Heterolytic splitting of allylic alcohols with palladium(O)-TPPTS in water. Stabilities of the allylphosphonium salt of TPPTS and of the ionic complex [Pd(η-allyl)(TPPTS)2]+

Article abstract of DOI:10.1021/om701273j

The Pd(TPPTS)₃ complex (TPPTS is the sodium salt of tris(m-sulfophenyl)phosphine) easily ionizes allyl alcohol in water over a wide range of pH: OH^- and TPPTS are released, and [Pd(η³- allyl)(TPPTS)₂]₊ is formed. The released TPPTS further reacts with the palladium cationic complex to reversibly produce both the allylphosphonium salt of TPPTS [(allyl)Ar₃P]⁺ and Pd(TPPTS)₃, the latter acting as the catalyst of the allylation of TPPTS by allyl alcohol. Primary allylic alcohols, such as butenol (trans-2-buten-1ol), prenol (3-methyl-2-buten-1-ol), geraniol, and cinnamyl alcohol, react with Pd(TPPTS)₃ to produce hydroxide ion, the corresponding hydrosoluble cationic palladium complex, and allylic phosphonium salts. At room temperature, [Pd(η³-allyl)(TPPTS) ₂]⁺ is stable up to pH 12, but beyond this value, palladium precipitates. The temperature has an adverse effect on the complex stability: palladium precipitates at 80 °C, even at pH 7, with the formation of a small amount of propylene. The addition of [(allyl)Ar₃P] ⁺ increases the stability of [Pd(η³-allyl)(TPPTS) ₂]⁺. Above pH 10, [(allyl)Ar₃P]⁺ decomposes into OTPPTS and propylene by reaction with OH. At lower pH, [(allyl)Ar₃P]⁺ is slowly isomerized into [(propenyl)Ar₃P]⁺, which further reduces its stability toward pH and temperature. These consecutive reactions of the TPPTS ligand could explain most of the catalyst instability. This study outlines the basis for a better understanding of the instability phenomenon of the catalytic system Pd(0)-TPPTS in reactions with allylic intermediates, e.g. the Tsuji-Trost reaction, and in the reaction of dienes in aqueous media in which palladium often precipitates.

Full text of DOI:10.1021/om701273j

4300
Organometallics 2008, 27, 4300–4309
Heterolytic Splitting of Allylic Alcohols with Palladium(0)-TPPTS
in Water. Stabilities of the Allylphosphonium Salt of TPPTS and of
the Ionic Complex [Pd(η3-allyl)(TPPTS)2]+
Jean-Marie Basset, Denis Bouchu,† Gregory Godard, Iyad Karame´, Emile Kuntz,*
Fre´de´ric Lefebvre, Nicolas Legagneux, Christine Lucas, Daniel Michelet,‡ and
Jean Bernard Tommasino§
UniVersite´ de Lyon, Institut de Chimie de Lyon, Lyon, France, and UniVersite´ Lyon 1, CPE Lyon, CNRS,
UMR C2P2, LCOMS, Baˆtiment 308 F - CPE Lyon, 43 BouleVard du 11 NoVembre 1918,
F-69616 Villeurbanne, France
ReceiVed December 20, 2007
The Pd(TPPTS)3 complex (TPPTS is the sodium salt of tris(m-sulfophenyl)phosphine) easily ionizes
allyl alcohol in water over a wide range of pH: OH- and TPPTS are released, and [Pd(η3-allyl)(TPPTS)2]+
is formed. The released TPPTS further reacts with the palladium cationic complex to reversibly produce
both the allylphosphonium salt of TPPTS [(allyl)Ar3P]+ and Pd(TPPTS)3, the latter acting as the catalyst
of the allylation of TPPTS by allyl alcohol. Primary allylic alcohols, such as butenol (trans-2-buten-1-
ol), prenol (3-methyl-2-buten-1-ol), geraniol, and cinnamyl alcohol, react with Pd(TPPTS)3 to produce
hydroxide ion, the corresponding hydrosoluble cationic palladium complex, and allylic phosphonium
salts. At room temperature, [Pd(η3-allyl)(TPPTS)2]+ is stable up to pH 12, but beyond this value, palladium
precipitates. The temperature has an adverse effect on the complex stability: palladium precipitates at 80
°C, even at pH 7, with the formation of a small amount of propylene. The addition of [(allyl)Ar3P]+
increases the stability of [Pd(η3-allyl)(TPPTS)2]+. Above pH 10, [(allyl)Ar3P]+ decomposes into OTPPTS
and propylene by reaction with OH-. At lower pH, [(allyl)Ar3P]+ is slowly isomerized into
[(propenyl)Ar3P]+, which further reduces its stability toward pH and temperature. These consecutive
reactions of the TPPTS ligand could explain most of the catalyst instability. This study outlines the basis
for a better understanding of the instability phenomenon of the catalytic system Pd(0)-TPPTS in reactions
with allylic intermediates, e.g. the Tsuji-Trost reaction, and in the reaction of dienes in aqueous media
in which palladium often precipitates.
Introduction
The sodium salt of tris(m-sulfophenyl)phosphine (TPPTS)
(Figure 1) was first synthesized 30 years ago.1 Its strong ability
to coordinate low-valent transition metals (Co, Ni, Ru, Rh, Pd,
and Pt) as strictly nonextractable coordination complexes opened
the field of aqueous biphasic catalytic reactions.2 In 1984, the
development of a commercial biphasic oxo synthesis with
rhodium TPPTS catalyst (Ruhrchemie/Rhoˆne-Poulenc process)
demonstrated the strong potential of this concept.3 On the
laboratory scale, this system fully matched the conditions that
allowed a development study: i.e., yield, selectivity, catalyst
separation by decantation, and very good stability for 150 h
Figure 1. TPPMS, TPPDS, and TPPTS.
with an excess of TPPTS.4 This stability was later improved in
the Ruhrchemie pilot plant5 and then in the first industrial unit.6
Before 1984,7 palladium catalysts based on the hydrosoluble
phosphines TPPTS, TPPDS, and TPPMS (Figure 1) were mainly
studied for the dimerization, functionalization, or telomerization
of dienes, such as butadiene and isoprene. An important
objective was the hydrodimerization of butadiene catalyzed by
palladium(0)-TPPTS to produce octadienol (eq 1),7a an inter-
* To whom correspondence should be addressed. Tel: +33 (0)4 72 43
17 94. Fax: +33 (0)4 72 43 17 95. E-mail: kuntz@cpe.fr.
† Current address: Centre Commun de Spectrome´trie de Masse (UMR
5181 - Universite´ Claude Bernard Lyon 1).
‡ Current address: TECMO, 68 rue Masse´na, 69006 Lyon, France.
§ Current address: Laboratoire Synthe`se, Electrosynthe`se et Re´activite´
des Compose´s Organiques Fluore´s (MR 5181 - Universite´ Claude Bernard
Lyon 1).
(5) Cornils, B.; Hibbel, J.; Lieder, B.; Much, J.; Schmidt, V.; Wiebus,
E.; Konkol, W. (Ruhrchemie) U.S. Patent 4 523 036, Sep 18, 1982 (prior
German patent).
(6) Cornils, B.; Falbe, J. Proceedings of the 4th International Symposium
on Homogeneous Catalysis; Leningrad, Sep 1984; p 487.
(7) (a) Kuntz, E. G. (Rhoˆne-Poulenc Ind.) French Patent 2 366 237, July
27, 1976. (b) Yochimura, N.; Tamura, M. (Kuraray) U.S. Patent
4 334 117, Sep 1, 1979 (prior Japan patent). (c) Yochimura, N.; Tamura,
M. (Kuraray) U.S. Patent 4 356 333,March 28, 1980. (d) Yochimura, N.;
Tamura, M.;(Kuraray) U.S. Patent 4 417 079,Feb 16, 1982. (e) Groult, A.;
Guy, A. Tetrahedron 1983, 39 (9), 1543–1545.
(1) Kuntz, E. G. (Rhoˆne-Poulenc Ind.), French Patent 2 314 910, June
25, 1975.
(2) (a) Joo, F.; Toth, Z. J. Mol. Catal. 1980, 8, 369–383. (b) Kuntz,
E. G. CHEMTECH 1987, 17, 570–575.
(3) Cornils, B.; Kuntz, E. J. Organomet. Chem. 1995, 502, 177–186.
(4) Jenck, J. (Rhoˆne-Poulenc Ind.), French Patent 2 478 078, Dec 3,
1980.
10.1021/om701273j CCC: $40.75
2008 American Chemical Society
Publication on Web 08/16/2008

Splitting of Alcohols with Pd(0)-TPPTS in Water
Organometallics, Vol. 27, No. 17, 2008 4301
mediate for octanol and C9 compounds (linear diacid, diamine,
and dialcohol). However, the Pd(0) TPPTS catalyst was not
stable at 80 °C: a precipitation of metallic palladium occurred
after only a few hours of reaction. The exact reasons for such
a precipitation are still not well understood. Studies on the
hydrodimerization of butadiene were unsuccessfully carried out
under similar conditions with various cocatalysts.8
recently for the allylation of ꢀ-diketones, amines, ethyl acetoac-
etate,18 arenethiols,19 and lactones20 and for the isoprenylation
of haloanilines with 1,1-dimethylallyl alcohol.21 The allylation
of aldehydes was also achieved with allylic alcohols with an
indium-promoted palladium-TPPTS catalyst.22 The allylation
of arenethiols was carried out at room temperature with the
recycling of the catalyst without an accumulation of byproduct
or a loss of catalytic activity.19 The use of an allylic alcohol as
starting materialsmore easily available than the corresponding
ester or carbonatesand the formation of water as the only
byproduct make this procedure highly interesting for green
chemistry.23
However, the introduction of an excess of the allylic
phosphonium salts of TPPMS in the presence of palladium9 led
to a reaction rate and a stability of the catalytic system good
enough for the implementation of a commercial process for the
production of octadienol (5000 tons/year). Unfortunately, the
complexity of the process is likely to limit its potential
applications.10
A second important reaction was the Tsuji-Trost Pd(0)-
catalyzed nucleophilic substitution of allylic compounds, a well-
established methodology in organic synthesis (eq 2).11 However,
it was shown only in 1989 that a catalytic allylation could be
performed in a two-phase systemswater/organic solventsusing
water-soluble Pd complexes with catalyst separation.12 Allyla-
tion in a two-phase system was then extended to a variety of
nucleophiles using mainly TPPTS and TPPMS ligands with
allylic acetates, carbonates, ethers, and chlorides.13 The instabil-
ity of the catalyst has already been highlighted.12,14
Oshima reported that [Pd(η3-allyl)(TPPTS)2]+ species and
TPPTS oxide were readily formed from a mixture of allyl
alcohol, palladium acetate, TPPTS, and sodium carbonate.18
[Pd(η3-allyl)(TPPTS)2]+ was also formed from palladium allyl
chloride dimer and TPPTS in D2O. We also observed that
[Pd(η3-allyl)(TPPTS)2]+ was formed from the allyl alcohol
reaction with the Pd(TPPTS)3 complex at pH 7.2. Moreover,
the allylphosphonium salt of TPPTS was stoichiometrically
produced in this reaction.24
In our study of the C-allylation of phenols by allyl alcohol,
we observed that palladium precipitation occurred in basic
medium at 80 °C (pH 9.5-12),24 conditions slightly different
from those of the hydrodimerization of butadiene into octadienol
in the presence of sodium hydrogenocarbonate and the
Pd-TPPTS catalyst.7a Recently, the deactivation of a homo-
geneous Pd catalyst in organic solvent was studied using various
Pd(1,1-dimethylallyl)(P-P ligand)OTf complexes as catalytic
intermediates and piperidine as nucleophile.25 The formation
of inactive Pd dimers and trimers, a possible first step in the
deactivation process of the Tsuji-Trost reaction, led to larger
palladium clusters and eventually palladium black. Therefore,
it seemed crucial to determine the conditions for the formation
of (π-allyl)palladium(II) complexes and phosphonium salts in
aqueous solution and their stability versus pH and temperature
in order to understand the reasons for the instability of the
catalytic system.
Allylic alcoholssmore available than their derivatives23s
were directly used in a two-phase system with palladium for
the allylation of aromatic amines,15 for the isoprenylation of
an amino acid (4-bromotryptophan) with 1,1-dimethylallyl
alcohol,16 for the C-allylation of phenol or guaiacol,17 and more
(8) (a) Yoshimura, N. Aqueous-Phase Organometallic Catalysis: Con-
cepts and Applications; Cornils, B., Herrmann, W. A., Eds.; Wiley-VCH:
Weinheim, Germany, 1998; pp 408-417. (b) Lee, B. I.; Lee, K. H.; Lee,
J. S. J. Mol. Catal., A 2001, 166, 233–242.
(9) Maeda, T.; Tokitoh, Y.; Yoshimura, N. (Kuraray) Eur. Patent
0 296 550, June 24, 1987 (prior Japan patent). Allylic phosphonium salts
of TPPMS (substituents: allyl; 2,7-octadien-1-yl; 2-buten-1-yl).
(10) (a) Yoshimura, N. , Aqueous-Phase Organometallic Catalysis:
Concepts and Applications; Cornils, B., Herrmann, W. A., Eds.; Wiley-
VCH,: Weinheim, Germany, 2003; pp 540-549. (b) Li, C.-J.; Chan, T.-H.
ComprehensiVe Organic Reactions in Aqueous Media; Wiley-VCH: Wein-
heim, Germany, 2007.
(11) (a) Trost, B. M.; Verhoeven, T. R. Organometallic Compounds in
Organic Synthesis and in Catalysis, ComprehensiVe Organometallic Chem-
istry; Pergamon Press: Oxford, U.K., 1982, Vol. 8, p 799. (b) Tsuji, J.
Palladium Reagents and Catalysts InnoVations in Organic Synthesis; Wiley-
VCH: Chichester, U.K., 1996. (c) Trost, B. M.; Van Vranken, D. L. Chem.
ReV. 1996, 96, 395–422. (d) Negishi, E. Handbook of Organopalladium
Chemistry for Organic Synthesis; Wiley: New York, 2002.
We report herein the allylation of Pd(TPPTS)3 with allylic
alcohols in water into the ionic compounds [Pd(η3-alkenyl)(T-
PPTS)2]+ and the allylic phosphonium salts of TPPTS.26 The
study of the stabilitysversus pH and temperaturesof ionic
compounds obtained from allyl alcohol was carried out and
(17) Kuntz, E. G.; Amgoune, A. French Patent 2 862 639, Nov 25, 2003.
(18) Kinoshita, V. H.; Shinokubo, H.; Oshima, K. Org. Lett. 2004, 6
(22), 4085–4088.
(19) Komine, N.; Sako, A.; Hirahara, S.; Hirano, M.; Komiya, S. Chem.
Lett. 2005, 34 (2), 246–247.
(20) Kinoshita, V. H.; Shinokubo, H.; Oshima, K. Angew. Chem., Int.
Ed. 2005, 44, 2397–2400.
(21) Yokoyama, Y.; Takagi, N.; Hikawa, H.; Kaneko, S.; Tsubaki, N.;
Okuno, H. AdV. Synth. Catal. 2007, 349, 662–668.
(22) Fontana, G.; Lubineau, A.; Scherrmann, M. Org. Biomol. Chem.
2005, 3, 1375–1380.
(23) (a) Muzart, J. Tetrahedron 2005, 61, 4179–4212. (b) Muzart, J.
Eur. J. Org. Chem. 2007, 3077–3089.
(24) Kuntz, E. G.; Amgoune, A.; Lucas, C.; Godard, G. J. Mol. Catal.
2006, 244, 124–136.
(25) Tromp, M.; Sietsma, J. R. A.; van Bokhoven, J. A.; van Strijdonck,
J. P. F.; van Haaren, R. J.; van der Eerden, A. M.; van Leeuween,
P. W. N. M.; Koningsberger, D. C. Chem. Commun. 2003, 128–129.
(26) (a) Homologous Pt(η3-allyl)(TPPTS)2+ was recently prepared from
PtCl2(TPPTS)2 and allyl alcohol with stoichiometric formation of
1-hydroxyacetone.26b We have shown that PtCl2(TPPTS)2 catalyzes a
selective C-allylation of guaiacol with allyl alcohol into p-eugenol and
o-eugenol (65/35) in water.26c The selectivity of the PtCl2(TPPTS)2 catalyst
in this allylation was similar to that obtained with the palladium(0) TPPTS
catalyst.24. (b) Helfer, D. S.; Phaho, D. S.; Atwood, J. D. Organometallics
2005, 17, 78–89. (c) Reference 17, patent example number 17.
(12) (a) Sinou, D. French Patent 2 652 514, June 22, 1989. (b) Safi, M.;
Sinou, D Tetrahedron Lett. 1991, 32 (18), 2025–2028.
(13) (a) Sinou, D.; Goux, C.; Massacret, M.; Lhoste, P. Organometallics
1995, 14, 4585–4593. (b) Dos Santos, S.; Tong, Y.; Quignard, F.; Choplin,
A.; Sinou, D.; Dustata, J. P. Organometallics. 1998, 17, 78–89. (c) Genet,
J. P.; Savignac, M. J. Organomet. Chem. 1999, 576, 305–317.
(14) (a) Dos Santos, S.; Quignard, F.; Sinou, D.; Choplin, A. Top. Catal.
2000, 13, 311–318. (b) Liao, M.; Duan, X.; Liang, Y. Tetrahedron Lett.
2005, 46, 3469–3472.
(15) Grosselin, J.; Kempf, H.; Lecouve, J. (Rhoˆne-Poulenc Chimie) Eur.
Patent 0 470 000 A1, Aug 2, 1990 (prior French patent).
(16) (a) Yokoyama, Y.; Hikawa, H.; Mitsuhashi, M.; Uyama, A.;
Murakami, Y. Tetrahedron Lett. 1999, 40, 7803–7806. (b) Yokoyama, Y.;
Hikawa, H.; Mitsuhashi, M.; Uyama, A.; Hiroki, Y.; Murakami, Y. Eur. J.
Org. Chem. 2004, 6, 1244–1253, Dimethylallyl alcohol reacted with
4-bromotryptophan to give clavicipitic acid by a Heck reaction under highly
basic conditions (NaOH), but with NaOOCCH3, isoprenylation of 4-bro-
motryptophan occurred.

4302 Organometallics, Vol. 27, No. 17, 2008
Basset et al.
provided essential results for a better understanding of catalytic
reactions in water with palladium and TPPTS.
Results and Discussion
1. Pd(TPPTS)3 Reaction with Allyl Alcohol. Pd(TPPTS)3
was prepared in aqueous solution by the reduction of palladium
acetate by an excess of TPPTS, made easier by NaOH addition
(TPPTS/Pd molar ratio of 4 or greater) (eq 3).27,28 By conven-
tion, 4 refers to a solution of the Pd(TPPTS)3 complex with
either a small excess or no excess of TPPTS and with sodium
acetate and TPPTS oxide. 1a-3a refer to [Pd(η3-allyl)(T-
PPTS)2]+, TPPTS oxide, and the allylphosphonium salt of
TPPTS, respectively.
Pd(OAc)2 + 4PAr3 + 2NaOH f Pd(PAr3)3 + OPAr3 +
Figure 2. Reaction rate between allyl alcohol and Pd(TPPTS)3:
H+/Pd ratio versus time (allyl alcohol/Pd ) 10, 5, 1). Conditions:
[Pd] initial concentration 0.0055 M in water at room temperature,
initial pH 7; after allyl alcohol addition, 0.2 M HClO4 was added
to keep the pH at 7.
2a
2NaOAc + H2O
PAr3 ) TPPTS (3)
The 31P NMR spectrum of a solution of 4 shows only two
signals corresponding to 2a (δ 35.14 ppm) and 4 (δ 23.85 ppm).
However, as previously described,29 a signal shift toward low-
field values was observed for 4 in the presence of an excess of
TPPTS due to an exchange reaction of TPPTS with the
complex.30
We did not observe the dissociation of Pd(TPPTS)3 into
Pd(TPPTS)2 and TPPTS in dilute medium by 31P NMR (eq 4;
palladium concentrations 2 × 10-4 and 1.2 × 10-4 M): the
signal was still close to 23.85 ppm (23.90 and 23.84 ppm,
respectively) in the absence of free TPPTS.31 The Pd(0) TPPTS
system can then mainly be considered as Pd(TPPTS)3 with an
eventual excess of phosphine (for n g 3 in this study).
This reaction consumes protons, which allows a kinetic study
by keeping the pH at a constant value by the addition of
perchloric acid. Figure 2 shows the curves obtained at room
temperature and pH 7 for allyl alcohol/Pd molar ratios of 10,
5, and 1. In the first two cases, the conversion of 4 into
palladium allyl complex 1a and phosphonium species 3a is
quantitative, as observed by 31P NMR. The intensity ratio of
the two signals (2/1) is in agreement with eq 5. At pH 7 with
an allyl alcohol/Pd molar ratio of 10, Pd(TPPTS)3 prepared with
sulfate or chloride exhibited the same reaction rate during the
whole reaction.28 With an excess of allyl alcohol, the curves of
Figure 2 show unambiguously that the reaction proceeds in two
steps: the proton consumption is very fast for the first few
minutes of the reaction and then becomes slower.
Pd(TPPTS)3 h Pd(TPPTS)2 + TPPTS
a. Reaction at pH 7. As shown previously by us,24 4 reacts
(4)
with allyl alcohol according to eq 5.
With allyl alcohol as limiting reagent, for example allyl
alcohol/Pd ) 1, the reaction was slower. After 5 h, the amount
of acid added corresponds to the amount of allyl alcohol added:
the allyl alcohol reacted quasi-quantitatively and was converted
into 1a (60%) and 3a (40%). The signal of 4 broadened and
shifted to lower chemical shifts, indicating the presence of free
TPPTS in the solution.
Under similar conditions with various allyl alcohol/palladium
ratios lower than 2, the amount of acid added is at least 95% of
the amount of allyl alcohol introduced in all cases. Allyl alcohol
then reacted quasi-quantitatively. All attempts to detect free
alcohol in the reaction medium by 1H NMR failed, which agrees
with the above assumption (the detection limit was estimated
at ca. 5 × 10-4 M). With a small excess of allyl alcohol (a
ratio of 2.09), both 1a and 3a are produced quantitatively (see
the Supporting Information, reaction of Pd(TPPTS)3 with allyl
alcohol as limiting reagent).
(27) (a) Kuntz, E.; Vittori, O. J. Mol. Catal. A, Chem. 1998, 129, 159–
171. (b) Herrmann, W. A.; Kellner, J.; Rielp, H. J. Organomet. Chem. 1990,
389, 103–128.
(28) Pd(TPPTS)3 solutions with sulfate (sulfate/Pd molar ratio 1) were
prepared from palladium sulfate according to ref 27a and Pd(TPPTS)3
solutions with chloride (chloride/Pd molar ratio 4) from potassium tetra-
chloropalladate(II) by the same method.
(29) Choplin, A.; Dos Santos, S.; Quignard, F.; Sigismondi, S.; Sinou,
D. Catal. Today 1998, 42, 471–478.
(30) With a Pd concentration of 0.011 M, the following values were
obtained: n (TPPTS/Pd(0) ratio) ) 3, δ 23.85 ppm; n ) 3.53, δ 21.75 ppm;
n ) 4.06, δ 17.97 ppm; n ) 8, δ 2.06 ppm (TPPTS alone: δ -5 ppm).
This variation allows an estimation of the ratio TPPTS/Pd(0) in the solution,
even during catalysis. Indeed, under some conditions, the allylation reaction
of 4 liberates TPPTS, which will be in equilibrium with the unreacted
complex. The chemical shift value of 4 will then allow an estimation of
the TPPTS/complex ratio in the solution.
(31) This result involves an equilibrium constant of dissociation K lower
than 10-5. It is in agreement with recent results of the dissociation of
Pd(TPP)3 (TPP ) triphenylphosphine) into TPP and Pd(TPP)2. The
dissociationconstantwasfoundtobe1.7×10-4 inorganicmedium:Amatore,
C.; Jutand, A.; Mensah, L.; Meyer, G.; Fiaud, J.; Legros, J. Eur. J. Org.
Chem. 2006, 1185–1192.
b. Reaction at pH 3, 7, and 9. Allylation was performed at
pH 3, 7, and 9 (Figure 3). At pH 3 and 7, the initial rate was
fast. At pH 3, one proton was consumed per Pd in 1 min, while
ca. 5 min is necessary at pH 7, but after 70-80 min the overall
reaction was complete. At pH 9, the reaction becomes very slow:
one proton was consumed per Pd in 2 h, and the rapid and slow
steps cannot be discriminated. These data combined with those

Splitting of Alcohols with Pd(0)-TPPTS in Water
Organometallics, Vol. 27, No. 17, 2008 4303
this feature is that TPPTS and the allyl alcohol compete to
generate Pd(TPPTS)3 and Pd(TPPTS)2(allyl alcohol), respec-
tively. This latter complex leads to cation 1a and OH- by
ionization.
d. Reversibility of the Formation of the Allylphospho-
nium Salt of TPPTS and Ionic Character of the Palladium
Complex. The allylphosphonium salt 3a reacted quickly with
complex 4 in solution at room temperature. After the addition
of 4 to a solution of 3a, 31P NMR spectroscopy showed that 4
was partially but rapidly converted into 1a (conversion 25%).
The equilibrium was reached after 20 min (eq 7). Therefore,
after the reaction, the concentration of 4 is 0.75c0 and that of
free TPPTS 0.5c0: i.e., the total ratio TPPTS/Pd(0) is 3.66, in
agreement with the signal at δ 19 ppm of the 31P NMR spectrum
attributed to complex 4 in the presence of a TPPTS excess.
Figure 3. Rate of allyl alcohol reaction with Pd(TPPTS)3: H+/Pd
ratio versus time at pH 3, 7, and 9 at room temperature. Conditions:
[Pd] initial concentration 0.0055 M in water at room temperature,
allyl alcohol/Pd ) 10.
above are in agreement with a pH-dependent formation of 1a
from 4 and a slower but pH-independent formation of 3a from
TPPTS released during the first reaction.
c. Reaction with an Excess of TPPTS .32,33 The above data
show that the transformation of TPPTS into 3a is catalyzed by
palladium species present in the reaction medium. This point
can be proven by doing the reaction in the presence of an excess
of TPPTS.
This reaction can be regarded as a redox reaction between
four components, including two ions. The phosphonium ion is
reduced into phosphine by the palladium(0) complex, which is
transformed into palladium(II) ion.
When allyl alcohol and TPPTS were in excess to palladium
(allyl alcohol/Pd(0) ≈ 11 and TPPTS/Pd(0) ) 7), all palladium
was converted into [Pd(η3-allyl)(TPPTS)2]+ and all the remain-
ing TPPTS molecules were also transformed into phosphonium
species 3a. The addition of 6 H+ per Pd was necessary to
maintain the pH at 7, in agreement with eq 6.
From the solution containing both 1a and 3a, it has not been
possible to identify signals from mass spectroscopy that could
establish the ionic character of 1a. The TPPTS, TPPDS, and
TPPMS ligands have already been investigated by ESMS in
the negative-ion mode and exhibited strong negative ions.35
The three homologous Pd(II) complexes of TPPTS (1a),
TPPDS (1b), and TPPMS (1c) were prepared by the dissolution
of [Pd(η3-allyl)(µ-Cl)]2 in a well-stirred solution of TPPTS,
TPPDS, or TPPMS in water or in D2O, with a P/Pd atomic
ratio of 2 (eq 8) (Supporting Information).36
The ionic character of the complexes was established by
electrospray mass spectroscopy (ESMS). The complexes 1a-c,
using 10-3 M water solutions diluted (1/100) with a mixture of
methanol, dichloromethane, and water (v/v 40, 45, 15) contain-
ing 0.1% of formic acid, also exhibited ions in their negative
form: one major ion with 1c and three major ions with 1b. The
complex 1a did not exhibit major ions but many signals. Only
two ions in their negative form at m/z 1170.7 and 1148.7 were
identified (Figure 4). The ion mass spectra exhibited an isotopic
The initial reaction rate at pH 7 was about 15 times less than
that without a TPPTS excess. Similar results obtained in
nonaqueous medium also showed a decrease of the reaction rate
in the presence of an excess of triphenylphosphine in the Heck
reaction with palladium catalyst.34 A possible explanation for
(32) In organic solvent, allylic triphenylphosphonium salt syntheses have
been reported from the reaction of TPP with allylic acetates32a or allylic
trifluoroacetates32b in the presence of Pd(TPP)4 catalyst: (a) Tsukahara, Y.;
Kinoshita, H.; Inomata, K.; Kotake, H Bull. Chem. Soc. Jpn. 1984, 57, 3013–
3014. (b) Granberg, K. L.; Ba¨ckvall, J.-E. J. Am. Chem. Soc. 1992, 114,
6858–6863.
(33) More recently, alkenyl and alkyltriphenylphosphonium salts
(RP+Ph3, X-) were prepared from different metal-catalyzed additions of
triphenylphosphine and strong acids to alkynes,33a 1,3-alkadienes,33b and
1-alkenes.33c Addition on alkynes with methanesulfonic acid with LiPF6,
X ) PF6-, catalyst Pd(TPP)4 or [RhCl(cod)]2: (a) Arisawa, M.; Yamamuchi,
M. J. Am. Chem. Soc. 2000, 122, 2387–2388. Addition on 1,3-alkadienes
with trifluoromethanesulfonic acid, X ) CF3SO3, catalyst RhH(TPP)4: (b)
Arisawa, M.; Momozuka, R. Chem. Lett. 2002, 272–73. Addition on alkenes
with bis(trifluoromethanesulfonyl)imide (Tf2NH), X ) (CF3SO2)2N-,
catalyst Pd2(dba)3, CHCl3: (c) Arisawa, M; Yamamuchi, M. J. Am. Chem.
Soc. 2006, 128, 272–273.
(35) Henderson, W.; Olsen, G. M. Polyhedron 1998, 17 (4), 577–588.
(36) When the ratio TPPTS/Pd was increased to 2.25 in the above
preparation of 1a, the compounds 1a, 4, and 3a appeared in the molar ratio
1a/4/3a ) 89/11/11, which showed that TPPTS in excess was partially
allylated by 1a. 4 appeared at 23.26 ppm (normally at 23.85 ppm but TPPTS
in excess made a shift of-0.59 ppm). A similar reaction in organic solvent
was already observed with allylpalladium chloride dimer and triphenylphos-
phine (TPP). In the presence of more than 2 equiv of TPP in methanol,
allyltriphenylphosphonium cation and palladium(0) TPP complex were
formed.36a An analogous reaction was observed in DMF without TPP excess
because the palladium(0) complex was stabilized with chloride anion and
one TPP molecule:36b (a) Shaw, B. L.; Powel, J J. Chem. Soc. A 1968,
774–777. (b) Amatore, C.; Jutand, A.; M’Barki, M. A.; Meyer, G.; Mottier,
L. Chem. Eur. J. 1999, 5, 466–473. (c) Amatore, C.; Jutand, A.; M’Barki,
M. A.; Meyer, G.; Mottier, L. Eur. J. Inorg. Chem. 2001, 873–880.
(34) Amatore, C.; Blart, E.; Geneˆt, J. P.; Jutand, A.; Lemaire-Audoire,
S.; Savignac, M. J. Org. Chem. 1995, 60, 6829–6839.

4304 Organometallics, Vol. 27, No. 17, 2008
Basset et al.
Figure 4. Selected negative-ion electrospray mass spectral data of ionic complexes 1c (TPPMS), 1b (TPPDS), and 1a (TPPTS).
the amount of TPPTS oxide 2a in the solution had increased
(2a/Pd molar ratio of 1), and propylene was detected in the gas
phase. All these data are in agreement with eq 9.
Starting from pH 11 under the same conditions, the pH was
11.76 after 2 days (theoretical pH 11.81). The conversion of 4
was 100% into 1a. 3a was not detected. The amount of TPPTS
Figure 5. Allyl alcohol reaction without acid addition: pH versus
oxide 2a in the solution had increased, but a new oxide
time. Conditions: [Pd] ) 0.0055 M in water at room temperature,
appeared: the disodium salt of propenylbis(m-sulfophenyl)phos-
allyl alcohol/Pd molar ratio 10, initial pH 7.14, final pH 11.65,
phine oxide (2b) with the ratio 2b/2a ) 23/77.
close to the theoretical pH 11.74, after reaction [OH-] ) [Pd] )
0.0055 M.
Thus, the lack of 3a in the synthesis of 1a from allyl alcohol,
palladium acetate, TPPTS, and sodium carbonate, as previously
published by Oshima,18 can be explained by its instability in
basic medium.
pattern that confirms the ionic character of these palladium
complexes in solution (see the Supporting Information: mass
spectra, NMR data).
The instability of hydrosoluble alkyl- and vinyl-substituted
triphenylphosphonium salts in basic medium has already been
described. The reaction of hydroxide ions with hydrosoluble
alkylphosphonium salts derived from TPPTS and ethyl acrylate
were reported to lead to a phosphine oxide (Scheme 1).37 It
e. Allylation with Allyl Alcohol without Acid Addition.
Under conditions similar to those above, but without acid
addition, the pH increased to basic values after allyl alcohol
addition (Figure 5). After 2 h of reaction, the 31P NMR showed
that the conversion of 4 was 15% into 1a and 4% into 3a (molar
ratio 3a/initial 4). One day later, the pH of the solution was
11.65, which corresponds to the formation of one hydroxide
ion per palladium. Moreover, 31P NMR showed that the
conversion of 4 was 100% into 1a. 3a was not detected, but
(37) (a) Larpent, C.; Patin, H. C. R. Acad. Sci. Paris 1987, 1055. (b)
Larpent, C.; Patin, H. Tetrahedron 1988, 44, 6107–6118.
(38) (a) Fenton, G. W.; Ingold, C. K. J. Chem. Soc. 1929, 2342. (b)
Grayson, M.; Keough, T. P. J. Am. Chem. Soc. 1960, 82, 3919–3924.

Splitting of Alcohols with Pd(0)-TPPTS in Water
Organometallics, Vol. 27, No. 17, 2008 4305
Table 1. Allylic Alcohols, Allylic Palladium Complexes, and Allylic
Scheme 1
Phosphonium Species (with L ) PAr3 ) TPPTS)
Scheme 2
a For the characterization of allylic compounds, see theSupporting
Information.
has long been known that phosphonium salts are decomposed
by base to produce tertiary phosphine oxides and hydrocar-
bons.38
Vinyl-substituted triphenylphosphonium salts have already
been prepared from TPPTS or TPPMS and various R-acetylenic
carboxylic acids. These salts are easily transformed into TPPTS
oxide, TPPMS oxide, or vinyl-substituted diphenylphosphine
oxide by NaOH addition.39
In our case, the addition of sodium hydroxide to a solution
of 3a (0.02 M) at room temperature produced mainly 2a +
propylene (70%) and 2b (30%). The phosphonium salt decom-
position consumed hydroxide ions according to the reaction
shown by eq 10.
as the limiting reagent compared to Pd(TPPTS)3, a high
conversion rate of allyl alcohol into [Pd(η3-allyl)(TPPTS)2]+
and [(allyl)Ar3P]+ was observed. There are reversible reactions
between the four phosphorus compounds, since it was shown
that [(allyl)Ar3P]+ quickly transforms Pd(TPPTS)3 into [Pd(η3-
allyl)(TPPTS)2]+ and 2 TPPTS. In the presence of an excess of
allyl alcohol, Pd(TPPTS)3 and the eventual TPPTS in excess
are fully transformed into [Pd(η3-allyl)(TPPTS)2]+ and
[(allyl)Ar3P]+. Palladium finally acts as a catalyst for the
transformation of TPPTS into phosphonium hydroxide by the
action of allylic alcohol. Under our conditions, this hydroxide
is neutralized by an acid during the reaction.
At a nonregulated pH, the reaction releases OH- and the
reversibility of reaction I could be evaluated by the possible
measure of an equilibrium for a high pH. However, the
instability of phosphonium ions above pH 10 made the system
irreversible and therefore removed any equilibrium measure.
The allyl alcohol added to the aqueous solution of Pd(TPPTS)3
at a neutral pH transforms it fully into [Pd(η3-allyl)(TPPTS)2]+
with a pH increase to 11.65 corresponding to 1 OH- produced
by palladium. At the same time, TPPTS is transformed into
[(allyl)Ar3P]+ by producing 1 OH- and then mainly into
OTPPTS and propylene by consuming 1 OH- (reaction III).
Palladium then looks like a catalyst for the oxidation of TPPTS
into TPPTS oxide, whereas allylic alcohol is reduced to
propylene.
2. Pd(TPPTS)3 Reactions with Allylic Alcohols: trans-2-
Buten-1-ol (C4), Prenol (C5), Cinnamyl Alcohol (C9), and
Geraniol (C10). The heterolytic cleavage observed with allyl
alcohol was extended to other substituted primary allylic
alcohols at a regulated pH. In all cases, complex 4 disappeared
totally; but with C5 and C10 the reaction was too slow at pH 7
and had to be completed at a more acidic pH. The complex 4
was transformed into a hydrosoluble cationic palladium complex
and a hydrosoluble allylic phosphonium species (Table 1). The
reactions with C5, C9 and C10 were very selective. The 31P NMR
spectra of the solutions were similar. Each solution showed only
one signal near +22 ppm, attributed to TPPTS phosphonium
species (5a1, 9a1, 10a1-geranyl) and two doublets between +25
Unfortunately, palladium appears as a catalyst for the
oxidation of TPPTS into TPPTS oxide by allyl alcohol which
is mainly reduced into propylene in basic medium. The study
of the stability of 3a solutions without palladium revealed the
existence of several transformation reactions of 3a (see section
3).
f. Overall Mechanism. The first reaction observed between
allyl alcohol and Pd(TPPTS)3 was the formation of the complex
[Pd(η3-allyl)(TPPTS)2]+ and the release of a TPPTS molecule
and OH- (Scheme 2, reaction I). This reaction was fast in acidic
medium, but its rate decreased greatly around pH 9-10. The
initial reaction rate was lower in the presence of an excess of
TPPTS. The equilibrium between Pd(TPPTS)3 and Pd(TPPTS)2
could not be demonstrated, but TPPTS and allyl alcohol would
be in competition to form the complex Pd(TPPTS)2(allyl
alcohol), which would then ionize into OH- and [Pd(η3-
allyl)(TPPTS)2]+. TPPTS, the most active nucleophile in the
aqueous solution, reacts then with [Pd(η3-allyl)(TPPTS)2]+ to
produce reversibly the allylphosphonium TPPTS species and
Pd(TPPTS)2 immediately stabilized into Pd(TPPTS)3 by a
second TPPTS molecule (reaction II). By using allyl alcohol
(39) Larpent, C.; Patin, H Tetrahedron Lett. 1988, 29 (36), 4577–4580.

4306 Organometallics, Vol. 27, No. 17, 2008
Basset et al.
Figure 6. Rate of cinnamyl alcohol reaction with Pd(TPPTS)3: H+/
Pd ratio versus time. Conditions: [Pd] ) 0.0055 M in water at 25
°C, cinnamyl alcohol/Pd ) 10, initial pH 7, immediately after
alcohol addition, pH was maintained at 7 by 0.2 M HClO4 addition.
Figure 7. Comparative reactivity of allylic alcohols. Conditions:
[Pd] ) 0.0055 M in water at 25 °C, initial pH 7, immediately after
allylic alcohol addition, pH was maintained at 7 by 0.2 M HClO4
addition. Geraniol and cinnamyl alcohol were not totally soluble
in water at the concentration 0.055 M; prenol and geraniol reactions,
too slow at pH 7, were completed at a lower pH.
ppm and +30 ppm, attributed to TPPTS palladium (II)
complexes [PdC5L2]+, [Pd(syn-C9)L2]+, [Pd(C10-geranyl)L2]+)
(Table 1).
In the presence of a large excess of allylic alcohols, it was
not possible to characterize the ionic complexes and phospho-
nium species directly in solution. The strategy was to prepare
Scheme 3. Decomposition of the Allylphosphonium Chloride
Salt of TPPTS47
and characterize the phosphonium salt solutions (1H NMR, 31
P
NMR, and mass spectra) and then to characterize the palladium
complex solutions in the presence of phosphonium salts, but
without an excess of allylic alcohol (Supporting Information).
With C4, complex 4 was transformed into the hydrosoluble
cationic palladium complex [Pd(syn-C4)L2]+ (90%),40 which
appeared as a singlet (+27.73 ppm), and into the hydrosoluble
phosphonium species 4a1 (95%). In order to characterize the
two ions, [Pd(syn-C4)L2]+ was prepared from the reaction of
crotylpalladium chloride dimer with 4 TPPTS and 4a1 was
prepared directly from the reaction of TPPTS with crotyl
chloride.
The reaction of cinnamyl alcohol is given here as an example
(Figure 6). Acid addition must be very fast at the beginning of
the reaction to keep the pH constant at 7. This behavior was
attributed to the preferential formation of the palladium complex.
Then, the reaction slowed down. This second step was attributed
to the formation of phosphonium salt.
A comparison of the relative reactivities of allylic alcohols
during the first step of the reaction (Scheme 2, reaction I) were
quantified by the initial rates of acid addition (expressed in (H+/
Pd) min-1), as depicted in Figure 6. It can be concluded that
the initial rates of cinnamyl alcohol (0.30 min-1), allyl alcohol
(0.20 min-1), and trans-2-buten-1-ol (0.10 min-1) were high.
presumably a steric hindrance effect. This phenomenon could
limit their reactivity in these conditions. In very basic media,
the ionization of substituted alcohols could then become the
limiting step in catalytic reactions.
More substituted allylic alcohols, such as prenol (0.01 min-1
)
and geraniol (0.0023 min-1), were less reactive at pH 7. At pH
values between 5 and 6, these alcohols were more reactive and
led to a complete conversion of 4 at pH 4.5 (Figure 7).
These data indicate that the ionization step in the Tsuji-Trost
reaction in water depends on the substitution in position 3 of
these primary allylic alcohols. Monosubstitution with the phenyl
group has an accelerating effect, while substitution by a methyl
has a slightly slowing effect. Disubstituted alcohols such as
prenol and geraniol react much more slowly at pH 7, due to
3. Stability of the Allylphosphonium Chloride Salt of
TPPTS: Effect of pH and Temperature. It appears from the
previous results that the stability of the catalytic system could
be related to the stability of the phosphonium cations produced.
An irreversible destruction of phosphonium salts should lead
to the decomposition of TPPTS by an equilibrium shift (Scheme
2, reaction II), and so, to the precipitation of palladium.
The study of the stability of 3a versus pH (from 4.75 to 13)
and temperature (from 25 to 80 °C) revealed the existence of
several transformations (Scheme 3). The transformation into 2a
and propylene consumed OH- (eq 10). 3a can also isomerize
into the vinylic species 3b, which can then be hydrated into 3c.
3b,c are also transformed into the oxides 2b,c, respectively
(OH- consumption results in a variation of pH). In this study,
by a judicious choice of pH, temperature, and time conditions,
(40) 31P NMR spectrum: major signal, singlet at 27.73 ppm (90%); other
signals, two doublets at 27.03 ppm (10%) (d, JPP ) 40 Hz) and 25.45 ppm
(d, JPP ) 40 Hz), probably the anti isomer [Pd(anti-C4)L2]+, which should
have signals similar to the signals of the complexes [Pd(C5)L2]+,
[Pd(C9)L2]+, and [Pd(C10)L2]+. In the case of the preparation of [Pd(syn-
C4)L2]+ from the reaction of crotylpalladium chloride dimer with four
TPPTS, the same spectrum was obtained.

Splitting of Alcohols with Pd(0)-TPPTS in Water
Organometallics, Vol. 27, No. 17, 2008 4307
Table 2. Selected Negative-Ion Electrospray Mass Spectrum Data of 3a–ca
a The m/z required is given in parentheses.
(31P NMR spectra remained unchanged), while a decomposition
was observed in basic medium.
At room temperature upon the gradual addition of 1 M sodium
hydroxide to a dilute solution of 1a, there was no decomposition
up to around pH 12. However, at pH 12.5, a blackening of the
solution occurred, and then metallic palladium precipitated. The
31P NMR analysis of the solution after decomposition showed
the disappearance of the signal of 1a at +25.60 ppm, the
appearance of the complex 4 Pd(TPPTS)3, corresponding to 50%
of palladium, and 2a (with the surface ratio 2a/4 ) 0.36). We
point out here that complex 1a, prepared by the reaction of allyl
alcohol in excess with Pd(TPPTS)3 without acidification, was
stable for several weeks at room temperature at pH 11.65.
At pH 13.87 (0.75 M NaOH solution) the reaction proceeded
instantaneously: a metallic palladium precipitate appeared and
1a was also transformed into 4 (48 mol %/Pd) and 2a (12 mol
%/Pd). The evolution of propylene (12 mol %/Pd) is in
agreement with a partial oxidation of TPPTS in 1a into 2a.
However, two unidentified phosphorus products also appeared
(broad signals at +27 and +30 ppm), showing that decomposi-
tion proceeds via another mechanism.
Figure 8. Stability of the allylphosphonium salt of TPPTS 3a. R1
corresponds to a 1% decomposition rate of 3a into 3b,c (80 °C pH
7.5; 50 °C pH 8.5; 25 °C pH 9). R2 corresponds to a 1%
decomposition rate of 3a into oxides (80 °C pH 9.5; 50 °C pH 10;
25 °C pH 10.6).
we were able to concentrate aqueous solutions of 3b,c and 2b,c
products for their characterization in solution by 31P, 13C, and
1H NMR. The ionic nature of species 3a-c was demonstrated
by mass spectroscopy (Table 2 and the Supporting Information).
In a highly basic medium (pH above 11), the main decom-
position product is TPPTS oxide 2a (with 2b,c if conditions
are less basic). At lower pH values (between 7 and 10), 3b,c
appear, whereas the decomposition into oxide 2a is much slower
(<1% h-1). In neutral or slightly acidic medium, there is still
a very low decomposition of 3a at 80 °C (0.32% h-1 at pH 7
and 0.03% h-1 at pH 4.7).
A possible reaction for the formation of the metallic palladium
is depicted in eq 11. Allylic alcohol is not observed in the
reaction medium, as it would be transformed immediately into
propanal upon an aldolization reaction.
These transformation reactions of phosphonium salt 3a can
be quantified in a simplified way by considering the curves that
allow a 1% h-1 decomposition of 3a (Figure 8). The curve R1
shows the decomposition of 3a into 3b,c also at 1% h-1, and
the curve R2 shows the transformation of 3a into oxides at 1%
h-1. In the area A located below R1, 3a can be considered stable
if the pH is below 9 at room temperature and below 7.5 at 80
°C. In the area C located above R2, 3a decomposed quickly
into oxides if the pH is above 10.6 at room temperature and
above 9.5 at 80 °C. In the area B located between R1 and R2
with a gap of 1.5 pH units, the decomposition into 2a is very
slow, but 3a is quickly converted into 3b,c on heating (see the
Supporting Information).
In conclusion, complex 1a is stable at room temperature in
aqueous medium over a fairly wide range of pH values, but
not beyond pH 12. Above this pH, a dark color appears followed
by the precipitation of palladium.
b. Stability at 80 °C. At 80 °C, complex 1a decomposes
slightly at pH 7.60. After 12 h, the pH has decreased from 7.60
to 3.97 (H+ released 1%/Pd) and propylene has evolved (2.5%)
with OTPPTS formation (1.5%) and an unidentified phosphorus
compound X at +27.80 ppm (1%). The precipitation of metallic
palladium was not observed in the Schlenk tube, but it should
correspond to a palladium concentration decrease of only 20
ppm. Other tests at 80 °C under similar conditions confirmed
the intrinsic instability with a pH decrease and sometimes the
appearance of a small amount of metallic palladium.
At this pH, the propylene released cannot be explained by
phosphonium salt 3a decomposition, since 3a is stable under
the reaction conditions. Another hypothesis is a direct reaction
of the complex with water involving propylene and 2a formation
4. Stability of [Pd(η3-allyl)(TPPTS)2]+. An aqueous solu-
tion of 1a was prepared by the reaction of allylpalladium
chloride dimer with TPPTS (solution S1 of 1a at 0.0825 M) in
the presence of chloride at the same concentration and oxide
2a (3-5 mol %/Pd, pH around 7).
a. Stability at Room Temperature. The solution S1 was
stable for several months at room temperature under argon. No
reaction occurred in acidic medium according to NMR analysis

4308 Organometallics, Vol. 27, No. 17, 2008
Basset et al.
Scheme 4
basic media and temperature on its stability. Therefore, the
disappearance of the excess of TPPTS through phosphonium
salt formation and degradation reduces the stability of the
cationic palladium complex. The metallic palladium precipitation
in the Tsuji-Trost reaction or diene telomerization has probably
several causes. However, this study showed that, over a narrow
range of pH and temperature, the phosphonium salt of TPPTS
and the π-allyl cationic palladium complex were stable.
A second important point is the synthesis of new hydrosoluble
cationic allylic palladium complexes from trans-2-buten-1-ol,
prenol, and geraniol that are stable in water at room temperature.
Formally, these complexes could also result from the reaction
of corresponding dienes with Pd(TPPTS)3 and a proton from
the aqueous solution. A knowledge of the formation conditions
and reactivity of these palladium(II) complexes toward nucleo-
philes or water could afford a better understanding of the
catalyzed functionalization or hydration of butadiene into
butenyl or octadienyl compounds. These complexes and their
isomers could also govern the selectivity of the functionalization
reactions of dissymmetric dienes, such as isoprene or myrcene,
and of the isomerization reaction of butenols as well as that of
Tsuji-Trost reactions in water.
and an acidity increase (Scheme 4). One part of palladium
precipitates, the other part remains as complex 4 or unidentified
dimers.
c. Stability at 80 °C in the Presence of 3a. A solution
obtained from the reaction of 4 with a small excess of allyl
alcohol at pH 7.4 according to eq 5 ([1a] ) [3a] ) 0.011 M,
[allyl alcohol]remaining ) 0.003 M) was heated at 80 °C for 14 h.
After the reaction, the pH did not change and no propylene was
observed in the gaseous phase. The concentrations of 1a and
2a remained the same, but 3a was transformed (21%) into
phosphonium salt 3b. This conversion of about 1.5% h-1 is in
agreement with the intrinsic stability of 3a previously observed
(estimated at 1% h-1 at pH 7.5) (Figure 8).
The better stability of complex 1a in the presence of 3a was
confirmed by two comparison tests without allyl alcohol. The
addition of 3a reduced both the pH variation and propylene
release without palladium precipitation at 80 °C (Figure 9).
Experimental Section
1. Chemical Reagents. All the following reagents were used
as supplied by the manufacturer: Pd(OAc)2, allylpalladium chloride
dimer, crotylpalladium chloride dimer, allyl alcohol, crotyl chloride,
trans-2-buten-1-ol, cinnamyl alcohol, 3-methyl-2-buten-1-ol (pre-
nol), geraniol (Aldrich); HClO4, H2SO4, NaOH, CH3COONa,
CH3COOH (Laurylab); D2O (SDS). pH buffers 4.01, 7.01, and
10.01 at 25 °C were supplied by Bioblock (France). TPPTS and
TPPDS were gifts from Ruhrchemie-Celanese (TPPTS with TPPDS
5%, TPPTS oxide 3-5.2%; TPPDS with TPPDS oxide 1%).
TPPMS (with TPPMS oxide 1%) was prepared according to the
method of Chatt et al.41 The trisodium salt of allyltris(m-sulfophe-
nyl)phosphonium chloride 3a (with TPPTS oxide 5.2%, disodium
salt of allylbis(m-sulfophenyl)phenylphosphonium chloride 5%,
trisodium salt of propenyltris(m-sulfophenyl)phosphonium chloride
3b 0.6%) was prepared according to our previous report.24 Water
was distilled and degassed with argon.
Figure 9. Beneficial effect of 3a on 1a stability. In the second
reaction, 3a isomerization into 3b was observed (1.6%/14 h: i.e.,
0.11% h-1). Without palladium, the 3a isomerization rate in a pH
range 6.5-7.0 was estimated at 0.1% and 0.38% h-1 (see the
Supporting Information).
Another solution obtained from the reaction, according to eq
5 at pH 9 (1a ) 3a ) 0.01 M, [allyl alcohol]remaining ) 0.045
M), was heated for 17 h, at 80 °C under the same conditions.
After heating, the pH was 8.3 (0.03 equiv of H+/Pd), propylene
was evolved (0.027 equiv/Pd), and a slight darkening of the
solution was observed, meaning that a very small amount of
metallic palladium appeared in suspension and precipitated after
cooling. These results show that decomposition of the complex
1a in the presence of 3a occurs more rapidly at pH 9 than at
pH 7. The 31P NMR spectrum of the solution shows that 3a
isomerizes into 3b (3b/3a ) 98/2) after heating. We can propose
that the disappearance of the beneficial effect of 3a at higher
pH is due to its rapid thermal isomerization into 3b.
2. Measurements. 31P NMR, 1H NMR, and 13C NMR spectra
were recorded on a Bruker Avance 300 MHz spectrometer (external
reference for phosphorus analyses H3PO4 85%). The samples of
palladium TPPTS complexes or phosphorus compounds were
dissolved in 0.5 mL of D2O or 0.4 mL of water with 0.1 mL of
D2O.
Mass spectra were recorded on a ThermoFinnigan LCQ Advan-
tage ion trap instrument, detecting positive ions (+) or negative
ions (-) in the ESI mode. Samples (in methanol/dichloromethane/
water, 45/40/15 v/v/v containing 0.1% of formic acid) were infused
directly into the source (5 µL min-1) using a syringe pump. The
following source parameters were applied: spray voltage 3.0-3.5
kV, nitrogen sheath gas flow 5-20 arbitrary units. The heated
capillary was held at 200 °C.
These results are in agreement with palladium precipitation
observed in basic medium during phenol allylation24 and
butadiene hydromerization.7a
The pH measurements were performed with a PHN 330 T pH
meter and a combined electrode (Tacussel, France).
All experiments were performed under an inert atmosphere using
argon and using GAN syringes with 100 mm needles (Bioblock,
France). The oxygen concentration in water was measured with a
Bioblock Scientific 9071 oxymeter and found to be less than 0.1
ppm.
Conclusion
We demonstrated that the reaction between Pd(TPPTS)3 and
allylic alcohols led to the formation of two cationic compounds:
a palladium complex and an allylic phosphonium salt. The
π-allyl cationic palladium complex is stable over a wide range
of pH values at room temperature but is highly unstable upon
heating. A detailed investigation of degradation reactions of the
allylphosphonium salt of TPPTS showed the harmful effect of
The propylene concentration in the gas phase was determined
by GC analysis using an HP-5890 instrument with an HP-5 capillary
(41) Ahrland, S.; Chatt, J.; Davies, N. R.; Williams, A. A. J. Chem.
Soc. 1958, 264–276.

Splitting of Alcohols with Pd(0)-TPPTS in Water
Organometallics, Vol. 27, No. 17, 2008 4309
column (30 m × 0.25 mm) and an HP GCD series mass
spectrometer.
(3a). The Pd(TPPTS)3 signal at 23.85 ppm fully disappeared, and
the amount of OTPPTS 2a remained unchanged. The ratio of the
intensity of the signals was close to 1.20/2.00/1.00 (2a/1a/3a).
3. Pd(TPPTS)3 Water Solution Preparation. Pd(TPPTS)3
solution (4) was prepared from Pd(OAc)2 + nTPPTS (4 < n <
4.5) by a redox process: the excess of TPPTS reduced palladium(II)
to palladium(0). The pH was controlled by NaOH additions.27a At
the end of the reaction, the pH was brought to 7.0-7.2. Conditions:
[Pd] ) 0.06-0.10 M, OTPPTS/Pd molar ratio 1.2, TPPTS/Pd )
3-3.5, sodium acetate/Pd ) 2 (using D2O and NaOH/D2O, a D2O
Pd(TPPTS)3 solution was prepared, [Pd] ) 0.053 M).
5. Reaction of the Allylphosphonium Chloride Salt of
TPPTS with Pd(TPPTS)3 at Room Temperature. In a 10 mL
Schlenk tube equipped with a magnetic stirrer, 0.1 mmol of 3a
was dissolved in 1.5 mL of water and 0.5 mL of D2O. A 0.5 mL
portion of Pd(TPPTS)3 aqueous solution (0.059 M) was added. The
31P NMR spectra of the aqueous solution remain unchanged after
20 min or 1 day or after heating for 3 h at 55 °C (equilibrium
among 3a, 4, 1a, and TPPTS).
4. Typical Allylic Compound Reaction with Pd(TPPTS)3
at Room Temperature. A 20 mL Schlenk tube was equipped with
a pH electrode. Distilled water and aqueous Pd(TPPTS)3 solution
were added,42 and the pH adjusted to 743 with 0.2 M NaOH or 0.2
M HClO4 (volume of the solution 5 mL, [Pd] ) 0.0055 M). The
allylic compound was then added with a syringe (except cinnamyl
alcohol, which is solid). The solution was homogeneous, except
for cinnamyl alcohol and geraniol, which are not soluble enough
at 0.055 M. Within a few seconds, the pH increased. The 0.2 M
HClO4 was added with GAN syringes to keep the pH at 744,45 with
variations between 6.8 and 7.2. After 100 min, the pH became
stable. A 0.026 mL portion of HClO4 (2 equiv/Pd) was poured in
(solution S1 with allyl alcohol).
6. Stability of [Pd(η3-allyl)(TPPTS)2]+(1a). a. Slow
Decomposition. A 0.1 mL portion of 1 M NaOH was added to 5.5
mL of a 0.0071 M 1a solution (pH 7.5) in 60 min to reach pH 12.2
and then 0.2 mL to reach pH 12.5. A blackening of the solution
occurred; palladium precipitated (see section 4a).
b. At pH 13.87 (0.75 M NaOH Solution). A 0.5 mL portion
of a 0.0825 M 1a solution in D2O was added to 1.5 mL of 1 M
NaOH in D2O. Palladium precipitated immediately. The gaseous
phase contained 12 mol % of propylene/1a (see section 4a).
c. At 80 °C. Portions of the solutions amounting to x mL were
poured in a Schlenk tube. After y hours at 80 °C and cooling to
room temperature, propylene was analyzed in the gaseous phase,
the pH was measured, and water was analyzed: (i) x ) 2 mL (1a
0.016 M, pH 7.6), y ) 14 h (final pH 3.97) (see section 4b); (ii) x
) 5 mL (1a 0.01 M + 3a 0.012 M from the reaction of 4 with
allyl alcohol in low excess [allyl alcohol] ) 0.003 M, pH 7.4), y
) 12 h (final pH 7.40) (see section 4c); (iii) x ) 5 mL (1a 0.01 M,
pH 6.93), y ) 14 h (final pH 5.91) (see section 4c); (iv) x ) 5 mL
(1a 0.01 M + 3a 0.01 M, pH 6.98), y ) 14 h (final pH 6.80) (see
section 4c); (v) x ) 5 mL (1a 0.01 M + 3a 0.01 M from the reaction
of 4 with allyl alcohol in excess [allyl alcohol] ) 0.045 M, pH
9.0), y ) 17 h (final pH 8.3) (see section 4c).
The 31P NMR spectrum of solution S1 (0.4 mL) + D2O (0.1
mL) exhibited two new signals: 25.61 ppm (1a)46 and 22.31 ppm
(42) With an excess of TPPTS (4 equiv of TPPTS/Pd), 6 equiv of HClO4/
Pd was added. Five equivalents of TPPTS/Pd was transformed into 3a.
(43) Or at pH 3 and 9 with allyl alcohol (pH influence study).
(44) Geraniol and prenol reactions were slow at pH 7. The reactions
were faster at pH 6 and 5. Total conversions were obtained at pH 4.5.
(45) With allyl alcohol (R ) 10) without acid addition, the pH was
increased and measured over 5.5 h. 31P NMR spectrum after 2 h: 1a/4,
15%; 3a/4, 4%. After 1 day, the pH of the solution was 11.65 and propylene
was identified in the gas phase. The conversion of 4 was 100% into 1a. 3a
had disappeared, but the amount of TPPTS oxide 2a present in the solution
had increased (2a/1a ) 2.20/1).
(46) Since TPPTS contained 5% of TPPDS, a small concentration of
[Pd(allyl)(TPPTS)(TPPDS)]+ was present: 31P NMR peaks at 25.98, 25.61
ppm (major) and 25.22, 24.97 ppm. By addition of TPPDS to the allylation
reaction of 4 by allyl alcohol (TPPDS/TPPTS ) 1.6/3), an increase of signals
due to TPPDS was actually observed. These peaks were also observed after
the reaction of allylpalladium dimer and TPPTS. .
(47) Homologous compounds with triphenylphosphine, 3a′-c′ and
2b′,c′, were already synthesized with available 1H NMR analysis (except
2c′). 3a′, allyltriphenylphosphonium chloride, was synthesized from the TPP
reaction with allyl chloride,47a 3b′, 2-propenyltriphenylphosphonium bro-
mide, from 3a′ bromide (prepared from TPP and allyl bromide) isomer-
ization in a refluxing 2-methoxyethanol solution,47b 3c′, 2-hydoxypropyl-
triphenylphosphonium hydroxide, from TPP reaction with propylene
oxide,47c 3c′ iodide, from TPP reaction with propylene oxide, NaI, and
carbon dioxide,47d (d); 2b′, propenyldiphenylphosphine oxide, from
(C6H5)2POCl and 1-bromopropene Grignard reagent,47e and 2c′, 2-hydox-
ypropyldiphenylphosphine oxide, from (C6H5)2HPO reaction with propylene
carbonate:47f (a) Buchi, G.; Wu¨est, H, HelV. Chim. Acta 1971, 188-189,
1681–1687. (b) Falvello, L. R.; Fernandez, S.; Larraz, C.; Llusar, R.;
Navarro, R.; Urriolabeitta, E. P. Organometallics 2001, 20, 1424–1436.
(c) Christol, H.; Christau, H.; Soleiman, M. Tetrahedron Lett. 1976, 37,
3321–3324. (d) Huang, J.; Sin, M. J. Org. Chem. 2003, 68 (17), 6705–
6709. (e) Ducan, M.; Gallagher, M. J. Org. Magn. Reson. 1981, 15 (1),
37–42. (f) Tarpey, P, K.; Martin, G. J. Org. Chem. 1962, 27, 1817–1823.
Acknowledgment. We gratefully acknowledge the Ecole
Supe´rieure de Chimie Physique Electronique de Lyon
(ESCPE Lyon) and CNRS for the financial support and
Ruhrchemie (Hoechst) for a gift of 200 g of TPPTS and 10 g
of TPPDS. We also thank Alessandra Quadrelli, Catherine
Santini, Christophe Cope´ret, Christine Clozel, and Olivier
Vittori for helpful discussions.
Supporting Information Available: Text and figures giving
experimental procedures and spectral and analytical data for the
products of the reaction of Pd(TPPTS)3 with trans-2-buten-1-ol,
prenol, cinnamyl alcohol, geraniol and allyl alcohol as limiting
reagent and for products obtained from the allylphosphonium salt
of TPPTS decomposition and data for the estimation of the
allylphosphonium salt of TPPTS decomposition rate and preparation
details and characterization data for homologous complexes with
TPPTS (1a), TPPDS (1b), and TPPMS (1c). This material is
available free of charge via the Internet at http://pubs.acs.org.
OM701273J