Phosphiranes as ligands for platinum catalysed hydrosilylations

Article abstract of DOI:10.1016/S0040-4020(99)00783-8

The stability of phosphiranes can be enhanced by inclusion of the PC₂ ring into a polycyclic framework. BABAR-Phos type phosphiranes are easily prepared in high yield, they are thermally stable and allow the synthesis of various platinum(0) complexes. In contrast to other platinum(0) phosphine complexes, these are active hydrosilylation catalyst precursors. The reversible insertion of an electron rich metal centre into one P-C bond of the phosphirane ring was demonstrated for the first time.

Full text of DOI:10.1016/S0040-4020(99)00783-8

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 143–156
Phosphiranes as Ligands for Platinum Catalysed Hydrosilylations
*
¨
¨
¨
Jurgen Liedtke, Sandra Loss, Christoph Widauer and Hansjorg Grutzmacher
¨
¨
Laboratory of Inorganic Chemistry, ETH-Zentrum, Universitatsstrasse 6, CH-8092 Zurich, Switzerland
Received 28 April 1999; accepted 2 August 1999
Abstract—The stability of phosphiranes can be enhanced by inclusion of the PC2 ring into a polycyclic framework. BABAR-Phos type
phosphiranes are easily prepared in high yield, they are thermally stable and allow the synthesis of various platinum(0) complexes. In contrast
to other platinum(0) phosphine complexes, these are active hydrosilylation catalyst precursors. The reversible insertion of an electron rich
metal centre into one P–C bond of the phosphirane ring was demonstrated for the first time. ᭧ 1999 Published by Elsevier Science Ltd. All
rights reserved.
Introduction
energy than the ligand donating orbitals and the electron
donating metal centred orbitals lie lower in energy than
the ligand centred acceptor orbitals.
The elemental steps, i.e. oxidative addition of E–H to an
unsaturated transition metal fragment (a), co-ordination of
an olefin (b), insertion (c), and reductive elimination (d), of
a catalysed addition of an element hydride bond, E–H, to a
CyC bond system are recalled in a simplified manner in
Scheme 1.
When the phosphorus becomes pyramidalised, the HOMO,
which is a pure pz orbital in D3h, decreases in energy. At the
same time the energy of the degenerate set of anti-bonding
orbitals strongly decreases with increasing pyramidalisation
which eventually become the LUMOs of an AH3 molecule
in C3v symmetry. These orbitals are responsible for the
p-accepting properties of the phosphine.2 Phosphiranes
are the most strongly pyramidalised phosphines known
(SЊ(P): 240–275)3 which, according to Scheme 2, should
make them excellent p-acceptors and promising ligands in
transition metal catalysed reactions where step (d) is rate
determining.
It is often assumed that the last step (d), i.e. the formation of
the E–C bond, has the highest activation barrier and
determines the overall rate and efficiency of the process.1
Reductive elimination is favoured when the metal centre has
a formally high oxidation state. For a given metal centre and
substrates E–H and CyC, the only way to influence the
electron density on the metal lies in the choice of the appro-
priate phosphine ligand. Under such conditions, an electron
withdrawing phosphine PR3 may accelerate the catalytic
reaction (note however that the first step, the oxidative addi-
tion (a), requires an electron rich metal centre and this rate
may become slower). Apart from using electronegative
substituents on the phosphorus centre one might think
about altering the bond angles.
Phosphiranes may be synthesised using different methods
which are briefly summarised in Scheme 3.3 The conden-
sation reaction (a) is the most obvious approach4a–f and has
recently been used to prepare chiral phosphiranes stereo-
selectively.4b Another elegant way of synthesising chiral
phosphiranes has been found by reacting oxiranes with
PR1 transfer reagents.4d–f
This approach is illustrated in Scheme 2, showing simple
MO diagrams for eight valence electron AH3 species, which
result when a planar form (D3h) is converted into a pyrami-
dal form (C3v).
It is clear that a planar phosphine (D3h symmetry) shall have
good s-donating but poor p-accepting properties since its
HOMO and LUMO lie quite high in energy. Remember that
the electron accepting orbitals on the metal lie higher in
Keywords: catalysis; hydrosilylation; phosphorus; phosphiranes; platinum.
* Corresponding author. Tel.: ϩ41-1-632-2855; fax: ϩ41-1-632-1090;
e-mail: gruetz@elwood.ethz.ch
Scheme 1. Simplified catalytic cycle showing the addition of an E–H bond
to a CyC bond.
0040–4020/00/$ - see front matter ᭧ 1999 Published by Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(99)00783-8

144
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
Functionalized phosphiranes were prepared using method
(b) in which a highly reactive phosphandiyl metal
complex generated in situ is added in a [2ϩ1] cyclo-
addition reaction to an olefin or to an allene.4g,h Related
to this synthesis is the addition of a carbene to a phospha-
alkene [eq. (c)].4i–p In these reactions, intermediates A
(when a diazo compound is used as carbene source)
and/or B which undergoes a conrotatory electrocyclic
ring closure frequently have been observed or even
isolated. Using method (c), alkylidene phosphiranes4k–m
and bicyclic phosphiranes4o were obtained. A cyclo-
addition–cycloreversion sequence was used to prepare
thio-phosphiranes with a PSC ring skeleton.4q Recently,
reaction (d) has been described which allows the synthesis
of C-functionalized phosphiranes.4r
Phosphiranes are quite reactive entities and are in particular
photolytically labile and may undergo either [2ϩ1] cyclo-
reversion reactions,3a–c,5a,b retro-electrocyclisation,5c or
ring–chain rearrangement reactions.4f,5d–h The ring opening
polymerisation of the phosphirane (MesءP)(CH2)2
(Mesءˆ2,4,6-tBuC6H2) gave polymers with molecular
weights up to 7800.5h The lability of phosphiranes may be
one reason why their use in transition metal catalysed
reactions was found to be limited although initial promising
results were obtained in asymmetric hydrogenations.6
Phosphiranes, in which the PC2-ring is embedded in a poly-
cyclic framework are considerably more stable7a–e and a
Scheme 2. Walsh diagram for the conversion of a trigonal planar AH3
molecule (D3h) into a pyramidal AH3 molecule (C3v).
Scheme 3. Representative examples for the synthesis of phosphiranes.

J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
145
chloride 2 to the amines 3a–c. After lithiation and subse-
quent reaction with PCl3 the dichloroaminophosphanes 4a–
c are isolated as colourless crystals in Ͼ90% yield. Using
commercially available magnesium turnings and THF as
solvent, dehalogenation of 4a,b leads to gram quantities
of the aminosubstituted phosphiranes 5a,b (Ͼ90%)
(Scheme 5). They form colourless crystals, which melt with-
out decomposition and can be handled in air. Neither the
pentafluorophenylsubstituted dichlorophosphane 4c nor the
phosphirane 5c are very stable compounds. Compound 4c
undergoes a noteworthy rearrangement reaction to give the
iminophosphorane 6. This reaction proceeds slowly in the
solid state at room temperature (several days) but is
complete in solution after several minutes at about 60ЊC.
Phosphirane 5c could not be isolated but was characterised
in solution by NMR spectroscopy. It polymerises upon
concentration of the reaction solution to yield an as yet
undefined colourless gum like material. We ascribe both
reactions to the enhanced polarity of the N–C bond in 4c
or P–C bonds in 5c, respectively (see Table 3 below).
Scheme 4. Representative examples of polycyclic cage compounds
containing PC2, P2C, and P3 rings.
representative collection of such phosphiranes is shown in
Scheme 4.
Results and Discussion
Formally the phosphiranes 5a–c are formed by an intramo-
lecular [2ϩ1] cycloaddition of an R2N–P phosphinidene
unit to the CyC double bond of the central seven-membered
ring of the dibenzotropylidene unit which has a rigid boat
conformation. Because of the low reactivity of triplet phos-
phinidenes, this approach has found little application in
intermolecular cycloaddition reactions for the preparation
of phosphiranes.3,5b,i In contrast to 1-phenylphosphirane,
5a,b do not react with strongly alkylating agents such as
MeOSO2CF3 to form phosphiranium salts.9 However,
Synthesis of phosphiranes (BABAR-Phos) and a
phosphiranium salt
We searched for a simple high yield synthesis of phosphir-
anes, in which the PC2-ring is embedded in a polycyclic
framework. In a preliminary communication we described
such phosphiranes, which we named BABAR-Phos.8 They
are easily prepared in excellent yields starting from the
dibenzoannulated tropolone 1, which is converted via the
Scheme 5. Synthesis of phosphiranes 5a–c.
Scheme 6. Synthesis of phosphiranium salt 8.

146
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
˚
Table 1. Selected B3LYP/6-31G(d) bond lengths [A] and angles [Њ] for the phosphiranes I–V
N–P
P–C4
P–C5
C4–C5
NPC4
NPC5
C4PC5
PC4C5
PC5C4
I
1.745
1.641
1.733
1.663
1.748
1.851
1.808
1.917
1.823
2.826
1.873
1.808
1.870
1.796
1.889
1.513
1.531
1.546
1.610
1.522
101.3
119.1
99.8
110.1
88.5
105.2
119.1
98.1
107.2
99.5
47.9
50.1
48.2
52.8
30.1
66.8
65.0
64.3
62.7
38.5
65.3
65.0
67.5
64.4
111.5
II
III
IV
V
Table 2. Selected atomic charges derived from the natural population analysis of the compounds III–Va
C1
C2
C3
C4
C5
C6
C7
N
P
III
IV
V
Ϫ0.131
Ϫ0.135
Ϫ0.112
Ϫ0.253
Ϫ0.207
Ϫ0.483
Ϫ0.230
Ϫ0.239
Ϫ0.244
Ϫ0.528
Ϫ0.511
Ϫ0.459
Ϫ0.540
Ϫ0.528
Ϫ0.577
Ϫ0.227
Ϫ0.231
Ϫ0.263
Ϫ0.214
Ϫ0.187
Ϫ0.262
Ϫ0.971
Ϫ0.962
Ϫ0.989
0.947
1.415
0.707
a The NBO analyses were calculated at the B3LYP/6-31G(d) level.
Figure 2. Two dimensional plots of the Laplace distribution for compounds
III and IV, in a plane containing the atoms of the three-membered ring P,
C4 and C5, respectively.‡
BABAR-Phos 5a–c and the phosphiranium salt 8, we
performed DFT calculations on the B3LYP/6-31G(d) level
for the model compounds I–V. A depiction of the structures
of I–V is given in Fig. 1 and pertinent bond lengths and
angles are listed in Table 1.
The results of the topological analysis of the charge density
of the parent amino substituted phosphirane I are in good
agreement with the results previously reported for the phos-
phirane HP(CH2)2.10 As was found there, the P–C bonds are
highly polar as can be seen from the location of the bond
critical points rb (Table 3) which are non-symmetrically
placed on the P–C bond paths being closer to the phospho-
rus centre. Protonation of the phosphorus lone pair leads to
II whereby only small variations of the P–C bond properties
are observed. The C–P bonds become slightly shorter
whereas the C–C bond length increases from 1.515 to
Figure 1. Schematic representation of the compounds I–V optimised at the
B3LYP/6-31G(d) level.†
chloride abstraction from 7, which was synthesised from
lithiated 3a and tBuPCl2, by AlCl3 in methylene chloride
as solvent, cleanly afforded the phosphiranium salt 8 as
colourless crystals (Scheme 6).
In analogy to the preparation of phosphiranes 5a–c, this
reaction corresponds to an original intramolecular [2ϩ1]
cycloaddition of a phosphenium ion, R2Pϩ, to a CyC double
bond. The reverse reaction, i.e. transfer of R2Pϩ units from
phosphiranium salts is documented.9a,b Addition of THF to a
solution of 8 reverses the halide abstraction and starting
material 7 is reformed.
˚
1.533 A (Table 1). In Table 2, atomic charges of III–V
derived from a NBO analysis are listed. In Fig. 2, the
Laplace fields in the PC2 plane of BABAR-Phos model III
are compared with the phosphiranium salt IV and the results
of a topological analysis of the charge densities of all
compounds I–V are given in Table 3.
Computational results
‡
The electron densities were calculated at the B3LYP/6-31G(d) level. In
the graphical representation Ϫ72r is drawn, so that local concentration is
expressed by a positive value (solid lines) and local depletion is expressed
by a negative value (dashed lines). Bond critical points are indicated by
black squares.
In order to gain insight into the electronic structure of
†
Selected geometrical parameters are given in Table 1.

J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
147
Table 3. Selected results of the topological analysis of the charge density of the compounds I–IVa
Ϫ3
Ϫ5
Ϫ3
72rb [e A
]
H(rb) [h A
]
eb
˚
˚
˚
˚
˚
Bond
A–B [A]
rb
A-rb [A]
rb [e A
]
I
P–N
1.745
1.870
1.891
1.515
1.641
1.828
1.828
1.533
1.733
1.930
1.887
1.547
1.663
1.844
1.814
1.611
0.383
0.391
0.395
0.498
0.387
0.397
0.397
0.500
0.383
0.424
0.403
0.498
0.385
0.408
0.391
0.500
0.669
0.731
0.746
0.755
0.636
0.726
0.727
0.766
0.664
0.818
0.761
0.771
0.640
0.751
0.709
0.805
1.02
0.98
0.94
1.65
1.22
1.09
1.09
1.62
1.05
0.88
0.96
1.54
1.22
1.07
1.13
1.38
7.54
Ϫ3.05
Ϫ3.16
Ϫ12.00
15.04
Ϫ5.44
Ϫ5.44
Ϫ11.60
8.40
Ϫ4.11
Ϫ4.49
Ϫ10.33
12.44
Ϫ6.44
Ϫ4.57
Ϫ7.88
Ϫ0.78
Ϫ0.91
Ϫ0.85
Ϫ1.34
Ϫ0.93
Ϫ1.08
Ϫ1.08
Ϫ1.30
Ϫ0.82
Ϫ0.71
Ϫ0.87
Ϫ1.18
Ϫ0.99
Ϫ1.04
Ϫ1.14
Ϫ0.98
0.10
0.49
0.62
0.23
0.09
0.65
0.65
0.18
0.13
0.58
0.39
0.24
0.24
0.48
0.42
0.25
P–C4
P–C5
C4–C5
P–N
II
P–C4
P–C5
C4–C5
P–N
III
IV
P–C4
P–C5
C4–C5
P–N
P–C4
P–C5
C4–C5
a rb is the location of the bond critical point given in the ratio of distances A-rb/A–B; A-rb is the distance of the bond critical point rb to the atom A; rb and
72rb give the electron density and the second derivative at the bond critical point; H(rb) gives the energy density at the bond critical point; eb the ellipticity at
the bond critical point.
However, the Laplacian 72rb of the P–N bond
type ligands that these contain rather positively charged
phosphorus centres that should make them electron-with-
drawing ligands. Furthermore, positive as well as negative
charge may be partially delocalised into the ligand
framework.
strongly increases from 7.5 to 15.0 e AϪ5. This reflects
˚
the increasing dipolar nature of the covalent Pdϩ–Nd –
bond due to the more electropositive P atom in II.
The corresponding results for III and IV are analogous with
one remarkable exception. Protonation of BABAR-Phos III
results in a significant increase of the C4–C5 bond length by
0.054 A from 1.546 to 1.610 A. This finding is in good
agreement with the experimental value of 1.601(6) in 8
(vide infra). Accordingly, the charge density rb at the ring
critical point of the C4–C5 bond decreases in the order
IIIϾIV. This can be rationalised by the contraction of the
P–N and C–P bonds in IV, which results in a higher ring
strain in the phosphirane moiety.
Synthesis of catalyst precursors
˚
˚
In our previous investigation, we studied the binding
qualities of BABAR-Phos ligands qualitatively by compar-
ing the isolated and fully characterised d10 complexes
[Cu(iPrBABAR-Phos)2(MeCN)2]ϩ
and
[Pt(iPrBABAR-
Phos)2(PPh3)2] (iPrBABAR-Phosˆ5a).8 While the phosphir-
ane 5a [SЊ(P)Ͻ250] remains tightly bonded to the platinum
centre it is readily displaced by Ph3P [SЊ(P)Ϸ303] from the
copper complex. This observation is in accord with the
assumption that BABAR-Phos ligands are worse s-donors
but better p-acceptors than Ph3P. Given that platinum(0)
olefin complexes are active hydrosilylation catalysts,1d the
mono BABAR-Phos complexes 11a,b were prepared.8
While 11a could be obtained in pure form as slightly yellow
crystals, the isolation of larger amounts of pure 11b failed.
However, 11b was characterised by multi-nuclear NMR
spectroscopy in solution in presence of excess divinyltetra-
methylsiloxane (dvtms). Furthermore, single crystals suita-
ble for an X-ray analysis were obtained (vide infra). In order
to investigate the effect of the co-ordinated olefin on the
catalyst performance, we also prepared the new maleic
anhydride 12 (malan) complex 13 as shown in Scheme 7.
As expected, protonation results in an increase of positive
charge at the P-atom that rises from 0.95 in III to 1.42 in IV.
The negative charges on the carbon atoms C4 and C5 are
only slightly diminished. However, if the charges on the
hydrogen atoms H4 and H5 are taken into account, a
decrease of the negative group charges C4–H4 (Ϫ0.261 in
III versus Ϫ0.197 in IV) and C5–H5 (Ϫ0.27 in III versus
Ϫ0.21 in IV) is observed. This means that s-charge transfer
from P to H4 and H5 occurs in IV.
It is well known that phosphiranes underlie P–C bond
rupture upon nucleophilic attack which takes place on the
phosphorus centre.3 We have therefore investigated the
effect caused by the smallest ‘nucleophile’ accessible to
chemists: a hydride. Indeed, in the corresponding anion V
the C4–P bond is broken when the geometry is optimised
taking the structure of III as a starting point. In the B3LYP/
6-31G(d) minimum structure the moiety C2–C3–C4 (C2–C3
This new complex forms two isomers, 13 and 130, in
solution. It crystallises in the form of isomer 13, which, in
solution, slowly gives back a 5/1 equilibrium mixture of 13
and 130. We assume that 130 differs from 13 by a 180Њ
rotation of the maleic anhydride ligand, however, a con-
comitant rotation of the BABAR-Phos ligand can not be
excluded. In contrast to 11a,b where both halves of the
3
4
˚
˚
1.400 A, C –C 1.403 A) shows the typical bond character-
istics of an allyl anion. Inspection of the atomic charges
listed in Table 2 show that a considerable part of the extra
negative charge is distributed over C2–C3–C4. The rest of
the additional negative charge is allocated on the N and C5
centres while the charge at the P-atom remains positive. We
conclude from our theoretical inspection of BABAR-Phos
1
1
ligand show identical H and 13C NMR signals, all H and
13C nuclei are different in 13, 130 giving rise to very
complex spectra but indicating that any exchange phenomena

148
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
Scheme 7. Preparation of BABAR-Phos Platinum(0) olefin complexes.
is slow on the NMR time scale. Crystalline 13 is colourless
and stable in air.
or benzene are kept at room temperature, a lemon yellow
powder starts to precipitate after about one hour. The new
compound 21 is sufficiently soluble in THF to obtain NMR
spectra . At room temperature very broad 31P NMR signals
are observed which sharpen at lower temperatures and split
31
In the P NMR spectra of 11a,b and 13,130, one resonance
signal at about Ϫ80 ppm is observed which is approxi-
mately D70 ppm shifted to higher frequencies when
compared to the unbounded BABAR-Phos ligands 5a–c
(ϷϪ150 ppm). The 1J(195Pt31P) coupling constants
(Ϸ4200 Hz) lie in the upper range of values observed
for platinum(0) complexes.11 However, they compare
well with couplings which have been observed for
the bis(phosphirene) complex {Pt(PPh3)2[PhP(CPh)2]}
(1J195Pt31Pˆ4222 Hz) which could be characterised in solu-
tion below Ϫ40ЊC.12 It seems that platinum(0) complexes
containing an intact phosphirane or phosphirene ligand
could never be isolated before.
1
into two AMX spin systems. However, the H NMR and
13C NMR spectra remain very complex. Layering a concen-
trated THF solution of 21 with n-hexane led to crystals
suitable for an X-ray analysis, which allowed unequivocally
the determination of the structure as a binuclear mixed
valent platinum(0) platinum(II) phosphido bridged complex
[Pt2(3,5CF3BABAR-Phos)2(h2Ϫ3,5CF3BABAR-Phos)(dvtms)]
as shown in Scheme 8.
The intermolecular insertion of carbene isolobal electron
rich metal centres, i.e. [Pt(0)(PR3)2] or [Pd(0)(PR3)2] frag-
ments, into one P–C bond of phosphiranes or phosphirenes
is a well documented reaction.12,13 In analogy, one might
propose the reaction of [Pt(3,5CF3BABAR-Phos)2] fragments
with 11b as a reaction path leading to 21. However, more
Rearrangement reaction of 11b
When solutions of 11b in a non-polar solvent like n-hexane
Scheme 8. Synthesis of 21.

J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
149
Figure 3. Molecular structure of dichlorophosphine 4a. Selected bond
˚
lengths [A] and angles [Њ]; P1–Cl1 2.078(1), P1–Cl2 2.117(1), P1–N1
1.640(2), N1–C1 1.503(3), N1–C16 1.500(3), C4–C5 1.337(4), Cl1–P1–
Cl2 95.11(4), Cl1–P1–N1 101.52(9), Cl2–P1–N1 103.83(9), C1–N1–P1
118.4(2).
Figure 5. Molecular structure of the cation of phosphiranium salt 8.
Selected bond lengths [A] and angles [Њ] are listed in Tables 4 and 5,
˚
respectively.
complex reaction steps involving ligand scrambling (which
obviously occurs) and intramolecular insertion of Pt centres
in P–C bonds cannot be ruled out at this stage of our inves-
tigations. Compound 21 can be prepared easily in high yield
when the components 5b, dvtms (10) and [Pt(nbn)3] (9) are
mixed in benzene in the correct stoichiometry. Remarkably,
when dvtms and 9 are added to a solution of 21, complete
but slow reconstitution of 11b is observed. This observation
demonstrates for the first time that the metal insertion
reaction into the P–C bond of a phosphirane can be revers-
ible. We believe that this may be due to the particular elec-
tronic and steric properties of BABAR-Phos leading to a
sufficiently small thermodynamic difference between the
ring opened compound and intact PC2 cycle and rather
small activation barriers for the insertion and extrusion
processes.
Preliminary two-dimensional NMR experiments lead us to
assume that the complex NMR spectra of 21 are due to the
formation of two different isomers in solution which inter-
change on the NMR time scale. Since the phosphorus nuclei
of the intact BABAR-Phos ligand (dϷϪ75) shows no cross-
peak with the 31P nucleus (dϷϪ45) of the ring opened
ligand, an equilibrium between these two can be excluded
and the exchange process very likely consists of a hindered
rotation around the Pt–P bonds of the BABAR-Phos ligands.
Molecular structures of 4a, 5b, 8, 11b, 13 and 21
We have characterised the dichlorophosphine 4a (Fig. 3),
BABAR-Phos 5b8 (Fig. 4), the phosphiranium salt 8 (Fig. 5),
and the platinum complexes 11a8 and 11b (Fig. 6), 13
(Fig. 7), and 21 (Fig. 8) by X-ray analyses.14
For comparison, selected bond lengths and angles of the
BABAR-Phos framework in these compounds are listed in
Tables 4 and 5, respectively. Further selected data are given
in the figure captions. The bond parameters show some
common expected trends: in all compounds the ring internal
P–C bonds (a1,a2) and P–N (c) bonds become shorter when
the P-lone pair is involved in bonding. At the same time, the
basal C–C bond, b, becomes longer and the sum of bond
angles, SꢀP†Њˆaϩb1ϩb2; at phosphorus becomes larger.
These effects are most pronounced in the phosphiranium
salt 8 where the lone pair at phosphorus becomes a covalent
Figure 4. Molecular structure of BABAR-Phos 5b. Selected bond lengths
[A] and angles [Њ] are listed in Tables 4 and 5, respectively.
˚

150
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
˚
Figure 6. Molecular structure of 11b. Further selected bond lengths [A] and angles [Њ] are listed in Tables 4 and 5, respectively; X1 lies in the centre of the bond
C24yC25, X2 in the centre of C30yC31, Pt1–C24 2.140(4), Pt1–C25 2.152(4), Pt1–C30 2.156(4), Pt1–C31 2.135(4), C24–C25 1.430(5), C30–C31 1.405(6),
P1–Pt1–X1 113.2(2), P1–Pt1–X2 113.9(2), X1–Pt1–X2 132.7(2).
P–C s-bond. The structural parameters of the BABAR-Phos
unit of all platinum complexes regardless of formal oxi-
dation state or co-ordination number are quite similar, i.e.
S(P)Њ lies in a narrow range of 251–257; about 10Њ larger
than in the unbound ligand 5b. The Pt–P bond lengths in the
three-co-ordinate platinum(0) complexes 11a,b and 13 are
somewhat shorter than the distances of the four-co-ordinate
platinum(II) centre to the phosphorus centres of the
˚
Figure 7. Molecular structure of 13. Further selected bond lengths [A] and angles [Њ] are listed in Tables 4 and 5, respectively. X1 lies in the centre of the bond
C25yC26; X2 in the centre of C32yC33. The 3,5-(CF3)2C6H3 substituent on N1 has been omitted for clarity. Pt1–C25 2.17(1), Pt1–C26 2.19(1), Pt1–C32
2.10(1), Pt1–C33 2.08(1), C25–C26 1.38(2), C32–C33 1.42(2), P1–Pt1–X1 108.0(4), P1–Pt1–X2 121.9(4) X1–Pt–X2 130.1(4).

J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
151
˚
Figure 8. Molecular structure of 21. Further selected bond lengths [A] and angles [Њ] are listed in Tables 4 and 5, respectively; X1 lies in the centre of the bond
C24yC25, X2 in the centre of C30yC31, P1–Pt1 2.333(2), P1–Pt2 2.274(2), P1–C4 2.542(2), P1–C5 1.832(8), Pt1–C4 2.120(8), C4–C5 1.550(11), N1–P1
1.738(7), Pt2–C24 2.133(10), Pt2–C25 2.152(9), Pt2–C30 2.150(9), Pt2–C31 2.122(9), C30–C31 1.410(12), C24–C25 1.418(13), P1A–Pt1–P1B 91.26(8),
C4–Pt1–P1A 98.0(2), P1–Pt1–P1B 101.17(8), P1–Pt1–C4 69.3(2), Pt1–P1–N1 102.1(2), Pt1–P1–C5 86.0(3), N1–P1–C5 97.1(4), Pt1–P1–Pt2 130.6(1),
P1–Pt2–X1 114.1(4), P1–Pt2–X2 111.6(4) X1–Pt2–X2 134.1(4). Sum of bond angles at P1: 285.2Њ.
˚
Table 4. Selected bond lengths of the BABAR-Phos entities in 5b, 8, 11a,b,
13 and 21
the Pt(0)– PPh3 distances [2.324(1) A] observed in
[Pt(iPrBARBAR-Phos)2(PPh3)2].8
˚
˚
˚
˚
˚
Z
a1 [A]
a2 [A]
b [A]
c [A]
d [A]
The structure of the ring opened phosphirane moiety in 21
merits some attention. Apart from the P–C distances, which
5ba
8
–
1.850(3)
1.863(3)
1.774(3)
1.821(7)
1.844(4)
1.532(4)
1.601(6)
1.576(9)
1.544(6)
1.738(2)
1.628(4)
1.684(6)
1.711(3)
–
tBu 1.774(3)
1.823(5)
2.241(2)
2.231(1)
˚
became very different [P1–C5 1.832(8) A; P1–C4
11aa Pt
1.812(6)
1.811(4)
˚
˚
˚
2.542(2) A versus 1.811(8) A and 1.825(8) A for a1 and
a2, respectively] as a consequence of the insertion reaction,
other structural parameters remain remarkably constant (i.e.
the N–P–C5 angles vary by 3Њ and the C–C distances
within the C–C skeleton do not change significantly).
11b
13
Pt
Pt
1.825(10) 1.827(10) 1.597(14) 1.682(10) 2.251(2)
1.825(8) 1.561(12) 1.687(7) 2.295(2)
21b
Pt1 1.811(8)
a See Ref. 8.
b Averaged data of both BABAR-Phos ligands are given; the largest
˚
difference being 0.04 between the Pt–P1A (2.313(2) A) and Pt–P1B
(2.277(2) A) distances.
˚
This is probably a consequence of the rigid polycyclic
structure of the BABAR-Phos ligand and may explain the
reversibility of the insertion reaction. The fold angle along
the P1–C4 axis within the Pt1–P1–C4–C5 cyclobutane
moiety amounts to 25.8Њ. Since the [Pt(3,5CF3BARBAR-
Phos)2(h2-3,5CF3BARBAR-Phos)] unit may be described
as a phosphine ligand co-ordinating to the Pt2 centre
˚
BABAR-Phos ligands in 21 [Pt–P1A (2.313(2) A), Pt–P1B
˚
(2.277(2) A], which reflects the sterical crowding in this
˚
complex. The longest Pt–P distance [2.333(2) A] is found
in 21 between the bridging phosphido centre P1 and the
platinum(II) centre Pt1 which falls in the range of
Table 5. Selected bond angles in 5b, 8, 11a,b, 13 and 21
a [Њ]
b1 [Њ]
b2 [Њ]
S [Њ]
g1 [Њ]
g2 [Њ]
g3 [Њ]
5ba
8
48.8(1)
53.6(2)
51.4(3)
50.0(2)
51.8(5)
50.9(4)
98.8(1)
108.2(2)
101.7(3)
100.5(2)
102.1(5)
101.0(4)
99.0(1)
108.2(2)
103.1(3)
101.4(2)
103.3(5)
102.7(4)
246.5
269.9
256.2
251.9
257.2
254.6
–
–
–
119.7(2)
119.8(2)
122.5(1)
121.8(3)
124.4(3)
124.9(2)
130.0(2)
125.4(1)
130.5(4)
128.1(3)
124.9(2)
131.4(2)
133.7(1)
126.8(4)
125.8(3)
11aa
11b
13
21b
a See Ref. 8.
b
˚
˚
Averaged data of both BABAR-Phos ligands are given; the largest difference being 0.04 between the Pt–P1A (2.313(2) A) and Pt–P1B (2.277(2) A)
distances.

152
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
Figure 9. Hydrosilylation of triethylvinylsilane 15 with triethylsilane 14. Conditions: 4 mmol 15, 2 mmol 14, 0.002 mmol catalyst ([Pt(nbn)3] (9)/phosphine
ratioˆ1:1.1). Catalyses: No. 1 (×): 9/5a; No. 2 (X): 9/5b; No. 3 (O): 9/P(OPh)3; No. 4 (O): 9/P(NiPr2)Ph2; No. 5 (O): 13; No. 6 (B): 9/PPh3.
we note the sum of bond angles around P1 which
amounts to 285.2Њ.
reaction no. 5). In the solutions of 9 and 5a,b, broad signals
in the 31P NMR spectra at about Ϫ70 ppm indicate
exchange reactions in which probably differently substituted
[Pt(BABAR-Phos)n(nbn)m] (m, n: 1–3) species are involved.
These complexes were found to be unstable and could not be
isolated. However, it can be clearly seen from Fig. 9, that
platinum(0) phosphirane complexes show considerable
catalytic activity in reactions (A) and (B). Specifically, no
conversions were observed in reaction (A) with P(OPh)3, or
P(NiPr2)Ph2 (reaction no. 3 and 4) or when the maleic anhy-
dride complex 13 was used as catalyst precursor (reaction
no. 5). When PPh3 is employed as ligand, little conversion
(Ϸ5%) is seen initially but the reaction does not proceed any
further.
Hydrosilylation reactions catalysed by platinum
BABAR-Phos complexes
We have reported the use of BABAR-Phos 5a,b as ligands
for catalyst precursors in hydrosilylation reactions (A) (see
Fig. 9) and (B) (see Fig. 10).8 In these experiments, we
mixed [Pt(nbn)3] (9) with an excess of dvtms (10) prior to
the addition of BABAR-Phos 5a,b to afford catalyst precur-
sors 11a,b. By means of NMR spectroscopy, the complexes
11a,b were the only platinum complexes detected in these
solutions.
The binuclear complex 21 also shows no catalytic activity in
reaction (A). There are different possible reasons why
neither platinum precursor complexes generated with phos-
phines bearing electron withdrawing substituents, i.e.
P(OPh)3 or P(NiPr2)Ph2, or electron poor olefins nor the
platinum complex 21 lead to catalytic active complexes.
In order to exclude the possibility that the catalytic activity
which we observed is not due to the high activity of
[Pt2(dvtms)3] (Karstedt catalyst)1d,15 which may be possibly
present in small amounts, we have now altered slightly the
protocol and performed the reactions without dvtms.
Instead, we used a mixture of 1 equivalent [Pt(nbn)3] (9)
and 1.1 equivalents of a phosphine [i.e. 5a in reaction no. 1,
5b in reaction no. 2, P(OPh)3 in reaction no. 3, P(NiPr2)Ph2
in reaction no. 4 and PPh3 in reaction no. 6] or isolated 13 (in
One may be that the use of the phosphines PPh3, P(OPh)3, and
P(NiPr2)Ph2 leads to oligo(phosphine) complexes—eventually

J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
153
Figure 10. Hydrosilylation of trimethoxyvinylsilane 18 with trimethoxysilane 19. Conditions: 1 mmol substrate, 0.5 mmol catalyst, C6D6 as solvent. Catalyses:
no. 1: 9/5b; No. 2: 9/5a; No. 3: [Pt2(dvtms)3].
in the course of the catalytic reaction—in which free
co-ordination sites for the substrate molecules are blocked.
Such [Pt(PR3)nLm] species are expected to be less substi-
tution labile than simple platinum(0) olefin complexes.
Also in line with this assumption is the observation that
the maleic anhydride complex 13 is inactive. Probably due
to the high affinity of the electron poor olefin 12 towards the
platinum centre, the oxidative addition and/or co-ordination
of the olefin becomes hampered [step (a) and (b) in Scheme
1, respectively].
mixtures of 9/5b (reaction no. 1); 9/5a (reaction no. 2),
and the Karstedt catalyst [Pt2(dvtms)3] (reaction no. 3)
were used. In the hydrosilylation reaction (B) two isomers,
19 and 20, are obtained as products. As can be seen from the
comparison, the Karstedt catalyst is still significantly more
active but, as expected, less selective than the platinum(0)
BABAR-Phos precursors. Note, however, that in contrast to
reactions with Karstedt catalyst, we did not observe the
formation of elemental platinum with BABAR-Phos
platinum complexes. This observation indicates the higher
stability of BABAR-Phos platinum complexes, which is an
important issue in view of efforts directed towards the
recycling of the catalyst. Qualitatively, all results obtained
in this work resemble the ones we observed when all cata-
lyses were carried out in the presence of dvtms whereby
platinum(0) phosphine dvtms complexes are formed as
precursors.8
Another reason, however, may indeed be that the particular
electronic properties, i.e. the small sum of bond angles at the
phosphorus centre in BABAR-Phos complexes, lead to more
active phosphine complexes. The (reproducible) drop off of
the reaction rate after approx. 2000 min using 9 and 5b as
the catalytic system may be explained by the inactivity of
21, which may form in the course of the catalysis reaction.
In this binuclear complex the ring opened phosphirane
behaves like an ‘ordinary’ phosphine [SЊ(P1)ˆ285] towards
the Pt(0)(dvtms) moiety. In preliminary work we have
already shown that the dvtms complex [Pt(PPh3)(dvtms)]
is catalytically inactive.8 Since the reconstitution of cata-
lytically active BABAR-Phos species from 21 in the
presence of other labile platinum(0) complexes—which
may be present in solution—is quite slow, a decrease of
the conversion rate is the consequence. Note that we have
so far no evidence that a phosphirane ring opening reaction
occurs when the N-isopropyl substituted BABAR-Phos
ligand 5a is employed and hence no unusual decrease in
activity is observed.
Conclusion
We have shown that phosphiranes are interesting as ligands
in transition metal catalysed reactions. Specifically, cata-
lytic reactions in which reductive elimination is the overall
rate determining step, may profit from the particular elec-
tronic properties of phosphiranes (i.e. a relatively high posi-
tive charge on the phosphorus centre and low lying
p-acceptor orbitals). The approach of incorporating the
PC2 cycle within a rigid polycyclic framework may prove
to be fruitful in order to suppress undesired side reactions.
Thus we could demonstrate that the insertion of electron
rich metal centres in one P–C bond is reversible. It will
be particularly interesting to search for BABAR-Phos
ligands in which the insertion and extrusion in one P–C
bond is even more facile. This would allow development
In Fig. 10 the results for the catalysed hydrosilylation reac-
tion of tri(methoxy)vinyl silane 18 with trimethoxysilane 17
are shown [reaction (B)]. As catalyst precursors 1:1.1

154
Table 6. Selected NMR and physical data of 4a–c, 5a–c, 6, 7, 8, 11a,b, 13 and 21a
31P 195Pt 1H PCHCH
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
d
d
d
d
13C PCHCH
m.p. [ЊC]
4a
4b
4c
5a
169
155
162
Ϫ154
7.12
7.21
7.25
130.7
130.6
130.6
114
122
–
2
1
2.65, d, JPHˆ22
21.3, d, JPCˆ38
85
147
2
5b
Ϫ146
3.12, d, JPHˆ20
1
22.4, d, JPCˆ38
4
2
5c
6
7
Ϫ146, t, JFPˆ41
3.23, d, JPHˆ21
5
7.05
131.5
–
125
50
160
7.11, 7.12, JABˆ12
130.5,d,
J
ˆ1;131.6, d,
PC
5JPCˆ2
2
1
8
Ϫ16
4.38, d, JPHˆ4.8
21.5, d, JPCˆ8
133
1
1
2
3
1
2
11a
11b
13
Ϫ74, JPtPˆ4203
Ϫ3688, d, JPPtˆ4203
2.85, JPH2ˆ5.7, JPt3Hˆ13.8
23.1, JPCˆ15, JPtCˆ18
(dec.)Ͼ95
(dec.) Ͼ85
(dec.)Ͼ165
1
1
1
2
Ϫ77, JPtPˆ4215
Ϫ3616, d, JPPtˆ4215
3.02, m, JPHˆ6.0, JPtHˆ12
25.3, d, JPCˆ18, JPtCˆ14
1
1
3
2
1
2
Ϫ77, JPtPˆ4247
Ϫ7169, d, JPPtˆ4247
3.20, m, JHHˆ4.5, JPHˆ11.7,
21.4, d, JPCˆ15, JPtCˆ16;
1 2
3JPtHˆ16; 3.7, m, JHHˆ4.5,
23.3, d, JPCˆ12, JPtCˆ12
3
2JPHˆ11.7, JPtHˆ12
3
21a
Ϫ46, m,
Ϫ5416, m
(dec.)Ͼ163
1JPtPˆ3201,
1JPtPˆ1599,
2JPPˆ211 (P1);
Ϫ80, m,
1JPtPˆ1960,
2JPPˆ211, JPPˆ16
2
(P1A); Ϫ88, m,
1JPtPˆ2090,
3JPtPˆ84 (P1B).
a Only the NMR data of the major isomer are given.
of catalysts in which the metal centre is parked within the
ligand during the resting state but pulled out by substrate
molecules during the active state.16
(s, Ca), 57.31, 66.21 (s,C1), 122.28, 125.44, 126.86, 127.59,
128.45, 128.66, 129.42, 129. 95, 130.45, 131.12 (s, C4, C8,
C9, C10, C11), 133.15, 133.88, 139.99, 140.66 (s,C2, C3).
3b (due to endo exo isomerisation all resonances are
1
exchange broadened at Tˆ298 K17): H NMR (CDCl3): d
Experimental
4.88 (broad, s, NH), 5.75 (broad, s, CH), 7.04 (s, CH olefin,
2H), 7.2–7.7 (m, CH aryl, 10 H). 13C NMR (CDCl3): d
63.89 (broad s, C1), 110.16 (broad s, Ca), 112.59 (broad s,
The preparation of BABAR-Phos 5a,b and complexes 11a,b
has been briefly reported before.8 Selected NMR and physi-
cal data of the compounds described in this work are listed
in Table 6.
1
Cb), 123.39 (q, JFCˆ272.7 Hz, Ce), 127.35 (broad, s, Cd),
128.74, 129.98, 130.58, 132,99, 137.22 (s, C4, C8, C9, C10,
2
C11), 131.74 (q, JFCˆ32.6 Hz, Cc), 147.31 (s, C2 or C3).
19F NMR (CDCl3): d Ϫ63.5 (s).
Dibenzotropylidene amines 3a, 3b, 3c. The chloride 2 was
prepared as described in the literature.17 To a solution of
5.41 g 2 (23.8 mmol) in 500 ml toluene, 47.6 mmol of the
corresponding primary amine RNH2 (RˆiPr, 3,5-(CF3)2C6H3,
F5C6) were added under a dry atmosphere. After about 1 h a
white precipitate of amine hydrochloride has formed.
Subsequently 200 ml of saturated aqueous Na2CO3 was
added to the mixture. The organic phase was separated
and the aqueous phase extracted three times with 200 ml
of diethyl ether. The combined organic layers were dried
over sodium sulphate and the solvents were removed
under reduced pressure. The crude products were
re-crystallised from hexane to give 85–95% of the
pure colourless amines.
3c (due to endo exo isomerisation all resonances are
1
exchange broadened at Tˆ298 K17): H NMR (CDCl3) d
4.82 (broad, s, NH), 5.91 (broad, s, CH), 7.32 (s, CH olefin,
2H), 7.34–7.77 (m, CH aryl, 8 H). 13C NMR (CDCl3): d
65.79 (broad s, C1), 121.80 (broad s, CF), 122.04 (broad s,
CF), 127.46 (s, Ca), 128.23, 128.35, 128.68, 129.00, 130.36
(s, C4, C8, C9, C10, C11), 132.63 (s, C2 or C3), 135.65 (m, CF),
137.72 (s, C2 or C3).
General procedure for the preparation of the amino-
dichloro phosphines 4a, 4b, 4c and 7. To a solution of
the amine 3a, 3b or 3c (40 mmol) in diethyl ether, 25 ml
of a 1.6 M hexane solution of nBuLi (40 mmol) was
added at Ϫ30ЊC. The reaction mixture was warmed to
room temperature and then added dropwise to a solution
of either PCl3 (5 ml, 57 mmol) or tBuPCl2 (3.8 g,
24 mmol) in 20 ml Et2O at Ϫ20ЊC, respectively. After
removing all solvents under vacuum, the residue was
dissolved in 50 ml toluene and filtered over Celite.
Evaporation of the toluene yielded nearly pure solid
3a (a slow equilibrium at Tˆ298 K between endo and exo
isomer17 of amine 3a is established by the doubling of the
1H and 13C resonances):1H NMR (CDCl3): d 0.92, 1.13 (d,
3JHHˆ6.20 Hz, CH3, 6H), 1.93 (broad, s, NH, 1H), 2.24,
3
2.91 (sept, JHHˆ6.20 Hz, CHiPr, 1H), 4.27, 4.94 (s, CH,
1H), 7.04, 7.16 (s, CH olefin, 2H), 7.25–7.45 (m, CH aryl,
8H). 13C NMR (CDCl3): d 22.45, 23.32 (s, Cb), 44.48, 45.84

J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
155
Donahue, P. E. Organometallics 1991, 10, 3750 and references
therein.
products which can be re-crystallised from toluene (4b, 4c
and 7) or Et2O (4a) (yields Ͼ90%).
2. (a) Albright, T. A.; Burdett, J. K.; Whangbo, M.-H. Orbital
Interactions in Chemistry; Wiley: New York, 1985; p 140. (b)
Dunne, B. J.; Morris, R. B.; Orpen, A. G. J. Chem. Soc., Dalton
Trans. 1991, 653 and references therein.
3. Recent reviews: (a) Dillon, K. B.; Mathey, F.; Nixon, J. F.
Phosphorus: The Carbon Copy; Wiley: New York, 1998; p 182.
(b) Mathey, F.; Regitz, M. Comprehensive Heterocyclic Chemistry
II; Padwa, A., Ed.; Elsevier: Amsterdam, 1996; p 277. (c) Mathey,
F. Chem. Rev. 1990, 90, 997.
BABAR-Phos 5a, 5b and 5d. A solution of 4a, 4b or 4c
(29 mmol) in 100 ml THF was stirred in presence of 1 g
(41 mmol) Mg turnings at room temperature. After 5 h,
5 ml of dry dioxane were added to precipitate MgCl2. The
solvents were evaporated, the residue was dissolved in
100 ml toluene and filtered over Celite. The toluene phase
was concentrated to 10% of its volume and hexane was
added to precipitate spectroscopically pure 5a or 5b.
Compound 5c could not be isolated prior to polymerisation
in concentrated solutions.
4. Only selected recent publications are given here which point out
the names of research groups involved the synthesis and chemistry
´
of phosphiranes: (a) Mezailles, N.; Fanwick, P. E.; Kubiak, C. P.
8. A solution of 1 g of 7 (2.7 mmol) in 5 ml CH2Cl2 was
added to a vigorously stirred suspension of 0.4 g AlCl3
(3 mmol) in 20 ml CH2Cl2 at Ϫ78ЊC. The solution was
filtered and the solvent removed in vacuo to give 1.3 g
(2.6 mmol) 7 (98%) as a white micro-crystalline powder.
Crystals suitable for an X-ray crystal structure determina-
tion were obtained by re-crystallisation from a saturated
CH2Cl2 solution.
Organometallics 1997, 16, 1526. (b) Li, X.; Robinson, K. D.;
Gaspar, P. P. J. Org. Chem. 1996, 61, 7702. (c) Hockless, D. R.;
Kang, Y.; McDonald, M. A.; Pabel, M.; Willis, A. C.; Wild, S. B.
Organometallics 1996, 15, 1301. (d) Marinetti, A.; Ricard, L.;
Mathey, F. Synthesis 1992, 157. (e) Marinetti, A.; Mathey, F.
Tetrahedron Lett. 1987, 28, 5021. (f) Tsuji, K.; Sasaki, S.;
Yoshifuji, M. Heteroat. Chem. 1998, 7, 607. (g) Review: Mathey,
F. Angew. Chem. 1987, 99, 285; Angew. Chem., Int. Ed. Engl.
1987, 26, 275. (h) Krill, S.; Wang, B.; Hung, J.-T.; Horan, C. J.;
Gray, G. M.; Lammertsma, K. J. Am. Chem. Soc. 1997, 119, 8432.
(i) Becker, P.; Brombach, H.; David, O.; Leuer, M.; Metternich,
H.-J.; Niecke, E. Chem. Ber. 1992, 125, 771. (j) Kolodiazhnyi,
O. I.; Gishkun, V. E. Heteroat. Chem. 1998, 9, 659. (k)
Yoshifuji, M.; Toyota, K.; Yoshimura, H.; Hirotsu, K.; Okamoto,
A. J. Chem. Soc., Chem. Commun. 1991, 124. (l) Manz, B.; Maas,
G. J. Chem. Soc., Chem. Commun. 1995, 25. (m) Lentz, D.;
Marschall, R. Z. Anorg. Allg. Chem. 1992, 617, 53. (n) Yoshifuji,
M.; Yoshimura, H.; Toyota, K. Chem. Lett. 1990, 827. (o) Schnurr,
11a and 11b. Platinum complex 11a was obtained in 70%
yield (97 mg) by mixing 53 mg (0.11 mmol) [Pt(nbn)3] (9),
50 mg (0.27 mmol) dvtms (10) and 60 mg (0.21 mmol) 5a
in 2 ml toluene. Crystals were grown from concentrated
hexane solutions. Complex 11b was prepared in the same
manner but its isolation failed because of decomposition in
the absence of excess dvtms. Although decomposition
during work-up took place, 11b could be fully characterised
by NMR-spectroscopy in the presence of excess dvtms.
Some crystals of 11b were grown out of hexane and could
be isolated beside oily products resulting from
decomposition.
¨
W.; Regitz, M. Tetrahedron Lett. 1989, 30, 3951. (p) Markl, G.;
Hohenweider, K.; Ziegler, M. L.; Nuber, B. Tetrahedron Lett.
¨
¨
1990, 31, 4849. (q) Markl, G.; Holzl, W. Tetrahedron Lett. 1989,
30, 4501. (r) Zablocka, M.; Miquel, Y.; Igau, A.; Majoral, J. P.;
Skowronska, A. J. Chem. Soc., Chem. Commun. 1998, 1177.
5. [2ϩ1] cycloreversions: (a) Gaspar, P. P.; Li, X.; Silverman, J.;
Haile, T.; Pae, D. H.; Xiao, M. Prog. Organosilicon Chem.
(Jubilee Int. Symp. Organosilicon Chem.), 10th ed.; Mariniec, B.,
Chojnowski, J., Eds.; 1995; p 247. (b) Li, X.; Weissman, S. I.; Lin,
T.-S.; Gaspar, P. P. J. Am. Chem. Soc. 1994, 116, 7899 and refer-
ences therein. (c) Retro-electrocyclisation: Chaquin, P.; Gherbi, A.
J. Org. Chem. 1995, 60, 3723. (d) Ring–chain rearrangements:
Nguyen, M. T.; Landuyt, L.; Vanquickenborne, L. G. J. Chem.
Soc., Faraday Trans. 1994, 90, 1771. (e) Recent reports on thermal
rearrangements of phosphirane complexes: Wang, B.; Lake, C. H.;
Lammertsma, K. J. Am. Chem. Soc. 1996, 118, 1690. (f) Huy, N. H.
T.; Mathey, F. Synlett. 1995, 353. (g) Flash pyrolysis of phosphir-
anes: Haber, LeFloch, P.; Mathey, F. Phosphorus, Sulfur and Sili-
con 1993, 75, 225. (h) Kobayashi, S.; Kadokawa, J.-I. Marcomol.
Rapid Commun. 1994, 15, 567; Surface phosphinidenes [RP(Mg]
have been proposed: Bock, H.; Bankmann, M. J. Chem. Soc.,
Chem. Commun. 1989, 1130.
13. To a solution of 100 mg (0.21 mmol) 9 and 21 mg
(0.21 mmol) maleic anhydride 12 in 2 ml toluene, 96 mg
(0.21 mmol) 5b dissolved in 2 ml toluene were added.
After 30 min the solvent and free nbn were removed in
vacuo. The resulting beige powder was washed 2 times
with 0.5 ml benzene to give 130 mg (1.6 mmol, 75%) 13
as white crystalline powder after drying in vacuo.
21. To a solution of 100 mg (0.21 mmol) 9 and 20 mg
(0.11 mmol) dvtms in 2 ml hexane was added dropwise a
solution of 5b in 2 ml benzene. After a few seconds 21 starts
to precipitate as yellow micro crystalline powder which was
filtered off and washed with 2 ml hexane. After drying in
vacuo 160 mg (0.08 mmol, 78%) of pure 21 were obtained.
Layering a concentrated THF solution with hexane yielded
crystals suitable for an X-ray analysis.
6. Marinetti, A.; Mathey, F.; Ricard, L. Organometallics 1993, 12,
1207.
References
¨
7. (a) C: Binger, P.; Leisinger, S.; Regitz, M.; Bergstrasser, U.;
1. For recent calculations see: (a) Bode, B. M.; Day, P. N.;
Gordon, M. S. J. Am. Chem. Soc. 1998, 120, 1552 and references
therein. (b) Sakaki, S.; Ogawa, M.; Musashi, Y.; Arai, T. J. Am.
Chem. Soc. 1994, 116, 7258. (c) Sakaki, S.; Ikeki, M. J. Am. Chem.
Soc. 1993, 115, 2373. (d) An overview about some experimental
aspects is given by: Lewis, L. N.; Sy, K. G.; Bryant, Jr., G. L.;
Bruckmann, J.; Kru¨ger, C. J. Organomet. Chem. 1997, 529, 215
¨
and references therein. (b) D: Hu, D.; Schaufele, H.; Pritzkow, H.;
Zenneck, U. Angew. Chem. 1989, 101, 929; Angew. Chem., Int. Ed.
Engl. 1989, 28, 900. (c) E: Avarvari, N.; Ricard, L.; Mathey, F.; Le
¨
Floch, P.; Lober, O.; Regitz, M. Eur. J. Org. Chem. 1998, 2039. (d)
¨
¨
F: Heydt, H.; Bergstrasser, U.; Fassler, R.; Fuchs, E.; Kamel, N.;

156
J. Liedtke et al. / Tetrahedron 56 (2000) 143–156
¨
Mackewitz, T.; Michels, G.; Rosch, W.; Regitz, M.; Mazerolles,
P.; Laurent, C.; Faucher, A. Bull. Soc. Chim. Fr. 1995, 132, 652. (e)
Also, P4 (G) is a quite stable molecule.
(Rint.ˆ0.0613); R1ˆ3.21%, wR2ˆ7.10% (based on F2) for 388
parameters and 5992 reflections with IϾ2s(I). 13: Triclinic,
space
group
P1(bar);
aˆ11.1323(5),
bˆ11.4105(4),
˚
8. Liedtke, J.; Loss, S.; Alcaraz, G.; Gramlich, V.; Gru¨tzmacher,
H. Angew. Chem. 1999, 111, 1724; Angew. Chem. Int. Ed. Engl.
1999, 38, 1623.
9. (a) Hockless, D. C. R.; McDonald, M. A.; Pabel, M.; Wild, S. B.
J. Organomet. Chem. 1997, 529, 189. (b) For p-exchange reactions
of phosphiranium salts see: Sølling, T. I.; McDonald, M. A.; Wild,
S. B.; Radom, L. J. Am. Chem. Soc. 1998, 120, 7063.
10. Nguyen, M. T.; Van Praet, E.; Vanquickenborne, L. G. Inorg.
Chem. 1994, 33, 1153.
11. Pregosin, P. S. Annual Reports on NMR Spectroscopy 1986,
17, 285.
12. Al Juaid, S. S.; Carmichael, D.; Hitchcock, P. B.; Marinetti,
A.; Mathey, F.; Nixon, J. A. J. Chem. Soc., Dalton Trans. 1991,
905.
cˆ13.3558(4) A, aˆ68.374(1), bˆ77.122(2), yˆ84.763(1)Њ;
3
˚
Vˆ1537.4(1) A ; Zˆ2, MoKa-radiation, 2Qmaxˆ56.6Њ. 6780
reflections, 5333 independent (Rint.ˆ0.0459); R1ˆ5.99%,
wR2ˆ14.18% (based on F2) for 386 parameters and 3598 reflec-
tions with IϾ2s(I). 21: Monoclinic, space group P21/n;
˚
aˆ12.7945(5), bˆ25.6583(10), cˆ24.2287(9) A, bˆ95.005(1)Њ;
3
˚
Vˆ7923.6(5) A ; Zˆ4, MoKa-radiation, 2Qmaxˆ52.7Њ. 48355
reflections, 16125 independent (Rint.ˆ0.0890); R1ˆ5.42%,
wR2ˆ10.17% (based on F2) for 993 parameters and 9944 reflec-
tions with IϾ2s(I). All structures were solved by direct methods
and refined against full matrix (versus F2) with SHELXTL
(Version 5.0). Non-hydrogen atoms were refined isotropically,
hydrogen atoms were refined on calculated positions using the
riding model. The isopropyl group in structure 8 as well as one
CF3-group in 11b and 21 were disordered over 2 or 3 positions,
respectively. In structure 21 five partially occupied atoms equiva-
lent to carbon were refined to describe one disordered THF mole-
cule. Crystallographic data (excluding structure factors) for the
structures reported in this paper have been deposited with the
Cambridge Data Centre as supplementary publication nos.
CCDC-119359 (4a); 106558 (8); 119356 (11b); 119357 (13);
119358 (21). Copies of the data can be obtained free of charge
on application to CCDC, 12 Union Road, Cambridge, CB21EZ,
UK (fax: ϩ44-1223-336-033; e-mail: deposit@ccdc.cam.ac.uk).
15. (a) Chandra, G.; Lo, P. Y.; Hitchcock, P. B.; Lappert, M. F.
Organometallics 1987, 6, 191. (b) Hitchcock, P. B.; Lappert, M. F.;
Warhurst, N. J. W. Angew. Chem. 1991, 103, 439; Angew. Chem.,
Int. Ed. Engl. 1991, 30, 438.
13. (a) Carmichael, D.; Hitchcock, P. B.; Nixon, J. F.; Mathey, F.;
Ricard, L. J. Chem. Soc., Dalton Trans. 1993, 1811 and references
therein. (b) Ajulu, F. A.; Hitchcock, P. B.; Mathey, F.; Michelin,
R. A.; Nixon, J. A.; Pombeiro, A. J. L. J. Chem. Soc., Chem.
Commun. 1993, 142.
14. 4a: Triclinic, space group P1(Bar); aˆ11.1247(3),
˚
bˆ11.9546(3), cˆ14.8553(4) A, aˆ105.010(1), bˆ101.051(1),
3
˚
yˆ106.338(1)Њ; Vˆ1755.00(8) A ; Zˆ2, MoKa-radiation,
2Qmaxˆ49.4Њ.
11185
reflections,
5935
independent
(Rint.ˆ0.0328); R1ˆ4.48%, wR2ˆ10.80% (based on F2) for 397
parameters and 4244 reflections with I Ͼ2s(I). 8: Orthorhombic,
space
group
˚
Pnma;
aˆ18.4162(6),
bˆ11.7414(4),
3
˚
cˆ11.9241(4) A; Vˆ2578.4(2) A ; Zˆ4, MoKa-radiation,
2Qmaxˆ46.5Њ.
13763
reflections,
1959
independent
(Rint.ˆ0.0555); R1ˆ4.33%, wR2ˆ9.91% (based on F2) for 166
16. A reaction catalysed by such a catalyst may be called
Park&Ride catalysis and finds prominent precedence in biological
systems.
17. Kessler, H. In Methoden der Organischen Chemie, (Houben-
Weyl); Mu¨ller, E., Ed.; Thieme: Stuttgart, 1972; Vol. 5/1d, p 301.
parameters and 1421 reflections with IϾ2s(I). 11b: Monoclinic,
space
group
˚
P21/c;
aˆ12.4110(3),
bˆ13.3071(3),
3
˚
cˆ19.9116(5) A, bˆ94.848(1)Њ; Vˆ3276.7(1) A ; Zˆ4, MoKa-
radiation, 2Qmaxˆ56.6Њ. 26 804 reflections, 8117 independent