Synthesis of new hemilabile amphiphilic phosphines. Complexing properties toward ruthenium(II) and catalytic activity for hydrogenation of prenal

Article abstract of DOI:10.1021/om9905916

New functionalized phosphines R-(C₆H₄)(OCH₂CH₂)nPPh₂ {1 (R = tert-octyl, n = 1), 2 (R = tert-octyl, n? = 5), 3 (R = tert-octyl, n? = 13), 4 (R = n? -nonyl, n? = 1.4), 5 (R = n-nonyl, n? = 5), 6 (R = n-nonyl, n? = 11)}, CH₃(OCH₂CH₂)₃PPh₂ (7), CH₃(OCH₂CH₂)₃PPhR (8, R = isopentyl), and HOCH₂CH₂(OCH₂CH₂)₂PPhR (9, R = n-octyl) have been synthesized. These ligands contain polyether and hydrophobic groups in the same molecule and, therefore, can lead to compounds that combine hemilabile (polyether groups) and amphiphilic (hydrophobic and hydrophilic groups) properties. Complexation studies with Ru(II) have been performed, and the unusual tridentate (O, O, P) coordination mode has been characterized. The hemilabile character has been shown by evolution from the tridentate (O, O, P) coordination mode to the bidentate (O, P) and monodentate (P) after a reaction between CO and [RuCl₂(8)-(PPh₃)]. Dinuclear [Ru₂Cl₄L₅] complexes have been obtained from ligands with a short polyether chain (1, 4). The catalytic selective hydrogenation of an α,β-unsaturated aldehyde (prenal) with ruthenium complexes prepared in situ from RuCl₃ and ligands 1-6 has been carried out in a 2-propanol/water mixture. Conversions to prenol to the order of 90% or higher are observed with selectivity of 90-96% after 20 min at P = 30 bar and T = 50°C. The most active systems have been observed with ligands containing long polyether chains.

Full text of DOI:10.1021/om9905916

Organometallics 1999, 18, 5475-5483
5475
Syn th esis of New Hem ila bile Am p h ip h ilic P h osp h in es.
Com p lexin g P r op er ties tow a r d Ru th en iu m (II) a n d
Ca ta lytic Activity for Hyd r ogen a tion of P r en a l
Esteve Valls and J oan Suades*
Departament de Quı´mica, Edifici C, Universitat Auto`noma de Barcelona,
08193 Bellaterra, Spain
Rene´ Mathieu
Laboratoire de Chimie de Coordination du CNRS, 205 Route de Narbonne,
31077 Toulouse Cedex 4, France
Received J uly 27, 1999
New functionalized phosphines R-(C6H4)(OCH2CH2)nPPh2 {1 (R ) tert-octyl, n ) 1), 2 (R
) tert-octyl, nj ) 5), 3 (R ) tert-octyl, nj ) 13), 4 (R ) n-nonyl, nj ) 1.4), 5 (R ) n-nonyl, nj )
5), 6 (R ) n-nonyl, nj ) 11)}, CH3(OCH2CH2)3PPh2 (7), CH3(OCH2CH2)3PPhR (8, R )
isopentyl), and HOCH2CH2(OCH2CH2)2PPhR (9, R ) n-octyl) have been synthesized. These
ligands contain polyether and hydrophobic groups in the same molecule and, therefore, can
lead to compounds that combine hemilabile (polyether groups) and amphiphilic (hydrophobic
and hydrophilic groups) properties. Complexation studies with Ru(II) have been performed,
and the unusual tridentate (O, O, P) coordination mode has been characterized. The
hemilabile character has been shown by evolution from the tridentate (O, O, P) coordination
mode to the bidentate (O, P) and monodentate (P) after a reaction between CO and [RuCl2(8)-
(PPh3)]. Dinuclear [Ru2Cl4L5] complexes have been obtained from ligands with a short
polyether chain (1, 4). The catalytic selective hydrogenation of an R,â-unsaturated aldehyde
(prenal) with ruthenium complexes prepared in situ from RuCl3 and ligands 1-6 has been
carried out in a 2-propanol/water mixture. Conversions to prenol to the order of 90% or
higher are observed with selectivity of 90-96% after 20 min at P ) 30 bar and T ) 50 °C.
The most active systems have been observed with ligands containing long polyether chains.
In tr od u ction
active agent, yielding metal complexes with similar
properties. Therefore, amphiphilic metal complexes can
accumulate at the interfaces and, at the same time, can
aggregate to form supramolecular systems.4 This micro-
heterogeneous arrangement of metal centers in the
reaction medium can be advantageous for some catalytic
processes.5 Some examples of ligands with amphiphilic
properties have been reported, most of them amphiphilic
phosphines.3,6
Another class of functionalized ligands that are
strongly related with homogeneous catalysis are hemi-
labile phosphorus-oxygen ligands.7 These can lead to
metal complexes with weak metal-oxygen bonds, which
may create reversible empty coordination sites favorable
for catalytic activity.
The design and synthesis of new functionalized
ligands, capable of improving certain properties in
transition metal complexes, has become a central topic
in the development of inorganic and organometallic
chemistry. Phosphines are one of the most versatile
ligands, since they can lead to stable metal-phosphorus
bonds and, at the same time, very different functional-
izations may be introduced by means of the groups
bound to the phosphorus atom. Consequently, function-
alized phosphine ligands have become a field of interest,
which has led to important industrial applications such
as water-soluble1 and asymmetric phosphines.2
In recent years, some authors have shown interest
in searching for amphiphilic ligands, which contain
hydrophilic and hydrophobic groups, since they may
lead to metal complexes with special and interesting
properties.3 These ligands can anchor a transition metal
atom in a molecule with the properties of a surface-
(4) Van Esch, J . H.; Stols, A. L. H.; Nolte, R. J . M. J . Chem. Soc.,
Chem. Commun. 1990, 1658-1659.
(5) (a) Hanson, B. E.; Ding, H.; Kohlpaintner C. W. Catal. Today
1998, 42, 421-429. (b) Oehme, G.; Grassert, I.; Ziegler, S.; Meisel, R.;
Fuhrman, H. Catal. Today 1998, 42, 459-470. (c) Selke, R.; Holz, J .;
Riepe, A.; Borner, A. Chem. Eur. J . 1998, 4, 769-771.
(6) (a) Fell, B.; Papadogianakis, G. J . Mol. Catal. 1991, 66, 143. (b)
Ding, H.; Hanson, B. E.; Bartik, T.; Bartik, B. Organometallics 1994,
13, 3761. (c) Brauer, D. J .; Fischer, J .; Kucken, S.; Langhans, K. P.;
Stelzer, O.; Weferling, N. Z. Naturforsch. 1994, 49b, 1511. (d) Okano,
T.; Harada, N.; Kiji, J . Chem. Lett. 1994, 6, 1057. (e) Ding, H.; Hanson,
B. E.; Bakos, J . Angew. Chem., Int. Ed. Engl. 1995, 34, 1645. (f) Gulyas,
H.; Arva, P.; Bakos, J . Chem. Commun. 1997, 2385-2386.
(1) Aqueous-Phase Organometallic Catalysis; Cornils, B., Herrmann,
W. A., Eds.; Wiley-VCH: Weinheim, 1998; Chapter 3.2.
(2) (a) Kagan, H. B. In Asymmetric Synthesis; Morrison, J . D., Ed.;
Academic Press: Orlando, 1985; Vol. 5, Chapter 1. (b) Brunner, H.;
Zettlmeier, W. Handbook of Enantioselective Catalysis with Transition
Metal Compounds; VCH: Weinheim, 1993; Vols. I and II.
(3) Papadoginakis, G.; Sheldon, R. A. In ref 1, Chapter 3.2.4.
10.1021/om9905916 CCC: $18.00 © 1999 American Chemical Society
Publication on Web 11/25/1999

5476 Organometallics, Vol. 18, No. 26, 1999
Valls et al.
Sch em e 1
In this context, we decided to explore the synthesis
of new functionalized phosphines that feature the
combination of both properties described above: amphi-
philic (bound to hydrophilic and hydrophobic groups)
and hemilabile (bound to ether groups). The synthesis
of these new ligands was designed using poly(ethylene
glycol) monoalkyl ethers as starting materials for the
following reasons. (1) These compounds are common
since they are used extensively as nonionic surfactants
and interesting ligands may be prepared by straight-
forward methods. (2) The hydrophilic/hydrophobic char-
acter and the structure of the hydrophobic group may
be easily modified by using poly(ethylene glycol)
monoalkyl ethers with different alkyl and polyether
chains as starting material with minor changes in the
method of synthesis. This last point is important in
order to modulate the characteristics of ligands with the
aim of optimizing a catalytic process. To our knowledge,
only a very few examples of ligands with related
characteristics have been reported: water-soluble cata-
lysts based on functionalized poly(ethylene glycol)8 and
ethoxylated tris(p-hydroxyphenyl)phosphines which ex-
hibit thermoregulated phase-transfer function.9
In this paper we report (1) the synthesis and char-
acterization of nine new amphiphilic phosphines, (2) the
study of complexing properties of synthesized ligands
toward ruthenium(II), (3) an unprecedented mode of
bonding of an ether-phosphine ligand, and (4) the use
of some reported ligands, associated with Ru(II), in the
catalytic selective hydrogenation of R,â-unsaturated
aldehydes. Part of this work has been communicated
previously.10
were characterized by NMR spectroscopy. The more
relevant data are the 31P spectra (a signal at δ ≈ -21
ppm) and the shifts and coupling with the phosphorus
nucleus observed in 1H and 13C spectra for methylene
groups placed in the R and â positions with respect to
the phosphorus atom. It should be pointed out that
ligand 1 was isolated as a single compound (although
the starting material was the nonionic surfactant IG-
EPAL CA210, which is a mixture of polyethethylene
glycol monoalkyl ethers (nj ) 1.5), only the pure com-
pound (CH3)3CCH2C(CH3)2C6H4O(CH2)2P(C6H5)2 was
obtained after crystallization). Ligands 2-6, however,
were obtained as mixtures of compounds with the same
structure but with different ethoxylation grades in
accordance with the nature of the initial nonionic
surfactants. The average n value for each ligand was
Resu lts a n d Discu ssion
Syn th esis of liga n d s. Ligands 1-6 were synthesized
in two steps starting from the corresponding commercial
nonionic surfactants IGEPAL11 as shown in Scheme 1.
In the first step, alkyl chlorides were obtained from
alcohols in high yields by the reaction of the poly-
(ethylene glycol) monoalkyl ethers with an excess of
triphenylphosphine in carbon tetrachloride. The remain-
ing triphenylphosphine and the triphenylphosphine
oxide were easily isolated as insoluble solids in the
reaction medium. Phosphines 1-6 were prepared by
reacting the appropriate alkyl chloride with lithium
diphenylphosphide at 0 °C in diethyl ether. Ligand 1
crystallizes as a white solid in ethanol, while ligands
2-6 were obtained as pale yellow oils. Ligands 1-6
1
established by H NMR spectroscopy and is shown in
Scheme 1.
Ligands 7-9 (Scheme 2) were prepared in order to
have ligands available that were related to 1-6 with
some modifications in their molecular structure. Ligand
7 is interesting for purposes of comparison since it is
similar to 1-6 but without the hydrophobic chain.
Ligands 8 and 9 have the phosphorus atom placed
between the polyether and the hydrophobic chains.
Ligand 7 was synthesized from triethylene glycol mono-
methyl ether by a method similar to ligands 1-6.
Ligands 8 and 9 were prepared in two steps as shown
in Scheme 2. In the first step, the isopentyl and n-octyl
diphenylphosphines were prepared by reacting lithium
diphenylphosphide with isopentyl bromide and octyl
chloride, respectively. Second, reductive cleavage of one
phosphorus-aryl bond with lithium led to the corre-
sponding alkylarylphosphide, which reacted with tri-
ethylene glycol monomethyl ether chloride and trieth-
ylene glycol monochlorohydrin to yield 8 and 9, respec-
tively. It should be emphasized that ligand 9 was
achieved in good yield from triethylene glycol monochlo-
rohydrin by direct reaction with butyllithium without
(7) (a) Bader, A.; Lindner, E. Coord. Chem. Rev., 1991, 108, 27, and
references therein. (b) Lindner, E.; Pautz, S.; Haustein, M. Coord.
Chem. Rev. 1996, 155, 145. (c) Lindner, E.; Pautz, S.; Fawzi, R.;
Steimann, M. J . Organomet. Chem. 1998, 555, 247-253. (d) Le Gall,
I.; Laurent, P.; Soulier, E.; Salau¨n, J .; Abbayes, H. J . Organomet.
Chem. 1998, 567, 13-20.
(8) Karakhanov, E. A.; Kardasheva, Y. S.; Maksimov, A. L.; Pre-
deina, V. V.; Runova, E. A.; Utukin, A. M. J . Mol. Catal. A: Chem.
1996, 107, 235-240.
(9) (a) J in, Z.; Zheng, X.; Fell, B. J . Mol. Catal. A: Chem. 1997, 116,
55. (b) Chen, R.; Liu, X.; J in, Z. J . Organomet. Chem. 1998, 571, 201-
204. (c) Zheng, X.; J iang, J .; Liu, X.; J in, Z. Catal. Today 1998, 44,
175-182.
(10) Valls, E.; Suades, J .; Donadieu, B.; Mathieu, R. Chem. Commun.
1996, 771.
(11) These products are ethoxylated alkylphenols with different
ethoxylation grades. Since they are not pure compounds, the length
of the polyether chains displayed in Scheme 1 are actually the mean
value of a mixture of compounds with the same structure but with
different ethoxylation grades.

New Hemilabile Amphiphilic Phosphines
Organometallics, Vol. 18, No. 26, 1999 5477
Sch em e 2
Sch em e 3
it can barely manifest amphiphilic properties since only
one oxygen atom is present in the polyether chain, and
finally, it is the sole ligand prepared that crystallizes
as a white solid. Thus, the reaction between ligand 1
and RuCl3‚3H2O, in the same conditions as previously
described for ligands 2-9, affords a brown microcrys-
talline solid. Its 31P NMR spectrum at room temperature
shows a singlet at 42.2 ppm and a broad AB2 pattern
at 23.7 and 29.5 ppm. Different recrystallizations of this
complex afforded solids, which showed identical spectra
with the same integration ratio between the three
resonances (2:1:2). These results are in accordance with
the assignment of these signals to different phosphorus
atoms of a unique compound and suggest a polynuclear
nature for this complex. The elemental analysis is also
in agreement with this hypothesis and suggests a
stoichiometry [Ru2Cl4(1)5]. Polynuclear ruthenium(II)
complexes have been extensively described in the lit-
erature,14 as well as the high stability of six-coordinated
Ru(II) complexes containing a triply bridge chloride,
which is particularly favorable with tertiary phosphine
ligands without bulky groups.15 Hence, we propose for
this complex a structure similar to those previously
reported for other dinuclear Ru(II) complexes (Scheme
3) such as [Ru2Cl4L5] (L ) PPh2Et, PPhEt2),15 [(PPh3)2-
(PF3)Ru(µ-Cl)3RuCl(PPh3)2],16 and [(L)(PkP)Ru(µ-Cl)3-
RuCl(PkP)] (PkP ) diphosphine).17
The broad AB2 pattern observed at room temperature
induced us to do low-temperature NMR experiments
(Figure 1). At 193 K we observed the splitting of the
singlet at 42.2 ppm to an AB system and the AB2
pattern evolves to an ABC spin system. This is consis-
tent with the proposed structure in Scheme 3 if we
admit that the ligand conformation suppresses the
equivalence of the two P(a) and the two P(b) nuclei. This
inequivalence could be favored by hydrophobic interac-
tions between tert-octyl groups as well as by steric
hindrance.
the use of alcohol protective groups.12 Ligands 7-9 were
obtained as colorless oils and were characterized by
NMR spectroscopy.
Ru th en iu m (II) Com p lexes. Previous studies of
ether phosphine complexing properties toward Ru(II)7
have shown their ability to act as mono (P) or bidentate
(O, P) ligands consistent with their hemilabile character.
In contrast to previously studied ligands, 2-9 contain
at least two oxygen atoms which can be simultaneously
coordinated to the metal atom so a tridentate (O, O, P)
or higher coordination mode is possible.
Since ligands 2-6 are mixtures of compounds with
the same structure but with different ethoxylation
grades, the study of their coordination properties has
been limited to NMR spectroscopy in solution. In
contrast, ligands 1, 7, 8, and 9 are not mixtures, so their
study is less restricted than that of 2-6. Moreover, their
study can supply information that can be useful to
understand the behavior of ligands 2-6.
Rea ction s w ith Ru Cl3‚3H2O. The reaction of RuCl3‚
3H2O with a 3-fold excess of 1-9 in 2-propanol/H2O (95:
5) did not lead to the isolation of solids with the
exception of 1. The 31P NMR spectra of the solutions
obtained from the reaction with 2-9 show multiple
resonances at 20-60 ppm, some of them being broad.
This observation suggests the formation of mixtures
after reaction with RuCl3‚3H2O and is consistent with
the complicated stereochemistry described in the lit-
erature for the hemilable P-O ligands with the RuCl2
fragment.7
Rea ction s w ith Ru Cl2(P P h 3)3. The complexity de-
tected in the reactivity of our ligands with RuCl3‚3H2O
induced us to search for another approach in order to
simplify the coordination study. With this aim, we
undertook the phosphine substitution in the ruthenium-
(II) complex RuCl2(PPh3)3 with ligands 1-9. Previous
studies of this reaction with ether phosphines revealed
the formation of trans-dichloro-cis-bis(ether phosphine)-
ruthenium(II) complexes as shown in Scheme 4.18
(13) Holz, J .; Quirmbach, M.; Borner A. Synthesis 1997, 985.
(14) Comprehensive Coordination Chemistry; Wilkinson, G., Ed.;
Pergamon: Oxford, 1987; Vol. 4, Chapter 45.5.5.1, and references
therein.
(15) Armit, P. W.; Boyd, A. S. F.; Stephenson, T. A. J . Chem. Soc.,
Dalton Trans. 1975, 1663.
The chemical behavior of ligand 1 is quite different
from ligands 2-9 in accordance with their physical and
chemical properties. For ligand 1 the bidentate (P, O)
coordination is hindered by an aryl group. In addition,
(16) Head, R. A.; Nixon, J . F. J . Chem. Soc., Dalton Trans. 1978,
901.
(12) Some examples of hydroxy phosphine preparation from hydroxy
compounds without the use of OH-protective groups have been
previously described.13 The experimental method used for the synthesis
of 9 from (n-octyl)PPh2 enables the phosphine to be obtained in one
step with good yield and purity.
(17) (a) Fogg, D. E.; J ames, B. R. Inorg. Chem. 1995, 34, 2557. (b)
MacFarlane, K. S.; Thorburn, P. W.; Cyr, P. W.; Daniel, E. K.; Chau,
Y.; Rettig, S. J .; J ames, B. R. Inorg. Chim. Acta 1998, 270, 130.
(18) Lindner, E.; Mo¨ckel, A.; Mayer, H. A.; Ku¨hbauch, H.; Fawzi,
R.; Steimann, M. Inorg. Chem. 1993, 32, 1266.

5478 Organometallics, Vol. 18, No. 26, 1999
Valls et al.
lished for ligand 8,10 but other possibilities with two
nonequivalent phosphorus atoms should also be con-
sidered.
In accordance with their chemical nature,19 ligands
1 and 4 show a substantially different behavior from
ligands 2, 3, 5, and 6. The 31P NMR spectrum of the
solution obtained after reaction of RuCl2(PPh3)3 with 1
denotes the presence of (a) free triphenylphosphine, (b)
characteristic signals of the previously observed [Ru2-
Cl4(1)5] complex in the direct reaction between RuCl3‚
3H2O and 1, and (c) traces of an unidentified complex
characterized by two doublets at δ ) 50.7 and 38.3 ppm
(J PP ) 39 Hz).
Ligand 4 leads to a complicated 31P NMR spectrum
that suggests the presence of a mixture of ruthenium
complexes. Some remarkable signals are a broad band
at 58 ppm (δ similar to those observed for ligands with
a long polyether chain 2, 3, 5, 6), broad resonances
between 25 and 45 ppm (δ similar to that observed with
ligand 1, which has no polyether chain), and a resonance
at 64 ppm (δ similar to trans-dichloro-cis-bis(ether
phosphine)ruthenium(II) complexes (Scheme 4). This
singular behavior may be related to the composition of
4,19 which contains molecules with short polyether
chains (in comparison with 2, 3, 5, and 6) coexisting with
the molecule with the phenolic oxygen as the only
oxygen atom (similar to 1).
In contrast to the fact that solid compounds could not
be isolated from the reaction between 1-6 and RuCl2-
(PPh3)3, ligands 7-9 lead to the formation of dark red
solutions, from which solid compounds were isolated.
Single crystals were only obtained from ligand 8, and
the X-ray structure was determined and has been
communicated previously.10 The structure of this com-
plex, [RuCl2(8)(PPh3)], is outlined in Scheme 6, showing
the coordination around the ruthenium atom with the
two chlorine atoms in trans position, the ligand 8 η3
bonded through the phosphorus and two oxygen atoms
in meridional position, and the triphenylphosphine in
the remaining coordination site.
F igu r e 1. 31P{1H} NMR spectra of [Ru2Cl4(1)5] at different
temperatures.
Sch em e 4
The most remarkable features of this molecular
structure are the unprecedented η3 (O, O, P) coordina-
tion mode of the hemilabile ligand 8 and the long Ru-O
distance (2.436(7) Å) between the metal and the oxygen
atom in trans position to the phosphorus atom of ligand
8. The significant difference between the two Ru-O
distances20 in the complex [RuCl2(8)(PPh3)] suggests
that this structural difference could be reflected in a
different labile character of the two oxygen atoms. To
check this hypothesis, the reactivity of [RuCl2(8)(PPh3)]
with PBu3 and CO was studied by 31P NMR spectros-
copy (Scheme 7). After adding 1 mol of PBu3 to [RuCl2(8)-
(PPh3)], the dichloromethane red solution immediately
turned green. However, the 31P NMR spectrum at room
temperature showed the resonances of [RuCl2(8)(PPh3)]
and free triphenylphosphine, suggesting that the newly
formed complexes were fluxional. By lowering the
temperature to 233 K, two groups of new sharp reso-
nances were observed: for one, two doublets of the same
intensity at δ 69.5 and 42.4 ppm (J PP ) 42 Hz), for
another, a triplet at δ 70.6 ppm and a doublet at δ 16.3
ppm (J PP ) 29 Hz) with a 1:2 intensity ratio, consistent
with the formation of [RuCl2(8)(PBu3)] and [RuCl2(8)-
As mentioned in the previous section, the study of this
reaction with 2-6 was limited to NMR spectroscopy on
account of the nature of these ligands. The reaction in
dichloromethane between 1-6 and RuCl2(PPh3)3 in a
molar ratio 2:1 leads to the formation of dark red
solutions. The 31P NMR spectra of these solutions show
a signal at -5 ppm assigned to free triphenylphosphine,
and there is no trace of free ligands 1-6. These results
are in agreement with the phosphine substitution
reaction described above. Nevertheless, previously re-
ported 31P NMR spectra of trans-dichloro-cis-bis(ether
phosphine)ruthenium(II) complexes19 show a single
resonance at δ ≈ 64 ppm, while with ligands 2, 3, 5,
and 6 a quite broad resonance at δ ≈ 58 ppm with the
characteristic shape of an AB system is observed (J ≈
40 Hz). Unfortunately, these compounds are waxy, and
it was not possible to determine an X-ray structure; so
based on the NMR results, the following possibilities
are tentatively proposed (Scheme 5). The η3 (POO)
coordination is attractive since it has been well estab-
(19) Ligand 1 is a pure compound with a sole oxygen atom in the
phenoxy group. Although ligand 4 is a mixture of molecules, it also
contains a molecule with a sole oxygen atom in the phenoxy group. In
addition, the other molecules with more oxygen atoms display polyether
chains shorter than 2, 3, 5, and 6.
(20) Ru-O (trans to PPh3) ) 2.191(6), Ru-O (trans to 8) ) 2.436(7).

New Hemilabile Amphiphilic Phosphines
Organometallics, Vol. 18, No. 26, 1999 5479
Sch em e 5
Sch em e 6
(PBu3)2], respectively, as shown in Scheme 7. The
unanticipated replacement of PPh3 by PBu3 should be
pointed out, since this result is contrary to the expected
addition of PBu3 to [RuCl2(8)(PPh3)] by displacement
of the most labile oxygen atom. Nevertheless, we cannot
rule out that the first step of the reaction mechanism
could be the opening of the Ru-O bond, consistent with
the hemilabile character of the oxygen atom, followed
by PBu3 addition.
tants.21 It has been postulated that hydrophobic inter-
actions perform a significant role when the hydrophobic
chain is constituted by five carbon atoms or more,22 so
this result may be perceived as a sign of the contribution
of hydrophobic interactions in the chemical behavior of
amphiphilic metal complexes.
In contrast to the unexpected substitution of PPh3 by
PBu3, the reaction of [RuCl2(8)(PPh3)] with CO is
consistent with the predicted hemilabile character for
ligand 8. Indeed, carbon monoxide was bubbled in a
dichloromethane solution of [RuCl2(8)(PPh3)] at room
temperature, and after a few minutes the red solution
turned yellow. The 31P NMR spectrum showed two
sharp doublets at δ 11.7 and 26.4 ppm (J PP ) 251 Hz),
indicating the existence of two different phosphorus
atoms in trans position. The IR spectrum shows a single
band in the ν(CO) region (2005 cm-1) consistent with
two trans CO ligands. All these facts support the
hypothesis that the two Ru-O bonds are open after the
addition of two CO ligands to [RuCl2(8)(PPh3)] to form
the all-trans-[RuCl2(CO)2(8)(PPh3)] complex (Scheme 7).
A dichloromethane solution of this complex was heated
to reflux, and the new 31P NMR spectrum also showed
two phosphorus atoms in trans position (J PP ) 292 Hz),
now at 45.8 and 48.0 ppm. The IR spectrum of this
solution showed a band at 1970 cm-1 in the ν(CO)
region, in agreement with a CO in trans position to
chloride.18 These data are consistent with the formation
of [RuCl2(CO)(8)(PPh3)], in which the two phosphines
are trans and the CO ligand is in trans position to
chloride (Scheme 7). The changes observed in the
coordination mode of ligand 8 after the reaction between
[RuCl2(8)(PPh3)] and CO should be emphasized. It
moves successively from η3 (POO) in the initial complex
to η1 (P) in [RuCl2(CO)2(8)(PPh3)] and to η2 (PO) in
[RuCl2(CO)(8)(PPh3)], revealing the coordination ver-
satility of ligand 8. The reactivity of [RuCl2(8)(PPh3)]
displayed in Scheme 7 can be associated with the steric
and electronic features of CO and PBu3.
With regard to the reaction between RuCl2(PPh3)3 and
9, a solid compound was obtained which shows a 31P
NMR spectrum in solution that is closely related with
the spectrum of the above complex [RuCl2(8)(PPh3)]:
two sharp doublets at δ ) 63.3 and 58.8 ppm (J ) 43
Hz). This similar chemical behavior can be understood
considering their similar chemical structure, since both
are dialkylaryl phosphines bonded to a hydrophobic
chain and to a short polyether chain. Finally, a solution
of the solid compound obtained after the reaction
between 7 and RuCl2(PPh3)3 displays a 31P NMR
spectrum with a singlet resonance at δ ) 62.3 ppm,
concordant with the results obtained with other ether
phosphines (Scheme 4) and in contrast to the behavior
described above for all the other ligands reported in this
work. This result should be emphasized because 7 is
the ligand that is most similar to previously reported
ether phosphines. It contains only two more ethylene
glycol groups with respect to Ph2PCH2CH2OCH3.18
Consequently, there could be a possible influence of the
hydrophobic group on the coordination properties of
amphiphilic ligands, in contrast with the role of the
polyether chain length. Thus, after reacting with RuCl2-
(PPh3)3, ligands with a hydrophobic group and different
polyether chains (2, 3, 5, and 6) tend to exhibit similar
31P NMR spectra, which are different from spectra
obtained with ligands without a hydrophobic group (7
and Ph2PCH2CH2OCH3).
Ca ta lysis. Selective hydrogenation of R,â-unsatur-
ated aldehydes is an attractive homogeneous catalytic
process because the corresponding unsaturated alcohols
The last salient feature of complex [RuCl2(8)(PPh3)]
is shown by the molecular arrangement in the crystal
structure. As can be seen in Figure 2, RuCl2(8)(PPh3)
molecules are disposed in such a way that hydrophobic
isopentyl groups are close together. The shortest inter-
molecular C-C distances (4.04 Å) are similar to those
previously reported for compounds that exhibit repre-
sentative hydrophobic interactions such as surfac-
(21) (a) Okuyama, Watanabe, H.; Shimomura, M.; Hirabayashi, K.;
Kunitake, T.; Kajiyama, T.; Yasuoka, N. Bull. Chem. Soc. J pn. 1986,
59, 3351. (b) Okuyama, K.; Niijima, N.; Hirabayashi, K.; Kunitake,
T.; Musunoki, M. Bull. Chem. Soc. J pn. 1988, 61, 1485.
(22) Desiraju, G. R. In Comprehensive Supramolecular Chemistry;
Atwood, J . L., Davies, J . E., MacNicol, D. D., Vo¨gtle; Lehn, J . M., Eds.;
Pergamon: New York, 1996, Vol. 6, p 5.

5480 Organometallics, Vol. 18, No. 26, 1999
Valls et al.
Sch em e 7
preparation of catalysts from ligands 1-6, in view of
the practical problems in isolating their corresponding
pure ruthenium complexes since 2-6 are mixtures of
molecules with the same structure but with a different
ethoxylation grade.
The substrate 3-methyl-2-butenal was catalytically
hydrogenated to 3-methyl-2-butenol in 2-propanol/water
under mild conditions (30 bar H2, 50 °C) by a catalyst
preparation from RuCl3 and 3 equiv of ligands 1-6.
Reactions were performed for 20-160 min. The data
from Table 1 demonstrate that high conversions of the
starting compound were achieved after 20 min, in
particular by ligands with long polyether chains. The
gas chromatography analysis of reaction products shows
that with the exception of prenal and prenol there are
only traces of minor compounds, 3-methyl-1-butanol
being the most abundant (e1%). All reactions performed
show reasonably high regioselectivity, allylic alcohol
percentages of the hydrogenated products being on the
order of 90% or higher. A possible hydrogen transfer
from 2-propanol was ruled out since no prenol was
obtained in a test reaction without hydrogen.
F igu r e 2. Schematic representation of molecular packing
in the [RuCl2(8)(PPh3)] complex. For the sake of clarity only
Ru, Cl, O (bonded to Ru), and P atoms and isopenytl groups
are presented.
Sch em e 8
The reported catalytic system shows higher efficiency
and selectivity than ruthenium catalysts based on
triphenylphoshine, and the results are similar to those
observed with TPPTS in biphasic medium (Table 2). As
mentioned in the Introduction, one of the most attrac-
tive aspects of amphiphilic ligands 1-6 is the plausible
modulation of their properties by changing the polyether
chain length and the hydrophobic group structure.
Hence, in this case the comparison between the results
obtained with the different ligands is particularly
relevant. For instance, ligands 1 and 4 show a notably
different behavior, which can be related to the limited
solubility of the corresponding ruthenium complexes in
the reaction medium. Thus, in both processes the
formation of solid compounds was observed, and the
solid obtained from ligand 1 could be identified as [Ru2-
Cl4(1)5] by 31P NMR. In the same way, ligand 4 afforded
are valuable products in the field of fragrance and flavor
chemistry.23 A limited number of catalytic systems
based on ruthenium complexes that allow CdO hydro-
genation over CdC have been reported.24 For the
purpose of appraising the performance of the am-
phiphilic ligands 1-6 in catalysis, we examined the
catalytic hydrogenation of 3-methyl-2-butenal (prenal)
to 3-methyl-2-butenol (prenol) in 2-propanol/water mix-
tures (Scheme 8).
The catalysts were generated in situ from RuCl3 with
a ligand excess.25 This is the simplest approach to the
i
the most distinct conversions with respect to the PrOH/
(23) (a) Bauer, K.; Garbe, D. Ullman Encyclopedia, 3rd ed.; VCH:
New York, 1988; Vol. A11, p 141. (b) Nomura, K. J . Mol. Catal. A:
Chem. 1998, 130, 12-13.
(24) J oo´, F.; Kova´cs, J .; Be´nyei, A. C.; Katho´, A. Catal. Today 1998,
42, 441, and references therein.
(25) Grosselin, J . M.; Mercier, C.; Allmang, G.; Grass, F. Organo-
metallics 1991, 10, 2126.
H2O ratio (t ) 20 min), which is also associated with
the low solubility of its complexes in the most polar
medium. On the other hand, ligands 2, 3, 5, and 6 lead
to soluble complexes, which result in more efficient
catalysts, and manifest a similar behavior. Indeed, all

New Hemilabile Amphiphilic Phosphines
Organometallics, Vol. 18, No. 26, 1999 5481
Ta ble 1. Ca ta lytic Hyd r ogen a tion of P r en a l to P r en ol w ith Ru Cl3 a n d Liga n d s 1-6
conversion (%)a
reaction time )
reaction time ) 20 min
reaction time ) 35 min
reaction time ) 50 min
reaction time ) 65 min
160 min
iPrOH/H2O iPrOH/H2O iPrOH/H2O iPrOH/H2O iPrOH/H2O iPrOH/H2O iPrOH/H2O iPrOH/H2O
iPrOH/H2O
(95:5)
ligand
(95:5)
(80:20)
(95:5)
(80:20)
(95:5)
(80:20)
(95:5)
(80:20)
1
2
3
4
5
6
38(85)
36(84)
92(90)
75(95)
74(94)
64(95)
93(96)
78(96)
75(95)
91(92)
35(94)
100(93)
100(95)
100(93)
95(93)
92(96)
100(95)
100(94)
100(91)
100(90)
74(92)
100(94)
99(89)
a
In parentheses, selectivity in % prenol.
Ta ble 2. Ca ta lytic Hyd r ogen a tion of P r en a l to
P r en ol w ith Sever a l Ru th en iu m Com p lexes25
were separated by filtration and washed with hexane (50 mL).
The resulting solution was evaporated under vacuum, and the
alkyl chloride was obtained as a colorless oil. Yield: 29.0 g
(≈99%).
P(H2)
T
time conversion, selectivity
catalyst
(bar) (°C) (h)
%
(prenol), %
(b) Syn th esis of (CH3)3CCH2C(CH3)2C6H4OCH2CH2P P h 2
(1). A hexane solution of n-butyllithium (1.6 M, 0.13 mol, 80
mL) was added dropwise over 30 min with stirring to a solution
of Ph2PH (0.10 mol) in diethyl ether (100 mL) and cooled at 0
°C, and the resulting red solution was stirred at this temper-
ature for a further 30 min. Next, to this cooled solution was
added dropwise via cannula, with stirring, a cooled solution
(0 °C) of alkyl chloride (CH3)3CCH2C(CH3)2C6H4(OCH2CH2)n-
Cl) (nj ≈ 1.5, ≈ 0.10 mol, 29.0 g) in diethyl ether (100 mL) and
stirred for a further 20 min at 0 °C. The cold bath was
removed, and the mixture was allowed to warm to room
temperature, stirred for 20 min, and brought to reflux for 3 h.
The resulting reaction mixture was cooled to room tempera-
ture, EtOH (2 mL) and deoxygenated water (2 mL) were added
to hydrolyze the remaining excess BuLi reagent, and the
solvent was removed by vacuum evaporation. Deoxygenated
water (70 mL) was added to the mixture and was extracted
with portions (3 × 60 mL) of hexane, the organic layer was
dried over Na2SO4, and the solvent was removed under
reduced pressure to give a colorless oil. After addition of EtOH
(100 mL) a white precipitate was formed, which was filtered
and washed with cold EtOH. Yield: 19.8 g (≈95%, based on
the 50% abundance of (CH3)3CCH2C(CH3)2C6H4OCH2CH2Cl in
the alkyl chloride mixture).
a
H2Ru(PPh3)4
30
30
30
30
20
20
50
50
50
50
35
35
2.5
3
3
7
9
100
81
33
60
69
91
80
38
73
93
97
a
RuCl2(PPh3)3
a
HRuCl(CO)(PPh3)3
a
HRu(OAc)(PPh3)3
a
RuCl3/4PPh3
RuCl3/5TPPTSb
1
100
a
i
b
Solvent ) PrOH/water (95:5). Biphasic medium ) toulene/
water.
these ligands display high conversions at 20-35 min
with selectivity to the order of 90-96%, and a slight
improvement is observed by increasing the medium
polarity. Although this preliminary catalytic study does
not lead to notable differences between ligands 2, 3, 5,
and 6, Table 1 suggests a slight improvement in ligands
with linear nonyl hydrophobic chains with respect to
ligands with bulky tert-octyl groups. If this trend could
be corroborated in future works, it will correlate with
better packing of linear hydrophobic groups in supra-
molecular structures.
Exp er im en ta l Section
Sign ifica n t NMR Da ta . 31P{1H} NMR (acetone-d6): -20.7.
All reactions were performed under nitrogen by standard
Schlenk tube techniques. Infrared spectra were recorded with
a Perkin-Elmer 1710 FT spectrometer. The NMR spectra were
recorded on a Bruker AC400 (Servei de Ressona`ncia Magne`tica
Nuclear de la Universitat Auto`noma de Barcelona) and AM250
(Laboratoire de Chimie de Coordination) instruments. All
chemical shift values are given in ppm and are referenced with
respect to residual protons in the solvent for proton spectra
to solvent signals for 13C spectra and to phosphoric acid for
phosphorus spectra.
The nonionic surfactants IGEPAL were purchased from
Aldrich Chemical Co., and RuCl3 was obtained from J ohnson
Matthey. The complex RuCl2(PPh3)3 was prepared by pub-
lished procedures.26 Microanalyses were performed in the
Servei d'Ana`lisi Qu´ımica de la Universitat Auto`noma de
Barcelona.
Syn th esis of 1. (a) Syn th esis of (CH3)3CCH2C(CH3)2C6H4-
(OCH2CH2)n Cl(nj ≈ 1.5). Triphenylphosphine (34.1 g, 0.13
mol) was added to a solution of the nonionic surfactant
IGEPAL CA210 (29.4 g, ≈ 0.11 mol) in carbon tetrachloride
(100 mL), and the mixture was heated to reflux for 1 h. In
this period, the precipitation of P(O)Ph3 as a white solid was
observed. The mixture was allowed to cool to room tempera-
ture, and hexane (100 mL) was added in order to complete
the precipitation of P(O)Ph3 and the PPh3 in excess. Solids
1H NMR (CDCl3; except phenyl resonances): 0.65 (s, (CH3)3),
3
1.27 (s, (CH3)2), 1.63 (s, CH2, tert-octyl), 2.52 (t, J HH ) 7.5
3
3
Hz, CH2-P), 4.03 (q(apparent), J PH ≈ J HH ) 7.5 Hz, CH2O).
13C{1H} NMR (CDCl3; except phenyl resonances): 28.3 (d, 1J PC
) 13.2 Hz, CH2P), 31.5 (s, (CH3)2), 31.6 (s, (CH3)3), 32.2 (s,
2
CMe3), 37.8 (s, CMe2), 56.4 (s, CH2, tert-octyl), 65.1 (d, J PC
25.8 Hz, CH2O).
)
Syn th esis of 2-6. All these ligands were prepared by
procedures analogous to that described above for ligand 1, and
the specific data of these preparations are as follows.
Liga n d 2. (a ) (CH3)3CCH2C(CH3)2C6H4(OCH2CH2)n Cl (nj
≈ 5): IGEPAL CA520 (42.7 g, ≈ 0.1 mols), PPh3 (34.1 g, 0.13
mol), CCl4 (90 mL). Yield: 44.5 g (≈99%).
(b ) (CH3)3CCH2C(CH3)2C6H 4(OCH2CH 2)n P P h 2 (nj ≈ 5)
(2): (CH3)3CCH2C(CH3)2C6H4(OCH2CH2)5Cl (44.5 g, ≈ 0.10
mol), Ph2PH (19.1 g, 0.10 mol), BuLi 1.6 M (69 mL, 0.11 mol).
Yield: 55.8 g (≈94%).
Sign ifica n t NMR Da ta . 31P{1H} NMR (acetone-d6): -20.4.
1H NMR (CDCl3; except phenyl resonances): 0.65 (s, (CH3)3),
3
1.27 (s, (CH3)2), 1.63 (s, CH2, tert-octyl), 2.34 (t, J HH ) 7.5
Hz, CH2-P), 3.4-4.1 (m, CH2O). 13C{1H} NMR (CDCl3; except
1
phenyl resonances): 28.6 (d, J PC ) 12.6 Hz, CH2P), 31.5 (s,
(CH3)2), 31.6 (s, (CH3)3), 32.2 (s, CMe3), 56.8 (s, CH2, tert-octyl),
2
67.1 (s, 18H, CH2O), 68.4 (d, J PC ) 25.8 Hz, PCH2CH2O),
69.6-70.7 (m, CH2O).
Liga n d 3. (a ) (CH3)3CCH2C(CH3)2C6H4(OCH2CH2)n Cl (nj
≈ 12): IGEPAL CA720 (29.4 g, ≈ 0.04 mols), PPh3 (20.5 g,
(26) Stephenson, T. A.; Wilkinson, G. J . Inorg. Nucl. Chem. 1966,
28, 945.

5482 Organometallics, Vol. 18, No. 26, 1999
Valls et al.
0.078 mols), CCl4 (90 mL). To remove the residual P(O)Ph3,
the oil obtained after evaporation under reduced pressure was
dissolved in MeOH/H2O (1:1). Water was slowly added until a
white cloudiness was observed, two phases were formed, and
the upper phase, which was rich in P(O)Ph3, was removed.
This purification method was repeated until 31P NMR spec-
troscopy showed insignificant levels of phosphine oxide in the
final product. Yield: 19.6 g (≈65%).
CH2)3OH (48.2 g, 0.29 mol), PPh3 (100 g, 0.38 mol), CCl4 (150
mL). Yield: 33.6 g (63%).
(b) CH3(CH2CH2O)3P P h 2 (7). This phosphine was pre-
pared by a procedure analogous to that described above for
ligand 1, and the specific data of this preparation are as
follows: CH3(OCH2CH2)3Cl (18.3 g, 0.10 mol), Ph2PH (19.1 g
0.10 mol), BuLi 1.6 M (69 mL, 0.11 mol). Yield: 30.2 g (91%).
This phosphine was also prepared starting with PPh3 instead
of Ph2PH by the procedure described in the synthesis of ligand
8 (yield ) 81%).
(b) (CH3)3CCH2C(CH3)2C6H4(OCH2CH2)n P P h 2 (nj ≈ 12)
(3): (CH3)3CCH2C(CH3)2C6H4(OCH2CH2)12Cl (31.1 g, ≈ 0.04
mol), Ph2PH (7.7 g, 0.04 mol), BuLi 1,6 M (32 mL, 0.05 mol).
Yield: 34.0 g (≈92%).
1
Sign ifica n t NMR Da ta . 31P{1H} NMR (CDCl3): -22.3. H
3
NMR (CDCl3; except phenyl resonances): 2.36 (t, J HH ) 7.5
Hz, CH2-P), 3.31 (s, CH3), 3.4-3.7 (m, CH2O). 13C{1H} NMR
Sign ifica n t NMR Da ta . 31P{1H} NMR (acetone-d6): -20.4.
1H NMR (CDCl3; except phenyl resonances): 0.65 (s, (CH3)3),
1
(CDCl3; except phenyl resonances): 28.5 (d, J PC ) 12.6 Hz,
CH2P), 58.6 (s, CH3), 68.2 (d, 2J PC ) 23.9 Hz, CH2CH2P), 69.8-
71.6 (m, CH2O).
3
1.28 (s, (CH3)2), 1.64 (s, CH2, tert-octyl), 2.34 (t, J HH ) 8.0
Hz, CH2-P), 3.4-4.1 (m, CH2O). 13C{1H} NMR (CDCl3; except
1
Syn th esis of 8. (a ) (CH3)2CH(CH2)2P P h 2. To a stirred
solution of triphenylphosphine (26.2 g, 0.10 mol) in THF (150
mL) was added finely cut lithium (2.0 g, 0.29 mol), and the
mixture was stirred for 3 h at room temperature. The obtained
dark red-brown solution was separated via cannula from the
excess lithium, and tert-butyl chloride (9.3 g, 0.10 mol) was
added dropwise with continuous stirring. The resulting red
solution was stirred at room temperature for a further 20 min.
This solution was cooled at -78 °C, and a solution of Br(CH2)2-
CH(CH3)2 (15.1 g, 0.10 mols) in THF (50 mL) was dropwise
added with vigorous stirring, yielding a colorless solution with
a white precipitate. Next, a drop of BuLi (1.6 M in hexane)
was added to the mixture, and the solution turned a slightly
red color, which faded after some seconds. This process was
repeated until a permanent slightly red color was obtained.
Then, a few milliliters of water was cautiously added to
hydrolyze any excess BuLi, and the solvent was removed in
vacuo. The residual oil was extracted in hexane (2 × 100 mL)/
water(100 mL). The organic phase was dried with Na2SO4, and
the resulting solution was evaporated to dryness. The product
was obtained as a colorless oil. Yield: (21.5 g) (84%). 31P{1H}
NMR (acetone-d6): -15.3.
phenyl resonances): 28.6 (d, J PC ) 12.6 Hz, CH2P), 31.5 (s,
(CH3)2), 31.6 (s, (CH3)3), 32.2 (s, CMe3), 56.8 (s, CH2, tert-octyl),
2
60.7 (s, CH2O), 67.1 (s, CH2O), 68.4 (d, J PC ) 25.8 Hz,
PCH2CH2O), 69.3-70.7 (m, CH2O).
Liga n d 4. (a ) CH3(CH2)8C6H4(OCH2CH2)n Cl (nj ≈ 1.4):
IGEPAL CO210 (31.7 g, ≈ 0.11 mol), PPh3 (34.1 g, 0.13 mol),
CCl4 (90 mL). Yield: 33.5 g (≈ 99%).
(b) CH3(CH2)8C6H4(OCH2CH2)n P P h 2 (nj ≈ 1.4) (4): CH3-
(CH2)8C6H4(OCH2CH2)1.4Cl (30.0 g ≈ 0.10 mol), Ph2PH (19.1
g, 0.10 mol), BuLi 1.6 M (75 mL, 0.12 mol). Yield: 44.9 g
(≈90%).
Sign ifica n t NMR Da ta . 31P{1H} NMR (CDCl3): -21.3. 1H
NMR (CDCl3; except phenyl resonances): 0.3-1.7 (m, n-nonyl),
2.36 (t, 3J HH ) 7.6 Hz, C(aryl)-OCH2CH2OCH2CH2-P), 2.50 (t,
3J HH ) 7.7 Hz, C(aryl)-OCH2CH2-P), 3.5-4.2 (m, CH2O).
13C{1H} NMR (CDCl3; except phenyl resonances): 8-55 (m,
1
nonyl), (28.2 (d, J PC ) 13.9 Hz, C(aryl)-OCH2CH2P), 28.7 (d,
2
1J PC ) 15.6 Hz, C(aryl)-OCH2CH2OCH2CH2-P), 64.5 (d, J PC
) 27.9 Hz, C(aryl)-OCH2CH2P), 67.0 (s, C(aryl)-OCH2CH2O),
2
68.5 (d, J PC ) 24.4 Hz, C(aryl)-OCH2CH2OCH2CH2-P), 69.1
(s, C(aryl)-OCH2CH2O).
Liga n d 5. (a ) CH3(CH2)8C6H4(OCH2CH2)n Cl (nj ≈ 5):
IGEPAL CO520 (20.0 g, ≈ 0.05 mol), PPh3 (16,0 g, 0.06 mol),
CCl4 (90 mL). Yield: 18.7 g (≈90%).
(b ) P P h [(CH 2)2CH (CH 3)2][(CH 2CH 2O)3CH 3] (8). This
ligand was prepared by a procedure analogous to that de-
scribed above for isopentyldiphenylphosphine but using iso-
pentyldiphenylphosphine as starting material instead of tri-
phenylphosphine and CH3(OCH2CH2)3Cl as the alkyl halide.
The specific data of this preparation are as follows: (CH3)2-
CH(CH2)2PPh2 (9.8 g, 0.038 mol), Li (0.8 g, 0.11 mol), (CH3)3-
CCl (4.2 mL, 0.039 mol), CH3(OCH2CH2)3Cl (7.0 g, 0.038 mol).
Yield: 10.2 g (82%) of colorless oil.
(b) CH3(CH2)8C6H4(OCH2CH2)n P P h 2 (nj ≈ 5) (5): (CH3-
(CH2)8C6H4(OCH2CH2)5Cl (18.7 g, ≈ 0.04 mol), Ph2PH (7.5 g,
0.04 mol in 80 mL of Et2O), BuLi 1.6 M (28 mL, 0.05 mol).
Yield: 22.4 g (≈ 90%).
Sign ifica n t NMR Da ta . 31P{1H} NMR (CDCl3): -21.7. 1H
NMR (acetone-d6; except phenyl resonances): 0.3-1.7 (m,
3
n-nonyl), 2.32 (t, J HH ) 7.6 Hz, CH2-P), 3.5-4.2 (m, CH2O).
1
Sign ifica n t NMR Da ta . 31P{1H} NMR (CDCl3): -30.0. H
13C{1H} NMR (CDCl3; except phenyl resonances): 8-55 (m,
NMR (CDCl3; except phenyl resonances): 0.6-1.8 (m, isopen-
tyl), 1.98 (t, 3J HH ) 7.6 Hz, CH2-P), 3.2-3.8 (m, CH2O, CH3O).
13C{1H} NMR (CDCl3; except phenyl resonances): 21.9-34.7
2
nonyl), 68.2 (d, J PC ) 26.2 Hz, OCH2CH2-P), 60-72 (m,
OCH2).
2
Liga n d 6. (a ) CH3(CH2)8C6H4(OCH2CH2)n Cl (nj ≈ 11):
IGEPAL CO720 (26.0 g, ≈ 0.04 mol), PPh3 (22.6 g, 0.086 mol),
CCl4 (90 mL). The same procedure previously described in the
synthesis of the alkyl chloride of ligand 3 should be performed
with the oil obtained after evaporation under vacuum. Yield:
16.5 g (≈62%).
(m, isopentyl), 58.7 (s, CH3O), 68.5 (d, J PC ) 22.0 Hz, OCH2-
CH2P), 69.7-71.8 (m, CH2O).
Syn th esis of 9. (a ) CH3(CH2)7P P h 2. This compound was
prepared by a procedure analogous to that described above for
isopentyldiphenylphosphine but using octyl chloride instead
of isopentyl bromide. The specific data are as follows: PPh3
(50.0 g, 0.19 mol in 200 mL THF), Li (3.5 g, 0.50 mol), (CH3)3-
CCl (17.7 g, 0.19 mol), CH3(CH2)7Cl (28.3 g, 0.19 mol). Yield:
48.2 g (85%) of a pale yellow oil. 31P{1H} NMR (acetone-d6):
-15.0.
(b) P P h [(CH2)7CH3][(CH2CH2O)2CH2CH2OH] (9). To a
stirred solution of Ph2P(CH2)7CH3 (30.3 g, 0.10 mol) in THF
(150 mL) was added finely cut lithium (2.0 g, 0.29 mmol), and
the mixture was stirred for 3 h at room temperature. The
obtained dark brown solution was separated via cannula from
the excess lithium, and tert-butyl chloride (9.5 g, 0.10 mol) was
dropwise added with continuous stirring. The resulting orange
solution was stirred at room temperature for a further 20 min.
This solution was cooled at -78 °C, and a solution of Cl(CH2-
CH2O)2CH2CH2OH (17.2 g, 0.10 mol) in THF (50 mL) and a
(b) CH3(CH2)8C6H4(OCH2CH2)n P P h 2 (nj ≈ 11) (6): CH3-
(CH2)8C6H4(OCH2CH2)11Cl (23.2 g, ≈ 0.03 mol), Ph2PH (5.6 g,
0.03 mol), BuLi 1.6 M (21 mL, 0.03 mol). Yield: 24.3 g (≈87%).
Sign ifica n t NMR Da ta . 31P{1H} NMR (acetone-d6): -20.4.
1H NMR (acetone-d6; except phenyl resonances): 0.4-1.8 (m,
3
n-nonyl), 2.34 (t, J HH ) 7.6 Hz, CH2-P), 3.4-4.2 (m, CH2O).
13C{1H} NMR (CDCl3; except phenyl resonances): 8-55 (m,
2
nonyl), 68.2 (d, J PC ) 24.4 Hz, OCH2CH2-P), 60-72 (m,
OCH2).
Syn th esis of 7. (a ) CH3(OCH2CH2)3Cl. This alkyl chloride
was prepared by a procedure analogous to that described above
for ((CH3)3CCH2C(CH3)2C6H4(OCH2CH2)2Cl), but the final
product was purified by distillation (T ) 65 °C, P ≈ 0.1 mmHg).
The specific data of this preparation are as follows: CH3(OCH2-

New Hemilabile Amphiphilic Phosphines
Organometallics, Vol. 18, No. 26, 1999 5483
solution of BuLi (1.6 M in hexane, 64 mL, 0.1 mol) were
simultaneously and dropwise added with vigorous stirring.
When both additions were finished, a colorless solution with
a white precipitate was obtained. Next, a drop of BuLi (1.6 M
in hexane) was added to the mixture, and the solution turned
a slightly red color, which faded after some seconds. This
process was repeated until a permanent slightly red color was
obtained. A few milliliters of water was cautiously added to
hydrolyze any excess BuLi, and the solvent was removed in
vacuo. The residual oil was extracted in hexane (3 × 100 mL)/
water (100 mL). The organic phase was dried over Na2SO4 and
the resulting solution was evaporated to dryness. The product
was obtained as a pale yellow oil. Yield: 29.4 g (83%).
tored by 31P{1H} NMR spectroscopy. 31P{1H} NMR (T ) 233
2
2
K): 16.3 (d, J PP ) 29.5 Hz, [RuCl2(8)(PBu3)2]), 42.4 (d, J PP
)
2
42 Hz, [RuCl2(8)(PBu3)]), 69.5 (d, J PP ) 42 Hz, [RuCl2(8)-
2
(PBu3)]), 70.6 (t, J PP ) 29.5 Hz, [RuCl2(8)(PBu3)2]).
CO. Carbon monoxide was bubbled through a solution of
[RuCl2(8)(PPh3)] (200 mg, 0.26 mmol) in CH2Cl2 (10 mL).
Almost immediately a color change was observed from red to
yellow, and the NMR and IR spectra from this solution were
obtained. 31P{1H} NMR (insert acetone-d6): 11.7 (d, 2J PP ) 251
2
Hz, [RuCl2(CO)2(8)(PPh3)]), 26.4 (d, J PP ) 251 Hz, [RuCl2-
(CO)2(8)(PPh3)]). IR (CH2Cl2, νCO, cm-1): 2005.
The above dichloromethane solution was allowed to reflux
for 45 min, and the 31P NMR and IR spectra were newly
Sign ifica n t NMR Da ta . 31P{1H} NMR (acetone-d6): -30.0.
1H NMR (CDCl3; except phenyl resonances): 0.8-1.8 (m, octyl),
2
obtained. 31P{1H} NMR (acetone-d6): 45.8 (d, J PP ) 292 Hz,
2
[RuCl2(CO)(8)(PPh3)], 48.0 (d, J PP ) 292 Hz, [RuCl2(CO)(8)-
3
2.02 (t, J HH ) 8.0 Hz, CH2-P), 3.4-3.8 (m, CH2O). 13C{1H}
(PPh3)]). IR (CH2Cl2, νCO, cm-1): 1970.
NMR (CDCl3; except phenyl resonances): 13.9 (s, CH3), 22.4
(s, CH2-CH3), 25.7 (d, 2J PC ) 13.2 Hz, PCH2CH2-hexyl), 28.1
(d, 1J PC ) 10.7 Hz, CH2P), 28.6 (d, 1J PC ) 12.6 Hz, CH2P), 29.0-
Rea ctivity of 1-7 a n d 9 w ith Ru Cl2(P P h 3)3. In
a
representative procedure, a solution of RuCl2(PPh3)3 (0.05
mmol) and the corresponding ligand (0.10 mmol) in dichlo-
rometane (20 mL) was stirred for 1 h. The resulting solution
was evaporated in vacuo, yielding a dark oil, which was used
to measure the 31P NMR spectrum. The same procedure was
2
31.6 (m, octyl), 61.4 (CH2O), 68.7 (d, J PC ) 22.0 Hz,
PCH2CH2O), 69.9-72.3 (m, CH2O).
Syn th esis of [Ru 2Cl4(1)5]. A deoxygenated solution of
RuCl3‚3H2O (0.5 g, 1.9 mmol) in methanol (95:5, 30 mL) was
heated to reflux for 5 min. To the resulting cold solution was
added 1 (11.5 mmol, 4.8 g), and the resulting solution was
heated to reflux for 3 h. The methanol solution was cooled,
and a dark brown solid crystallized, which was recrystallized
in hot 2-propanol. Yield: 0.95 g (41%).
i
used for ligand 7, but the final oil was dissolved in PrOH and,
after cooling, a dark solid was isolated. The oil obtained from
the preparation with ligand 9 was vigorously stirred with
pentane (the complex is not soluble in this solvent) for several
minutes. The pentane solution was rejected, fresh pentane
added, and the process repeated until the dark oil was
transformed into a dark solid. 31P{1H} NMR (acetone-d6).
•
Anal. Calcd for C140H175Cl4O5P5Ru2 : C, 69.01; H, 7.24.
Found: C, 68.51; H, 7.52. 31P{1H} NMR (CD2Cl2, T ) 293 K):
42.2 (s), 29.5 (t, b), 23.7 (d, b). 1H NMR (CDCl3; except phenyl
resonances): 0.66 (s, (CH3)3), 0.67 (s, (CH3)3), 1.26 (s, (CH3)2),
1.62 (s, CH2, tert-octyl), 2.5-3.7 (b, CH2-P and CH2O). 13C{1H}
NMR (CDCl3; except phenyl resonances): 26.5 (b, CH2P), 31.5
(s, CH3), 31.7 (s, CH3), 32.2 (s, CH3), 37.8 (s, CMe2), 56.8 (s,
CH2, tert-octyl), 64.5 (b, CH2O).
2
2
1: {22.9 (b), 28.4 (t, J PP ) 30 Hz), 41.0 (s), 38.3 (d, J PP
)
2
39 Hz), 50.7 (d, J PP ) 39 Hz)}.
2: {58.4 (d, J PP ) 40 Hz), 58.7 (d, J PP ) 40 Hz)}.
3: {58.6 (b)}.
2
2
4: {28.0 (b), 31.0 (b), 38.6 (b), 41.0 (b), 53.9 (b), 58.8 (b), 64.2
(s)}.
5: {58.4 (d, J PP ) 40 Hz), 58.7 (d, J PP ) 40 Hz)}.
2
2
Rea ctivity of 2-9 w ith Ru Cl3‚3H2O. In a representative
procedure, a deoxygenated solution of RuCl3‚3H2O (0.05 g, 0.19
mmol) in 2-propanol/water (95:5, 30 mL) was heated to reflux
for 5 min. To the resulting cold solution was added 0.57 mmol
of ligand, and the mixture was heated to 50 °C for 3 h. Solvents
were evaporated in vacuo, and the resulting dark oil was
extracted with hexane (10 mL). The resulting oil was evapo-
rated to dryness in vacuo. The 31P NMR was measured with a
portion of this final oil.
2
2
6: {58.4 (d, J PP ) 43 Hz), 58.8 (d, J PP ) 43 Hz)}.
7: {62.2 (s)}.
2
2
9: {58.8 (d, J PP ) 43 Hz), 63.3 (d, J PP ) 43 Hz)}.
Ca ta lysis. The products of the catalytic reactions were
identified by GC and mass analysis. Gas chromatography was
run on a KONIK KNK-3000-HRGC equipped with a Chro-
mosorb W-HP packed column. The peaks were identified by
GC/MS techniques. The GC detector sensitivity was calibrated
by comparison with authentic samples.
Hyd r ogen a tion P r oced u r e. A typical experiment was
performed in a glass-lined stainless steel autoclave with
magnetic stirring (1000 rpm). Hydrated ruthenium chloride
(146 mg, 0.5 mmol) and 1.5 mmol of ligand were placed in the
autoclave vessel, which was closed. Air was evacuated under
vacuum and replaced with nitrogen. A purged solution of 2
mL (20 mmol) of 3-methyl-2-butenal in 15 mL of 2-propanol/
water was transferred to the autoclave by means of a cannula.
The reactor was placed in a thermostated oil bath for 20 min
with vigorous stirring to allow formation of the catalyst
precursor and to reach the desired temperature in the reaction
vessel. Next, H2 (30 bar) was introduced in the reactor.
Samples were extracted using a syringe with purge each time.
Syn th esis of [Ru Cl2(8)(P P h 3)]. A solution of RuCl2(PPh3)3
(1.24 g, 1.4 mmol) and 8 (0.90 g, 2.8 mmol) in dichlorometane
(20 mL) was stirred for 1 h. The resulting dark solution was
evaporated in vacuo, yielding to a dark oil, which was extracted
with pentane in order to eliminate free phosphines. The
extraction was repeated until the dark oil was transformed
into a dark solid. The resulting dark solid was dried in vacuo
and crystallized in MeOH. Yield: 0.20 g (38%).
Anal. Calcd for
C
36H46Cl2O3P2Ru•: C, 56.84; H, 6.10.
Found: C, 56.53; H, 5.96. 31P{1H} NMR (CD2Cl2, T ) 293 K):
58.8 (d), 63.3 (d), (J PP ) 43.3 Hz). 1H NMR (CD2Cl2; except
phenyl resonances): 0.6-1.6 (m, isobutyl), 2.3-2.5 (m, CH2P),
2.6-2.8 (m, CH2P), 3.17 (s, CH3O), 3.2-4.4 (m, CH2O).
Rea ctivity of [Ru Cl2(8)(P P h 3)] w ith P Bu 3 a n d CO.
P Bu 3. A solution of crystalline [RuCl2(8)(PPh3)] (20 mg, 0.026
mmol) in CD2Cl2 (1 mL) was prepared, and 0.5 mL of this
solution were placed in a NMR tube and frozen by liquid
nitrogen. A solution of PBu3 (6.5 µL, 0.026 mmol) in CD2Cl2
(1 mL) was prepared, and 0.5 mL was added to the NMR tube
and frozen again in liquid nitrogen. The reaction was moni-
Ack n ow led gm en t. This research was supported by
the Direccion General de Investigacio´n Cient´ıfica y
Te´cnica.
OM9905916