Synthesis of sulfated mono- and ditertiary phosphines, complex chemistry and catalysis

Article abstract of DOI:10.1139/v01-040

Cyclic and bicyclic sulfates have been prepared from commonly available alcohols. Nucleophilic cleavage of the cyclic sulfates affords a new type of water-soluble mono- and ditertiary phophines bearing -OSO₃Li groups in distinguished positions in the molecular framework. Both phosphines have amphiphilic character. Reactions of the chiral 2 and the dppp analogue 5 with [Rh(COD)Cl]₂ and Pt(PhCN)₂Cl₂ provide novel zwitterionic complexes. Rhodium complexes of 2 and 5 have been successfully applied in liquid biphasic hydroformylation of styrene and octene-1. When the rhodium complex of 5 was used as catalyst in hydroformylation of styrene, less then 4 ppm rhodium could be detected in the organic phase.

Full text of DOI:10.1139/v01-040

1040
Synthesis of sulfated mono- and ditertiary
phosphines, complex chemistry and catalysis
H. Gulyás, A. Dobó, and J. Bakos
Abstract: Cyclic and bicyclic sulfates have been prepared from commonly available alcohols. Nucleophilic cleavage of
the cyclic sulfates affords a new type of water-soluble mono- and ditertiary phophines bearing -OSO3Li groups in dis-
tinguished positions in the molecular framework. Both phosphines have amphiphilic character. Reactions of the chiral 2
and the dppp analogue 5 with [Rh(COD)Cl]2 and Pt(PhCN)2Cl2 provide novel zwitterionic complexes. Rhodium com-
plexes of 2 and 5 have been successfully applied in liquid biphasic hydroformylation of styrene and octene-1. When
the rhodium complex of 5 was used as catalyst in hydroformylation of styrene, less then 4 ppm rhodium could be de-
tected in the organic phase.
Key words: cyclic sulfates, water soluble phosphines, amphiphilic character, Rh complexes, Pt complexes,
hydroformylation.
Résumé : On a préparé des sulfates cycliques et bicycliques à partir d’alcools disponibles facilement. Le clivage nu-
cléophilique des sulfates cycliques conduit à la formation d’un nouveau type de phosphines mono- et ditertiaires, solu-
bles dans l’eau, portant des groupes -OSO3Li dans des positions définies du squelette moléculaire. Les deux types de
phosphines présentent un caractère amphiphilique. Les réactions du [Rh(COD)Cl]2 et du Ph(PhCN)2Cl2 avec composé
chiral 2 et de son analogue à la dppp, 5, conduisent à la formation de nouveaux complexes zwitterioniques. On a pu
appliquer avec succès les complexes 2 et 5 du rhodium à l’hydroformylation biphasique liquide du styrène et de l’oct-
1-ène. Quand on a utilisé le complexe de rhodium 5 comme catalyseur dans l’hydroformylation du styrène, on n’a dé-
tecté que quatre ppm de rhodium dans la phase organique.
Mots clés : sulfates cycliques, phosphines solubles dans l’eau, caractère amphiphilique, complexes du rhodium, com-
plexes du platine, hydroformylation.
[Traduit par la Rédaction] Gulyás et al. 1048
Introduction
many aqueous organometallic catalytic systems to improve
the catalytic activity and (or) selectivity of the reaction (4).
In our case the sulfate group is attached to the ligand itself,
rendering it not only water-soluble, but also amphiphilic. To
our knowledge, only a very few ligands with amphiphilic
properties have been reported (5). These ligands are ex-
pected to increase the activity of organic–aqueous biphasic
catalytic systems by improvement of the miscibility of the
two liquid phases (6). Also quite few ligands are known,
where the coordinating functional group is not affected by
the group providing the water solubility (7). In these mole-
cules the -PPh2 unit is remote enough from the -OSO3Li
function so that the electronic nature of the -PPh2 group will
be unaffected.
Much effort has been directed toward the synthesis of new
catalytic species with recent focus on water-soluble catalysts
(for recent reviews see ref. 1). Transition-metal complexes
can easily be rendered water soluble by incorporating polar
substituents (2) e.g., -SO3Na, -CO2Na, -OH, -PMe+3, -NMe3+,
-PO(ONa)2, and polyethers.
Recently we introduced a novel route to a new class of
water-soluble phosphines (3). We have reasoned that this
methodology will be of particular interest since: (i) a wide
range of starting compounds (e.g., diol, triol, tetrol) is avail-
able; (ii) the synthesis is simple; and (iii) provides an easy
access to
a series of amphiphilic or water-soluble
phosphines. Water solubility of these phosphines is provided
by the -OSO3Li group.
Salts of sulfuric acid hemiesters are known to be excellent
detergents. For example sodium dodecylsulfate is used in
Hemilabile ligands have recently come into focus. Weak
coordination of a functional group can stabilize certain con-
formations of the catalytically active complex and improve
the selectivity of a reaction (8). Another advantage of a
Received September 5, 2000. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on July 14, 2001.
Dedicated to Professor Brian James on the occasion of his 65th birthday.
H. Gulyás. Research Group for Petrochemistry, Hungarian Academy of Science, H-8201 Veszprém, P.O. Box 158, Hungary.
A. Dobó. Central Research Institute for Chemistry of the Hungarian Academy of Sciences H-1025 Budapest, Pusztaszeri u. 59-67,
Hungary.
J. Bakos.1 Department of Organic Chemistry, University of Veszprém, H-8201 Veszprém, P.O. Box 158, Hungary.
1Corresponding author (telephone: (36) 88 422022/4355; fax: (36) 88 427492; e-mail: bakos@almos.vein.hu).
Can. J. Chem. 79: 1040–1048 (2001)
© 2001 NRC Canada
DOI: 10.1139/cjc-79-5/6-1040

Gulyás et al.
1041
weakly coordinating functional group is that it can be easily
replaced by the substrate resulting in an increase in the ac-
tivity of the catalyst. A hard sulfate group is not expected to
have strong coordination ability toward soft metal centers,
but its labile coordination should not be overlooked.
We report here detailed synthetic route for the preparation
of mono- and ditertiary phosphines. Complex formation and
catalytic application of the model compounds are presented.
75 deg from the upright position, with a horizontal ampli-
tude of 3 cm. The reaction was monitored by the change in
pressure.
Leaching of rhodium was determined as follows: The to-
luene phase was separated. 0.5 mL of cH2SO4 and 1 mL of
cH2O2 were added. Further 5x1mL of cH2O2 was added dur-
ing one week. All volatiles were evaporated from the ho-
mogenous solution. 10 mL of deionised water and 0.1 g of
Na2SO4 were added. Analysis was undertaken on a Perkin
Elmer 5100 PC atom absorption spectrometer in conjunction
with a series of rhodium standards.
Experimental
General information
Synthesis of the phosphines and their complexes were car-
ried out under a purified argon atmosphere using standard
Schlenk-type techniques. Tetrahydrofuran (THF), diethyl
ether (Et2O), toluene, and benzene were distilled from so-
dium benzophenone ketyl under argon. Methylene chloride
(CH2Cl2) and pyridine were distilled from CaH2 and metha-
nol (MeOH) was distilled from Mg(OMe)2 under argon. Wa-
ter (H2O) was distilled twice under argon. Deuterated
methanol and water (CD3OD and D2O) were deoxygenated
by bubbling with argon for 10 min. Deuterated chloroform
(CDCl3) was deoxygenated by distillation under argon.
Pentaerythritol was purchased from Aldrich and used as re-
ceived. (2R,4R)-2,4-Pentanediol was prepared by the slightly
modified literature method (9). GC analyses were performed
using a Hewlett–Packard 5830A gas chromatograph (SPB-1
30 m column, film thickness (0.1 µm), N2 carrier gas
(2 mL min–1)). Optical rotations were obtained using a
Schmidt Haensch 21245 polarimeter. Nuclear magnetic reso-
nance spectra were recorded on a Varian Unity 300 and
Bruker DRX-500 spectrometer using CDCl3, CD3COCD3,
Cyclic sulfates
Cyclic sulfate 1 was prepared using the general literature
method described by Gao and Sharpless (10).
Synthesis of (4R,6R)-4,6-dimethyl-2,2-dioxide-1,3,2-
dioxathiane (1)
Thionyl chloride (4.6 mL, 0.064 mol) was added, via a
dropping funnel, to (2R,4R)-2,4-pentanediol (5.5 g, 0.053 mol)
in 50 mL CCl4. The resulting light brown solution was
refluxed for 1 h to complete the sulfite formation. HCl was
removed under reduced pressure (the water aspirator was ad-
justed to a quite gentle flow). CH3CN (50 mL), NaIO4 (16 g,
0.075 mol), and RuCl3 hydrate (0.008 g, 0.032 mmol, 40%
ruthenium) were added to the solution. The resulting brown
mixture was cooled to 0°C and 75 mL of water was added
slowly. The mixture turned orange within 5 min. The reac-
tion was allowed to stir at ambient temperature for 1 h. The
mixture was then diluted with Et2O (400 mL). The organic
phase was separated, washed with water (20 mL), saturated
NaHCO3 solution (2 × 20 mL), and brine (20 mL). After be-
ing dried over MgSO4 the solution was filtered through a
pad of SiO2 to remove the ruthenium. The solvent was re-
moved in vacuum. Recrystallization of the product from
Et2O–hexane yielded a white crystalline solid (8.85 g, 82%),
1
D2O, and CD3OD as solvents. The H and 13C NMR chemi-
cal shifts are referenced to internal or external standard
SiMe4, and to residual portio-solvent peaks (i.e., CHCl3).
The 31P NMR spectra were recorded with 85% H3PO4 as an
external standard. The 195PtNMR (64 MHz) chemical shift is
given in ppm relative to Na2PtCl6 as an external standard.
Positive chemical shifts are downfield from the standard.
Adjustments of pH were performed using an OP-211/1 Lab-
oratory Digital pH Meter. Rhodium content of the organic
1
mp 34 to 35°C. [α]2D4 = 37.5° (c 5.02, CHCl3). H NMR
(300.15 MHz, CDCl3) δ: 1.64 (d, JHCCH = 6.6 Hz, 6H, CH3),
2.03 (t, JHCCH = 5.6 Hz, 2H, CH2), 5.03 ppm (pseudo-sext
due to signal overlap, 2H, CH). 13C NMR (75.42 MHz,
CDCl3) δ (ppm): 20.0 (s, CH3), 35.5 (s, CH2), 80.3 (s, CH).
phase was determined by atom absorption using
a
PerkinElmer 5100 PC atom absorption spectrometer,
electrothermal atomization HGA 600, AS 70 autosampler,
and GEM software. Fast atom bombardment (FAB) mass
spectrometric measurements were performed using a VG-
ZAB-2SEQ reverse geometry instrument operating at an ac-
celerating voltage of 8 kV. A cesium ion gun producing
30 keV ions was utilized for ion bombardment. Glycerol was
used as the FAB matrix.
Synthesis of 2,4,8,10-tetraoxa-3,9-
dithiaspiro[5.5]undecane 3,9-dioxide (3)
A solution of SOCl2 (17.5 mL, 0.24 mol) in pyridine
(20 mL) was added dropwise to a solution of pentaerythritol
(13.6 g, 0.1 mol) in pyridine (250 mL). The reaction mixture
was stirred for 3 h. The solvent, remaining SOCl2, and most
of the pyridine–hydrochloride were removed under reduced
pressure. The brownish-white solid residue was washed with
water, NaHCO3 solution, water again, and with diethyl ether.
After drying, the product was obtained as a white,
microcrystalline solid (19.3 g, 85%), mp 152°C. H NMR
(500.13 MHz, CD3COCD3) δ: 3.929 (d, JHCH = 12.0 Hz, 1H,
CH2), 3.934 (d, JHCH = 12.0 Hz, 1H, CH2), 4.406 (d, JHCH
12.1 Hz, 1H, CH2), 4.411 (d, JHCH = 12.1 Hz, 1H, CH2, 4.61
(d, JHCH = 12.1 Hz, 2H, CH2), 4.94 ppm (d, JHCH = 12.3 Hz,
2H, CH2). 13C NMR (75.42 MHz, CD3COCD3) δ (ppm):
35.68 (s, C), 59.95 (s, CH2), 59.99 (s, CH2).
Catalysis
Two-phase hydroformylation reactions of styrene and 1-
octene with rhodium complexes of 2 and 5 were carried out
in a 20 mL stainless steel autoclave. The catalysts were
made in situ. Mixing the water, toluene, alkene, phosphine,
and [Rh(COD)Cl]2 resulted in an orange emulsion which
was transferred into the autoclave under argon atmosphere.
The autoclave was pressurized with CO–H2 and then placed
in a heating mantle preheated to the desired reaction temper-
ature. The autoclave was shaken at a frequency of 180 min–1,
1
=
© 2001 NRC Canada

1042
Can. J. Chem. Vol. 79, 2001
Scheme 1.
Synthesis of 2,4,8,10-tetraoxa-3,9-
dithiaspiro[5.5]undecane 3,3,9,9-tetraoxide (4)
NaIO4 (54.3 g, 0.25 mol) and RuCl3 hydrate (27.2 mg,
0.107 mmol, 40% ruthenium) were added to the bicyclic sul-
fite 3 (19.3 g, 0.085 mol) in 400 mL of CH3CN. Water
(240 mL) was added slowly to the brown suspension. The
mixture turned orange and a considerable amount of white
solid precipitated. The reaction was stirred further for 1.5 h.
The white solid was filtered and washed exhaustively with
water. The white residue was dried at ambient temperature
and extracted with hot acetone. Removal of the solvent un-
der reduced pressure yielded a white, microcrystalline solid
(17.6 g, 68% calculated for pentaerythritol), mp 280–282°C.
1H NMR (300.15 MHz, CD3COCD3) δ (ppm): 5.01 (s, 8H,
CH2). 13C NMR (75.42 MHz, CD3COCD3) δ (ppm): 34.30
(s, C), 74.4 (s, CH2).
1 equiv of LiPPh2
THF
O
S
O
O
O
Ph2P
OSO3Li
2
1
to give a white solid foam (5.42 g, 73%). FAB-MS–: 637 (M
– Li+). H NMR (500.13 MHz, D2O) δ (ppm): 2.28 (s, 4H,
1
CH2), 3.89 (s, 4H, CH2), 6.89 (m, 8H, Ar), 6.95 (t, JHCCH
=
7.1 Hz, 4H, Ar), 7.09 (m, 8H, Ar). 31P NMR (202.45 MHz,
D2O) δ (ppm): –28.9. 13C NMR (125.77 MHz, D2O) δ
(ppm): 33.05 (m, CH2), 42.20 (t, JPCC = 12.1 Hz, C), 71.62
(m, CH2), 128.98 (d, JPCCC = 3.8 Hz, Ar), 129.29 (s, Ar),
133.15 (d, JPCC = 21.4 Hz, Ar), 138.35 (d, JPC = 7.5 Hz, Ar).
Synthesis of (2R,4S)-4-(diphenylphosphino)-pent-2-yl
sulfate lithium salt (2)-butyl sulfate lithium salt
(4R,6R)-4,6-Dimethyl-2,2-dioxide-1,3,2-dioxathiane (1)
(2.0 g, 0.012 mol) in THF (15 mL) was added to
LiPPh2·dioxane adduct (4.4 g, 0.016 mol) in THF (50 mL) at
0°C, meanwhile the bright red color turned light brown. The
mixture was stirred at ambient temperature for 1 h. The THF
was removed under reduced pressure, then deoxygenated
water (40 mL) and Et2O (20 mL) were added. The aqueous
phase was separated and washed with Et2O (20 mL) again.
The pH of the solution was adjusted to 7.0 using diluted
H2SO4 solution. The water was removed under reduced
pressure, then CH2Cl2 (20 mL) was added. Li2SO4 was re-
moved by filtration. Evaporation of the solvent under vac-
uum yielded a white, solid foam (3.1 g, 72%). FAB-MS–:
Results and discussion
Preparation of cyclic sulfates
Cyclic sulfate of (2R,4R)-2,4-pentanediol (1) (Scheme 1)
was obtained by a simple one pot reaction following the pro-
cedure developed for 1,2-cyclic sulfates by Gao and
Sharpless (10). The synthesis involves the treatment of diols
with thionyl chloride followed by ruthenium-catalyzed oxi-
dation. Sulfate 1 could be obtained in this way with high
yield (82%).
1
351 (M – Li+). H NMR (300.15 MHz, CDCl3) δ (ppm):
The bicyclic sulfate of pentaerythritol (4) could not be
prepared by the general literature method. Pentaerythritol
and SOCl2 react with negligible reaction rate in CCl4, which
is the recommended solvent for the sulfite formation. Obvi-
ously, this fact is due to the poor solubility of the tetrol in
CCl4. Although, upon addition of dioxane to the CCl4, the
bicyclic sulfate could be isolated after the oxidation step, the
yield was still very poor (17%).
0.84 (dd, JPCCH = 15.1 Hz, JHCCH = 6.6 Hz, 3H, CH3), 1.01
(d, JHCCH = 6.0 Hz, 3H, CH3), 1.02 (overlapped by the signal
of CH3, 1H, diastereotopic CH2), 1.57 (br m, 1H,
diastereotopic CH2), 2.55 (br s, 1H, CH-PPh2), 4.49 (br m,
1H, CH-OSO3Li), 6.91 (br m, 6H, Ar), 7.23 (br m, 4H, Ar).
31P NMR (121.42 MHz, CDCl3) δ (ppm): 1.13. 13C NMR
(125.77 MHz, D2O) δ: 17.99 (d, JPCC = 14.7 Hz, CH3-CH-
PPh2), 23.40 (s, CH3-CH-OSO3Li), 27.37 (d, JPC = 6.9 Hz,
CH3-CH-PPh2), 42.63 (d, JPCC = 15.8 Hz, CH3-CH-
OSO3Li), 77.71 (d, JPCCC = 12.3 Hz, CH3-CH-OSO3Li),
131.03 (s, Ar), 131.50 (d, JPCC = 6.3 Hz, Ar), 135.99 (d,
JPCC = 17.6 Hz, Ar), 138.41 (d, JPCC = 32.7 Hz, Ar).
To gain the product in a satisfying yield, a different proce-
dure has been developed for the synthesis of 4 (Scheme 2).
(i) The formation of sulfite and the oxidation were carried
out in consecutive manner by the isolation of the intermedi-
ate sulfite 3. Reaction of pentaerythritol and SOCl2 in
pyridine provided the corresponding bicyclic sulfite 3 with
good yield (85%). Even though this spiro-compound is a
known molecule of great practical importance (11), its
stereochemistry has not been discussed in the literature in
details. In a double chair conformation this molecule should
Preparation of 1,3-bis(diphenylphosphino)-2,2-
bis(lithiumsulfatemethyl)-propane (5)
2,4,8,10-Tetraoxa-3,9-dithiaspiro[5.5]undecane
3,3,9,9-
tetraoxide (4) (3.0 g, 0.012 mol) in THF (20 mL) was added
to LiPPh2·dioxane adduct (8.4 g, 0.035 mol) in THF (60 mL)
at 0°C, meanwhile the red color of the reaction mixture
faded. The mixture was stirred at ambient temperature for
1 h. The THF was removed under reduced pressure, then
deoxygenated water (40 mL) and Et2O (20 mL) were added.
After separation of the phases, the aqueous phase was
washed with Et2O (20 mL) again. The pH of the solution
was adjusted to 7.0 using a diluted H2SO4 solution. The wa-
ter was removed under reduced pressure, then CH2Cl2
(20 mL) and CH3OH (6 mL) were added. Li2SO4 was re-
moved by filtration. The solvent was evaporated in vacuum
1
have a C2 axis as the only symmetry element, but its H
NMR spectrum exhibits a more complex picture. Details of
the stereochemical investigation of the bicyclic sulfite will
be presented in a subsequent publication. (ii) Racemic
bicyclic sulfite was dissolved in CH3CN. NaIO4 and a cata-
lytic amount of RuCl3·3H2O were added. Slow addition of
water to this mixture initiated the oxidation. The crude prod-
uct was precipitated from the reaction mixture spontane-
ously. The modified procedure afforded the bicyclic sulfate
ester 4 in a 68% overall yield after purification.
© 2001 NRC Canada

Gulyás et al.
1043
Scheme 2.
LiO3SO
OSO3Li
O
S
O
O
O
O
O
O
O
O
O
O
O
NaIO , RuCl3
2 equiv of LiPPh2
THF
4
S
S
S
CH3CN-H2O
O
O
Ph2P
PPh2
3
4
5
Preparation, stability, and characterization of the novel
phosphines
Coordination chemistry of the ligands with rhodium
The monotertiary phosphine was reacted with
2
The monotertiary phosphine 2 has been prepared by the
nucleophile cleavage of the cyclic sulfate of (2R,4R)-2,4-
pentanediol 1 with LiPPh2 in THF (Scheme 1). Nucleophilic
attack occurred with complete inversion at the stereogenic
center.
[Rh(COD)Cl]2 at a 2:1 phosphorus:Rh ratio in CD3OD and
D2O (Scheme 3). In both solvents two complexes were
formed and some of the ligand remained uncoordinated. The
31P NMR spectrum (CD3OD) exhibited resonances at
36.8 ppm (d, JRh-P = 147 Hz) and 30.4 ppm (d, JRh-P
=
The bicyclic sulfate ester 4 reacts smoothly with 2 equiva-
lents of LiPPh2 to give the new water-soluble “sulfated”
phosphine 5, which has two hydrophilic tails attached to the
bridgehead carbon atom (Scheme 2). It is interesting to note
that the ditertiary phosphine 5 is a structural analogue of the
well-known ligands dppp (1,3-(diphenylphosphino)propane)
and dpppts (tetrasulfonated dppp). Both dppp and dpppts
have important catalytic applications (12).
147 Hz). The coordination shifts (∆δ = δcomplex – δfree ligand
)
are 29.5 and 35.9 ppm, respectively. Relative intensity of the
two doublets (signal at 30.4 ppm compared to the signal at
36.8 ppm) was 0.3. The broad signals were indicative of a
dynamic equilibrium between the complexes and the free
ligand, which was more significant in D2O than in CD3OD.
For example doublet character of the signal at 30.4 ppm was
not perceptible in D2O.
Although it is known that an ionic sulfate group can act as
a leaving group (13), it was clear that its anionic nature ren-
ders it kinetically much less reactive than a ROSO2OR′. The
negative charge decreases the polarity of C—O bond, and
the electrostatic repulsion between the “leaving group” and
an anionic attacking group also increases the activation en-
ergy of the substitution. Accordingly, no replacement of the
-OSO3Li was experienced in THF at ambient temperature,
even when an excess of LiPPh2 was used. The sulfated
phosphines proved to be stable in hot, neutral, or slightly ba-
sic aqueous solution also. For example, when the aqueous
solution of 5 was heated at 80°C for 1 h, or the aqueous so-
lution of 2 was heated at 90°C, pH = 10 for several hours no
appreciable amount of decomposition or hydrolysis could be
detected.
NMR and FAB-MS data are in agreement with the for-
mula in the case of both phosphines. Both phosphines fea-
ture amphiphilic character. For example, at ambient
temperature solubility of 2 is 490 g L–1 in H2O and 290 g L–1
in CH2Cl2. For a comparison, the water solubility of TPPMS
(monosulfonated triphenylphosphine) having about the same
size and one sulfonate group is 12 g L–1 (14). Ditertiary
phosphine 5 with two sulfated tails attached to the bridge-
head carbon atom has a great water solubility and is also sol-
uble in a number of organic solvents (alcohols, THF).
We presumed that [Rh(COD)Cl]2 was cleaved by 2 equiv
of the phosphine to give complex I, but the substitution of
chloride leading to complex II did not take place completely.
To prove our hypothesis, the experiment was repeated in
CD3OD saturated with KCl. Under these conditions, chlo-
ride substitution didn’t take place at all, which resulted in
the disappearance of the doublet at 30.4 ppm and an increase
in the intensity of the signal of the free ligand.
When ligand 2 was reacted with [Rh(COD)Cl]2 at a 1:1 or
lower phosphorus:Rh ratio in CD3OD the doublet at
30.4 ppm was absent from the spectrum again. When phos-
phorus:Rh ratio was increased to 4, the relative intensity of
the two signals (30.4 vs. 36.8 ppm) was found to increase to
0.66, and a new species appeared at about 50 ppm. All sig-
nals were broad probably as a result of an exchange between
the free and coordinated phosphine. Doublet character of the
signal at 36.8 ppm and multiplicity of the new resonance
were not observable. Finally, [Rh(COD)Cl]2 was treated
with AgNO3 in CD3OD to obtain [Rh(COD)(CD3OD)2]NO3.
This complex, with two weakly coordinating ligands, was
reacted with 2 equiv of the phosphine. Now the spectrum
displayed only a doublet at 30.4 ppm.
These results clearly indicate that the doublet at 36.8 ppm
belongs to the complex I, and the other doublet at 30.4 ppm
can be assigned to the complex II. Chemical shifts are con-
sistent with the proposed structures. Since chloride is ex-
pected to be a weaker σ-donor and a stronger π-acceptor
ligand than phosphine 2, electron density of the central atom
should be larger in II than in I resulting a more downfield
resonance for complex I.
When chelating phosphine 5 was reacted in a 2:1 ratio
with [Rh(COD)Cl]2 in CD3OD or D2O–EtOAc, complex III
formed as the only species (Scheme 4). 31P NMR data of III
(δ 13.9 ppm (d, JRhP = 141 Hz), in CD3OD or D2O) do not
significantly differ from that of the analogue complex of
Characterization of Rh(I) and Pt(II) complexes
Rh(I) and Pt(II) complexes are often applied catalysts in
homogenous catalysis (15). To study the coordination fea-
tures of the new phosphines, reactions of 2 and 5 with
[Rh(COD)Cl]2 (COD = 1,5-cyclooctadiene) and Pt(PhCN)2Cl2
were carried out. Reactions were monitored by 31PNMR
spectroscopy since the coordination shifts and metal–phos-
phorus coupling constants are diagnostic features for the
identification of the products.
© 2001 NRC Canada

1044
Can. J. Chem. Vol. 79, 2001
Scheme 3.
LiO3SO
PPh2
Rh
+
Ph2P
OSO3Li
Cl
I
4 equivof 2
CD3OD
-LiCl
[Rh(COD)Cl]2
LiO3SO
-O3SO
PPh2
Rh +
PPh2
II
Scheme 4.
-
-
LiO3SO
Ph2P
OSO3
LiO3SO
OSO3
2 equivof 5
CD3OD or D2O
1 atm H2, RT
CD3OD, slow
[
Rh(COD)Cl]2
PPh2
Ph2P
PPh2
Rh+
Rh+
OCD3
D
D3CO
D
IV
III
dppp (16). MS-FAB+ and MS-FAB– spectra of the complex
showed parent ions [M + Li+] (m/z = 855) and [M – Li+]
(m/z = 848). A zwitterionic structure for the complex is sug-
of the complex is characteristic of a square-planar, dsp2 hy-
brid PtP2X2 species (δ 27.7 (s, JPt-P = 2544 Hz) in CD3OD).
The value of JPt-P clearly indicates the presence of two phos-
phine ligands trans to each other (Scheme 5).
–
gested in which one of the -OSO3 groups of the ligand acts
as a counterion for the rhodium.
Compound 5 forms three complexes when reacted in a 1:1
ratio with Pt(PhCN)2Cl2 in a mixture of CH2Cl2–D2O.
Complex VI (δ –1.8 ppm (s, JPt-P = 3467 Hz), in D2O), VII
(δ –3.0 ppm (s, JPt-P = 2215 Hz), in D2O), and VIII (δ –6.8 ppm
(s, JPt-P = 3438 Hz), in D2O) are generated in a 1:0.81:0.73
ratio (Scheme 6). 31P NMR data clearly indicated that both
of the phosphorus on the chelating ligand are coordinated
and that they are equivalent (i.e., the other two coordination
sites should be occupied by the same species in each com-
plex). The coupling values of JPt-P of VI and VIII suggested
chloride or solvent coordination trans to the phosphorus,
whereas JPt-P of VII is characteristic of Pt(II) phosphine
complexes with phosphorus atoms trans to each other.
Complex VII could be obtained as the only product from
the reaction mixture when the experiment was carried out in
When H2 was bubbled into the solution of III, a minor
amount (13% based on 31P NMR) of complex IV arose (δ
37.5 ppm (d, JRhP = 187 Hz) in CD3OD). The relative resis-
tance of III against oxidative addition of H2 is not surpris-
ing. Analogous complexes behave in a similar way as
described in the literature (16). Complexes containing NBD
(NBD = 2,5-bicyclo[2.2.1]heptadiene) instead of COD tends
to react with hydrogen without difficulty, since NBD is eas-
ily saturated under a hydrogen atmosphere (17).
Coordination chemistry of the ligands with Pt(II)
When amphiphilic phosphine 2 was reacted in a 2:1 ratio
with Pt(PhCN)2Cl2 in CH2Cl2 or in a mixture of CH2Cl2 and
MeOH, complex V formed exclusively. 31P NMR spectrum
© 2001 NRC Canada

Gulyás et al.
1045
Scheme 5.
LiO3SO
PPh2
Pt
Cl
2 equiv of 2
Pt(PhCN)2Cl2
CH2Cl2-CD3OD
Cl
Ph2P
OSO3Li
V
Scheme 6.
-O3SO
-
OSO3
a, 1 equiv of 5
b, 2 equiv of AgNO3
1 equiv of 5
CH2Cl2-H2O
VI + VII + VIII
Pt(PhCN)2Cl2
a, CH2Cl2-CH3OH
b, D2O
Ph2P
D2O
PPh2
Pt+2
1 equiv of 5
CH2Cl2-CH3OH
1 or 2 equiv of 5
CH3OH
OD2
VIII
LiO3SO
OSO3Li
Ph2
P
Ph2
P
-
LiO3SO
-O3SO
OSO3
Pt +2
P
Ph2
P
Ph2
OSO3Li
Ph2P
Cl
PPh2
Cl
Pt
VI
VII
CD3OD at a 1:1 or 1:2 ligand:metal ratio. Some of the start-
ing complex remained unreacted when a 1:1 ligand to plati-
num ratio was used. Based on the 31P NMR data and the
experimental evidence, either a dimeric or a monomeric
square-planar PtP4 structure could be imagined for complex
VII. The 195Pt{1H}-NMR spectrum clearly showed the pres-
ence of four equivalent phosphorus atoms coordinating to
the platinum (δ –5011 ppm (q, JPt-P = 2215 Hz), in D2O)
(Fig. 1). This fact confirms the square-planar, zwitterionic
one complex was formed which was assumed to be complex
VI based on 31P NMR data (CD3OD, δ –1.5 ppm (s, JPt-P
3445 Hz), in CD3OD). The slight difference in the spectro-
scopic data is probably due to the different nature of the sol-
vent.
Dichloro-complex VI could be transformed into the aqua
complex VIII by treatment with AgNO3 in H2O (D2O) or
with CF3SO3Ag in a THF–H2O (D2O) solvent mixture. A
negligible amount of VII was also formed in both cases.
Since the NMR data of the complex VIII are independent of
the applied silver salt, strong coordination of NO–3 or CF3SO3−
can be excluded. The broad resonances of complex VIII
were indicative of an exchange of H2O molecules in the co-
ordination sphere.
=
–
PtP4 structure, in which two OSO3 groups act as counter
ions for the central atom. Further evidence of the structure
was obtained by FAB-MS spectroscopy. The parent ion
[M – Li+] was detected and the measured isotopic distribu-
tion showed nice agreement with the simulated one (Fig. 2).
Note that the formation of VII as a single product even at a
1:1 ligand-to-metal ratio under the conditions above is prob-
ably due to the poor solubility of the precursor in MeOH and
as a consequence an excess of the ligand was present in the
solution.
To inhibit the substitution of chlorides either by a second
phosphine ligand or solvent, the ligand was dissolved in
MeOH and was added to a CH2Cl2 solution of the precursor
[Pt(PhCN)2Cl2] in a 1:1 ratio (Scheme 6). In this case only
Hydroformylation of styrene and octene-1
Rhodium complexes of
2 and 5 were used for
hydroformylation of styrene and octene-1 (Scheme 7). Both
systems proved to be suitable catalysts for hydroformylation
(Table 1).
Ligands 2 and 5 gave complete chemoselectivity toward
hydroformylation of styrene: no hydrogenation of the sub-
strate was observed.
© 2001 NRC Canada

1046
Can. J. Chem. Vol. 79, 2001
Fig. 1. 195Pt NMR spectrum of the complex VII.
- 4920
- 4930
- 4940
- 4950
- 4960
- 4970
- 4980
- 4990
- 5000
- 5010
- 5020
- 5030
- 5040
- 5050
- 5060
- 5070
- 5080
- 5090
Scheme 7.
CHO
CO/H2 (1:1)
Rh-complexes
*
CHO
CHO
*
CHO
CO/H2 (1:1)
Rh-complexes
( )
n
( )
m
m = 0,1,2
n + m = 2
Up to 90% regioselectivity towards the more valuable
branched aldehyde was observed when 2 was used as the
ligand. At the same time, though, no optical induction was
detected. It is not surprising, since optically active mono-
dentate ligands usually provide very poor enantioselectivity
even in relatively apolar medium. Additionally, the
enantiomeric excess of the chiral products obtained in aque-
ous media is generally much lower than in organic solvents
(17). An explanation for the lower enantiomeric excess
could be the partial racemization via an enol intermediate,
which is known to proceed swiftly in protic media (18). The
evaluation of the suitability of the chiral ligand 2 for other
asymmetric catalysis is in progress.
finding was that after separation of the phases, less than
4 ppm of the rhodium with respect to the amount of the rho-
dium catalyst employed was detected in the colorless or-
ganic phase. The aqueous phase could be recycled. The
decrease in activity is probably due to the partial oxidation of
the phosphine and a small loss of the catalyst during phase
separation. In the absence of organic solvent, the regio-
selectivity toward the branched aldehyde increased to 86%.
No hydrogenation side product was observed in the
hydroformylation of octene-1 either, although both ligands
allow a moderate isomerization of the substrate (Table 2).
Both ligands provide rather low selectivity under the applied
conditions toward the more valuable normal aldehyde. In-
creasing the ligand 2:rhodium ratio from 2 to 4 led to a
slight increase in selectivity to n-aldehyde, but the activity
The rhodium complex of 5 gives a lower regioselectivity
(81%) in a toluene–H2O two-phase system. An important
© 2001 NRC Canada

Gulyás et al.
Table 1. Hydroformylation of styrene.
1047
Conv.
(%)
Selectivity b/n
(%)
Ligand
Subst./Rh
Solvent
T (°C)
t (h)
2a
250
300
300
300
300
PhMe–H2O
PhMe–H2O
PhMe–H2O
PhMe–H2O
H2O
60
80
80
80
80
3
4
12
12
10
100
100
95
84
87
90/10
81/19
81/19
79/21
86/14
5b, c
5d
5e
5f
Note: Conditions: catalyst: [Rh(COD)Cl]2 + ligand; initial pressure: 70 bar (CO:H2 = 1:1); P/Rh = 2.2; water (3mL);
toluene (3 mL) (if used).
a[Rh] = 1.0 × 10–2 M.
b[Rh] = 2.2 × 10–2 M.
cLess than 4 ppm rhodium could be detected in the organic phase.
dFirst recycling of the aqueous phase.
eSecond recycling of the aqueous phase.
f[Rh] = 2.9 × 10–2 M.
Table 2. Hydroformylation of octene-1.
Select. (%)
isoma
Ligand
Subst./Rh
P/Rh
Solvent
T (°C)
t (h)
Conv. (%)
n-ald
b-aldb
n/b
2c
2c
2c
5d
5d
250
250
250
300
300
2.05
2.20
4.00
2.40
2.40
H2O
60
60
60
80
80
6
6
8
5
17
100
100
79
99
85
2.6
2.8
7.1
5.8
16.0
53.7
53.4
59.1
46.7
53.5
43.7 (4.7)
43.8 (5.3)
33.8 (-)
47.5 (11.5)
30.5(0.3)
1.23
1.21
1.75
0.98
1.75
PhMe–H2O
PhMe–H2O
PhMe–H2O
H2O
Note: Conditions: catalyst: [Rh(COD)Cl]2 + ligand; initial pressure: 70 bar (CO:H2 = 1:1), 3.5 mL of water, 3.5 mL of toluene (when used).
aPercentage of 2-, 3-, 4-octenes formed.
bPercentage of 2-ethylheptanal and 2-propylhexanal formed.
c[Rh] = 1.5 × 10–2 M.
d[Rh] = 2.9 × 10–2 M.
Reaction of 2 with [Rh(COD)Cl]2 resulted in an equilibrium
of Rh(COD)Cl(2) and zwitterionic Rh(COD)(2)2 complexes,
meanwhile 5 gave the zwitterionic Rh(COD)(5) complex as
the only product. Treatment of 2 with Pt(PhCN)2Cl2 gave
the square-planar trans-PtCl2(2)2 complex. Reaction of 5
with Pt(PhCN)2Cl2 in a D2O–CH2Cl2 two-phase system pro-
vided three complexes, which could be identified as
PtCl2(5), and zwitterionic complexes Pt(5)2 and Pt(D2O)2(5).
Rhodium complexes of 2 and 5 could be successfully ap-
plied as catalysts for hydroformylation of styrene and
octene-1. The use of 2 in the hydroformylation of styrene re-
sulted in high regioselectivity toward the branched aldehyde
(up to 90%), however, no optical induction was observed.
When the rhodium complex of 5 was applied as a catalyst in
the biphasic hydroformylation of styrene, only 4 ppm of rho-
dium could be detected in the organic phase.
Fig. 2. Measured and simulated isotopic distribution of the par-
ent ion of the complex VII (FAB-MS–:M – Li+).
The application of these sulfated phosphines in other
biphasic and aqueous transition-metal catalyzed reactions is
being studied.
decreased. The complex modified with 5 provided the high-
est normal:branched ratio in a substrate–water two-phase
system, but a considerable amount of isomerized octenes
was formed.
Acknowledgements
Financial support from Hungarian National Science Foun-
dation (OTKA T 032 004) is gratefully acknowledged. We
are grateful to Mr. Béla Édes for skillful assistance in the
synthetic and catalytic experiments.
Conclusions
We have discovered a versatile route for the synthesis of
water-soluble mono- and ditertiary phosphines. The starting
materials are readily available, and the transformations
highly selective. The synthetic route is sufficiently flexibile
to allow fine tuning of the structure of water-soluble ligands.
References
1. Recent reviews see: (a) F. Joó and Á. Kathó. J. Mol. Catal. A:
Chem. 116, 3 (1997); (b) B. Cornils, W.A. Herrmann, and
© 2001 NRC Canada

1048
R.W. Eckl. J. Mol. Catal. A: Chem. 116, 27 (1997); (c) I.T.
Horváth and F. Joó. Aqueous organometallic chemistry and
catalysis. Kluwer Academic Publishers, Dordrecht. 1995.
2. (a) W.A. Herrmann and C.W. Kohlpainter. Angew. Chem. Int.
Ed. Engl. 32, 1524 (1993); (b) P. Kalck and F. Monteil. Adv.
Organometal. Chem. 34, 219 (1992); (c) M. Barton and J.M.
Atwood. J. Coord. Chem. 24, 43 (1991); (d) M.J. Russel. Plati-
num Met. Rev. 32, 179 (1988).
3. H. Gulyás, P. Árva, and J. Bakos. Chem. Commun. 2385 (1997).
4. (a) G. Oehme, E. Paetzold, and R. Selke. J. Mol. Catal. 71, L1
(1992); (b) A. Kumar, G. Oehme, J.P. Roque, M. Schwarze,
and R. Selke. Angew. Chem. 106, 2272 (1994); (c) S.
Trinkhaus, R. Kadyrov, R. Selke, J. Holz, L. Götze, and A.
Börner. J. Mol. Catal. A: Chem. 144, 15 (1999), and
refs. therein.
5. (a) H. Ding, B.E. Hanson, T. Bartik, and B. Bartik.
Organometallics, 13, 3761 (1994); (b) D.J. Bauer, J. Fischer,
S. Kuchen, K.P. Langhans, O. Stelzer, and N. Werferling. Z.
Naturforsch. B: Chem. Sci. 49, 1511 (1994); (c) T. Okano, N.
Harada, and J. Kiji. Chem. Lett. 6, 1057 (1994); (d) E. Valls, J.
Suades, B. Donadieu, and R. Mathieu. Chem. Commun. 771
(1996); (e) M.S. Goedheijt, B.E. Hanson, J.N.H. Reek, P.C.J.
Kamer, and P.W.N.M. van Leeuwen. J. Am. Chem. Soc. 122,
1650 (2000).
Can. J. Chem. Vol. 79, 2001
9. (a) K. Ito, T. Harada, A. Tai, and Y. Izumi. Chem. Lett. 1049
(979); (b) K. Ito, T. Harada, A. Tai. Bull. Chem. Soc. Jpn. 53,
3367 (1980).
10. Y. Gao and K.B. Sharpless. J. Am. Chem. Soc. 110, 7538
(1988).
11. (a) A. Guenther, T. Koenig, W.D. Habicher, and K.
Schwetlick. Polym. Degrad. Stab. 55, 209 (1997); (b) Polym.
Degrad. Stab. 60, 385 (1998); Polym. Degrad. Stab. 64, 115
(1999).
12. (a) E. Drent, J.A.M. van Broekhoven and M.J. Doyle. J.
Organomet. Chem. 417, 235 (1991); (b) E. Drent. Pat. Appl.
EP 121 965 (1984) (to Shell); (c) G. Verspui, G. Papadogianakis,
and R.A. Sheldon. Chem. Commun. 401 (1998).
13. M.J. Burk, A. Pizzano, J.A. Martin, L.M. Liable-Sands, and
A.L. Rheingold. Organometallics, 19, 250 (2000), and
refs. therein.
14. B. Salvesen and J. Bjerrum. Acta Chem. Scand. 16, 735 (1962).
15. B. Cornils and W.A. Herrmann. Applied homogeneous cataly-
sis with organometallic compounds. Wiley–VCH, Weinheim.
1996.
16. (a) D.A. Slack, J. Greveling, and M.C. Baird. Inorg. Chem. 18,
3125 (1979); (b) D. Heller, S. Borns, W. Baumann, and R.
Selke. Chem. Ber. 129, 85 (1996).
17. (a) I. Tóth, I. Guo, and B.E. Hanson. J. Mol. Catal. A: Chem.
116, 217 (1997); (b) R.W. Eckl, T. Priermeier, and W.A.
Herrmann. J. Organometal. Chem. 532, 243 (1997); (c) M.D.
Miquel-Serrano, A.M. Masdeu-Bulto, C. Claver, and D. Sinou.
J. Mol. Catal. A: Chem. 143, 49 (1999).
6. B. Fell and G. Papandogianakis. J. Mol. Catal. 66, 143 (1991).
7. (a) H. Ding, B.E. Hanson, and J. Bakos. Angew. Chem. Int.
Ed. Engl. 34, 1645 (1995); (b) H. Ding, J. Kang, B.E. Hanson,
and C.W. Kohlpaintner. J. Mol. Catal. A: Chem. 124, 21 (1997).
8. (a) N. Nomura, J. Jin, H. Park, and T.V. RajanBabu. J. Am.
Chem. Soc. 120, 459 (1998); (b) M. Nandi, J. Jin, and T.V.
RajanBabu. J. Am. Chem. Soc. 120, 459 (1998), and
refs. therein.
18. I. Tóth and B.E. Hanson. J. Mol. Catal. 71, 365 (1992).
© 2001 NRC Canada