Sterically crowded diphosphinomethane ligands: Molecular structures, UV-photoelectron spectroscopy and a convenient general synthesis of tBu2PCH2PtBu2 and related species

Article abstract of DOI:10.1039/b210114a

A series of highly crowded symmetric and unsymmetric diphosphinomethanes R₂PCH₂PR′₂, important ligands in transition metal chemistry and catalysis, namely ^tBu₂PCH₂p^tBu₂ (dtbpm, 1¹), Cy₂PCH₂PCy₂ (dcpm, 2), ^tBu₂PCH₂PCy₂ (ctbpm, 3), ^tBu₂PCH₂PⁱPr₂ (iptbpm, 4) and ^tBu₂PCH₂PPh₂ (ptbpm, 5), has been prepared in high yields, using a general and convenient route, which is described in detail for 1. Other than 4, which is a colourless liquid, these compounds are crystalline solids at room temperature. Their molecular structures have been determined by single crystal X-ray diffraction, along with that of the higher homologue of 1, ^tBu₂CH₂CH₂^tBu ₂ (dtbpe, 6). The solid-state structures of the dioxide of 1, ^tBu₂P(O)CH₂P(O)^tBu₂ (7), and of two phosphonium cations derived from 1, protonated [^tBu₂P(H)CH₂P^tBu₂] ⁺ (8⁺) and the chlorophosphonium ion [^tBu₂P(Cl)CH₂P^tBu₂] ⁺ (9⁺), are also described and show a distinct structural influence of the tetracoordinate P centres. The gas phase UV-photoelectron spectra of the diphosphines 1-6 have been measured. Their first two ionisation potentials are found to be nearly degenerate and all are in the low energy range from 7.5 to 7.8 eV. Comparison with related mono- and bidentate phosphines demonstrates that 1-6 are excellent σ-donors towards metals, in accord with their known coordination chemistry. Molecular geometries and electronic structures of the diphosphine systems have been studied by quantum chemical calculations and are compared to experiment. Unlike standard semiempirical methods (AM1, PM3, MNDO), which give rather poor minimum structures and seem inadequate for such sterically crowded systems, ab initio calculations (RHF/6-31G^**) predict molecular geometries with reasonable accuracy and reflect the observed trends in experimental ionisation potentials.

Full text of DOI:10.1039/b210114a

View Article Online / Journal Homepage / Table of Contents for this issue
Sterically crowded diphosphinomethane ligands: molecular structures,
UV-photoelectron spectroscopy and a convenient general synthesis of
tBu2PCH2PtBu2 and related speciesyz
a
a
a
a
a
¨
¨
Frank Eisentrager, Alexander Gothlich, Irene Gruber, Helmut Heiss, Christoph A. Kiener,
Carl Kru¨ger,a J. Ulrich Notheis,a Frank Rominger,a Gunter Scherhag,a Madeleine Schultz,a
Bernd F. Straub,a Martin A. O. Vollanda and Peter Hofmann*a
a
Organisch-Chemisches Institut der Ruprecht-Karls-Universita¨t Heidelberg, Im Neuenheimer
Feld 270, D-69120, Heidelberg, Germany. E-mail: ph@phindigo.oci.uni-heidelberg.de
Max-Planck-Institut fu¨r Kohlenforschung, D-45466, Mu¨lheim an der Ruhr, Germany
b
Received (in Montpellier, France) 14th October 2002, Accepted 5th December 2002
First published as an Advance Article on the web 14th February 2003
A series of highly crowded symmetric and unsymmetric diphosphinomethanes R2PCH2PR02 , important
ligands in transition metal chemistry and catalysis, namely tBu2PCH2PtBu2 (dtbpm, 11), Cy2PCH2PCy2 (dcpm,
t
t
t
2), Bu2PCH2PCy2 (ctbpm, 3), Bu2PCH2PiPr2 (iptbpm, 4) and Bu2PCH2PPh2 (ptbpm, 5), has been prepared
in high yields, using a general and convenient route, which is described in detail for 1. Other than 4, which
is a colourless liquid, these compounds are crystalline solids at room temperature. Their molecular structures
have been determined by single crystal X-ray diffraction, along with that of the higher homologue of 1,
t
t
tBu2CH2CH2 Bu2 (dtbpe, 6). The solid-state structures of the dioxide of 1, Bu2P(O)CH2P(O)tBu2 (7), and of
two phosphonium cations derived from 1, protonated [tBu2P(H)CH2PtBu2]+ (8+) and the chlorophosphonium
ion [tBu2P(Cl)CH2PtBu2]+ (9+), are also described and show a distinct structural influence of the
tetracoordinate P centres. The gas phase UV-photoelectron spectra of the diphosphines 1–6 have been
measured. Their first two ionisation potentials are found to be nearly degenerate and all are in the low energy
range from 7.5 to 7.8 eV. Comparison with related mono- and bidentate phosphines demonstrates that 1–6 are
excellent s-donors towards metals, in accord with their known coordination chemistry. Molecular geometries
and electronic structures of the diphosphine systems have been studied by quantum chemical calculations and
are compared to experiment. Unlike standard semiempirical methods (AM1, PM3, MNDO), which give rather
poor minimum structures and seem inadequate for such sterically crowded systems, ab initio calculations
(RHF/6-31G**) predict molecular geometries with reasonable accuracy and reflect the observed trends in
experimental ionisation potentials.
function between two metal centres.4–8 Sterically demanding
Introduction
alkyl substitution also stabilises ligand-to-metal bonding by
increasing the s-donor strength of diphosphinomethane
ligands relative to those bearing smaller alkyl substituents, and
protects the metal centre through the steric bulk of the substitu-
ents, kinetically hampering side reactions and decomposition.9
Coordination of bidentate phosphines with a small bite-angle
often leads to transition metal complexes that show unusual
behaviour.2 In particular, diphosphinomethane-k2P complexes
incorporating metal centres into four-membered MPCP-che-
late rings have been shown to exhibit unprecedented properties
and reactivity patterns. Sterically bulky phosphorous substitu-
tion stabilises such four-membered chelate structures in several
ways. As a consequence of the so-called geminal dialkyl effect,3
the use of sterically demanding substituents at the phosphor-
ous atoms of diphosphinomethane ligands engaged in k2P
bonding allows for a reduction of ring strain, which is
inherently present in four-membered MPCP chelate structures
and normally favours their ring opening. Moreover, substi-
tution with four bulky alkyl groups hampers the formation of
binuclear species with the diphosphinomethane ligand in its
otherwise preferred and very common m-PCP bridging
t
Using the extremely bulky ligand system Bu2PCH2PtBu2
[bis(di-tert-butylphosphino) methane, dtbpm, 1] in such a
bidentate mode, novel coordination and reaction chemistry
has been studied extensively for various transition metals, oxi-
dation states and d-electron counts. We and others have been
able to generate highly strained nickel triad 14 valence electron
fragments [(dtbpm-k2P)M(0)] with
a
d10 electron count
(M ¼ Ni, Pd, Pt) as reactive, high energy intermediates
in situ.8,10–12 In these intermediates, the P–M–P angle within
the MPCP chelate unit deviates by around 100ꢀ from the linear
equilibrium geometry of non-chelated d10-ML2 systems, which
leads to unusual bond activation chemistry13–15 and bonding
capabilities.16–20 Employing 1 as a ligand for metals with par-
tially filled d valence shells has led to the development of novel
y Dedicated to Professor Roald Hoffmann on the occasion of his 65th
birthday.
types of ruthenium carbene complexes [(dtbpm-k2P)(Cl)
z Electronic supplementary information (ESI) available: photoelec-
tron spectra for compounds 1–6, ORTEP diagrams for compounds
1–3 and 5–9+, calculated minimum structures (UB3LYP/6-31G**)
and structural parameters for two forms of the radical cation dmpm.
See http://www.rsc.org/suppdata/nj/b2/b210114a/
+
Ru CHR] as highly active olefin metathesis catalysts for
=
ROMP,21,22 observation of the first reversible C C bond acti-
=
vation of a ketene using [(dtbpm-k2P)Ir]+ as a template,23,24
and characterisation of the first stable three-coordinate alkyl
540
New J. Chem., 2003, 27, 540–550
DOI: 10.1039/b210114a
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2003

View Article Online
complex of rhodium, (dtbpm-k2P)Rh(CH2 Bu).25 Cationic
bis(di-tert-butyl)phosphinoethane (dtbpe, 6), another impor-
t
complexes [(dtbpm-k2P)M(R)]+ (M ¼ Ni, Pd) and related sys-
tems with other bulky diphosphinomethanes have been shown
to be efficient catalysts for ethylene/CO and ethylene polymer-
isation.26–31 Thus, the use of bulky diphosphinomethanes, and
of 1 in particular, has led to many unusual structures and
reactions, making these molecules attractive ligands. Parenthe-
tically, we note that chiral variants of bulky diphosphino-
alkanes R0RP(CH2)nPRR0 (n ¼ 1, 2) have recently been
synthesised by Imamoto et al.32 and have been shown to be
versatile and attractive ligands for enantioselective metal-
catalysed hydrogenation.
Not surprisingly, the utility of these ligands has initiated
many efforts to develop synthetic routes for various represen-
tatives of this class of compounds. For 1, the first preparation
of a small, spectroscopically identified sample, which was not
purified, was reported by Karsch in 1983.33 The need in our
group to easily access reasonable quantities of 1 resulted in
the development of a convenient preparative procedure, which
was first disclosed in a patent in 199134 and which allows the
synthesis of 1 on a 50–100 g scale in one run. A slightly mod-
ified version of this synthesis was included in a polyolefin
patent disclosed by Brookhart et al. in 1997.35 An updated
and optimised version of our procedure is described in the
Experimental of this paper as a representative case of a gen-
eral, high yield route to R2PCH2PR02 ligand systems including
1–5 (Fig. 1), which can easily be extended to a variety of other
phosphine-containing molecules.
tant ligand, and to those of its dioxide Bu2P(O)CH2P(O)tBu2
t
(7) as well as the cations of two phosphonium derivatives of
this diphosphine, namely protonated 1, [tBu2P(H)CH2PtBu2]+
(8+), and the chlorophosphonium salt [tBu2P(Cl)CH2PtBu2]+
(9+).
In phosphorous chemistry, photoelectron spectroscopy
(UV-PES) has been used to characterise the valence electronic
structure of neutral phosphine ligands.43–46 It has been shown
that the lower the first (lone pair) ionisation energy of phos-
phine ligands, the better they are as donors in a coordinative
bond. Photoelectron spectroscopy is thus one of the few phy-
sical measurements that can assess—at least in a qualitative
sense—the electronic properties and behaviour of ligands in
the absence of metals. The PE spectra of the ligands 1–6
described here are reported and compared with the corre-
sponding values for monodentate and less bulky bidentate
phosphines in the literature.
Theoretical studies of the geometries and ionisation poten-
tials of the bidentate phosphine systems 1–6 are also reported.
Standard semiempirical methods (AM1, PM3, MNDO47) fail
and ab initio calculations (Gaussian 98, RHF/6-31G**) are
employed in order to evaluate the applicability and the predic-
tive value of such calculations for this specific and unusual
class of ligand systems. A direct comparison is made between
the solid-state structural data and the optimised geometries.
The measured ionisation energies are compared with those pre-
dicted from the calculated orbital energies, testing the validity
of Koopmans’ theorem for these systems.
Several compounds in this series have also been prepared by
other groups.33,36–38 These reported synthetic routes, however,
are often hampered by low yields of sometimes impure pro-
ducts or are relatively inconvenient. For bis(dicyclohexylphos-
phino)methane (dcpm, 2), which is commercially available, an
early mention of its preparation and its use as a ligand by
Jonas in 198839 formed the impetus for our own dtbpm synth-
esis. Ligand 2 was also independently prepared by Roundhill
et al.in 1991,40 and later work of Rothwell described the ‘‘cat-
alytic’’ hydrogenation41 of Ph2PCH2PPh2 (dppm) to 2.42
The only general synthesis for R2PCH2PR02 systems avail-
able in the literature was reported recently by Werner
et al..38 Among other diphosphinomethanes and some
arsino(phosphino)methanes, 2, dicyclohexylphosphinodi-tert-
butylphosphinomethane (ctbpm, 3) and di-iso-propylphosphi-
nodi-tert-butylphosphinomethane (iptbpm, 4) are reported
using a route via stannylated monophosphine precursors. We
include 2, 3 and 4 in the present study because preparation
by our route is more general, avoids tin intermediates and
yields all compounds except iptbpm (4), which is a liquid at
room temperature, as white, crystalline solids [not liquids as
has been reported for 1,33 338 and diphenylphosphinodi-tert-
butylphosphinomethane (ptbpm, 5)36,37]. This has allowed
their structural investigation by X-ray diffraction, and we
report here the solid-state structures of 1, 2, 3 and 5. The
structure of 1 is compared to that of its higher homologue
Results and discussion
Synthesis
Our general route to methylene-bridged diphosphines, origin-
ally developed for 1,34 follows Scheme 1 (full details are
provided in the Experimental). Using this route, the related
bidentate phosphine ligands 2, 3, 4 and 5 (Fig. 1) were also
easily prepared by employing the corresponding dialkyl or
diaryl chlorophosphines.
Although the use of Grignard reagents to prepare dialkyl-
chlorophosphines is long established,48,49 the yields and purity
of the resulting magnesium complexes were not sufficient for
efficient synthesis in the quantities required. Therefore, the
procedure has been modified by using dioxane after the reac-
tion to precipitate the MgCl2-dioxane complex, which allows
ready filtration followed by distillation of the phosphine chlor-
ide in high yield. In the subsequent steps, accurate titration of
the lithium reagents immediately before use proved key to
improving yields.50 The water for all aqueous solutions must
be held at reflux under argon for at least several hours to
ensure complete removal of O2 , which otherwise leads to
contamination by phosphine oxides.
Fig. 1 Bidentate phosphine derivatives used in this study, their acro-
nyms and numbering scheme.
Scheme 1 General synthetic route to methylene-bridged bidentate
phosphine ligands.
New J. Chem., 2003, 27, 540–550
541

View Article Online
As chelating ligands, the diphosphines have to deviate dras-
The key step to the synthesis is the lithiation of the dialkyl-
methylphosphine, reported first by Karsch and Schmidbaur.51
As this reaction proceeds only slowly, even at high tempera-
ture, the commercially available pentane solution of tert-butyl
tically from these solid-state geometries. The P–C–P bond
angle must reduce by around 20ꢀ to reach the values of around
95–100ꢀ typical in structures containing
a methylene-
t
lithium must be taken to dryness and the solid BuLi redis-
bridged diphosphine ligand, and the dihedral torsion is
removed.9,16,17,53 The existence of a large body of solid-state
structure determinations in which 1 acts as chelating ligand
shows that the distortions required for k2-ligation to a transi-
tion metal centre are easily accomplished.54
solved in heptane before addition of the dialkylmethylphos-
phine as a heptane solution and heating to reflux. The
formed lithium methanide derivative is pyrophoric, and to
avoid possible contact with air, a short, wide neck containing
a frit and ending in a ground glass joint was attached before-
hand by blown glass to a round-bottomed flask, which is used
for this reaction. The joint is capped during the reaction, and
after the 18 h at reflux are complete, the cap is replaced by
another Schlenk flask. The insoluble product is then separated
from the liquor by tilting the flask so that the solution flows
through the frit.52 The final step yields the product diphos-
phines, which are very oxygen-sensitive but water stable and
not thermally sensitive. Other than 1, the diphosphines can
be crystallised from pentane; 1 itself can be most cleanly
obtained from methanol.
The melting point of 2 (95–97 ꢀC) is much higher than that
of the other diphosphines studied in this work. The calculated
density of the crystal is 10% higher than those of the other
ligands at 1.113 g cmꢁ3 (Table 2), although not as high as
those of the oxidised phosphines, which indicates that this
molecule packs more efficiently than 1, 3 and 5. For compari-
son, the melting point of dppm is 120–122 ꢀC,55,56 and the
calculated density in the crystal is 1.254 g cmꢁ3 57
A possible
.
structural reason for the high density of 2 can be seen in the
packing diagram shown in Fig. 3. The cyclohexyl groups are
arranged in a dense manner within the crystal, interleaving
with cyclohexyl groups from adjacent molecules. The flat phe-
nyl groups of dppm presumably allow still denser packing in
that molecule, while the bulkier tert-butyl and iso-propyl sub-
stituents in 1, 3, 4 and 5 prevent it. Fig. 4 shows a packing dia-
gram of 5 for comparison; it can be seen that the individual
molecules do not interleave, resulting in a lower density.
The sum of the three C–P–C angles of a phosphine gives an
indication of the hybridisation of the phosphorous atom and is
related, although not directly, to the Tolman cone angle. The
pyramidalisation of the phosphines in the structures presented
here can be compared with that in the crystal structures of
PMe3 and P(tBu)3 , in which the sums of the angles about P
Solid-state structures
All of the ligands except the colourless liquid 4 are white, crys-
talline solids at ambient temperature, and the crystal structures
of the solid diphosphines have been determined. The molecular
structure of 2, prepared according to Scheme 1, was mentioned
in a footnote of a paper by one of the present authors some
time ago.10 Details of this structure are now provided for
comparison with the other diphosphines reported. In the case
of 1, severe disorder in the crystal leads to inaccuracy in the
structural data; the structural problem is described in the
Experimental.
58
are 298ꢀ and 322ꢀ,59 respectively. The sums of angles about
the phosphorous atoms bearing tert-butyl groups in the biden-
tate ligands 1, 3 and 5 are close to 314ꢀ in each case. This
reflects the steric bulk of two tert-butyl groups; the phosphor-
ous geometry is not as flattened as in P(tBu)3 but less
pyramidalised than phosphorous atoms with three smaller
substituents. When 1 acts as a chelating ligand, the phosphor-
ous atoms are flatter; the sums of C–P–C angles about the
phosphorous atoms are above 320ꢀ,16,17 due to the smaller
repulsive effect of electrons in a dative P–M bond relative to
a lone pair. The phosphorous atoms in the bidentate phos-
phines 2, 3 and 5 bearing cyclohexyl or phenyl groups are
somewhat more pyramidalised, with sums of angles of around
304ꢀ.
The ethylene-bridged analogue of 1, dtbpe (6), has been used
to support interesting chemistry on nickel,60–66 palladium,67
platinum68 and ruthenium.69–71 In order to compare this
ligand with 1, its structure was also determined crystallogra-
phically and an ORTEP diagram is shown in Fig. 5. The centre
of the ethylene bridge lies on a special position, resulting in a
crystallographically imposed 180ꢀ dihedral angle between the
lone pairs on phosphorous. The bond distances and angles
as well as the pyramidalisation of the phosphorous atoms
are similar to those found in 1. As a chelating ligand, 6 binds
with P–M–P angles of around 85–90ꢀ, which is easily achieved
through rotation around P–C bonds and bending the flexible
ethylene backbone.
Fig. 2 shows, as a representative example, an ORTEP
diagram of 3. ORTEP diagrams for all crystallographically
characterised molecules are provided in the electronic supple-
mentary information (ESI). Important bond distances and
angles for all of the structures are contained in Table 1, while
relevant data collection and structure solution parameters are
provided in Table 2. In each molecule, the lone pairs of the two
phosphorous atoms point away from each other, with dihedral
angles of between 56ꢀ and 100ꢀ. This dihedral angle of the
diphosphines was determined by situating dummy atoms (X
and X0) at the calculated centroids of the three carbon atoms
bound to each phosphorous atom (P and P0), and then measur-
ing the torsion angle X–P–P0–X0. Rotation about the P–C
bonds is expected to be facile in solution, and the variation
in dihedral angle in the observed structures is unlikely to be
chemically significant. The angles about the bridging carbon
atom range from 110ꢀ to 120ꢀ; these angles in 1, 2 and 3 of
around 120ꢀ are significantly higher than the ideal tetrahedral
angle. This distortion is presumably related to the steric bulk
of the adjacent groups and is reproduced in calculations (see
below).
Oxidation of 1 with hydrogen peroxide led to isolation of the
t
dioxide Bu2P(O)CH2P(O)tBu2 (7). This compound was also
structurally characterised and an ORTEP stereoview diagram
is shown in Fig. 6, while selected bond distances and angles
can be found in Table 1 and data collection and structure solu-
tion parameters are in Table 2. In comparison to 1, the geome-
try of the molecule changes somewhat upon oxidation; the
P–C–P angle opens further to 128ꢀ and 125ꢀ in the two unique
molecules, presumably related to the more bulky P(O)(tBu)2
groups on the bridging carbon atom relative to the P(tBu)2
groups in the parent ligand, and electrostatic repulsion
Fig. 2 ORTEP diagram (50% probability ellipsoids) of 3. Hydrogen
atoms have been omitted for clarity.
542
New J. Chem., 2003, 27, 540–550

View Article Online
+
ꢀ
˚
Table 1 Selected bond distances (A) and angles ( ) in the molecular structures of 1–3 and 5–9
1
2
3
5
6
7a
8+
9+
P–C (bridge)b
P–C (bridge)
1.834(9)
1.832(9)
118.4(7)
95
1.859(2)
1.859(2)
120.5(1)
99
1.855(3)
1.863(3)
120.3(2)
91
1.861(2)
1.852(3)
109.7(2)
56
1.860(1)
1.837(2), 1.833(2)
1.833(2), 1.831(2)
128.1(1), 124.8(1)
94, 77
1.879(3)
1.806(3)
118.6(1)
65
1.906(7)
1.803(7)
118.8(4)
46
Special position
P–C–C 112.7(1)
180
P–C (bridge)–P
Dihedral angle of lone pairsc
P–X
mean P–O
1.48
322, 322
P–H
P–Cl
1.34(2)
313
339
1.918(4)
312
338
Sum of C–P–C anglesd
310
314
303
304
312
303
314
304
315
322, 322
a
b
Data for two unique molecules in the unit cell presented as molecule 1, molecule 2. For the asymmetrical molecules, the first P–C distance
involves the P atom bearing tert-butyl groups (3, 5) or the P(III) atom (8+, 9+). The dihedral angle is the torsion angle between a dummy atom
located at the calculated centroid of the three carbon atoms bound to P1, P1, P2 and a dummy atom located at the calculated centroid of the three
carbon atoms bound to P2. For consistency, for 7, 8+ and 9+ the same procedure was used although a lone pair is not always present. For the
c
d
asymmetrical molecules, the first value listed corresponds to the phosphorous atom bearing tert-butyl groups (3, 5) or the P(III) phosphorous atom
(8+, 9+).
between the two oxygen atoms. The P–C (bridge) distances
are, however, unchanged from those of 1. The dihedral angles
between the oxygen atoms in the two unique molecules are dif-
ferent from each other: 94ꢀ and 77ꢀ. This difference is mostly
likely due to packing effects, as the sums of the C–P–C angles
in 7 are larger than in 1 but similar in both molecules of the
asymmetric unit at 322ꢀ. The decrease in pyramidalisation of
the phosphorous centres upon oxidation is a consequence of
the smaller repulsive effect of the electrons in the P–O bond
relative to the repulsive effect of the electrons in the lone pairs
in 1; the corresponding sum of angles in (O)P(tBu)3 is 339ꢀ.72
In the context of hydroformylation studies using ligand 1,
protonated 1 was isolated as an unintended side product from
a reaction mixture as the triflate [tBu2P(H)CH2PtBu2]+-
[CF3SO3]ꢁ. X-Ray quality crystals of the colourless salt were
obtained upon crystallisation from diglyme and a molecular
structure determination was performed. As the structural
details of the cation of this compound (8+) seemed important
in comparison to 1, and as a complete spectroscopic char-
acterisation was also desirable, the [BArF]ꢁ {BArF ¼ B[3,5-
(CF3)2C6H3]4} salt of 8+ was prepared independently by
protonating 1 in diethyl ether with [H(OEt2)2]BArF and preci-
pitating the salt with hexane. Good quality crystals were
obtained by crystallisation of [8+][BarF]ꢁ from a 1:1 diethyl
ether–hexane mixture at ꢁ20 ꢀC. Bond distances and angles
are listed in Table 1 and data collection and structure solution
parameters are given in Table 2.
Most strikingly, the triply coordinated phosphorous atom of
8+ has a much longer P–C distance to the bridging carbon
˚
atom (1.88 A) than that with a proton (1.81 A). The corre-
˚
˚
sponding value in 1 is 1.83 A. This is most likely due to the
electron-withdrawing effect of the P(V) making a shorter bond
to the bridging carbon atom, which leads to a weaker, longer
bond to the P(III). In agreement with Bent’s rules,73,74 the cen-
tral carbon concentrates more p-character towards the more
electronegative phosphonium group. The P–C–P angle is the
same as that in the parent structure, and the torsion angle is
smaller than in 1. The sum of C–P–C angles about the phos-
phorous bearing the proton is very high (337ꢀ), with distinctly
less pyramidalisation of the phosphonium PC3 frame than at
the P centres in 1 or 7, again due to the smaller repulsive effect
of electrons in a P–H bond compared with a lone pair. The
sum of angles about the P(III) centre is 312ꢀ, typical for the
CH2–PtBu2 fragment as in neutral 1 and 6.
A second dtbpm derivative with one phosphonium group
was obtained accidentally when iron derivatives of dtbpm were
synthesised. In a reaction of (CO)4FeCl2 with 1, the complex
(dtbpm-k2P)Fe(CO)3 was isolated in low yield75 and a second,
Table 2 Selected crystal data and structure solution parameters for 1–3, 5–7, [8+][B{3,5-(CF3)2C6H3}4]ꢁ and [9+][Fe2Cl6]1/2ꢁꢂCH2Cl2
[8+]
[9+]
1
2
3
5
6
7
[B{3,5-(CF3)2C6H3}4]ꢁ [Fe2Cl6]1/2ꢁꢂCH2Cl2
Formula
FW
C17H38P2
304.41
Orthorhombic Triclinic
C25H46P2 C21H42P2
C21H30P2
344.39
Monoclinic
C18H40P2
318.44
Monoclinic
C17H38O2P2 C49H51BF24P2
336.41
Monoclinic
C18H40Cl6FeP2
586.99
Monoclinic
408.56
356.49
Monoclinic
1168.65
Monoclinic
Crystal
system
Space group
¯
P1 (#2)
9.718(2)
Pna21(#33)
9.3565(1)
17.3925(2)
12.5085(2)
90
P21/n (#14) P21/c (#14) P21/n (#14) P21/c (#14) P21/n (#14)
P21/c (#14)
18.2861(3)
9.9887(2)
17.3272(3)
90
˚
a/A
11.1753(5)
6.2975(2)
33.5911(4)
10.2346(2)
90
8.5001(1)
10.1304(2)
12.9111(2)
90
9.5711(2)
25.0934(4)
17.6357(1)
90
16.7907(1)
19.1459(2)
18.2699(2)
90
˚
b/A
10.423(2) 9.8062(4)
12.775(1) 20.3041(8)
˚
c/A
a/ꢀ
b/ꢀ
g/ꢀ
98.21(1)
96.84(1)
105.23(1) 90
90
90
90
94.633(1)
105.314(1)
90
106.176(1)
90
105.061(1)
90
113.552(1)
90
113.857(1)
90
3
˚
U/A
2035.54(5)
4
233
1219.0(4) 2217.8(2)
2088.15(8)
4
200
1067.75(3)
2
200
4090.1(1)
8
200
5384.02(9)
4
200
2894.47(9)
4
200
Z
T/K
2
100
4
200
m(Mo-Ka)calc/mmꢁ1 0.204
0.186
31 999
5611
0.0868
4742
0.196
18 216
3913
0.1198
2341
0.207
15 312
3599
0.197
10 678
2444
0.216
30 348
7114
0.20
47 177
9828
1.190
28 351
6617
Total reflection
Unique reflections
Rint
16 625
3534
0.0310
2874
0.0538
2435
0.0331
1954
0.0357
5422
0.0671
5980
0.1251
3776
Obs. reflections
2
[Fo > 2s(Fo2)]
R1 (obs. data)
wR2 (obs. data)
0.0405
0.1103
0.045
0.129
0.053
0.095
0.046
0.088
0.034
0.086
0.041
0.098
0.047
0.098
0.102
0.231
New J. Chem., 2003, 27, 540–550
543

View Article Online
Fig. 5 ORTEP diagram (50% probability ellipsoids) of 6. Hydrogen
atoms have been omitted for clarity.
Fig. 3 Packing diagram of 2 (50% thermal ellipsoids). Hydrogen
atoms have been omitted for clarity.
chlorine atom above this plane, due to the smaller repulsive
effect of a bonding electron pair relative to a lone pair on
phosphorous.
less soluble product was identified by X-ray crystallography to
be [tBu2P(Cl)CH2PtBu2]+[Fe2Cl6]1/2ꢁꢂCH2Cl2 . In the cation of
this salt, 9+, the chloro analogue of 8+, the phosphorous atom
bound to chlorine has also been oxidised to P(V). The structure
of this cation provides another comparison with 1, 7 and 8+
and thus has been included here. Important bond distances
and angles are listed in Table 1 and data collection and struc-
ture solution parameters are given in Table 2. The structural
features of cation 9+ are very similar to those of cation 8+;
in particular, the distance from the triply coordinated phos-
phorous atom to the bridging carbon atom C10 is again very
long, even longer than in 8+, in line with expectations from
Electronic structure and photoelectron spectroscopy
As described above, some aspects of the solid-state structures
are of chemical significance, as they relate to the hybridisation
at phosphorous, to the character of the lone pairs, to the bite-
angle and thus to the coordination and ligand properties. In an
attempt to quantify and understand the significance of the
structures in the crystalline state, and to relate them to gas
phase structures and gas phase photoelectron spectroscopy,
quantum chemical calculations have been carried out.
Semiempirical calculations at the AM1, PM3 and MNDO
level47 on the diphosphine systems 1–9+ revealed that the P–
C–P angles at the methylene bridges are generally badly over-
estimated (by 10–20ꢀ) in the optimised geometries at this level
of theory. For the compounds with P(V) centres, this angle is
also far too high in the calculated structures (ca. 140ꢀ).
Furthermore, the sum of the angles about the P(III) atoms is
overestimated by about 10ꢀ in every case, making the pyrami-
dalisation at phosphorous distinctly smaller than what is
observed. The large difference in P–C bond lengths to the brid-
ging carbon atoms found in the crystal structures of the phos-
phonium cations 8+ and 9+ is reproduced in the calculations,
but the calculated bond length to the P(III) atom is too short.
The pronounced deviations from experimental geometries
probably have to do with the fact that highly crowded phos-
phines—unlike sterically less congested, ‘‘normal’’ ones, where
semiempirical models perform better—fall outside the range of
systems utilised in parametrising these methods. Therefore,
geometric predictions are not valid and ab initio calculations
at an appropriately high level of theory (RHF/6-31G**) were
performed for all systems, 1–9+.
˚
˚
Bent’s rules (1.91 A as compared to 1.83 A in the parent 1),
˚
while that from the P(V) centre is short (1.80 A). The P–C–P
angle of 9+ as in 8+ is not different from that of 1, but the
phosphorous atom bearing the chloride is also in the centre
of a much flatter alkyl environment (sum of C–P–C angles
338ꢀ), as described for 8+. The sum of C–P–C angles about
the other phosphorous atom is also typical for the CH2–PtBu2
fragment (312ꢀ).
Comparing all seven solid-state molecular structures with a
P–C–P skeleton, the basic features of these bulky diphosphino-
methane ligands and the related oxidised phosphines are a
P–C–P angle that is larger than for an idealised sp3 geometry
at carbon (except 5), a twisted core that points the lone pairs
(or substituents) on the phosphorous atoms away from each
other, and relatively flat phosphorous geometries. The last
feature can be readily explained by the steric bulk of the sub-
stituents on the phosphorous. In the cases where a fourth sub-
stituent is bound to an oxidised P(V) atom, the three alkyl
substituents form an even less pyramidal (although not planar)
coordination environment, with the oxygen, hydrogen or
Fig. 6 Stereoview ORTEP diagram (50% probability ellipsoids) of
dioxide 7. Only one of the two independent molecules in the asym-
metric unit is shown. Hydrogen atoms have been omitted for clarity.
Fig. 4 Packing diagram of 5 (50% thermal ellipsoids). Hydrogen
atoms have been omitted for clarity.
544
New J. Chem., 2003, 27, 540–550

View Article Online
+
ꢀ
˚
Table 3 Selected bond distances (A) and angles ( ) of optimised structures (RHF/6-31G**) of the bidentate phosphines 1–9
1
2
3
4
5
6
7
8+
9+
P–C (bridge)a
P–C (bridge)
P–C (bridge)–P
1.876
1.877
120.2
1.869
1.869
118.9
1.878
1.878
118.4
1.875
1.860
117.0
1.861
1.852
109.7
1.876
1.876
P–C–C
112.7
178
1.850
1.850
126.0
1.899
1.827
122.0
1.908
1.828
125.3
Dihedral angle of lone pairsb
P–X
86
92
92
83
56
91
84
88
P–O
1.48
323
P–H
1.39
314
338
P–Cl
2.01
314
338
Sum of angles about Pc
314
306
314
317
314
311
314
304
316
a
For the asymmetrical molecules, the first P–C distance involves the phosphorous atom bearing tert-butyl groups (3, 4 and 5) or the P(III)
phosphorous atom (8+, 9+). b The dihedral angle is the torsion angle between a dummy atom located at the calculated centroid of the three carbon
atoms bound to P1, P1, P2 and a dummy atom located at the calculated centroid of the three carbon atoms bound to P2. For consistency, for 7, 8+
and 9+ the same procedure was used although a lone pair is not always present. For 7, the mean P–O bond distance is given, for 8+ the P–H distance
and for 9+, the P–Cl distance. c For the asymmetrical molecules, the first value listed corresponds to the P atom bearing tert-butyl groups (3, 4 and
5) or the P(III) atom (8+, 9+).
Selected distances and angles of the RHF/6-31G** opti-
mised geometries of 1–9+ are included in Table 3, and a dia-
gram of the minimum energy geometry calculated for ctbpm
(3) is shown in Fig. 7, which can be compared with the crystal-
lographically determined structure in Fig. 2. It can be seen that
the calculated geometries generally resemble the solid-state
structures. Although the absolute values of the distances and
angles calculated for the gas phase minimum structures are
not (and cannot be expected to be) precisely identical to the
values found in the solid state, the deviations are small or
insignificant, and the trends in bond lengths, torsion angles
and P–C–P angles within the six crystallographically charac-
terised neutral molecules and the two cations are reproduced
remarkably well by the calculations. The P–C (bridge) dis-
tances are slightly overestimated for most of the ligands, but
the calculated P–C–P angles are close to those observed. Even
the predicted dihedral angles of the lone pairs are similar to the
observed values, except in the two cations, for which the pre-
dicted dihedral angles are higher than those found in the mole-
cular structures. In addition, for the two cations, the P–X
distance is calculated to be somewhat higher than that
observed, while in the dioxide (7) the prediction is accurate.
The pyramidalisation about the P centres is predicted extre-
mely well for both the P(III) and P(V) atoms of all 8 complexes
studied crystallographically.
monodentate or chelating phosphine ligands, the UV-photo-
electron spectra of 1–6 were measured. The ionisation poten-
tials, taken as the maximum of the first broad ionisation
band in the spectrum of each compound (which actually repre-
sents the two lowest energy ionisation events) are given in
Table 4. Note that in the PE study of Puddephatt et al. invol-
ving the uncongested diphosphines dmpm and dmpe, a single
‘‘ionisation energy of the phosphorus lone pair’’ was
described; the PE spectra also showed only one unresolved first
ionisation peak, which was deconvoluted to an ideal Lorent-
zian–Gaussian shape but must contain the two ionisation
events from both P lone pairs at slightly different energies.45
The spectra of 1–6 are all very similar, with the first two ionisa-
tion potentials around 7.6 eV; a representative spectrum of 1 is
shown in Fig. 8. The second ionisation potential is not resolved
experimentally (except in the case of 5, where two peaks at 7.5
and 7.6 eV can be assigned), as it lies within 0.2 eV of the first
ionisation potential in each case. All PE spectra (compounds
1–6) are provided in the ESI.
The first two ionisations of the diphosphines correspond to
ionisation of the phosphorous lone pair electrons. Assuming
the validity of Koopmans’ theorem (ꢁej ¼ IPj), which should
be allowed in these organic systems, the first two ionisations
correspond to the two canonical, non-degenerate, symmetry-
adapted highest occupied molecular orbitals, which are the
linear combinations of the two P lone pairs, delocalised to
some extent into the P–C and C–C bonds. Comparison with
the values reported for monodentate PMe3 (8.59 eV)46 and
P(tBu)3 (7.70 eV)43 as well as the less bulky bidentate ligands
bis(dimethylphosphino)methane (dmpm) (8.4346 or 8.51 eV45)
and bis(dimethylphosphino)ethane (dmpe) (8.4546 or 8.47
eV45) shows that the diphosphines reported here are much
more readily ionised than their methyl-substituted analogues,
Calculations at the RHF/6-31G** level therefore provide a
reasonable method to assess the structural characteristics of
this type of ligand in the gas phase and even give an approxi-
mate picture of the conformations adopted in the solid state by
these rather unpolar molecules.
In order to quantify the s-donor character of these
electron-rich diphosphines, especially in comparison to other
Table 4 Measured ionisation potentials (eV) and calculated orbital
energies (HOMO, HOMO–1) from the optimised geometries of
diphosphines 1–6 (RHF/6-31G**)
1
2
3
4
5
6
Measureda
HOMO
HOMO–1
7.8
7.7
7.6
7.7
7.5, 7.6 7.8
8.24 8.39 8.34 8.30 7.90
8.39 8.46 8.45 8.45 8.22
8.19
8.46
0.27
D(HOMO–1) ꢁ (HOMO) 0.15 0.07 0.11 0.15 0.32
a
The value given corresponds to the centre of the first broad
ionisation band in the PE spectrum, which comprises the first two
ionisation events. For 5, the first two ionisation potentials
are resolved (see Fig. 10).
Fig. 7 Calculated minimum geometry for
Hydrogen atoms have been omitted for clarity.
3 (RHF/6-31G**).
New J. Chem., 2003, 27, 540–550
545

View Article Online
Fig. 10 UV-photoelectron spectrum of 5.
Fig. 8 UV-photoelectron spectrum of 1.
Koopmans’ theorem breaks down.78 The broad band observed
above 10 eV in all of the photoelectron spectra reported here is
due to ionisations from P–C bond orbitals and from the alkyl
substituents on the phosphorous atoms.
and resemble P(tBu)3 in their s-donor capabilities. The
reported values for phenyl-substituted phosphines such as
PPh3 (7.8876 or 7.80 eV77), dppm (7.79 eV) and bis(diphenyl-
phosphino)ethane (dppe) (7.86 eV)45 are closer to those
observed for the ligands here, for reasons discussed below.
Ab initio RHF/6-31G** calculations predict energies of the
HOMO and HOMO-1, corresponding to the first two ionisa-
tion energies of the diphosphines 1–6. Values have been
obtained using completely optimised geometries (NIMAG ¼ 0)
and the results are included in Table 4. Fig. 9 shows plots of
the calculated HOMO and HOMO-1 for 1. As expected, the
two frontier MOs correspond to the in- and out-of-phase
combinations of the lone pairs. The data in Table 4 shows
that the calculations consistently overestimate the ionisation
potentials by, on average, 0.5 eV. However, the energy split
between calculated HOMO and HOMO-1 is always small,
which agrees well with the experimental observations.
Apart from ionisations at 7.5 and 7.6 eV, the photoelectron
spectrum of 5 (Fig. 10) exhibits bands at 8.8–10 eV, which are
not observed in the other spectra. They are attributed to the
ionisation from the phenyl rings in that compound. This inter-
pretation, which is in line with interpretations of the PE spec-
tra of other phenyl-substituted phosphines,45,76–78 is verified by
the RHF/6-31G** HOMO-2 and HOMO-3 wave functions
shown in Fig. 11 for the optimised geometry of 5. Clearly,
these two orbitals do involve electron density mostly in the
aromatic rings, so the assignment seems to be reasonable.
The HOMO and HOMO-1 of 5 have similar appearances to
those orbitals displayed for 1 and again represent ionisation
from the two phosphorous lone pairs. It is worth noting that
phenyl-substituted phosphines like PPh3 (7.80 eV)77 and dppm
(7.79 eV)45 have rather low ionisation potentials, in the same
range as those found for 1–6, but in this case they are caused
by conjugative interactions between the P centre and the aryl
rings,that is by a particular stabilisation of the radical cations
formed upon ionisation. In these fully phenylated systems, the
difference between one-electron (MO energies) and collective
properties (ionisation) becomes particularly relevant and thus
A referee has pointed out that the consistent deviation of
observed ionisation potentials for 1–6 from those values
derived via Koopmans’ theorem from HOMO and HOMO-1
energies might be a consequence of a stabilisation of the radi-
cal cations through geometric relaxation and concomitant
formation of cyclic structures with a weak, 2-centre/3-electron
P–P interaction. We refrain from such an interpretation
because, unlike in the condensed phase, electron removal in a
PE experiment is a fast process that does not lead to a relaxation
of the molecular structure. We therefore see the discrepancy of
(on average) 0.5 eV between measured ionisation potentials and
calculated orbital energies as resulting from electronic reorga-
nisation, caused by removing electrons from strongly localised
(P lone pair) orbitals, as is well known for d-orbitals in transi-
tion metal compounds, where such Koopmans’ defects often
forbid the correlation of ground state orbital energies with
ionisation energies. The large organic groups attached to the
P centres in 1–6 are potent electron reservoirs and particularly
suited for efficient electron density relaxation towards the
ionised centres. The assumption of an incipient P–P interaction
upon ionisation and of stabilised cyclic radical cations is also
inconsistent with diphosphinoethane (6) showing exactly the
same behaviour as 1–5. A change from the ground state geo-
metry of 6 (Fig. 5, Tables 1, 3) to a four-membered ring struc-
ture with P–P 2-centre/3-electron bonding during the PE
ionisation event seems highly improbable. Moreover, in a cyc-
lic radical cation of this type, through-bond coupling79 would
lead to a SOMO of n+ character and to no net P–P attraction
upon ionisation. For diphosphinomethanes with a methylene
unit between the two P atoms this is different; calculations with
complete geometry optimisation for the radical cation of
dmpm at the UB3LYP/6-31G** level of theory lead to two
separate minima (NIMAG ¼ 0), one with a closed structure
Fig. 9 Plots of the calculated HOMO (left) and HOMO–1 for 1
(RHF/6-31G**).
Fig. 11 Contour plots of the calculated HOMO–2 and HOMO–3 for
5 (RHF/6-31G**).
546
New J. Chem., 2003, 27, 540–550

View Article Online
2
ꢀ
˚
(P–C–P ¼ 89.6 , P–P ¼ 2.62 A) and one with an open geome-
with Xe and Ar. The resolution of the P3/2 line of Ar was
20 meV.
ꢀ
˚
try (P–C–P ¼ 113.3 , P–P ¼ 3.12 A). The latter is only 1.36
kcal molꢁ1 more stable than the former (for details see ESI).
The difference in energies between the geometry optimised
ground state structures of 1–6 and their radical cation ground
states computed for the same geometry and at a sufficiently
high level of theory would give the first ionisation potentials
for these systems without having to resort to Koopmans’
theorem. We did not perform such calculations for these large
molecules, because despite the systematic small deviations in
absolute values, the ground state ab initio calculations at
the RHF/6-31G** level allow a consistent interpretation of
all PE spectra.
t
i
The chlorophosphines Bu2PCl, Cy2PCl and Pr2PCl were
prepared from freshly distilled PCl3 and the corresponding
Grignard reagent in diethyl ether.43,48,49 A modification of
the published procedure was used and the precise conditions
used are described below for Bu2PCl. BuCl was dried over
Na/Pb and distilled freshly before use. Methyl lithium and
tert-butyl lithium were purchased in solution from Fluka and
their exact concentrations were determined by titration before
each use. Ph2PCl was purchased from Aldrich and used with-
out further purification. Ethylene-bridged dtbpe (6) was
prepared according to the method of Po¨rschke et al..80
Pd(OAc)2 was purchased from Aldrich. (CO)4FeCl2 was
prepared according to the literature.81
t
t
Conclusions
Syntheses
A convenient, flexible and clean synthetic route to highly
substituted, methylene-bridged diphosphine ligands has been
established. The crystal structures of compounds prepared by
this route show some characteristic common features. The
P–C–P angles lie in the range 110–120ꢀ, which is distinctly
larger than the ideal tetrahedral angle about carbon; this
distortion is related to the steric bulk on the two adjacent
phosphorous atoms. The phosphorous atoms have relatively
flat coordination substituent environments compared with
methyl-substituted phosphines, which lead to greater p charac-
ter in their lone pairs. This can be observed in the photoelec-
tron spectra, which show these bulky diphosphinomethanes
to have very low, nearly degenerate first and second ionisation
potentials for the two P lone pairs, making them different from
smaller, bidentate phosphines such as dmpm. The ease of oxi-
dation predicted by the photoelectron spectra is illustrated by
the air sensitivity of the ligands and the preparation of the
dioxide of 1, 7.
Ab initio theory at the RHF/6-31G** level is a fast and quite
reliable tool to predict the molecular and electronic structures
of these phosphines, yielding optimised geometries that closely
reflect the solid-state structures. The trends of structural
changes induced by oxidation of both P atoms [dtbpm (1) vs.
(7)] or by transforming one of the P centres into a phospho-
nium unit, as in 8+ and 9+, are reflected correctly. Assuming
the validity of Koopmans’ theorem, the ionisation potentials
are predicted with an error of (on average) 0.5 eV for all
ligands in this series.
t
dtbpm (1). Step 1: BuMgCl. Mg turnings (97.4 g, 4.0 mol)
were weighed into a 2 L three-necked flask equipped with a
reflux condenser, a graduated, pressure-equalising dropping
addition funnel and a large magnetic stirrer bar. To activate
the Mg turnings, they were heated under vacuum and then stir-
red overnight at room temperature under Ar. Diethyl ether
(200 mL) was added. tBuCl (250 mL, 3.7 mol) was transferred
to the addition funnel via cannula and as much added to the
reaction flask (ca. 20 mL) as required for initiation of the
t
Grignard reaction. After the reaction began, the BuCl in the
addition funnel was diluted with 200 mL diethyl ether. This
solution was added at a rate that caused light reflux of the sol-
vent (ca. 90 min). The reaction mixture was then heated to
reflux for 1 h. After cooling, the grey-black solution was fil-
tered through a D3 frit with a Celite pad into a graduated
Schlenk flask. The residual solid was washed with diethyl ether
(2 ꢃ 100 mL) and the washings were filtered into the flask.
Titration was performed by the Gilman method.50 Yield: 600
mL of a 3.19 M solution (1.91 mol) (note that Step 2 functions
well on a maximum scale of 2.0 mol)
Step 2: tBu2PCl. Diethyl ether (1.6 L) was transferred to a 4 L
three-necked flask equipped with a gas inlet, overhead stirrer
and a 1 L graduated addition funnel. PCl3 (70 mL, 110 g,
t
0.80 mol) was added. BuMgCl (600 mL of a 3.19 M solution
in diethyl ether, 1.91 mol) was cannulated into the addition
funnel and diluted with a further 300 mL diethyl ether. This
solution was then added to the PCl3 solution at a rate of 100
mL hꢁ1 at ꢁ33–ꢁ41 ꢀC (dry ice/ethanol bath). After addition
of 450 mL, the solution was gradually allowed to warm to
room temperature as addition continued at the same rate.
The resulting pale viscous suspension was stirred overnight,
then the addition funnel was exchanged for a condenser and
the mixture was heated to reflux for 30 min. Dioxane (300
mL) was cannulated into the solution through the condenser
and the mixture was heated to reflux again for an hour. The
suspension became noticeably less thick. This dioxane addition
method facilitates the removal of MgCl2 as its less soluble
dioxane complex. After cooling to room temperature the mix-
ture was filtered through a frit to separate the ethereal solution
from the MgCl2-dioxane complex. The precipitate was washed
with diethyl ether (2 ꢃ 250 mL). The diethyl ether was then dis-
tilled off at atmospheric pressure and the residue was fraction-
ally distilled under vacuum. The product was collected at
64 ꢀC/12 mbar. Yield: 87.4 g of a colourless liquid (484 mmol,
The photoelectron spectra and calculations illustrate the
special nature of diphosphinomethane ligands with bulky sub-
stituents. These compounds are much better donors (for exam-
ple towards Lewis acidic metal centres) than diphosphine
ligands with smaller substituents, due to the high energy and
high p character of their P lone pairs, and they form strongly
stabilised structures when they bind to a metal to form a
four-membered chelate ring system. This provides a physical
and theoretical basis for understanding the resulting unique
organometallic and metal complex chemistry displayed by
these ligands.
Experimental
General
1
All reactions and product manipulations were carried out
under dry argon using standard Schlenk and dry box techni-
ques unless otherwise indicated. Dry, oxygen-free solvents
were employed throughout except where noted. The elemental
analyses were performed by the Mikroanalytisches Laborator-
ium der Chemischen Institute der Universita¨t Heidelberg.
Photoelectron spectra were measured at room temperature
on a PS18 spectrometer from Perkin–Elmer and calibrated
60% based on PCl3). H NMR (300 MHz, CDCl3 , 298 K): d
3
1.18 (d, JPH ¼ 12 Hz). 31P{1H} NMR (121.5 MHz, CDCl3 ,
298 K): d 146.2 (s).
Step 3: (tBu)2PMe. Di(tert-butyl)chlorophosphine (76.1 g,
0.421 mmol) was measured into a 1 L flask equipped with a
graduated dropping addition funnel and containing a magnetic
stirrer bar. The colourless liquid was dissolved in pentane
(285 mL) and cooled to ꢁ78 ꢀC. Methyl lithium (300 mL of
New J. Chem., 2003, 27, 540–550
547

View Article Online
a 1.6 M solution in diethyl ether, 0.48 mol) was added over 2.5
h and washed through with a further 100 mL diethyl ether. The
resulting white slurry was stirred and allowed to warm gradu-
ally to room temperature overnight. The reaction was then
quenched with an argon-saturated aqueous solution of NH4Cl
(33 mL of a 54.4 g Lꢁ1 solution). Upon stirring, all precipitate
dissolved. The organic phase was cannulated into a flask con-
taining MgSO4 , along with two pentane washings (30 mL
each) of the aqueous layer. After drying over MgSO4 for 1
h, the product was filtered into a round-bottomed flask
equipped with a magnetic stirrer, along with two pentane
washings (50 mL each) of the MgSO4 . The organic solvents
were distilled off at room pressure to yield a colourless liquid
(64.6 g, 0.403 mol, 96%). 1H NMR (500 MHz, CDCl3 , 298
details are not included as an alternative synthesis has been
published.38
4. Di(iso-propyl)chlorophosphine was used in step 3 and the
product was isolated as a colourless liquid in 92% yield. 31P
2
NMR (122 MHz, CDCl3 , 298 K): d 19.31 [d, JPP ¼ 98 Hz,
2
P(tBu)2], 3.03 [d, JPP ¼ 98 Hz, P(iPr2)]. MS (CI+):
m/z ¼ 277 (M + H)+ with the correct isotope pattern. Further
characterisation details are not included as an alternative
synthesis has been published.38
5. Diphenylchlorophosphine was used in step 3 and the pro-
duct was isolated from pentane in 88% yield as white crystals,
which melt at 40 ꢀC. Two other routes are also published.36,37
Anal. calcd for C21H30P2 : C, 73.2; H, 8.78; P, 18.0; found: C,
3
t
2
K): d 1.02 (d, JPH ¼ 11 Hz, 18H, Bu), 0.87 (d, JPH ¼ 4
Hz, 3H, Me). 31P{1H} NMR (202.5 MHz, CDCl3 , 298 K): d
12.2 (s).
1
73.4; H, 8.87; P, 17.7. H NMR (300 MHz, CDCl3 , 298 K): d
7.54 (m, 4H, m-C6H5), 7.36 (m, 6H, o/p-C6H5), 2.12 (s, 2H,
Step 4: (tBu)2PCH2Li.51. tert-Butyl lithium (280 mL of a 1.7
M solution in pentane, 0.476 mol) was transferred to a 500
mL round-bottomed flask with a blown-glass frit equipped
with a magnetic stirrer, and the solvent was removed under
vacuum. The (tBu)2PMe from the previous step (56.0 g,
0.348 mol) was dissolved in heptane (200 mL) and added.
The mixture was heated to reflux for a total of 18 h, resulting
in an orange solution and pale yellow precipitate. A further
200 mL of heptane was added and the slurry was filtered
through the attached frit at room temperature, then washed
with pentane (3 ꢃ 300 mL) and dried under vacuum. Total
yield of pale tan solid: 52.2 g, 0.314 mol (90%). 31P{1H}
NMR (121.5 MHz, THF-d8 , 298 K): d 45.2 (s).
CH2), 1.14 (d, JPH ¼ 10.9 Hz, 18H, tBu). 31P{1H} NMR
3
2
(122 MHz, CD2Cl2 , 298 K): d 15.9 [d, JPP ¼ 139.5 Hz,
2
P(tBu)2], ꢁ16.0 (d, JPP ¼ 139.5 Hz, PPh2). 13C{1H} NMR
1
(75 MHz, CD2Cl2 , 298 K): d 140.3 (dd, JPC ¼ 16.6 Hz,
2
3JPC ¼ 7.6 Hz, Cipso), 133.4 (d, JPC ¼ 19.4 Hz, Co), 128.8
3
1
(s, Cp), 128.5 (d, JPC ¼ 6.9 Hz, Cm), 32.0 [dd, JPC ¼ 24.2
3
2
Hz, JPC ¼ 6.9 Hz, C(CH3)3], 29.8 ([dd, JPC ¼ 13.8 Hz,
4JPC ¼ 2.1 Hz, C(CH3)3], 20.6 (dd, JPC ¼ 20.8 Hz,
1
1JPC ¼ 33.9 Hz, CH2). IR (KBr, cmꢁ1): 3069 w, 3055 w,
2947 s, 2893 m, 2860 m, 1585 w, 1465 m, 1433 s, 1387 w,
1365 m, 1176 m, 1093 w, 1067 w, 999 w, 809 m, 767 w, 736 s,
695 s, 507 m, 476 m. MS (CI+): m/z ¼ 345 (M + H)+ with the
correct isotope pattern.
Step 5: (tBu)2PCH2P(tBu)2. The solid lithium methanide
from the previous step (52.2 g, 0.314 mol) was weighed into
a round-bottomed flask equipped with a magnetic stirrer bar
in an argon-filled dry box and suspended in THF (1 L). After
7. The diphosphine 1 (2.0 g, 6.6 mmol) was dissolved in pen-
tane (20 mL) in a Schlenk tube equipped with a reflux conden-
ser and a magnetic stirrer bar. At reflux, 15 mL of a 30%
hydrogen peroxide solution was added. After stirring for 2 h
at reflux and cooling to room temperature, the pentane phase
was removed via cannula and the aqueous phase was washed
with pentane (2 ꢃ 2 mL). The combined organic phases were
dried over MgSO4 , filtered and concentrated under reduced
pressure. The oily product was then distilled under vacuum
at 230–240 ꢀC and the distillate was crystallised from a pen-
tane–diethyl ether mixture at 0 ꢀC. Yield: 1.9 g, 85%. Mp:
122 ꢀC. Anal. calcd for C17H38P2O2 : C, 60.67; H, 11.39; found:
C, 60.39; H, 11.42. 1H NMR (300 MHz, CDCl3 , 298 K): d 2.10
t
cooling the flask to ꢁ78 ꢀC, Bu2PCl (52.1 mL, 56.4 g, 0.313
mol) was added and the mixture was stirred and allowed to
warm gradually to room temperature overnight. The THF
was distilled off at room pressure, and the product was
extracted into pentane (100 mL). Hydrolysis with NH4Cl (80
mL of a 4% solution in argon-saturated water) caused the solu-
tion to warm. The organic phase, including two 100 mL pen-
tane washings of the aqueous phase, was dried over MgSO4
and filtered, and the pentane was distilled off at room pressure
resulting in a yellow oil. Dissolution in methanol and cooling
to ꢁ10 ꢀC led to the crystallisation of the product as colourless
needles. Yield: 74.0 g (0.243 mol, 77%). Anal. calcd for
C17H38P2 : C, 67.07; H, 12.58; P, 20.35; found: C, 67.11; H,
12.66; P, 20.33. Mp: 45–46 ꢀC. Bp: 75–80 ꢀC/10ꢁ3 mm. 1H
2
3
(t, JPH ¼ 12.8 Hz, 2H, CH2), 1.31 (d, JPH ¼ 14.0 Hz, 36H,
tBu). 31P{1H} NMR (121.5 MHz, CD2Cl2 , 298 K): d 59.7
(s). 13C{1H} NMR (75 MHz, CD2Cl2 , 298 K): d 36.7 [dd,
3
1JPC ¼ 34 Hz, JPC ¼ 6.2 Hz, PC(CH3)3], 26.9 [s, C(CH3)3],
2
1
NMR (300 MHz, C6D6 , 298 K): d 1.61 (t, JPH ¼ 1.8 Hz,
16.1 (‘‘t’’, JPC ¼ 40.5 Hz, CH2). IR (KBr, cmꢁ1): 2954 (vs,
3+5
t
2H, CH2), 1.19 (‘‘t’’,
J
PH
¼ 10.3 Hz, 36H, Bu). 31P{1H}
CH), 2903 (vs, CH), 2871 (vs, CH), 1158 (vs, PO). MS
NMR (121.5 MHz, C6D6 , 298 K): d 21.3 (s). 13C{1H} NMR
(CI+): m/z ¼ 337.2 (M)+ with the correct isotope pattern.
1
(75.7 MHz, C6D6 , 298 K): d 32.6 (dd, JPC ¼ 10.8 Hz,
3JPC ¼ 9.8 Hz, CMe3), 30.1 (‘‘t’’,
J
PC
¼ 15.7 Hz, CH3),
[8+][B(3,5-(CF3)2C6H3)4]ꢁ. In a glove box, [H(OEt2)2][BArF]
(108 mg, 0.107 mmol) and 1 (37 mg, 0.122 mmol) were weighed
together into a 20 mL Schlenk flask equipped with a magnetic
stirrer bar. Diethyl ether (2 mL) was added and the resulting
clear, colourless solution was stirred at room temperature for
10 min. Addition of hexane (2 mL) resulted in the precipitation
of a white solid, which was isolated and washed twice with hex-
ane. The fine white powder was dried under dynamic vacuum.
The compound can be recrystallised from ether–hexane (1:1) at
ꢁ20 ꢀC. Yield: 98 mg, 70%. Mp: 110–111 ꢀC (decomposition).
1H NMR (300 MHz, THF-d8 , 298 K): d 7.82 (s, 8H, o-BArF),
2+4
12.7 (t, JPC ¼ 36.2 Hz, CH2). IR (KBr pellet, cmꢁ1): 2900
(br, vs), 1459 (br, vs), 1385 (s), 1362 (vs), 1174 (br, vs), 1130
(m), 1017 (s), 982 (w), 930 (m), 808 (vs), 785 (s), 714 (s), 672
(s), 588 (m), 461 (m), 425 (m).
1
2. This ligand can be prepared in an analogous manner to 1
via Cy2PCH2Li, with a yield in the final step of 83%. Two other
routes are published.38,40 Mp: 95–97 ꢀC. Anal. calcd for
C25H46P2 : C, 73.49; H, 11.35; P, 15.16; found: C, 73.10; H,
11.34; P, 15.00.
1
3
7.60 (s, 4H, p-BArF), 5.97 (dt, JH,P ¼ 461.5 Hz, JH,H ¼ 6.4
3
2
3. Dicyclohexylchlorophosphine was used in step 3 and the
product was isolated in 53% yield as white crystals from pen-
tane. Mp: 46–47 ꢀC. Anal. calcd for C21H42P2 : C, 70.75; H,
11.87; found: C, 70.52; H, 11.88. 31P{1H} NMR (202.5
Hz, 1H, HPR3), 2.27 (ddd, JH,H ¼ 6.4 Hz, JP,H ¼ 14.5 Hz,
t
2
2JP,H ¼ 2.0 Hz, 2H, Bu2HPCH2PtBu2), 1.63 (d, JP,H ¼ 16.5
Hz, 9H, tBu2HPCH2PtBu2), 1.29 (d, JP,H ¼ 16.5 Hz, 9H,
2
tBu2HPCH2PtBu2). 31P{1H} NMR (THF-d8 , 121.5 MHz,
2
2
t
MHz, CD2Cl2 , 300 K): d 18.9 [d, JPP ¼ 108 Hz, P(tBu)2],
298 K): d 26.6 (d, JP,P ¼ 40.0 Hz, Bu2HPCH2PtBu2), 47.6
2
2
ꢁ5.7 (d, JPP ¼ 108 Hz, PCy2). Further characterisation
(d, JP,P ¼ 40.0 Hz, tBu2HPCH2PtBu2). 13C{1H} NMR
548
New J. Chem., 2003, 27, 540–550

View Article Online
(THF-d8 , 75.5 MHz, 298 K): d 162.8 (q, JBC ¼ 50.1 Hz, i-C
dcpm (2). The structure was measured on an Enraf–Nonius
¯
CAD-4 diffractometer. The space group P1 (#2) was con-
firmed by successful solution and refinement of the structure.
No absorption correction was applied. Hydrogen atoms were
refined isotropically.
3
BArF), 135.9 (br, o-C BArF), 129.6 (q, JC,F ¼ 49 Hz, m-C
1
BArF), 124.6 (q, JC,F ¼ 272 Hz, CF3), 118.5 (s, p-C BArF),
2
35.7 [m, PC(CH3)3], 34.5 [m, PC(CH3)63], 30.0 [d, JPC
¼
2
14.5 Hz, PC(CH3)3], 28.4 [d, JPC ¼ 4.15 Hz, PC(CH3)3], 7.1
1
1
(dd, JPC ¼ 41.5 Hz, JPC ¼ 53.6 Hz, HPCH2P).
dtbpe (6). The asymmetric unit consists of half the molecule,
with the middle of the ethylene bridge lying on an inversion
centre.
[9+][Fe2Cl6]1/2ꢁ. This salt was isolated as a by-product in an
attempt to prepare (dtbpm-k2P)FeCl2 ; no attempts were made
to optimise the synthesis of 9+. (CO)4FeCl2 (180 mg, 0.754
mmol) was suspended in toluene (10 mL) at ꢁ25 ꢀC. A solution
of 1 (250 mg, 0.821 mmol, 1.1 equiv.) was added via cannula.
Gas evolution was immediately observed and the colour chan-
ged from yellow to red and then to green. After addition of 1
was complete the solution was stirred for 30 min at ꢁ25 ꢀC,
then allowed to warm to room temperature and stirred for a
further 3 h. During this time, a yellow precipitate was
observed. The mother liquor was removed via cannula and
cooled to ꢁ25 ꢀC, which led to the crystallisation of yellow
(dtbpm-k2P)Fe(CO)3 . The residual yellow precipitate was dis-
solved in CH2Cl2 and cooling led to the formation of a
small quantity of pale yellow X-ray quality crystals of
[tBu2P(Cl)CH2PtBu2]+[Fe2Cl6]1/2ꢁꢂCH2Cl2 . No further char-
acterisation was conducted.
tBu2P(O)CH2P(O)tBu2 (7). Two independent, geometrically
similar molecules form the asymmetric unit.
[tBu2P(H)CH2PtBu2]+[B(3,5-(CF3)2C6H3)4]ꢁ (the cation of
which is 8+). The hydrogen atom on P1 was clearly seen in
the difference map and allowed to refine freely. Some disorder
of the CF3 groups of the anion was observed, as is character-
istic with this anion.
[tBu2P(Cl)CH2PtBu2]+[Fe2Cl6]1/2 CH2Cl2 (the cation of
ꢁ
.
which is 9+). The counter ion lies on a special position and is
present in the crystal as one [Fe2Cl6]2ꢁ with two bridging
chlorides per two phosphine cations.
Computational methods
Crystallographic studies
Complete geometry optimisations for 1–9+were carried out
using 2.5 GHz Dell PCs at the restricted Hartree–Fock
level using program systems Gaussian 98 as implemented in
Gaussian 98W,85 and Jaguar (Jaguar 3.5, Schro¨dinger, Inc.
Portland, Oregon, 2001) as implemented in Titan 1.05
(Wavefunction, Inc., Schro¨dinger, Inc, 2001). Minimum struc-
tures have been verified by vibrational frequency calculations
(NIMAG ¼ 0).
For each compound studied by X-ray diffraction, crystals were
grown by slow cooling of a concentrated pentane or methanol
solution of the compound. A crystal was mounted on a glass
fibre with perfluoropolyether. All measurements other than
for dcpm (2) were made on a Siemens SMART diffractometer
with graphite monochromated Mo-Ka radiation using a CCD
detector. Frames corresponding to a sphere of data were col-
lected using the o-scan technique; in each case, 20 s exposures
of 0.3ꢀ in o were taken. The reflections were integrated using
SAINT.82 An absorption correction was applied to each struc-
Acknowledgements
ture other than
1 (for reasons described below) using
SADABS83 and the data were corrected for Lorentz and polar-
isation effects. The space groups were determined by the sys-
tematic absences of hkl values using XPREP (Siemens,
SHELXTL 5.04). The structures were solved by direct meth-
ods and expanded using Fourier techniques. All non-hydrogen
atoms were refined anisotropically; hydrogen atoms were
included in calculated positions and not refined except for in
the structure of 2. The final cycle of full-matrix least-squares
refinement converged. The function minimised was
Sw[(Fo)2 ꢁ (Fc)2]2. All calculations were performed using the
SHELXTL crystallographic software package of Bruker.84
CCDC reference numbers 199933–40. See http://
www.rsc.org/suppdata/nj/b2/b210114a/ for crystallographic
files in CIF or other electronic format.
This work was supported by the Fonds der Chemischen
Industrie and the Deutsche Forschungsgemeinschaft. We
thank Dr. S. Ko¨stelmaier for preliminary MNDO calculations,
and we are grateful to Prof. R. Gleiter for providing access to
UV-PES instrumentation and to A. Flatow for carrying out
the PES measurements. We thank the Alexander von Hum-
boldt Foundation for a postdoctoral research fellowship (MS).
References
1
As in earler work, we use the shorthand notation dtbpm for
bis(di-tert-butylphosphino)methane (1) and analogously derived
acronyms for the other systems.
2
3
4
P. Dierkes and P. W. N. M. van Leeuwen, J. Chem. Soc., Dalton
Trans., 1999, 1519–1529.
R. M. Beesley, C. K. Ingold and J. F. Thorpe, J. Chem. Soc.,
1915, 107, 1080–1106.
B. Chaudret, B. Delavaux and R. Poilblanc, Coord. Chem. Rev.,
1988, 86, 191–243.
dtbpm (1). The skeleton of the molecule is disordered over
two orientations with almost the same occupancy, and the
bulky tert-butyl groups are located in roughly the same space
in the unit cell for both orientations. This disorder leads to a
pseudo-mirror plane. Therefore, the structure could be mod-
elled either in the centrosymmetric space group Pnma (#62)
with the mirror plane, or in the less-symmetric space group
Pna21 (#33). The latter was chosen, as use of the higher sym-
metry space group led to problems with special positions.
The disorder was modelled as two independent molecular geo-
metries, constrained to have the same bond lengths and angles,
which refined with occupancies of 54 and 46%. The structural
details described above correspond to the higher-occupancy
orientation. However, the disorder means that the accuracy
of the solution is low, and no absorption correction was
applied.
5
6
J. T. Mague, J. Cluster Sci., 1995, 6, 217–269.
G. M. Ferrence, P. E. Fanwick, C. P. Kubiak and R. J. Haines,
Polyhedron, 1997, 16, 1453–1459.
7
8
9
B. F. Straub, F. Rominger and P. Hofmann, Inorg. Chem., 2000,
39, 2113–2119.
S. M. Reid, J. T. Mague and M. J. Fink, J. Am. Chem. Soc., 2001,
123, 4081–4082.
P. Hofmann, C. Meier, U. Englert and M. U. Schmidt, Chem.
Ber., 1992, 125, 353–365.
10 T. Nickel, R. Goddard, C. Kruger and K.-R. Po¨rschke, Angew.
¨
Chem., Int. Ed. Engl., 1994, 33, 879–882.
11 B. F. Straub, F. Rominger and P. Hofmann, Inorg. Chem.
Commun., 2000, 3, 358–360.
12 N. Simhai, C. N. Iverson, B. L. Edelbach and W. D. Jones,
Organometallics, 2001, 20, 2759–2766.
New J. Chem., 2003, 27, 540–550
549

View Article Online
53 P. Hofmann, C. Meier, W. Hiller, M. Heckel, J. Riede and M. U.
13 P. Hofmann, H. Heiss, P. Neiteler, G. Muller and J. Lachmann,
¨
Angew. Chem., Int. Ed. Engl., 1990, 29, 880–882.
14 P. Hofmann and G. Unfried, Chem. Ber., 1992, 125, 659–661.
15 G. Scherhag, Dissertation, Ruprecht-Karls Universita¨t Heidel-
berg, Heidelberg, 2001.
Schmidt, J. Organomet. Chem., 1995, 490, 51–70.
54 Solid-state structures containing a k2-dtbpm chelate ligand bound
to Ni, Pd, Pt, Fe, Ru, Rh and Ir have been determined in our
group.
16 P. Hofmann, H. Heiss and G. Muller, Z. Naturforsch., B: Chem.
55 K. Issleib and D.-W. Muller, Chem. Ber., 1959, 92, 3175–3182.
¨
¨
Sci., 1987, 42, 395–409.
17 P. Hofmann, L. A. Perez-Moya, M. E. Krause, O. Kumberger
56 W. Hewertson and H. R. Watson, J. Chem. Soc., 1962,
1490–1494.
57 H. Schmidbaur, G. Reber, A. Schier, F. E. Wagner and
and G. Muller, Z. Naturforsch., B: Chem. Sci., 1990, 45, 897–908.
¨
18 P. Hofmann, L. A. Perez-Moya, O. Steigelmann and J. Riede,
Organometallics, 1992, 11, 1167–1176.
G. Muller, Inorg. Chim. Acta, 1988, 147, 143–150.
¨
58 J. Bruckmann and C. Kruger, Acta Crystallogr., Sect. C, 1995, 51,
¨
19 B. F. Straub and P. Hofmann, Inorg. Chem. Commun., 1998, 1,
350–353.
20 B. F. Straub, F. Rominger and P. Hofmann, Inorg. Chem.
Commun., 2000, 3, 214–217.
21 S. M. Hansen, M. A. O. Volland, F. Rominger, F. Eisentra¨ger
and P. Hofmann, Angew. Chem., Int. Ed., 1999, 38, 1273–1276.
22 M. A. O. Volland, C. Adlhart, C. A. Kiener, P. Chen and
P. Hofmann, Chem.-Eur. J., 2001, 7, 4621–4632.
23 D. B. Grotjahn, G. A. Bikzhanova, L. S. B. Collins, T. Concolino,
K.-C. Lam and A. L. Rheingold, J. Am. Chem. Soc., 2000, 122,
5222–5223.
1155–1158.
59 J. Bruckmann and C. Kruger, Acta Crystallogr., Sect. C, 1995, 51,
1152–1155.
60 F. M. Conroy-Lewis, L. Mole, A. D. Redhouse, S. A. Litster
and J. L. Spencer, J. Chem. Soc., Chem. Commun., 1991,
1601–1603.
61 I. Bach, K.-R. Po¨rschke, R. Goddard, C. Kopiske, C. Kru¨ger, A.
Rufinska and K. Seevogel, Organometallics, 1996, 15, 4959–4966.
62 I. Bach, K.-R. Po¨rschke, B. Proft, R. Goddard, C. Kopiske,
¨
C. Kruger, A. Rufinska and K. Seevogel, J. Am. Chem. Soc.,
¨
1997, 119, 3773–3781.
24 H. Urtel, G. A. Bikzhanova, D. B. Grotjahn and P. Hofmann,
Organometallics, 2001, 20, 3938–3949.
63 I. Bach, R. Goddard, C. Kopiske, K. Seevogel and K.-R.
Po¨rschke, Organometallics, 1999, 18, 10–20.
25 H. Urtel, C. Meier, F. Eisentra¨ger, F. Rominger, J. Joschek and
P. Hofmann, Angew. Chem., Int. Ed., 2001, 40, 781–784.
26 F. Lippert and A. Ho¨hn and P. Hofmann, World Pat. WO
96/37537, 1996 (BASF AG).
64 D. J. Mindiola and G. L. Hillhouse, J. Am. Chem. Soc., 2001, 123,
4623–4624.
65 D. Mindiola and G. Hillhouse, J. Am. Chem. Soc., 2002, 124,
9976–9977.
27 F. Lippert, A. Ho¨hn and P. Hofmann, World Pat. WO 96/37536,
1996 (BASF AG).
66 R. Melenkivitz, D. Mindiola and G. Hillhouse, J. Am. Chem.
Soc., 2002, 124, 3846–3847.
28 F. Lippert, A. Ho¨hn and P. Hofmann, World Pat. WO 95/03353,
1995 (BASF AG).
29 F. Lippert, A. Ho¨hn and E. Schauss, World Pat. WO 96/37522,
1996 (BASF AG).
67 R. Trebbe, R. Goddard, A. Rufinska, K. Seevogel and K.-R.
Po¨rschke, Organometallics, 1999, 18, 2466–2472.
68 N. Carr, B. J. Dunne, L. Mole, A. G. Orpen and J. L. Spencer,
J. Chem. Soc., Dalton Trans., 1991, 863–871.
30 N. A. Cooley, S. M. Green, D. F. Wass, K. Heslop, A. G. Orpen
and P. G. Pringle, Organometallics, 2001, 20, 4769–4771.
31 M. Schultz, F. Eisentra¨ger, C. Regius, F. Rominger and
P. Hofmann, in preparation.
32 I. D. Gridnev, Y. Yamanoi, N. Higashi, H. Tsuruta, M. Yasutake
and T. Imamoto, Adv. Synth. Catal., 2001, 343, 118–136.
33 H. H. Karsch, Z. Naturforsch., B: Anorg. Chem. Org. Chem.,
1983, 38, 1027–1030.
34 P. Hofmann and H. Heiss, Ger. Pat. DE 41 34 772, 1991 (BASF
AG).
35 J. Feldman, E. Hauptman, E. F. McCord and M. S. Brookhart,
World Pat. WO 98/47934, 1998 (Du Pont).
69 T. Gottschalk-Gaudig, J. C. Huffman and K. G. Caulton, J. Am.
Chem. Soc., 1999, 121, 3242–3243.
70 T. Gottschalk-Gaudig, J. C. Huffman, H. Ge´rard, O. Eisenstein
and K. G. Caulton, Inorg. Chem., 2000, 39, 3957–3962.
71 M. A. O. Volland, B. F. Straub, I. Gruber, F. Rominger and
P. Hofmann, J. Organomet. Chem., 2001, 617–618, 288–291.
72 D. W. H. Rankin, H. E. Robertson, R. Seip, H. Schmidbaur and
G. Blaschke, J. Chem. Soc., Dalton Trans., 1985, 827–830.
73 H. A. Bent, J. Chem. Educ., 1960, 37, 616–624.
74 H. A. Bent, Chem. Rev., 1961, 61, 275–311.
75 M. A. O. Volland, Dissertation, Ruprecht-Karls Universita¨t
Heidelberg, Heidelberg, 2001.
36 S. O. Grim, P. H. Smith, I. J. Colquhoun and W. McFarlane,
Inorg. Chem., 1980, 19, 3195–3198.
76 T. P. Debies and J. W. Rabalais, Inorg. Chem., 1974, 13,
308–312.
37 L. Weber and D. Wewers, Chem. Ber., 1984, 117, 1103–1112.
38 J. Wolf, M. Manger, U. Schmidt, G. Fries, D. Barth, B.
Weberndo¨rfer, D. A. Vicic, W. D. Jones and H. Werner, J. Chem.
Soc., Dalton Trans., 1999, 1867–1875.
77 S. Ikuta, P. Kebarle, G. M. Bancroft, T. Chan and R. J.
Puddephatt, J. Am. Chem. Soc., 1982, 104, 5899–5902.
78 D. E. Cabelli, A. H. Cowley and M. J. S. Dewar, J. Am. Chem.
Soc., 1981, 103, 3286–3289.
39 K. Jonas, preparation of dcpm, personal communication, 1988.
40 F. L. Joslin, J. T. Mague and D. M. Roundhill, Polyhedron, 1991,
10, 1713–1715.
41 I. P. Rothwell, the hydrogenation requires well over one equiva-
lent of catalyst (personal communication), 2000.
42 J. S. Yu and I. P. Rothwell, J. Chem. Soc., Chem. Commun., 1992,
632–633.
43 M. F. Lappert, J. B. Pedley, B. T. Wilkins, O. Stelzer and
E. Unger, J. Chem. Soc., Dalton Trans., 1975, 1207–1214.
44 B. L. Shaw, J. Chem. Soc., Chem. Commun., 1979, 104–105.
45 G. M. Bancroft, L. Dignard-Bailey and R. J. Puddephatt, Inorg.
Chem., 1986, 25, 3675–3680.
46 D. L. Lichtenberger and M. E. Jatcko, Inorg. Chem., 1992, 31,
451–455.
47 M. J. S. Dewar, E. G. Zoebisch, F. Eamonn and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902–3909.
48 K. Issleib and W. Seidel, Chem. Ber., 1959, 92, 2681–2694.
49 M. Fild, O. Stelzer and R. Schmutzler, Inorg. Synth., 1973,
14, 4–9.
50 H. Gilman and F. K. Cartledge, J. Organomet. Chem., 1964, 2,
447–454.
79 R. Hoffmann, Acc. Chem. Res., 1971, 4, 1–9.
80 K.-R. Po¨rschke, C. Pluta, B. Proft, F. Lutz and C. Kruger,
¨
Z. Naturforsch., B: Chem. Sci., 1993, 48, 608–626.
81 W. Hieber and G. Bader, Chem. Ber., 1928, 61, 1717–1722.
82 SAINT, v. 4.024: Siemens Industrial Automation, Inc, Madison,
WI, 1995.
83 G. M. Sheldrick, SADABS, Bruker AXS, Inc., Madison,
WI, 1996.
84 G. M. Sheldrick, SHELXTL Crystal Structure Determination
Package, Bruker AXS, Inc, Madison, WI, 1997.
85 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery,
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C.
Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G.
Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S.
Replogle and J. A. Pople, Gaussian 98, revision A.9, Gaussian,
Inc., Pittsburgh, PA, 1998.
51 H. H. Karsch and H. Schmidbaur, Z. Naturforsch., B: Anorg.
Chem. Org. Chem., 1977, 32, 762–767.
52 A large centrifuge, which can be used with the exclusion of air
from the flask, could also be used if available.
550
New J. Chem., 2003, 27, 540–550