Accelerated biphasic hydroformylation by vesicle formation of amphiphilic diphosphines

Article abstract of DOI:10.1021/ja9925610

The synthesis, aggregation behavior, and catalytic activity of rhodium complexes of a series of Xantphos derivatives with surface-active pendant groups, -4-C₆H₄O(CH₂)(n)C₆H₄(SO₃Na)- (n = 0,

Full text of DOI:10.1021/ja9925610

1650
J. Am. Chem. Soc. 2000, 122, 1650-1657
Accelerated Biphasic Hydroformylation by Vesicle Formation of
Amphiphilic Diphosphines
Marcel Schreuder Goedheijt,‡ Brian E. Hanson,† Joost N. H. Reek,‡
Paul C. J. Kamer,‡ and Piet W. N. M. van Leeuwen*,‡
Contribution from the Institute of Molecular Chemistry, Homogeneous Catalysis, UniVersity of Amsterdam,
Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands, and Virginia Polytechnic Institute and
State UniVersity, Department of Chemistry, Blacksburg, Virginia 24061
ReceiVed July 19, 1999
Abstract: The synthesis, aggregation behavior, and catalytic activity of rhodium complexes of a series of
Xantphos derivatives with surface-active pendant groups, -4-C6H4O(CH2)nC6H4(SO3Na)- (n ) 0, 3, 6) is
described. Electron microscopy experiments show that these ligands and their complexes form vesicles in
water if the hydrophobic part of the ligand is large enough (n ) 3, 6). The formed aggregates are stable at
elevated temperatures (90 °C), and their presence leads to a significant enhancement of the solubility of organic
substrates in aqueous solution. This enhanced solubility results directly in a higher reaction rate in the rhodium-
catalyzed hydroformylation of 1-octene. Furthermore, recycling experiments show that the TOF and the high
selectivity toward the more valuable linear aldehyde remains approximately the same in four consecutive
runs. This indicates that the aggregates stay intact during the recycling and the active rhodium complex is
retained in the water-phase quantitatively.
Introduction
the catalyst is dissolved or at the phase boundary. The main
drawbacks of two-phase systems are the low reaction rates due
to phase-transfer limitations caused by poor substrate solubility
in the water phase. For substrates that are moderately soluble
in water, two-phase catalysis is a commercially feasible process
as is apparent from the Ruhrchemie/Rhoˆne Poulenc process for
the hydroformylation of propene.
The fast-growing insight into organometallic chemistry has
resulted in an increasing number of applications of homogeneous
catalysis for the mild and highly selective production of
chemicals.1 While fundamental knowledge of ligand effects on
the selectivity and reactivity of transition metal complexes has
resulted in the development of numerous tailor-made homoge-
neous catalysts, practical applications often involve hetereogeous
catalysts because of their ease of separation from the reaction
mixture. Therefore, much research has been devoted to the
development of novel techniques to separate homogeneous
catalysts from the product, an important one being two-phase
catalysis.2-8 In a two-step process the catalysis is performed in
one-phase (organic) after which the catalyst is extracted with a
second immiscible phase (water) from the mixture.4 In a “one-
step-two-phase” system the catalyst is located in a solvent,
usually water, that is immiscible with the second solvent that
contains the substrate. The reaction occurs in the phase in which
An elegant example of two-phase catalysis in nonaqueous
environment that circumvents phase-transfer limitations is the
fluorous-phase catalysis, originally developed by Horva´th.9
While reaction rate and catalyst recovery is excellent, leaching
of the fluorous phase into the product phase remains a major
drawback. Another system that lacks phase-transfer limitations
is immobilizing the aqueous phase on a silica support in the
supported aqueous phase catalysis, which results in an enormous
increase of the contact surface.10 These immobilization studies,
however, showed that the high selectivity of the homogeneous
catalyst is not retained in such a heterogenized catalyst system.
The addition of cosolvents, like methanol, to the two-phase
system leads to a higher solubility of the organic substrates in
the water-phase but it often results in a cumbersome phase
separation and the cosolvent will leach into the product phase.2
Another approach is the addition of surface-active compounds
that enhance the solubility of the substrate.3,6,7 For a Diels-
Alder reaction this, together with the addition of a Lewis acid,
resulted in a million-fold rate acceleration.11 The formation of
an emulsion can disturb the easy separation in such a system.
Furthermore, if the catalyst and the substrates are both dissolved
in the organic part of the aggregate, the catalyst cannot be easily
separated from the organic product.
* To whom correspondence should be addressed.
† Virginia Polytechnic Institute and State University.
‡ Institute of Molecular Chemistry, University of Amsterdam.
(1) Cornils, B.; Herrmann, W. A. Applied Homogeneous Catalysis with
Organometallic Compounds; Wiley-VCH: Weinheim, 1996.
(2) Herrmann, W. A.; Kohlpaintner, C. W. Angew. Chem., Int. Ed. Engl.
1993, 32, 1524.
(3) Fell, B.; Schobben, C.; Papadogianakis, G. J. Mol. Catal. 1995, 101,
179.
(4) Buhling, A.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Elgersma
J. W., J. Mol. Catal. 1997, 116, 297.
(5) (a) Cornils, B.; Herrmann, W. A. Aqueous-Phase Organometallic
Chemistry -Concepts and Applications; Wiley-VCH: Weinheim, 1998.
(b) Cornils, B. Org. Process Res. DeV. 1998, 2, 121.
(6) (a) Monteil, F.; Queau, R.; Kalck, P. J. Organomet. Chem. 1994,
480, 177. (b) Buriak, J. M.; Osborn, J. A. Organometallics 1996, 15, 3161.
(c) Selke, R.; Ohff, M.; Riepe, A. Tetrahedron 1996, 52, 15079.
(7) (a) Ding, H.; Hanson, B. E. J. Mol. Catal. 1995, 99, 131. (b) Bartik,
T.; Ding, H.; Bartik, B.; Hanson, B. E. J. Mol. Catal. 1995, 98, 117.
(8) Fremy, G.; Monflier, E.; Carpenter, J. F.; Castanet, Y.; Mortreux,
A. Angew. Chem., Int. Ed. Engl. 1995, 34, 1474.
(9) Horva´th, I. T.; Ra´bai, J. Science 1994, 266, 72.
(10) Arhancet, J. P.; Davis, M. E.; Merola, J. S.; Hanson, B. E. Nature
1989, 339, 454.
(11) (a) Breslow, R.; Rideout, D. C. J. Am. Chem. Soc. 1980, 102, 7816.
(b) Breslow, R.; Guo, T. J. Am. Chem. Soc. 1988, 110, 5613. (c) Otto, S.;
Boccaletti, G.; Engberts, J. B. F. N. J. Am. Chem. Soc. 1998, 120, 4238.
10.1021/ja9925610 CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/10/2000

Hydroformylation by Diphosphine Vesicle Formation
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1651
A possible new way to improve the solubility of substrates
is the use of surface-active reagents that can form micelles12 or
vesicular structures. In this way the substrate and the catalyst
can be brought in close proximity in the hydrophobic interior
of the aggregate.13
of higher 1-alkenes, however, was very slow because of the
low solubility of the organic substrates in the aqueous phase.
Here we report on the syntheses and characterization of a
series of amphiphilic diphosphines based on a xanthene-type
backbone with pendant groups -C6H4O(CH2)nC6H4SO3Na- (n
) 0,3,6) (2, 3, 4, respectively) that are used in the rhodium-
catalyzed hydroformylation of 1-octene in aqueous solution. The
ligands and their rhodium complexes form aggregates in water.
Electron microscopy and fluorescence experiments evidenced
the formation of vesicles that are remarkably stable at elevated
temperatures. The formation of aggregates has a pronounced
effect on the solubility of organic substrates which results in a
higher activity of the catalyst, whereas the high selectivity is
retained. Furthermore, recycling can be done without any
formation of emulsions during the phase separation, and no
catalyst deactivation nor leaching is observed in four consecutive
cycles.
Rhodium-catalyzed hydroformylation is one of the most
important applications of homogeneous catalysis in industry.1,14
High selectivities in the hydroformylation of 1-alkenes have
been obtained for both diphosphite and diphosphine-modified
catalysts.15 A new, selective catalyst based on binuclear rhodium
complexes and tetraphosphine ligands has been developed by
Stanley and co-workers.16 Casey and co-workers developed the
concept of the natural bite angle of diphosphine ligands, which
was found to have a pronounced effect on the selectivity in the
hydroformylation.17 Mechanistic studies of Casey and Van
Leeuwen have contributed greatly to the understanding of steric
and electronic ligand effects on rate and selectivity of the
hydroformylation reaction.17,18
Results and Discussion
Synthesis. The phosphines containing the sulfonated pendant
groups were prepared starting from the 2,7-dialkylated xanthene
backbones (Scheme 1). The alkyl groups prevent sulfonation
of the backbone. Lithiation of the xanthene backbone and
subsequent reaction with ClP(OEt)2 gives the diphosphonite 8
in high yield (77%). This compound is made to react with the
lithiated analogues of 5, 6, and 7 to yield compounds 9, 10,
and 11 respectively. Selective sulfonation of the terminal phenyl
rings can take place under mild conditions according to a
previously developed procedure,12a giving 2, 3, and 4, respec-
tively.
Due to the relatively mild reaction conditions for sulfonation
only very small amounts of phosphine oxides (<5%) are formed.
The latter, moderately soluble in hot EtOH, are removed by
refluxing the sulfonated diphosphine in EtOH and subsequent
hot filtration. Neither the unsulfonated nor the sulfonated
diphosphines are very susceptible to oxidation; they can be
stored in air for months without noticeable oxidation.
We have shown previously that catalysts of the water-soluble
diphosphine 2,7-bis(SO3Na)Xantphos (1) induces high regiose-
lectivity for the formation of n-butyraldehydes in the rhodium-
catalyzed hydroformylation of propene.19 The hydroformylation
Aggregation Behavior of the Amphiphiles. The main
advantage of using amphiphilic diphosphines as ligands com-
pared to previously reported water soluble ligands19 is an
anticipated increase in solubility of organic substrates during
two-phase catalysis. Before the aggregation behavior of the
novel ligands 2-4 was examined, the ability of these ligands
to solubilize such organic substrates in water was studied. The
dissolution of the poorly water-soluble dye Orange OT was used
as a model compound since it can be measured easily by UV-
vis or fluorescence spectroscopy.20a Relatively high concentra-
tions of Orange OT could be dissolved in aqueous solutions of
3 and 4 (5 and 7 mM, respectively, at a diphosphine concentra-
tion of 2 mM) as was evident from the UV spectra, whereas
aqueous solutions of 1 or TPPTS do not dissolve Orange OT at
all (<0.1 mM). Orange OT was only moderately soluble ( ∼1
mM) in an aqueous solution containing 2. The amphiphilic
character of 3 and 4 has a positive effect on the solubility of
organic substrates which is required for efficient catalysis.
(12) Ding, H.; Hanson, B. E.; Bartik, T.; Bartik, B. Organometallics 1994,
13, 3761. (b) Hanson, B. E.; Ding, H.; Kohlpaintner, C. W. Catal. Today
1998, 42, 421. (c) Oehme, G.; Grassert, I.; Ziegler, S.; Meisel, R.; Fuhrmann,
H. Catal. Today 1998, 42, 459. (d) Ding, H.; Hanson, B. E.; Bakos, J.
Angew. Chem., Int. Ed. Engl. 1995, 34, 1645.
(13) Christian, S. D.; Scamehorn, J. F. Solubilization in Surfactant
Aggregates; Marcel Dekker: New York, 1995.
(14) Parshall, G. W. Homogeneous Catalysis: the applications and
chemistry of catalysis by soluble transition metal complexes; Wiley: New
York, 1980.
(15) (a) Moasser, B.; Gladfelter, W. L.; Roe, D. C. Organometallics 1995,
14, 3832. (b) van Rooy, A.; Orij, E. N.; Kamer, P. C. J.; van Leeuwen, P.
W. N. M. Organometallics 1995, 14, 34.
(16) (a) Broussard, M. E.; Juma, B.; Train, S. G.; Peng, W. J.; Laneman,
S. A.; Stanley, G. G. Science 1993, 260, 1784. (b) Matthews, R. C.; Howell,
D. K.; Peng, W. J.; Train, S. G.; Treleaven, W. D.; Stanley, G. G. Angew.
Chem., Int. Eng. Ed. 1996, 35, 2253.
(17) (a) Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L.
M.; Gavney, J. A. Jr.; Powell, D. R. J. Am. Chem. Soc. 1992, 114, 5535.
(b) Casey, C. P.; Petrovich, L. M. J. Am. Chem. Soc. 1995, 117, 6007. (c)
Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.; Petrovich,
L. M.; Matter, B. A.; Powell, D. R. J. Am. Chem. Soc. 1997, 119, 11817.
(18) (a) Kranenburg, M.; van der Burgt, Y. E. M.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M. J. Organometallics 1995, 14, 3081. (b) van der
Veen, L. A.; Boele, M. D. K.; Bregman, F. R.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J.; Schenk, H.; Bo, C. J. Am.
Chem. Soc. 1998, 120, 1161. (c) Kamer, P. C. J.; Reek, J. N. H.; van
Leeuwen, P. W. N. M. CHEMTECH 1998, September, 27.
Compounds 3 and 4 show line broadening in the 1H and 31
P
(∆ν1/2 ≈ 240 Hz) NMR spectra in D2O, suggesting that these
compounds form large aggregates in water. Several techniques
were used to study their aggregation behavior in more detail.
Dynamic light-scattering experiments on aqueous solutions of
3 and 4 showed the presence of aggregates with average
(20) (a) Adamson, A. W. In Physical Chemistry of Surfaces; Wiley-
Interscience: 1990; p 509. (b) Kabanov, A. V.; Bronich, T. K.; Kabanov,
V. A.; Yu, K.; Eisenberg, A. J. Am. Chem. Soc. 1998, 120, 9941.
(19) Schreuder Goedheijt, M.; Kamer, P. C. J.; van Leeuwen, P. W. N.
M. J. Mol. Catal. 1998, 134, 243.

1652 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Goedheijt et al.
Scheme 1. Synthesis of the Amphiphilic Diphosphines 2-4
Table 1. Dynamic Light-scattering Experiments Performed on
Aqueous Solutions of 1-4
might be due to the very specific structure of the amphiphile;
that is, the rigid hydrophobic backbone with four tails having a
hydrophilic headgroup. It must be noted that these molecules
can also act as bola-amphiphiles spanning the bilayer. If a
fraction of amphiphiles indeed is incorporated into the aggregate
in a different way, this might be the second component required
for the spontaneous formation of vesicles.
The average vesicle size formed by 3 is approximately the
same as that formed by 4, which suggests that the slightly longer
tails of 4 do not have much influence on the structure of the
formed aggregates.
These novel ligands (2-4) were developed for the two-phase
rhodium-catalyzed hydroformylation of larger substrates as
1-octene. Therefore, we also studied their aggregation behavior
under conditions that resemble that of the two-phase catalysis.
Thus, the effect of the addition of the rhodium metal, addition
of octene substrate, and temperature on the aggregation was
investigated. The aggregation behavior of complex 12, which
is the complex between ligand 3 and RhH(CO)PPh3 and closely
resembles the resting state of the active rhodium catalyst, was
studied. The concentration of 12 was the same as that used in
the catalytic experiments. Electron microscopy showed that
vesicles formed from 12 are very similar to those formed from
3. This indicates that the complexation of the metal does not
influence the structural features of the aggregates which is not
surprising in view of the apolar character of the metal complex
of 12. More interestingly, the addition of 1-octene to an aqueous
solution of 12 and subsequent sonication for several minutes,
resulted in the formation of larger vesicles (sonication of a
solution of only 12 did not change the vesicles). Their average
diameter increased from 140 nm to 500-600 nm as was
estimated from electron microscopy analysis (TEM, platinum
shadowing, Figure 1d). Apparently the larger vesicles formed
concentration
(M)
average hydrodynamic
radius (nm)
ligand
1
2
3
4
8.3 × 10-3
8.3 × 10-3
1 × 10-2
1.0
2.1
63
8.3 × 10-3
67
hydrodynamic radii of 63 and 67 nm, respectively (Table 1).
The observed, small hydrodynamic radii (1 and 2 nm, respec-
tively) for ligand 1 and 2 indicate that these ligands do not form
aggregates, presumably because the hydrophobic part of these
molecules is too small. To visualize the formed aggregates,
aqueous solutions of 1-4 were studied by transmission electron
microscopy (TEM). The micrographs clearly show that 3 and
4 form vesicles in solution of sizes varying from 50 to 250 nm
(Figure 1a, platinum shadowing technique), which is in the same
range as the hydrodynamic radii measured in the light scattering
experiments. Samples of 1 and 2 did not show the formation of
aggregates confirming the observations in the light-scattering
experiments. A few larger vesicles (up to 600 nm) were also
observed in the solution of 4 (Figure 1b, platinum shadowing
technique). Samples of 3 and 4 prepared with the freeze-fracture
technique substantiates the formation of vesicles (Figure 1c).
According to these measurements the vesicle structures are
single-walled, since no layers could be determined.21 After
addition of methanol (2-5%) to an aggregate solution no
vesicular structures were observed, indicating the disruption of
the aggregates.
Remarkably, the formation of the vesicles was spontaneous;
they were formed within 30 min from a 2 mM aqueous solution,
with or without sonification. Furthermore the vesicles are stable
at least for several days. Spontaneous formation of vesicles is
rare;22 an input of energy or an additional component is normally
required.23 In this particular case spontaneous vesicle formation
(22) A few examples: (a) Bandyopadhyay, P.; Bharadwai, P. K.
Langmuir 1998, 14, 7537. (b) Ghosh, P.; Khan, T. K.; Bharadwai, P. K.
Chem. Commun. 1996, 189. (c) Kimizuka, N.; Wakiyama, T.; Miyauchi,
H.; Yoshimi, T.; Tokuhiro, M.; Kunitake, T. J. Am. Chem. Soc. 1996, 118,
5808. (d) Kimizuka, N.; Kawasaki, T.; Hirata, K.; Kunitake, T. J. Am. Chem.
Soc. 1998, 120, 4094. (e) Kabanov, A. V.; Bronich, T. K.; Kabanov, K.;
Eisenberg, S. Kimizuka, N.; Wakiyama, T.; Miyauchi, H.; Yoshimi, T.;
Tokuhiro, M.; Kunitake, T. J. Am. Chem. Soc. 1998, 120, 9941.
(23) Lasic, D. D. J. Colloid Interface Sci. 1990, 140, 302.
(21) (a) Kunitake, T. Angew. Chem., Int. Ed. Engl. 1992, 31, 709. (b)
Fuhrhop, J.-H.; Mathieu, J. Angew. Chem., Int. Ed. Engl. 1984, 23, 100.
(c) Ringsdorf, H.; Schlarb, B. Venzmer, J. Angew. Chem., Int. Ed. Engl.
1988, 27, 7, 113.

Hydroformylation by Diphosphine Vesicle Formation
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1653
Figure 1. Electron microscopy pictures of the vesicles formed by 3 and 4. A uniform distribution of vesicles of 3; platinum shadowing technique
(a). A large and an average sized vesicle formed by 4 (b). Freeze-fractured sample of a solution of 3 showing the formation of monolayered vesicles
(c). Larger aggregates observed shortly after sonication of a solution of 12 and 1-octene, platinum shadowing technique (d). (Bar size 200 nm)
after incorporation of 1-octene are thermodynamically more
favorable. The increased solubility of 1-octene induced by these
aggregates is also demonstrated by 1H NMR spectroscopy. The
1H NMR spectrum of a saturated solution of 1-octene in D2O
did not show any octene signals at all, but after the addition of
3, 4, or 12 (2 mM) the olefinic signals of 1-octene were clearly
observed.
To study the temperature stability of the vesicles derived from
3 and 4, samples were prepared after heating the aqueous
solution to reflux for several minutes. Analysis by electron
microscopy showed that similar vesicles were present which
suggests that they are stable at elevated temperatures. The
thermostability of the aggregates was further studied by
fluorescence experiments.
Pyrene is an excellent probe to study aggregated systems in
situ due to its photophysical properties.24 Its fluorescence
spectrum is very well resolved and the relative intensity of two
different vibronic bands (I and III) depends on the solvent
polarity.24 This has been used to study micellar aggregates, the
Figure 2. The I(1)/I(3) band ratio in the fluorescence spectra of pyrene
critical micelle concentration and the extent of water penetration
dissolved in various aggregates as a function of the temperature showing
in a micelle.25,26 Here we use pyrene as a polarity probe to study
the stability of the vesicles formed by 3 and 4.
the stability of the vesicles at higher temperatures. Therefore,
the III/I band ratio of pyrene was measured as function of the
temperature (in the range from 298 to 363 K) in pure water
and in aqueous solutions of SDS (Sodium Dodecyl Sulfate) and
2-4 (in all cases above the critical aggregation concentration
the temperature range, indicating that pyrene is in the water
phase. In the presence of SDS or amphiphiles 3 or 4 the band
III/I band ratio is much higher at room temperature (between
0.74 and 0.85), showing that pyrene is dissolved in the organic
(CAC)). In an aqueous saturated solution of pyrene the III/I
phase of the surfactant (literature values are between 0.73 and
ratio is between 0.55 at 298 K and 0.63 at 353 K (lit. 0.64).24
0.96, dependent on the micelle forming surfactant).24 For SDS
a large drop is observed in the temperature range between 335
For ligand 2 an average ratio of 0.64 is observed throughout
and 355 K, corresponding to the phase transition temperature
(24) Kalyanasundaram, K.; Thomas, J. K. J. Am. Chem. Soc. 1977, 99,
of SDS (Figure 2). Above these temperatures the micelles
formed by SDS are disrupted, and pyrene is in the water phase.
For comparison these experiments were also carried out using
a vesicle solution of dipalmitoyl-phosphatidylcholine (dppc).
In these experiments the III/I band ratio was above 0.9 at all
2039.
(25) (a) Stam, van J.; Depaemelaere, S.; De Schryver, F. C. J. Chem.
Educ. 1998, 75, 93. (b) Damas, C.; Naejus, R.; Coudert, R.; Frochot, C.;
Brembilla, A.; Viriot, M.-L. J. Phys. Chem. B 1998, 102, 10917.
(26) Fendler, J. H.; Fendler, E. J. Catalysis in Micellar and Macromo-
lecular Systems; Academic Press: New York, 1975.

1654 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Goedheijt et al.
Table 2. Two-Phase Rhodium-Catalyzed Hydroformylation of
1-Octene with 1-4 in Watera
TOFb
70 °C
TOFb
90 °C
TOFb
120 °C
linear
aldehyde (%)
ligand
l/b
1
2
3
4
0-1
0-1
2
7
5
9
13
21
97.1
96.3
96.9
94.7
97.1
96.3
97.0
97.4
99/1
98/2
98/2
97/3
98/2
97/3
98/2
99/1
1-2
6
12
7
8
7
3 (cycle 1)
3 (cycle 2)
3 (cycle 3)
3 (cycle 4)
7
a Reaction conditions: incubation time, 15 h; reaction time, 24 h;
initial pressure, 210 psi (at 25 °C); stirring rate is 700 rpm; [Rh] ) 1.7
× 10-3 M; substrate/Rh ratio is 3820/1, ligand/Rh ratio ) 5. b TOF )
mol aldehyde/(mol Rh × h).
Figure 3. The relative reaction rate in the rhodium-catalyzed hydro-
formylation of 1-octene in water using ligands 1-4 at 343 and 393 K.
The TOFs measured for ligand 1 are set to 1, since this ligand is not
able to form aggregates and does not increase the solubility of substrates.
temperatures measured between 298 and 353 K, indicating that
the probe was well-dissolved in the aggregates in this temper-
ature range. The relatively large increase in the ratio (0.92-
1.07) observed in these experiments going from 298 to 348 K
was in agreement with the change observed by Lisse et al. and
Thomas et al.27 In the pre-transition temperature range this
increase was attributed to a deeper incorporation of the probe
into the bilayer.27 At the phase-transition temperature a sharp
increase was observed, indicating that the increased fluidity of
the bilayer allowed a greater access of the pyrene. We also
measured the fluorescence at higher temperatures and observed
a drop in III/I band ratio at temperatures higher than 353 K
which is likely due to an increased water penetration into the
aggregates. For amphiphiles 3 and 4 a very small increase in
III/I band ratio was observed from 298 to 320 K, and this ratio
showed a very small drop going to temperatures above 333 K.
These small fluctuations can also be attributed to a deeper
incorporation of the pyrene at 323 K and a possible phase
transition at 333 K. The most important observation with respect
to catalysis, however, is the fact that these vesicles are stable
at higher temperatures, especially when compared to micelle-
forming amphiphiles such as SDS, and organic substrates such
as pyrene are still located within the organic phase of the
aggregates at these elevated temperatures.
After the addition of methanol (2-5%) to a solution of 4 an
abrupt drop of the band III/I-ratio is observed (from 0.83 to
0.69), suggesting that the vesicles are disrupted. This is in
agreement with the TEM experiments. Apparently these am-
phiphilic diphosphines form remarkably thermostable aggre-
gates, but they are sensitive to the addition of methanol. We
ascribe this unusual thermostability to the typical structure of
the molecule: a rigid organic backbone and the presence of
four polar headgroups.
Catalysis. To investigate the effect of aggregate formation
with 3 and 4 on catalysis, hydroformylation experiments were
performed. 1-Octene was used since this substrate is not
sufficiently soluble in water to give substantial yields in the
rhodium-catalyzed hydroformylation using 1 as ligand.19 The
catalysis experiments were conducted, after an incubation time
of 15 h to ensure complete catalyst formation, at 70, 90, and
120 °C (210 psi of 1:1 CO/H2, with a ligand-to-rhodium ratio
of 5 [Rh] ) 1.7 mM).
vesicle solutions of 4. This is attributed to the increased
solubility of 1-octene. At 70 and 90 °C very similar catalytic
activities were found. At higher temperatures (120 °C) signifi-
cant activities were found using 1 and 2, indicating that at these
temperatures 1-octene is slightly soluble in the water phase even
in the absence of aggregates. The reaction rates observed for 3
and 4 were also higher at these elevated temperatures, but the
beneficial effect of the amphiphilic ligands was less pronounced.
To visualize the effect of aggregate formation the relative
reaction rates of 1-4 at 70 and 120 °C are plotted (the reaction
rate of 1 is set at 1; see Figure 3). At 120 °C the rate observed
for the surfactant ligand 4 is only a factor 3 higher than that for
1. The smaller difference in reaction rate at higher temperatures
suggests that the aggregates are partially disrupted at temper-
atures above 100 °C. In a separate catalytic experiment
performed in aqueous methanol (1:1 H2O/MeOH), a medium
in which no aggregates were formed, the activity of the catalyst
derived from 4 was only three times higher than that derived
from 1.
The local 1-octene concentrations in the vesicles is much
higher compared to the concentration usually used in homoge-
neous systems. Such a concentration effect in micellar media
can accelerate reaction rates.28 If the reaction is first order in
substrate, which is the case for the homogeneous systems of
this ligand type, a higher reaction rate will be obtained. Since
this is not observed, it is likely that for this system the rate-
limiting step is the octene transfer from the organic phase to
the vesicle. This is supported by the observed TOFs for 3 at 70
°C and 90 °C which are almost similar. Furthermore, the TOF
was dependent on the stirring rate and decreased dramatically
at lower stirring rates. Further experiments aiming at the
optimalization of the reactions rate will focus on the mass-
transfer limitation.
Thus far we explained the difference in catalytic performance
observed for the water soluble ligands 1 and 2 on one hand
and the vesicle-forming ligands 3 and 4 on the other hand.
Although the structures of the aggregates formed by 3 and 4
are very similar, a significant difference in reaction rate is
observed. Since the rate-limiting step is the octene transfer from
the octene bulk phase to the aggregates, the difference observed
between 3 and 4 must involve this transfer. We believe that the
higher solubility of organic substrates (as was directly measured
for orange OT) in the aggregates formed by 4, results in a larger
overall vesicle concentration. This concentration effect will
enhance the octene transfer from the bulk to the vesicle. The
reaction rate is, of course, also directly dependent on the amount
Table 2 shows that the catalytic activity is indeed much higher
when 3 and 4 are used as the ligands compared to 1 and 2. At
70 °C the average TOF, calculated after 24 h, was approximately
12-14 times higher when catalysis is performed using the
(27) (a) Lissi, E. A.; Abuin, E.; Saez, M.; Zanocco, A.; Disalvo, A.
Langmuir 1992, 8, 348. (b) Morris, D. A. N.; McNeil, R.; Castellino,
Thomas, J. K. Biochim. Biophys. Acta 1980, 599, 380.
(28) Tascioglu, S. Tetrahedron 1996, 11113.

Hydroformylation by Diphosphine Vesicle Formation
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1655
Experimental Section
of octene dissolved in the vesicle. Furthermore, the organic
fluorescent probe pyrene was shown to be in a more hydro-
phobic environment when dissolved in aggregates of 4 compared
to that in 3. This increased hydrophobicity might also enhance
the rate of octene transfer.
In all experiments high selectivities (96%) toward the linear
aldehydes are found, which is in line with the previously
reported results in organic solvents and aqueous media.18,19,29
The large bite angle, which is responsible for this high
selectivity, also results in the formation of mainly linear products
in the catalyst systems based on organized aggregates. This
indicates that the rhodium complexes formed in the aggregates
are similar to those formed by the monomeric species in water
and organic solution. The strong chelating effect of the pre-
organized diphosphine is very important to retain the high
selectivity.
Recycling Experiments. The recyclabilty of the aggregated
systems was studied by performing a series of consecutive runs.
The use of 3 as the ligand (Table 2) led to an easy separation
and, to our surprise, no formation of emulsions was observed.
After each run an immediate separation of the organic phase
takes place after cooling the reaction mixture. The fact that these
amphiphiles do not form emulsions is remarkable, and we
ascribe this to the stability and solubility of the aggregates in
water. Both the TOF and the selectivity are nearly unchanged
even after four cycles. This indicates that the catalyst is stable
and stays completely in the aqueous phase. No rhodium could
be detected in the organic phase by graphite furnace atomic
absorption spectrophotometry (GFAAS) (Rh in organic layer
, 1 ppb), confirming that these vesicles are very efficient in
keeping rhodium in the aqueous phase.
General. All reactions and measurements were carried out using
standard Schlenk techniques under an atmosphere of argon.
Solvents and Materials. Solvents were dried and deoxygenated by
distillation under nitrogen prior to use; water was distilled under
nitrogen. 4-Bromophenol, 1-bromo-3-phenylpropane, 2,7-di-tert-butyl-
4,5-dibromo-9,9-dimethylxanthene, diethyl chlorophosphite, 4-bromo-
diphenyl ether, 1-octene, nonane, and sulfuric acid were purchased from
Aldrich (Milwaukee, WI). Rh(acac)(CO)2 was purchased from Strem
Chemicals (Newburyport, MA). The CO/H2 (50/50) was purchased
from Airco. TPPTS,30 2,7-bis(SO3Na)Xantphos (1)19 and (PPh3)3RhH-
(CO)31 were prepared according to literature methods.
1
Product Analysis. H NMR (360 MHz), 13C NMR (90.6 MHz),
and 31P NMR spectra (145.8 MHz, referenced to external 85% H3PO4)
were recorded on a Bruker AM 360 spectrometer. Mass spectrometry
(FAB) was performed on a VG 7070E-HF spectrometer. Melting points
were measured on a Mel-Temp melting apparatus. Gas chromatography
was performed on a Varian 3300 gas chromatograph equipped with a
HP1 column 25 m × 0.32 mm × 0.52 µm and FID detector, He was
the carrier gas; the temperature program was from 50 °C (5 min) to
220 °C (1 min), at a heating rate of 2 °C/min and a HP 6890 series
equipped with a HP-5 (5% PH ME Siloxane) column 30 m × 0.25
µm. Molecular size detection was performed on a Biotage dp-801
molecular size detector. Microanalyses were performed by Atlantic
Microlab. Inc., Norcross, Georgia. Rhodium analysis was performed
on a Buck Scientific 200-A graphite furnace atomic absorption
spectrophotometer.
p-Bromophenoxy-1-phenyl-propane (6). To a solution of p-
bromophenol (8.66 g, 50 mmol) and NaOH (2.50 g, 62.5 mmol) in
DMSO (40 mL) was added 1-bromo-3-phenyl-propane (9.96 g, 50
mmol), and the mixture was heated at 100 °C for 5 h. The pale yellow
reaction mixture was cooled to room temperature. Water (100 mL) and
Et2O (100 mL) were added, and the mixture was vigorously stirred
after which the Et2O layer was separated. The water layer was washed
twice with Et2O, and the combined organic layers were dried over
MgSO4. Evaporation of the solvent yielded a pale yellow oil. Distillation
Conclusions
1
gave 6 as a colorless oil. Yield: 10.9 g (37.5 mmol; 75%). H NMR
New amphiphilic diphosphines 2-4 were developed for the
two-phase rhodium-catalyzed hydroformylation of organic
substrates as 1-octene. Surprisingly, ligands 3 and 4 and their
rhodium complexes spontaneously form vesicles (average size
∼140 nm) in aqueous solution as was evidenced by electron
microscopy (shadowing and freeze-fracture techniques), light
scattering, and fluorescence studies. The aggregates are remark-
ably thermostable; the presence of vesicular structures is also
observed at temperatures as high as 90 °C. With respect to
catalysis performed at higher temperatures such as the hydro-
formylation this high stability, especially compared to that of
(CDCl3; ppm) δ 7.39-7.24 (m, 7H, ArH), 6.79 (d, J ) 9.0 Hz, 2H,
ArH), 3.93 (t, J ) 6.3 Hz, 2H, OCH2), 2.83 (t, J ) 7.3 Hz, 2H, ArCH2),
2.11 (dt, J ) 6.3 Hz, J ) 7.3 Hz, 2H, CCH2C). 13C NMR (CDCl3;
ppm) 158.4 (s, 1C), 141.6 (s, 1C), 132.5 (s, 2C), 128.8 (s, 4C), 126.3
(s, 1C), 116.6 (s, 2C), 112.9 (s, 1C), 67.2 (s, 1C), 32.3 (s, 1C), 30.9 (s,
1C). Bp 137 °C (2‚10-5 Torr).
2,7-di-tert-butyl-4,5-bis(diethoxyphosphino)-9,9-dimethylxan-
thene (8). To a solution of 2,7-di-tert-butyl-4,5-dibromo-9,9-dimeth-
ylxanthene (5.00 g, 10.4 mmol) in 80 mL of THF at -70 °C was
n
added a solution of BuLi (20.8 mmol, 2.5 M) in hexanes. After the
addition was complete the mixture was stirred for 2 h at -70 °C. Then
a solution of diethyl chlorophosphite (20.8 mmol) in 20 mL of
THF was added dropwise, and the resulting pale yellow solution was
allowed to warm to room temperature overnight. The solvent was
removed in vacuo which gave a pale yellow waxy solid. CH2Cl2 (100
mL) and water (50 mL, degassed) were added, and the mixture was
vigorously stirred. The organic layer was separated and dried over
MgSO4. Evacuation of the solvent yielded 8 as a pale yellow oil.
Yield: 4.5 g (8.0 mmol; 77%). 1H NMR (CDCl3; ppm) δ 7.72 (s, 2H,
ArH), 7.59 (s, 2H, ArH), 3.95 (m, 8H, OCH2), 1.57 (s, 6H, C(CH3)2),
1.24 (t, J ) 6.4 Hz, 12 H, CH3), 1.22 (s, 18H, C(CH3)3). 13C{1H} NMR
(CDCl3; ppm) δ 149.9 (t, 2C), 144.9 (s, 2C), 128.8 (s, 2C), 128.0 (s,
2C), 125.3 (s, 2C), 124.4 (s, 2C), 62.9 (t, 4C), 34.4 (s, 2C), 34.2 (s,
1C), 31.5 (s, 2C), 31.3 (s, 6C), 17.2 (4C). 31P{1H} NMR (C6D6; ppm)
δ 152.2.
1
micelles, is a major advantage. From H NMR spectroscopy,
UV-vis, and fluorescence experiments it is evident that the
solubility of organic substrates increases significantly in the
presence of these aggregates, which results in higher reaction
rates in the rhodium-catalyzed hydroformylation of 1-octene.
These ligands result in the highest reported regioselectivities
and activities in the aqueous phase thus far for chelating
phosphines. The observed TOFs in the hydroformylation of
1-octene using ligands that form vesicles are up to 14 times
higher compared to ligands that do not form aggregates.
Furthermore, no decrease in activity is observed upon catalyst
recycling. These ligands are unique in aqueous phase catalysis
since the formation of the vesicles is spontaneous, the vesicles
are thermostable, they increase the solubility of organic
substrates and also the rate of conversion of these substrates,
and the catalytically active vesicles allow an easy separation of
the product and the catalyst.
2,7-di-tert-Butyl-4,5-bis{di[p-phenylphenoxy]phosphino)-9,9-di-
methylxanthene (9). To a solution of p-bromodiphenyl ether (4.9 g,
19.5 mmol) in 40 mL of THF at - 70 °C was added nBuLi (19.5 mmol,
2.5 M in hexanes). After the addition was complete, the mixture was
(30) Bartik, T.; Bartik, B.; Hanson, B. E.; Glass, T.; Bebout, W. Inorg.
Chem. 1992, 31, 2667.
(29) (a) van der Veen, L. A.; Kamer, P. C. J.; van Leeuwen, P. W. N.
M. Organometallics 1999, 18, 4765. (b) van der Veen, L. A.; Kamer, P. C.
J.; van Leeuwen, P. W. N. M. Angew. Chem., Int. Ed. 1999, 38, 336.
(31) Hallman, P. S.; Evans, D.; Osborn, J. A.; Wilkinson, G. Chem.
Commun. 1967, 305.

1656 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Goedheijt et al.
stirred for 2 h at - 70 °C. Then a solution of 8 (3.9 mmol, 2.2 g) in
10 mL of THF was added dropwise, and the resulting orange solution
was allowed to warm to room-temperature overnight after which the
solvent was evaporated. CH2Cl2 (50 mL) and water (50 mL, degassed)
were added, and the mixture was vigorously stirred. The organic layer
was separated and dried over MgSO4. The solvent was then removed
by vacuum. The resulting pale orange viscous oil was purified over a
silica gel column (hexanes). Yield: 1.6 g (1.5 mmol; 39%) of a colorless
(3)RhH(CO)(PPh3) (12). To a mixture of 3 (50 mg, 0.029 mmol)
and (PPh3)3RhH(CO) (26.9 mmol, 0.029 mmol) was added 1 mL of
DMSO. The resulting yellow solution was stirred at 45 °C overnight.
After evaporation of the solvent under high vacuum the yellow residue
was washed (four times) with Et2O (5 mL) to remove PPh3. The
compound was obtained as a yellow solid in quantitative yield. 31P
NMR (CD3OD; ppm): δ 23.4 (JRhP ) 147 Hz, JPP ) 125 Hz), 40.2
(JRhP ) 148 Hz, JPP ) 126 Hz).
1
oil. H NMR (C6D6; ppm): δ 7.52 (s, 2H, ArH), 7.49 (bq, 8H, ArH),
6-(p-Bromophenoxy)-1-phenyl-hexane (7). To a solution of p-
bromophenol (8.66 g, 50 mmol) and NaOH (2.5 g, 62.5 mmol) in
DMSO (40 mL) was added 1-bromo-6-phenyl-hexane (12.1 g, 50
mmol), and the mixture was heated at 100 °C for 5 h. The pale yellow
reaction mixture was cooled to room temperature. Water (100 mL) and
Et2O (100 mL) were added, and the mixture was vigorously stirred
after which the Et2O layer was separated. The water layer was washed
twice with Et2O, and the combined organic layers were dried over
MgSO4. Evaporation of the solvent yielded a pale yellow oil. Purifica-
tion by chromatography over a silica gel column with Et2O/hexane (1/
20) as the eluent gave 7 as a colorless oil. Yield: 12.0 g (35.2 mmol;
7.07-7.03 (m, 16H, ArH), 7.02-7.00 (m, 8H, ArH), 6.94 (dt, 4H, ArH),
6.69 (b, 2H, ArH), 1.64 (s, 6H, C(CH3)2), 1.23 (s, 18H, C(CH3)3). 13C-
{1H} NMR (C6D6; ppm) δ 157.8 (s, 4C), 156.9 (s, 4C), 150.4 (t, J )
9.3 Hz, 4C), 135.4 (s, 8C), 135.1 (s, 2C), 132.3 (s, 4C), 132.2 (s, 2C),
129.9 (s, 8C), 127.0 (s, 8C), 126.5 (s, 2C), 123.9 (s, 8C), 119.4 (s,
2C), 118.5 (s, 4C), 34.6 (s, 2C), 34.3 (s, 1C), 32.2 (s, 2C), 31.5 (s,
6C). 31P{1H} NMR (C6D6; ppm) δ -18.8. M.p. 58 °C. Anal.
(C71H64O5P2) calcd: C, 80.5; H, 6.1; found; C, 79.7; H, 6.2.
2,7-di-tert-Butyl-4,5-bis{di[p-(p-sulfonatophenyl)phenoxy]phos-
phino)-9,9-dimethyl-xanthene (2). Compound 9 (0.24 g, 0.23 mmol)
was cooled to - 70 °C, and 1 mL of H2SO4 (96%) was added. The
mixture was then warmed to room temperature and stirred for 5 days.
The brown mixture was carefully neutralized with aqueous NaOH (27%,
w/w) at 5 °C until a pH of 8 was reached. The resulting pale yellow
solution was evaporated to dryness which gave a white solid. MeOH
(40 mL) was added, and the suspension was refluxed for 1 h, cooled
to room temperature, and carefully decanted. The clear pale yellow
solution was evaporated to dryness which gave a pale yellow solid.
After several washings with EtOH pure 2 was obtained as a white solid.
Yield: 0.23 g (0.15 mmol; 67%). 1H NMR (CD3OD; ppm) δ 7.80 (d,
J ) 8.8 Hz, 8H, ArH), 7.50 (s, 2H, ArH), 7.24 (b, 8H, ArH), 7.04 (d,
J ) 8.8 Hz, 16H, ArH), 6.70 (b, 2H, ArH), 1.67 (s, 6H, C(CH3)2),
1.17 (s, 18H, C(CH3)3). 13C{1H} NMR (CD3OD; ppm) δ 159.8 (s, 4C),
157.9 (s, 4C), 150.3 (t, J ) 9.3 Hz, 4C), 143.4 (s, 4C), 136.3 (s, 4C),
135.1 (s, 2C), 132.1 (s, 4C), 132.0 (s, 2C), 129.7 (s, 8C), 127.0 (s,
8C), 126.5 (s, 2C), 123.7 (s, 8C), 118.3 (s, 2C), 118.0 (s, 4C), 34.5 (s,
2C), 34.3 (s, 1C), 32.3 (s, 2C), 31.4 (s, 6C). 31P{1H} NMR (CD3OD;
ppm) δ -13.3. Mass spectrometry (FAB, from a glycerol matrix yielded
the phosphine) 1489 (M + Na+). Anal. (C71H60O17P2S4Na4‚4H2O)
calcd: C, 55.4; H, 4.5; found: C, 54.9; H, 4.3.
1
72%). H NMR (CDCl3; ppm) δ 7.37 (d, J ) 8.2 Hz, 2H), 7.27 (d, J
) 8.1 Hz, 2H), 7.19 (m, 3H), 6.76 (d, J ) 8.4 Hz, 2H), 3.91 (t, J )
6.5 Hz, 2H), 2.64 (t, J ) 7.8 Hz, 2H), 1.76 (m, 2H), 1.67 (m, 2H),
1.45 (m, 4H). 13C {1H} NMR (CDCl3; ppm) δ 158.6 (s, 1C), 142.8 (s,
1C), 132.4 (s, 2C), 128.6 (s, 2C), 128.5 (s, 2C), 125.9 (s, 1C), 116.5
(s, 2C), 112.8 (s, 1C), 68.3 (s, 1C), 36.2 (s, 1C), 31.7 (s, 1C), 29.2 (s,
1C), 26.1 (s, 1C).
2,7-di-tert-Butyl-4,5-bis{di[p-(6-phenylhexyl)phenoxy]phosphino)-
9,9-dimethyl-xanthene (11). This compound was prepared as described
for 9. The resulting pale yellow oil was purified by column chroma-
tography (eluent Et2O/pentane 1:9). Compound 11 was obtained as an
1
colorless oil. H NMR (C6D6; ppm) δ 7.56 (bm, 8H, ArH), 7.51 (s,
2H, ArH), 7.19-7.11 (m, 22H, ArH), 6.85 (d, J ) 8.5 Hz, 8H, ArH),
6.65 (b, 2H, ArH), 3.61 (t, J ) 6.5 Hz, 8H, OCH2), 2.49 (t, J ) 7.6
Hz, 8H, ArCH2), 1.66 (s, 6H, C(CH3)2), 1.53 (m, 16H, CH2), 1.28 (m,
16H, CH2), 1.23 (s, 18H, C(CH3)3). 13C{1H} NMR (CDCl3; ppm) δ
159.4 (s, 4C), 150.5 (t, J ) 9.2 Hz, 2C), 145.1 (s, 4C), 141.5 (s, 4C),
135.4 (s, 8C), 132.3 (s, 2C), 129.1 (s, 2C), 128.7 (s, 2C), 128.6 (s,
8C), 128.1 (s, 4C), 128.0 (s, 2C), 126.1 (s, 8C), 122.7 (s, 2C), 114.2
(s, 8C), 67.8 (s, 4C), 34.5 (s, 1C), 34.2 (s, 2C), 32.4 (s, 4C), 32.3 (s,
2C), 31.2 (s, 6C), 30.7 (s, 4C), 28.9 (s, 8C), 26.1 (s, 8C). 31P{H} NMR
(C6D6; ppm) δ -18.3. Mass spectrometry (FAB, in a nitrobenzyl alcohol
matrix yielded the phosphine) 1397 (M + H). Anal. (C95H112O5P2)
calcd: C, 81.8; H, 8.1; found: C, 81.2; H, 8.2.
2,7-di-tert-Butyl-4,5-bis{di[p-(6-p-sulfonatophenylhexyl)phenoxy]-
phosphino)-9,9-dimethylxanthene (4). This compound was prepared
as described for 2. Compound 4 was obtained as an off-white solid.
1H NMR (CD3OD; ppm) δ 7.55 (d, J ) 8.3 Hz, 8H, ArH), 7.45 (s,
2H, ArH), 7.26 (d, J ) 8.3 Hz, 8H, ArH), 7.07 (b, 8H, ArH), 6.74 (d,
J ) 8.6 Hz, 8H, ArH), 6.59 (s, 2H, ArH), 3.30 (t, J ) 6.1 Hz, 8H,
OCH2), 2.72 (m, 8H), 1.71-1.63 (m, 32H), 1.62 (s, 6H, C(CH3)2, 1.12
(s, 18H, C(CH3)3).13C{1H} NMR (CD3OD; ppm) δ 161.9 (s, 4C), 152.3
(s, 4C), 146.7 (s, 4C), 145.4 (s, 4C), 144.0 (s, 4C), 141.2 (s, 4C), 136.3
(s, 2C), 132.1 (s, 4C), 131.6 (s, 2C), 131.4 (s, 4C), 130.8 (s, 2C), 130.1
(s, 2C), 129.8 (s, 8C), 129.5 (s, 4C), 128.9 (s, 2C), 127.5 (s, 8C), 126.9
(s, 2C), 124.2 (s, 2C), 116.2 (s, 8C), 68.6 (s, 4C), 36.3 (s, 2C), 35.7 (s,
1C), 33.3 (s, 4C), 32.6 (s, 4C), 32.3 (s, 6C), 30.8 (s, 4C), 30.4 (s, 4C),
28.5 (s, 8C). 31P{1H} NMR (C6D6; ppm) δ -15.5. Mass spectrometry
(FAB, from a glycerol matrix yielded the phosphine) 1827 (M + Na+).
Anal. (C95H108O17P2S4Na4‚4H2O) calcd: C, 60.8; H, 6.2; found: C,
60.1; H, 6.1.
2,7-di-tert-Butyl-4,5-bis{di[p-(3-phenylpropyl)phenoxy]phosphino)-
9,9-dimethyl-xanthene (10). This compound was prepared as described
1
for 9. Compound 10 was obtained as a white solid. H NMR (C6D6;
ppm) δ 7.52 (bm, 8H, ArH), 7.17-7.06 (m, 22H, ArH), 6.82 (d, J )
8.5 Hz, 8H, ArH), 6.59 (b, 2H, ArH), 3.58 (t, J ) 6.2 Hz, 8H, OCH2),
2.60 (t, J ) 7.4 Hz, 8H, ArCH2), 1.82 (m, J ) 6.2 Hz, J ) 7.4 Hz, 8H,
CCH2C), 1.67 (s, 6H, C(CH3)2), 1.24 (s, 18H, C(CH3)3). 31P{1H} NMR
(C6D6; ppm) δ - 18.3. 13C{1H} NMR (CDCl3; ppm) δ 159.2 (s, 4C),
150.4 (t, J ) 9.3 Hz, 2C), 145.1 (s, 4C), 141.7 (s, 4C), 135.4 (s, 8C),
132.6 (s, 2C), 129.2 (s, 2C), 128.9 (s, 2C), 128.6 (s, 8C), 128.2 (s,
4C), 128.1 (s, 2C), 126.1 (s, 8C), 122.7 (s, 2C), 114.4 (s, 8C), 67.0 (s,
4C), 34.9 (s, 1C), 34.6 (s, 2C), 32.3 (s, 4C), 32.3 (s, 2C), 31.5 (s, 6C),
30.9 (s, 4C). Mp 84 °C. Mass spectrometry (FAB, in a nitrobenzyl
alcohol matrix yielded the phosphine) 1228 (M + H). Anal. (C83H88O5P2)
calcd, C, 81.2; H, 7.2; found: C, 80.8; H, 7.3.
2,7-di-tert-Butyl-4,5-bis{di[p-(3-p-sulfonatophenylpropyl) phe-
noxy]phosphino)-9,9-dimethylxanthene (3). This compound was
prepared as described for 2. Compound 3 was obtained as an off-white
solid. 1H NMR (CD3OD; ppm) δ 7.72 (d, J ) 8.3 Hz, 8H, ArH), 7.44
(s, 2H, ArH), 7.28 (d, J ) 8.3 Hz, 8H, ArH), 7.04 (b, 8H, ArH), 6.81
(d, J ) 8.6 Hz, 8H, ArH), 6.55 (s, 2H, ArH), 3.95 (t, J ) 6.2 Hz, 8H,
OCH2), 2.84 (t, J ) 7.3 Hz, 8H, ArCH2), 2.08 (m, 8H, CCH2C), 1.63
(s, 6H, C(CH3)2, 1.12 (s, 18H, C(CH3)3). 13C{1H} NMR (CD3OD; ppm)
δ 161.5 (s, 4C), 152.1 (s, 4C), 146.9 (s, 4C), 145.9 (s, 4C), 144.0 (s,
4C), 141.6 (s, 4C), 136.6 (s, 2C), 132.2 (s, 4C), 131.9 (s, 2C), 131.8
(s, 4C), 130.7 (s, 2C), 130.3 (s, 2C), 130.0 (s, 8C), 129.7 (s, 4C), 128.9
(s, 2C), 127.4 (s, 8C), 126.8 (s, 2C), 124.1 (s, 2C), 115.8 (s, 8C), 68.1
(s, 4C), 36.3 (s, 2C), 35.7 (s, 1C), 33.3 (s, 4C), 32.6 (s, 4C), 32.3 (s,
6C). 31P{1H} NMR (CD3OD; ppm) δ -15.4. Mass spectroscopy (FAB,
from a glycerol matrix yielded the phosphine) 1659 (M + Na+). Anal.
(C83H84O17Na4P2S4‚4H2O) calcd, C, 58.3; H, 5.4; found: C, 58.1; H,
5.3.
Electron Microscopy. Unless noted otherwise all samples were
prepared by sonication (or just stirring) of a ligand solution of 3, 4,
and 12 in water (2‚10-3 M) at room temperature. The samples were
aged overnight at room temperature prior to use. To the solution of 12
1-octene (1 mL) was added. After sonication for 5 min. the sample
was analyzed. The platinum-shadowed samples were prepared by
bringing a drop of the solution onto a Formvar-coated copper grid.
The excess of the solution was blotted off with filter paper after 1 min,
and the sample was shadowed under an angle of 45° by evaporation of
Pt (layer thickness 2 nm). The samples were shadowed with a Edwards
306 vacuum coater. Freeze-fracture samples were prepared by bringing

Hydroformylation by Diphosphine Vesicle Formation
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1657
a drop of 0.1% (w/w) amphiphile dispersion onto a golden microscope
grid (150 mesh), placed between two copper plates, and frozen in liquid
propane. Sample holders were placed in a Balzers freeze etching system
BAF 400 D at 10-6 Torr and heated to -105 °C. After fracturing, the
samples were etched for 1 min. (∆T ) 75 °C), shaded with Pt (layer
thickness 2 nm), and covered with carbon (layer thickness 20 nm).
Replicas were allowed to warm to ambient temperature, left on 20%
chromic acid overnight, and rinsed with water. All samples were studied
with a Philips TEM 201 microscope (60 Kv).
into the reaction vessel under nitrogen. Nonane, 0.40 mL, was added
as an internal standard for gas chromatography analysis. The total
volume of the organic phase was 1.0 mL. The 1-octene/Rh ratio was
3820/1 in all catalytic runs. The reaction vessel was pressurized with
CO/H2 to 210 psi. The reaction was initiated by placing the reaction
vessel into a temperature-controlled bath preheated to the desired
reaction temperature. The temperature of the oil bath was controlled
by an Omega CN 2000 temperature process controller. The reaction
mixture was stirred with a magnetic stirring bar at 700 rpm. Catalytic
reactions were terminated by removing the vessel from the oil bath
and depressurizing after cooling in an ice-water bath. In all cases the
colorless organic layer was readily separated from the aqueous layer
after the reaction. The reaction product distribution was analyzed by
gas chromatography.
Fluorescence Spectroscopy. Emission spectra were recorded on a
SPEX Fluorolog 1680 0.22 m double spectrometer. The excitation
wavelength was set at 335 nm, and spectra were recorded from 350 to
450 nm. A saturated aqueous solution of pyrene (estimated to be <10-6
M in probe) was prepared at room temperature by stirring the probe in
degassed water overnight, followed by filtration to remove excess probe.
Pyrene saturated aqueous solutions of SDS, 2, 3, and 4 at 2 × 10-3
M
Acknowledgment. We thank Professor Dr. Harold M.
McNair for the assistance with GC measurements and Dr.
Richey M. Davis for the use of the molecular size detector
instrument. Dr. Gary Long and U. Chatreewongsin are acknowl-
edged for the use of the GFAAS. Peter J. J. A. Buijnsters and
Mr. Huub M. P. Geurts are gratefully acknowledged for their
assistance with the electron microscopy experiments and Profes-
sor Dr. R. J. M. Nolte, Dr. N.A.J.M. Sommerdijk, and Professor
Dr. J. B. F. N. Engberts for fruitful discussions. Financial support
of this work was provided by The Netherlands Institute for
Catalysis Research (NIOK) and The Netherlands Foundation
for Scientific Research (NWO). We thank DuPont Education
Foundation for support of the work at Virginia Tech.
and the emission spectra were recorded 15 min after mixing at 298,
323, 333, 343, 353, and 363 K.
Catalysis. Two-phase hydroformylation reactions of 1-octene with
rhodium complexes of 1-4 were carried out in a 30 mL capacity
stainless steel reaction vessel. The catalysts were made in situ by
mixing 0.25 mL of 0.01 M Rh(CO)2(acac) in MeOH and the indi-
cated amount of ligand. Rh(CO)2(acac) (0.25 mL 0.01 M in MeOH)
and the required amount of ligand were mixed, and the MeOH was
evaporated after which 1.5 mL of water was added to the residue. The
mixture was then transferred into the reaction vessel under nitrogen.
Catalyst formation: the reaction vessel with the catalyst solution only
was pressurized with CO/H2 to 210 psi and placed in a bath preheated
to the desired reaction temperature and stirred for the desired time,
cooled to room temperature with an ice-water bath, and slowly
depressurized. The substrate, 0.60 mL 1-octene, was then transferred
JA9925610