The results of the
hydroformylation of such a naphtha-model mixture at
80 ꢀC, 55 atm CO/H2, [total olefins]:[Rh] = 600 are de-
picted in Fig. 5(c). It is clearly noted that after an induc-
tion period of about 1h, a first rapid hydroformylation
100
80
60
40
20
0
60
45
30
15
0
%
0
15 30 45
t (min)
60 75 90
0
70 140 210 280 350 420
t (min)
(a)
100
(b)
100
80
60
40
20
0
80
60
40
20
0
%
%
0
70 140 210 280 350 420
t (min)
0
60
120 180 240 300
t (min)
(c)
(d)
100
80
60
40
20
0
0
70 140 210 280 350 420
t (min)
(e)
Fig. 6. Hydroformylation by RhH(CO)(TPPMS)3 in H2O/n-heptane; 80 ꢀC, 55 atm CO/H2, [subst.]:[Rh] = 600. (a) (j) Hex-2-ene; (m) 1-heptanal;
(r) 2-Me-hexanal; (·) 2-Et-pentanal. (b) (r) 2,3-Dimethyl-1-butene; (j) 3,4-dimethylpentanal. (c) (r) Styrene; (j) 2-phenyl-propanal; (m) 3-
phenyl-propanal. (d) (r) Cyclohexene; (j) cyclohexanecarboxaldehyde. (e) Total conversions for (r) hex-1-ene; (j) 2,3-dimethyl-1-butene; (m)
styrene; (·) cyclohexene.

3790
P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
HX1E and STY (amounting to about 20% of the mix-
ture) have been totally consumed, hydroformylation of
DM1B takes place almost exclusively, which represents
an overall lower rate accounting for the change in slope
observed in Fig. 5(a). CHX only begins to react after 3.5
h, when an important proportion of DM1B has been
converted. At the end of the 7 h run, the final composi-
tion shows that HX1E and STY have been completely
transformed into aldehydes, DM1B has been converted
in ca. 75% and CHX in about 30%. The general trend in
reactivity observed for the equimolar mixture of olefins
is maintained in this case, but concentration effects obvi-
ously also come into play. The importance of the latter
experiment is to demonstrate the practical potential of
the biphasic system, which can reduce the overall olefin
content in a naphtha-model mixture by 70% in 7 h, con-
verting the unsaturated compounds almost exclusively
into oxygenates.
ble with homogeneous systems except that low pressures
cannot be employed in biphasic conditions due to gas
solubility limitations.
For an equimolar mixture of the four alkenes, the
behavior of the TPPMS catalyst is also similar to that
of the TPPTS analogue (Fig. 7). HX1E reacts much fas-
ter than the other alkenes, STY is more rapidly con-
verted than DM1B and CHX only starts to react at
long reaction times, when most of the other substrates
have been consumed. Addition of 50 ppm of benzothio-
phene had no influence on the hydroformylation of the
four-component alkene mixture under the conditions
of Fig. 7.
3.4. Re-using the aqueous biphasic hydroformylation
catalysts
One of the main reasons for using aqueous biphasic
catalysts is the possibility of easily separating them from
the products at the end of the reaction and re-using
them. In order to test the recycling capacity of our cat-
alysts, a first series of experiments were carried out with
HX1E as the substrate and RhH(CO)(TTPMS)3 as the
catalyst precursor: the reactor was opened under a
stream of nitrogen at the end of a catalytic run (80 ꢀC,
55 atm CO/H2, [HX1E]:[Rh] = 600, 2 h), the organic
phase was syringed out, a new charge of fresh substrate
in n-heptane was added, and the reactor was re-sealed,
pressurized and heated for a new hydroformylation
run. In four consecutive cycles performed in this way
with the same catalyst solution, the conversion values
were 100%, 100%, 100% and 80%. The lowering in the
catalyst activity observed after the fourth cycle can be
attributed to decomposition of the catalyst solution by
contact with air. Indeed, more satisfactory results were
obtained by use of an alternative procedure: a similar
reaction was performed to 100% HX1E conversion;
the reactor was then cooled to room-temperature but
not opened, a new charge of substrate was introduced
through a high-pressure burette, and a new catalytic
run was performed; in this case the conversion of
HX1E was consistently greater than 97% in four consec-
utive cycles, for both the TPPMS and the TPPTS
catalysts.
3.3. Aqueous biphasic hydroformylation of C6 alkenes and
alkene mixtures by use of RhH(CO)(TPPMS)3
Because of its extreme water-solubility the TPPTS-
derived catalyst is only used industrially for the hydro-
formylation of short-chain alkenes that are moderately
soluble in water; the reaction in those cases is thought
to occur predominantly in the aqueous phase [1–4].
Longer-chain alkenes like the ones used in the present
study are very sparingly soluble in water and therefore
their reactions take place preferentially in the organic
phase or at the liquid–liquid interface. The TPPMS
complex RhH(CO)(TPPMS)3, being less soluble in
water than the trisulfonated analogue, is probably a bet-
ter suited catalyst for the hydroformylation of C6 alke-
nes, since its concentration at the liquid–liquid interface
is higher and it actually has surfactant properties [13].
As shown in Fig. 6, this catalyst is also efficient for
the hydroformylation of HX1E, DM1B, STY and
CHX in water/n-heptane mixtures under reaction condi-
tions similar to the ones used for the TPPTS derivative
(80 ꢀC, 55 atm CO/H2, [substrate]:[Rh] = 600). From
the linear sections of the plots in Fig. 6(e), the TOF val-
ues collected in Table 3 could be calculated. It is ob-
served that these values are similar to those obtained
for the TPPTS catalyst, and both of them are compara-
3.5. Stability of the aqueous biphasic hydroformylation
catalysts
Table 3
Aqueous
biphasic
hydroformylation
of
alkenes
by
It is known that both the active species and the excess
ligand of the Rh-TPPTS system in industrial use are
subject to slow decomposition, requiring repeated addi-
tion of fresh ligand to maintain the [Rh]:[P] ratio and of
sulfuric acid to avoid large variations of the pH of the
solution. If these precautions are not taken, faster hyd-
roformylation rates may be observed after some time
due to changes in the composition of the catalyst or to
RhH(CO)(TPPMS)3 (conditions: 80 ꢀC, 54 atm H2/CO (1:1),
[alkene]:[Rh] = 600)
Alkene
t (h)
% conv.
TOF (minÀ1 a
)
Rel. Act.
HX1E
DM1B
STY
CHX
a
1
3
5
7
100
100
100
100
14.1
4.4
2.9
2.0
1.0
0.3
0.2
0.14
Values calculated from the maximum slopes of the plots in Fig. 6.

P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
3791
150
100
80
60
40
20
0
120
90
60
30
0
0
100 200 300 400 500
t (min)
0
100 200 300 400 500
t (min)
(a)
(b)
Fig. 7. Hydroformylation of a HX1E : STY: CHX : DM1B 1:1:1:1 mixture by RhH(CO)(TPPMS)3 in H2O/n-heptane; 80 ꢀC, 55 atm CO/H2,
[subst.]:[Rh] = 600. (a) Total aldehyde production for (r) hex-1-ene; (m) styrene; (j) 2,3-dimethyl-1-butene; (·) cyclohexene. (b) Total alkene
conversion.
the formation of Rh colloids [1–4]. The hydroformyla-
tion of HX1E and of the four-component mixtures were
carried our in the presence of an excess of mercury un-
der optimized conditions for the two biphasic catalysts
(80 ꢀC, 55 atm CO/H2, [substrate]:[Rh] = 600, [Hg]:[Rh]
ca. 7000). In all cases the reaction profile was essentially
the same as in the absence of mercury, within experi-
mental error (data not shown), indicating that there is
no participation of metallic particles or colloids in the
catalysis under our reaction conditions. Analyses of
the final mixtures by 31P NMR spectroscopy were not
informative because the samples appeared to decompose
rapidly after the reactor pressure was released and dur-
ing their preparation for the spectroscopic measure-
ment. An in situ high-pressure NMR analysis would
probably help to clarify this point, but it is in any case
unlikely that extensive catalyst decomposition would
take place under our moderate conditions during the rel-
atively short reaction times employed in this work.
since in this case both linear and branched products
would be equally useful. Moreover, the isomerization
side-reaction is not necessarily a serious drawback for
this type of application, as final gasoline specifications
allow up to 12% of olefins and the internal isomers are
actually convenient since they provide a much higher oc-
tane rating than the terminal ones [2]. The properties of
the two biphasic systems reported here did not change
significantly when the catalysts were recycled up to four
times under anaerobic conditions, and addition of sul-
fur-containing compound like benzothiophene in con-
centrations up to 50 ppm did not cause any important
inhibition of the catalysis. In this respect, the aqueous
biphasic systems constituted by RhH(CO)(TPPTS)3
and RhH(CO)(TPPMS)3 can be considered as promis-
ing alternatives for naphtha treatment and related appli-
cations. In future papers we will describe similar work
using low-temperature ionic liquids as the immobiliza-
tion media for hydroformylation catalysts.
3.6. Conclusion
Acknowledgements
The results presented here demonstrate that aqueous
catalysts
biphasic
derived
from
RhH(CO)L3
We thank FONACIT (Caracas) for financial support
through Project CONIPET 97-003777, CODECIHT-
UC (Valencia) through Project 94017, and CYTED
(Madrid) through Project V.9.
(L = TPPTS or TPPMS) can hydroformylate C6 alkenes
and alkene mixtures with reasonable rates under moder-
ate reaction conditions. The main problem associated
with these catalysts is their strong tendency to isomerize
C‚C bonds from the terminal to an internal position if
the effective syngas concentration at the interface is low,
but this side-reaction can be suppressed by using higher
but still reasonable pressures around 50 atm. The mod-
est selectivity of this system for the production of linear
aldehydes from hex-1-ene or for branched aldehydes
from styrene may limit their applicability to classical
oxo uses, but this would not be a great disadvantage
for other applications involving olefin mixtures, in par-
ticular for the important problem of naphtha upgrading,
References
[1] P.W.N.M. van Leeuwen, C. Claver (Eds.), Rhodium catalyzed
hydroformylationB.R. James, P.W.N.M. van Leeuwen (Eds.),
Catalysis by Metal Complexes Series, vol. 22, Kluwer Academic
Publishers, Dordrecht, 2000.
[2] C.D. Frohning, C.W. Kohlpaintner, H.-W. Bohnen, second ed.,
in: B. Cornils, W.A. Herrmann (Eds.), Applied Homogeneous
Catalysis with Organometallic Compounds, vol. 1, Wiley-VCH,
Weinheim, 2002, pp. 31–103 (Chapter 2.2.1).

3792
P.J. Baricelli et al. / Journal of Organometallic Chemistry 689 (2004) 3782–3792
[3] B. Cornils, W.A. Herrmann, in: B. Cornils, W.A. Herrmann
(Eds.), Applied Homogeneous Catalysis with Organometallic
Compounds, second ed., Wiley-VCH, Weinheim, 2002, pp. 603–
633 (Chapter 3.1.1.1).
[8] N. Ahmad, J.J. Levinson, S.D. Robinson, M.F. Uttley, Inorg.
Synth. 28 (1990) 82.
´ ´ ´ ´
[9] F. Joo, J. Kovacs, A. Katho, A. Benyei, A.C. Decuir, D.J.
Darensbourg, Inorg. Synth. 32 (1998) 2.
[10] A.F. Borowski, D.J. Cole-Hamilton, G. Wilkinson, Nouv. J.
Chim. 2 (1978) 137.
[4] B. Cornils, W.A. Herrmann (Eds.), Aqueous-phase Organome-
tallic Catalysis. Concepts and Applications, Wiley-VCH, Wein-
heim, 1998;
[11] (a) R.H. Crabtree, Organometallics 1 (1983) 855;
(b) J.A. Widegreen, R.G. Finke, J. Mol. Catal. A: Chem. 198
(2003) 317.
C.W. Kohkpaintner, R.W. Fischer, B. Cornils, Appl. Catal. A:
General 221 (2001) 219.
[5] A.G. Abatjoglou, D.R. Bryant, U.S. Pat. (to Union Carbide)
4,731,486 (1988); A.G. Abatjoglou, 209th ACS Natnl. Meeting
(1995); Chem. Eng. News 1995, April 17th, p. 25.
[6] D. He, D. Pang, T. Wang, Y. Chen, Y. Liu, J. Liu, Q. Zhu, J.
Mol. Catal. 174 (2001) 21.
[12] Technical Report from ‘‘El Palito’’ Refinery, PDVSA, Venezuela.
´ ´
[13] (a) A. Andriollo, A. Bolivar, F.A. Lopez, D.E. Paez, Inorg.
Chim. Acta 238 (1995) 187;
´ ´
(b) R.A. Sanchez-Delgado, M. Medina, F. Lopez-Linares, A.
Fuentes, J. Mol. Catal. 116 (1997) 167;
´
(c) A. Andriollo, J. Carrasquel, J. Marin˜o, F.A. Lopez, D.E.
´
[7] P.J. Baricelli, E. Lujano, M. Rodrıguez, A. Fuentes, R.A.
´
Sanchez-Delgado, Appl. Catal. A: General 263 (2004) 187.
´
Paez, I. Rojas, N. Valencia, J. Mol. Catal. 116 (1997) 157.