Cationic chelate phosphane substituted carbyne tungsten complexes

Article abstract of DOI:10.1016/S0022-328X(00)00864-0

Reaction of W(CMes)Cl(CO)₂(dppe) (1) and W(CMes)Cl(CO)(dppep) (8) (Mes=2,4,6-trimethylphenyl, dppe=Ph₂PCH₂CH₂PPh₂, dppep=(Ph₂PCH₂CH₂)₂PPh) with AgSbF₆ or AgOTf (OTf=CF₃SO₃^-) as halide abstraction reagents affords [W(CMes)(CO)_n(Y)][X]; (2): Y=dppe, n=2; (9): Y=dppep, n=1; (a) X=SbF₆^-; (b): X=OTf. In solution, the metal moieties of compounds 2a, 9a and 9b exist as naked cations, which possess a square pyramidal coordination. The carbyne ligand occupies the apical position, and a free coordination site is located trans to the CMes group. The reaction of 2a with a variety of donor ligands leads to [W(CMes)(CO)₂(dppe)L][SbF₆] compounds, in which the ligand L binds trans to the carbyne unit (L=THF (4a), CH₃CN (4b), PMe₃ (5a), PEt₃ (5b)). In the carbonyl adducts [W(CMes)(CO)_n(Y)][SbF₆]; Y=dppe, n=3, (3); Y=dppep, n=2, (10); the chelating ligands rearrange so that a phosphorus donor atom coordinates trans to the carbyne ligand. The reaction of 2a with P(OMe)₃ and P(OEt)₃ results in the phosphonate complexes [W(CMes)(CO)₂(OPH(OR)₂)(dppe)][SbF₆]; R=Me (6a), Et (6b), which are formed in an Arbuzov dealkylation reaction. When 1 is reacted with a deficient amount of silver salt, the bimetallic complex [{W(CMes)(CO)₂(dppe)}₂Cl][SbF₆] (7) is obtained. The unusual molecular geometries of 6a and 7 were confirmed by a crystal structure determination. A theoretical analysis reveals the importance of electrostatic interaction for ligand bonding with the cationic metal fragments.

Full text of DOI:10.1016/S0022-328X(00)00864-0

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 622 (2001) 97–111
Cationic chelate phosphane substituted carbyne tungsten
complexes
Ewald Bannwart, Heiko Jacobsen, Rainer Hu¨bener, Helmut W. Schmalle,
Heinz Berke *
Anorganisch-Chemisches Institut der Uni6ersita¨t Zu¨rich, Winterthurerstrasse 190, CH–8057 Zu¨rich, Switzerland
Received 15 June 2000; received in revised form 21 September 2000; accepted 25 October 2000
Abstract
Reaction of W(CMes)Cl(CO)2(dppe) (1) and W(CMes)Cl(CO)(dppep) (8) (Mes=2,4,6-trimethylphenyl, dppe=
Ph2PCH2CH2PPh2, dppep=(Ph2PCH2CH2)2PPh) with AgSbF6 or AgOTf (OTf=CF3SO−3 ) as halide abstraction reagents affords
[W(CMes)(CO)n(Y)][X]; (2): Y=dppe, n=2; (9): Y=dppep, n=1; (a) X=SbF−6 ; (b): X=OTf. In solution, the metal moieties
of compounds 2a, 9a and 9b exist as naked cations, which possess a square pyramidal coordination. The carbyne ligand occupies
the apical position, and a free coordination site is located trans to the CMes group. The reaction of 2a with a variety of donor
ligands leads to [W(CMes)(CO)2(dppe)L][SbF6] compounds, in which the ligand L binds trans to the carbyne unit (L=THF (4a),
CH3CN (4b), PMe3 (5a), PEt3 (5b)). In the carbonyl adducts [W(CMes)(CO)n(Y)][SbF6]; Y=dppe, n=3, (3); Y=dppep, n=2,
(10); the chelating ligands rearrange so that a phosphorus donor atom coordinates trans to the carbyne ligand. The reaction of
2a with P(OMe)3 and P(OEt)3 results in the phosphonate complexes [W(CMes)(CO)2(OPH(OR)2)(dppe)][SbF6]; R=Me (6a), Et
(6b), which are formed in an Arbuzov dealkylation reaction. When 1 is reacted with a deficient amount of silver salt, the bimetallic
complex [{W(CMes)(CO)2(dppe)}2Cl][SbF6] (7) is obtained. The unusual molecular geometries of 6a and 7 were confirmed by a
crystal structure determination. A theoretical analysis reveals the importance of electrostatic interaction for ligand bonding with
the cationic metal fragments. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Triple bonded ligand system; Metal center; Carbyne tungsten complexes
electronic deficiency of 16e− species, and their chelating
properties were sought to stabilize the coordination
1. Introduction
sphere around the transition metal center.
In transition metal coordination chemistry, the car-
byne moiety CR acts as a triple bonded ligand system,
and establishes very close contacts to the metal center
[1]. As a consequence, it exhibits a strong influence on
ligands Lt-CR disposed in trans position [2]. The trans-
effect substantially labilizes the M–Lt-CR bond, or even
prevents coordination of a trans-ligand [1], resulting in
electronically unsaturated 16e− carbyne complexes. In
order to exploit the characteristic features of the CR
group, we approached the synthesis of bi- and triden-
tate phosphane substituted tungsten carbyne complexes
[3,4]. The electron donating phosphorus ligands were
supposed to account for some compensation of the
The underlying motivation for the design of such
complexes originated from the idea that catalytic trans-
formations mediated by transition metal complexes
generally require labile coordination sites. The carbyne
ligand might provide the lability needed, and render the
tungsten complex catalitically active.
In this work, we describe the synthesis of the cationic
tungsten carbyne complexes [W(CMes)(CO)2(dppe)]+
and [W(CMes)(CO)(dppep)]+ as potential precursors
for catalytic conversions (Mes=2,4,6-trimethylphenyl,
dppe=Ph2PCH2CH2PPh2, dppep=(Ph2PCH2CH2)2-
PPh)). We further initiated an investigation of their
reactivity by exploring the coordination behavior of
these cations towards prototypical donor ligands. The
reactivity studies are supported by theoretical calcula-
tions based on density functional theory (DFT) [5].
* Corresponding author. Tel.: +41-1-635-4680; fax: +41-1-635-
6802.
E-mail address: hberke@aci.unizh.ch (H. Berke).
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00864-0

98
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
2. Results and discussion
The reaction of 1 with AgSbF6 can be monitored in
the IR spectrum through the appearance of two new,
intense w(CO) bands at 2006 and 1940 cm−1. These
signals are shifted to higher wavenumbers as compared
2.1. Synthesis and reacti6ity of
[W(CMes)(CO)2(dppe)]+ complexes
to those for 1, which are found at 1997 and 1913 cm−1
,
respectively. The high field shift observed in the car-
bonyl stretching frequencies is caused by electrostatic
contributions to the [M]–CO bond [8,9], and provides a
first indication to the cationic nature of the transition
metal moiety. The intensity pattern of the w(CO) bands
suggests that the CO groups are coordinated in cis
position to each other.
A synthetic route to dppe substituted tungsten car-
byne cations has been described before in the literature.
The groups headed by Templeton [6] and Gimeno [7]
introduced the carbyne unit by protonation of the
corresponding vinylidene compounds. These systems,
however, seemed not to be suitable precursors for coor-
dinatively unsaturated species, since one might suspect
that the inevitable presence of a-hydrogens at the car-
byne residue could eventually cause complications in
the preparation of 16e− complexes. We therefore in-
tended to develop aryl carbyne species, which were
obtained in the reaction of carbyne chloride complexes
with silver salts, acting as halide abstraction reagents.
The only signal observed in the 31P{1H}-NMR spec-
1
trum is a singlet resonance at 50.7 ppm with a JPW
coupling constant of 237 Hz; a value characteristic of
phosphorus donors in trans position to a carbonyl
ligand [10,11]. This points to the fact that the P-atoms
of the dppe group and the CO ligands are located in the
same coordination plane. Further evidence for this
finding is provided by the 13C{1H}-NMR spectrum, in
which the chemically equivalent CCO atoms are iden-
tified as doublet of a doublet at 207.5 ppm. The two
2JCP values are 36.5 and 5.0 Hz, and are due to P nuclei
in trans and in cis-position, respectively. The resonance
for the CCMes atom is found at 300 ppm, and is signifi-
cantly shifted to lower field by 29 ppm, compared to
the corresponding signal measured for 1. These spectro-
scopic findings, together with a dominant M+ peak at
e/z=769 in the mass spectrum, unequivocally confirm
the square pyramidal coordination geometry of the
16e− cation 2a. The prominent axial position is occu-
pied by the carbyne ligand.
2.1.1. Preparation of [W(CMes)(CO)2(dppe)]+
Starting point for the synthesis of tungsten cations
containing the dppe ligand was the complex
W(CMes)(CO)2Cl(dppe) (1), which we have described
earlier in the context of the synthesis of carbyne hy-
dride species [3]. The reaction of 1 with a slight excess
of AgSbF6 in CH2Cl2 affords the naked cation
[W(CMes)(CO)2(dppe)]+ (2a) as SbF−6 salt (Scheme 1).
The conversion proceeds smoothly at room tempera-
ture, and is complete within minutes. When keeping 2a
in solution, decomposition is observed already after a
short period of time. One of the decomposition prod-
ucts could be identified as the triply carbonyl substi-
tuted cation 3 (vide infra). In the solid state, however,
2a appears to be more stable. It can be stored without
decomposition for an indefinite period of time at
−20°C.
A coordination of hexafluoroantimonate in 2a can be
definitely excluded. 31P{1H}-NMR spectra recorded at
2
various temperatures do not reveal any JFP coupling,
as would be expected in the case of coordination of the
counter ion [12]. Furthermore, only one band for the
Sb–F stretches is observed at 665 cm−1 in the IR
spectrum. A monodentate SbF−6 group, however, is
characterized by the presence of three w(SbF) bands
[13].
When silver trifluoromethanesulfonate AgOTf
(OTf=CF3SO−3 ) is used as halide abstraction reagent,
compound 2b is obtained as yellow powder in good
yield (Scheme 1). The w(CO) stretching frequencies
appear at 1998 and 1932 cm−1, and the value for the
13C-resonance of the CCMes atom is determined as 289.6
ppm. When these values are compared with the corre-
sponding data for compounds 1 and 2a, the metal
center in 2b seems to have less cationic character than
in 2a, but still carries a higher amount of positive
charge than in 1. These findings point to the presence of
a weakly coordinated OTf group in 2b; a fact that is
supported by three additional observations. When
NaOTf is added to a solution of 2b, a new signal
appears in the 19F-NMR spectrum. This new resonance,
Scheme 1.

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
99
Scheme 2.
which is attributed to a free triflate anion, is shifted to
higher field compared to the 19F-resonance at −78
ppm for 2b, which is due to coordinated OTf. Secondly,
for the free triflate anion, the S–O stretching frequen-
cies have been determined as was(SO3)=1273 and
ws(SO3)=1032 cm−1 [14], whereas for 2b, values of
1204 and 1007 cm−1 have been measured. The shift to
lower wavenumbers indicates a weakening of the S–O
bond, caused by coordination of the OTf group to the
metal center. Finally, a peak at e/z=920 in the mass
spectrum confirms the existence of a cation with the
composition C39H35F3O5P2SW. The patterns for the
dppe and carbonyl ligands in the 31P{1H}- and
13C{1H}-NMR spectra are in qualitative agreement
with those for 2a. Compound 2b, therefore, possesses a
pseudo octahedral coordination geometry with the car-
byne and the triflate groups in trans-, and the carbonyl
ligands in cis-position.
phosphorus atoms is too small to be resolved in the
31P{1H}-NMR spectrum; an observation which was
made before by Templeton [6] and Gimeno [7] for
analogous compounds.
From the spectroscopic data discussed above to-
gether with the 13C-resonances of the carbonyl ligands,
one can deduce the coordination geometry for the
cation 3. The complex possesses a pseudo octahedral
coordination sphere, in which one of the P atoms of the
dppe ligand is located trans to the carbyne group. The
remaining P atom and three carbonyl groups define a
meridional plane in this complex.
The fact that a carbon monoxide molecule does not
occupy the free coordination site trans to the carbyne
moiety in 2a underlines the strong trans influence of the
CMes group. Due to the strong [W]ꢀCMes triple bond,
the tungsten center cannot provide enough electron
density for p-backdonation to support a stable [W]–CO
bond in trans position.
The incoming CO ligand does not show a preference
for one specific coordination site. When the reaction of
1 with AgSbF6 is performed under an atmosphere of
13C-labeled carbon monoxide, both carbonyl reso-
nances in the 13C-NMR spectrum reveal the presence of
13CO ligands. The intensity ratio of the two signals is
approximately 1:1. When the 13C-enriched complex 3 is
treated with tetramethylammonium chloride, an equal
mixture of labeled and unlabeled 1 is obtained. This
reaction sequence offers the possibility of synthesizing
isotopically labeled starting material.
2.1.2. Reaction of [W(CMes)(CO)2(dppe)]+ with carbon
monoxide
The triply carbonyl-substituted cation 3, which ap-
pears during the decomposition of 2a in solution, can
be obtained in almost quantitative yield, when the
reaction of 1 with AgSbF6 is carried out under an
atmosphere of carbon monoxide (Scheme 1). The for-
mation of the new compound is seen in the IR spec-
trum by the appearance of three new w(CO) bands at
2074, 1993, and 1922 cm−1. Alternatively, 2a might be
dissolved in a non-coordinating solvent, and stirred for
a few minutes under 1 atm of CO. 3 is isolated as
bright-yellow powder, and contrary to 2a, it is stable
also in solution at room temperature.
2.1.3. Sol6ent complexes of [W(CMes)(CO)2(dppe)]+
When compounds 2a or 3 are dissolved in deuterated
THF, the NMR spectra reveal the existence of a new
compound 4a (Scheme 2), which is present in solution
together with cations 2a and 3.
The yield of product formation has been determined
as 20% by integration of the 31P{1H}-NMR signals.
The small concentration of 4a in the reaction mixture
might be one of the reasons why the CCMes resonance
for the new compound could not be detected. Neverthe-
less, the new signals for the dppe and CO ligands at
47.1 and 218.5 ppm, as well as the presence of a
coordinated THF molecule, indicated by low field shifts
The 13C-resonance of the CCMes atom appears at
2
308.6 ppm as doublet of a doublet with JCP values of
20.4 and 7.7 Hz, where the large coupling constant is
due to interaction with a P-nucleus in trans position. In
the 31P{1H}-NMR spectrum, two signals at 37.9 and
23.9 ppm are observed, which both exhibit tungsten
1
satellites with JPW couplings of 237 and 91 Hz, respec-
1
tively. The remarkably small JPW value for the signal
at higher field can be attributed to a phosphorus donor
coordinated in trans position to a carbyne group [6b,7]
2
The JPP coupling of the chemically non equivalent

100
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
of 1.2 and 2.7 ppm of the methylenic quintet reso-
nances, provide evidence for a [W(CMes)(CO)2(dppe)-
(THF)]+ cation, in which the solvent molecule occupies
the coordination site in trans position to the carbyne
group.
It was not possible to isolate complex 4a. Evapora-
tion of excess solvent always led to loss of coordinated
THF, and various attempts to crystallize or to precipi-
tate 4a were not successful.
Expectedly, a stable solvent complex can be obtained
with the stronger donor acetonitrile. When the reaction
of 1 and AgSbF6 is carried out in a 3:1 mixture of
dichloromethane and CH3CN, the cationic complex
2.1.4. Reaction of [W(CMes)(CO)2(dppe)]+ with
phosphorus donors
The reactivity of 2a towards a variety of phosphanes
was also investigated. In these experiments, a solution
of 2a in toluene was treated with one equivalent of the
PR3 donor at −20°C, and slowly warmed to room
temperature.
The reactions with the trialkylphosphanes PMe3 and
PEt3 proceed as expected. The new phosphorus donor
occupies the coordination site trans to the CMes moiety
(Scheme 2). The compounds 5a and 5b have been
characterized by NMR and IR spectroscopy. Besides
the now familiar pattern for a meridial coordination of
the dppe and CO ligands, new triplets for the PR3
donors appear in the 31P{1H}-NMR spectrum at
−58.9 and −26.2 ppm for 5a and 5b, respectively. The
[trans-W(CMes)(CO)2(dppe)(NCCH3)]
[SbF6]
4b
(Scheme 2) is isolated as yellow powder in almost
quantitative yield. Even in solution, in the absence of
stronger donor molecules, no dissociation of acetoni-
trile is observed. This reaction provides further oppor-
tunity to protect the cationic complex 2a against
decomposition.
1
small JPW coupling constants of 83 and 84 Hz again
corroborate the trans disposition of the CMes and PR3
groups.
Through inspection of the NMR spectra of com-
plexes 6a and 6b, which were obtained in the reaction
of 2a with P(OMe)3 and P(OEt)3, respectively, it be-
came clear that in these cases, the P-donors display an
unusual coordination chemistry (Scheme 2). A single
crystal analysis of compound 6a proved helpful in
clearifying the structure of this compound. The geome-
try of the cation 6a is depicted in Fig. 1.
The tungsten center in 6a possesses a pseudo octahe-
dral coordination, and the motif of meridional coordi-
nation of the dppe and CO ligands is well established.
The [W]ꢀCMes bond length amounts to 185.5(10) pm,
and lies within the expected range for tungsten carbyne
complexes.[16] The structure reveals that the P(OMe)3
molecule has been transformed oxidatively into
OPH(OMe)2. This ligand coordinates via an oxygen
atom to the tungsten center. The P-atom possesses
tetrahedral coordination geometry; the H-atom position
at this group has been located by difference electron
density calculations.
The spectral data are in accord with the results of the
crystal structure analysis. In the 31P{1H}-NMR spec-
trum, the signal for the dppe ligand now appears as
Fig. 1. The molecular structure of cation 6a in the crystal. Disordered
atoms and hydrogen atoms, except for H3 bound to P3, are omitted
for clarity. Displacement ellipsoids are shown at the 30% level. H3 is
dislayed as a sphere of arbitrary size. Selected bond distances (pm)
and angles (°): W–C(5), 185.1(10); W–C(1), 202.3(12); W–C(2),
202.9(14); W–O(3), 225.0(8); W–P(2), 254.8(3); W–P(1), 256.4(3);
P(3)–O(3), 146.5(8); P(3)–O(4), 155.5(8); P(3)–O(51), 158.0(10);
P(3)–O(52), 170(8); P(3)–H(3), 144.9(3); O(1)–C(1), 113.9(14); O(2)–
C(2), 112.7(15); O(4)–C(6), 146.3(16); O(51)–C(7), 142.3(16); O(52)–
C(7), 141(9); C(5)–W–C(1), 84.3(4); C(5)–W–C(2), 86.9(5);
C(1)–W–C(2), 90.9(5); C(5)–W–O(3), 172.4(4); C(1)–W–O(3),
88.2(4); C(2)–W–O(3), 92.7(4); C(5)–W–P(2), 103.0(3); C(1)–W–
P(2), 95.8(3); C(2)–W–P(2), 168.5(4); O(3)–W–P(2), 78.3(2); C(5)–
W–P(1), 105.5(3); C(1)–W–P(1), 169.6(3); C(2)–W–P(1), 93.2(4);
O(3)–W–P(1), 82.1(2); P(2)–W–P(1), 78.69(9); O(3)–P(3)–O(4),
115.0(5); O(3)–P(3)–O(51), 108.5(5); O(4)–P(3)–O(51), 107.8(5);
O(3)–P(3)–O(52), 129(3); O(4)–P(3)–O(52), 111(3); O(3)–P(3)–H,
108.5(4); O(4)–P(3)–H, 108.2(4); O(51)–P(3)–H(3), 108.8; O(52)–
P(3)–H(3), 75(3); P(3)–O(3)–W, 142.7(5).
3
doublet at 49.1 ppm, with a coupling constant JPP of 6
1
Hz, and tungsten satellites with JPW=236 Hz. A new
triplet signal with half the integrated intensity of the
3
doublet arises at 19.1 ppm, again with a JPP value of 6
Hz. The signal does not possess any tungsten satellites;
an indication that the corresponding P-nucleus is not
coordinated to the metal center. In the 31P-NMR spec-
1
trum, this resonance splits into two signals with a JPH
coupling of 752 Hz. This value is in the range of the
1JPH values reported for the free OPH(OMe)2
molecule.[16] The P-bounded proton appears as dou-
1
blet at 5.22 ppm, with the same JHP coupling constant
measured in the 31P-NMR experiment. These observa-
tions confirm the presence of the dimethyl phosphonate
ligand.

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
101
Scheme 3.
Fig. 2. The molecular structure of cation 7 in the crystal. Hydrogen atoms are omitted for clarity. Displacement ellipsoids are shown at the 30%
level. Selected bond distances (pm) and angles (°): W(1)–C(1), 185(2); W(1)–C(2), 202(2); W(1)–C(5), 188(2); W(1)–P(1), 252.2(6); W(1)–P(2),
255.7(8); W(1)–Cl(1), 261.7(6); W(2)–C(3), 197(3); W(2)–C(4), 196(4); W(2)–C(6), 176.9(19); W(2)–P(3), 254.9(7); W(2)–P(4), 252.4(7);
W(2)–Cl(1), 263.1(6); O(1)–C(1), 126(3); O(2)–C(2), 115(2); O(3)–C(3), 119(3); O(4)–C(4), 117(3); C(1)–W(1)–C(5), 88.0(9); C(1)–W(1)–C(2),
93.5(8); C(5)–W(1)–C(2), 86.6(8); C(1)–W(1)–P(1), 91.7(7); C(5)–W(1)–P(1), 98.8(7); C(2)–W(1)–P(1), 172.6(6); C(1)–W(1)–P(2), 170.0(6);
C(5)–W(1)–P(2), 97.4(7); C(2)–W(1)–P(2), 95.2(6); P(1)–W(1)–P(2), 79.2(2); C(1)–W(1)–Cl(1), 84.4(6); C(5)–W(1)–Cl(1), 169.6(7); C(2)–W(1)–
Cl(1), 86.7(5); P(1)–W(1)–Cl(1), 88.6(2); P(2)–W(1)–Cl(1), 91.2(2); C(6)–W(2)–C(4), 87.1(11); C(6)–W(2)–C(3), 91.6(9); C(4)–W(2)–C(3),
92.8(11); C(6)–W(2)–P(4), 97.2(7); C(4)–W(2)–P(4), 91.3(9); C(3)–W(2)–P(4), 170.4(7); C(6)–W(2)–P(3), 98.0(7); C(4)–W(2)–P(3), 170.0(9);
C(3)–W(2)–P(3), 95.7(8); P(4)–W(2)–P(3), 79.5(2); C(6)–W(2)–Cl(1), 173.3(7); C(4)–W(2)–Cl(1), 86.9(9); C(3)–W(2)–Cl(1), 85.7(7); P(4)–
W(2)–Cl(1), 85.9(2); P(3)–W(2)–Cl(1), 88.4(2); W(1)–Cl(1)–W(2), 126.4(3).
2.2. The v-Cl bridged dinuclear cation
[{W(CMes)(CO)2(dppe)}2Cl]+
The spectroscopic data for compound 6b comply
with those of 6a. This leads to the conclusion that a
OPH(OEt)2 ligand is present in the coordination
sphere around the tungsten center.
The reaction of 2a with halide abstraction reagents
has also been performed with a deficient stoichiometry
of the silver salt. During the reaction of 2a with 0.5
equivalents of AgSbF6, a complex product mixture
develops after a few minutes. Besides the known com-
pounds 1, 2a, and 3, a new species 7 is detected in the
31P{1H}-NMR spectrum (Scheme 3).
The yield of product formation, determined by inte-
gration of the 31P-signals, amounts to ca. 40%. From a
benzene solution of the product mixture, colorless
crystals of 7 formed, which were suitable for an X-ray
structure analysis.
The unexpected outcome of the reaction between 2a
and P(OR)3 poses a question as to the mechanism of
this transformation. Arbuzov-like [17] phosphite
dealkylations of P(OR)3 ligands have been reported
before in the literature [18]. New, however, is the
occurrence of phosphonic acid derivatives. NMR stud-
ies did not provide information about the fate of the
cleaved methyl group, nor could the origin of the
P-bounded proton be clarified as of yet. Traces of
water as responsible impurities seem to be improbable
since the reaction proceeds in a selective and repro-
ducible manner, independent of the choice of solvent.
Experiments in deuterated solvents also rule out the
possibility that solvent molecules might act as proton
donors.
The molecular geometry of cation 7 is displayed in
Fig. 2. The dinuclear system comprises two
[W(CMes)(CO)2(dppe)]+ units, bridged by a chloride
anion.

102
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
2.3. Synthesis and reacti6ity of
The W–Cl distances amount to 261.7(6) and 263.1(6)
[W(CMes)(CO)(dppep)]+ complexes
ppm. These bonds are rather long when compared to
the average [W]–(mCl) separation of 251 ppm [15]. The
large tungsten–chloride distances indicate a weak W–
Cl linkage, which is easily cleaved. It is not possible to
detect 7 in solution without the presence of 1 and 2a.
The two tungsten moieties are not equivalent in the
crystal; a fact that is suggested by the slight difference
in the W–Cl lengths, and which is clearly seen when the
tungsten–carbyne triple bonds are compared. The two
[W]ꢀCR bonds of 188(2) and 176.9(19) pm differ by
more than 10 pm. Also, the four [W]–CO) distances
vary from 202(2) to 185(5) pm. The latter value is
extremely short for a tungsten carbonyl bond, in com-
parison with average values of 202 and 200 pm for
prototypical tungsten carbyne chlorides (vide infra). We
will at a later point return to the molecular structure of
7. One should however keep in mind that the poor
quality of the crystals led to high R-values during the
course of the structure determination, and that there-
fore the geometric parameters should be looked at with
caution.
2.3.1. Preparation of [W(CMes)(CO)(dppep)]+
A second carbyne tungsten chloride with chelating
phosphorus donors prepared earlier [3] is trans-
W(CMes)(CO)2Cl(dppep) 8. When this complex is re-
acted with either AgSbF6 or AgOTf, a precipitate of
AgCl is observed after a few minutes. In the IR spec-
trum, a new w(CO) band appears, which in both cases
is shifted by 90 cm−1 to higher wavenumbers. The
spectroscopic data reveal the presence of the naked
cation [W(CMes)(CO)(dppep)]+, which can be isolated
either as the hexafluoroantimonate salt 9a, or as the
triflate salt 9b (Scheme 4).
In contrast to compound 2b, the OTf anion does not
coordinate to the tungsten center in cation 9b. Besides
the identical band for the w(CO) stretch in the IR
spectra of 9a and 9b, further evidence is furnished by
the w(SO) stretches at 1262 and 1032 cm−1, which
correspond to the literature values for the uncoordi-
nated OTf ion [13] (vide supra). Furthermore, reso-
nance of −78.7 ppm appears in 19F-NMR spectrum,
indicating the presence of free triflate. No new signals
can be seen when sodium triflate is added to the NMR
solution of 9b. The five phenyl groups of the dppep
introduce steric encumbrance, so that the free coordina-
tion site in the cation 9a or 9b is protected from
coordination of sterically more demanding ligands.
Neither complex 9a nor 9b could be isolated in the
purity required to obtain satisfactory elemental analy-
sis. The complexes decompose, and in the reaction
mixture, a doubly carbonyl substituted cation can be
identified as a new compound.
An asymmetric coordination around the bridging
Cl− is clearly supported by the 31P{1H}-NMR spec-
trum. At low temperatures, two singlets are detected
1
with JPW couplings around 230 Hz. The resonance at
40.7 ppm can be assigned to a complex with a short
W–Cl bond, and is comparable to the value of 40.3
ppm for 1a. The other signal appears at 44.7 ppm. The
observed low field shift indicates a more cationic metal
center, and a longer W–Cl separation. At temperatures
around 0°C, coalescence sets in, and at room tempera-
ture, the two different sets of dppe ligands are equiva-
lent on the NMR time scale. The fluctional behavior of
the complex leads to one single 31P-resonance at 42.5
ppm.
2.3.2. Reactions of [W(CMes)(CO)(dppep)]+
We conclude our experimental section with a concise
report on exemplary reactivity studies that have been
undertaken with compound 9a. When the reaction be-
tween 8 and AgSbF6 is carried out in the presence of
carbon monoxide, complex 10 containing two carbonyl
ligands is obtained (Scheme 4). The presence of two
intense w(CO) stretches reveal the cis-coordination of
the CO ligands. In the 31P{1H}-NMR spectrum, three
signals for the chemically non-equivalent P-nuclei are
The preparation of 7 under an atmosphere of 13CO
provides further evidence for facile W-Cl bond break-
age. All detected products 1, 2a 3, and 7 contain labeled
carbonyl ligands. The presence of 13C-1 might be ex-
plained by the reaction of 13C-2 with excess 1 to form
the labeled compound 13C-7. Dissociation of the dinu-
clear complex might then lead to 13C-enriched starting
material.
Scheme 4.

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
103
coupling constants with a small value of 75 Hz is
indicative of a P-nucleus in trans position to the car-
byne group. The spectroscopic data suggest that the
dppep ligand changes is coordination mode from a
meridional to facial arrangement. The driving force is
once again the pronounced trans influence of the car-
byne moiety, which prevents a coordination of a strong
p-acceptor trans to the CR ligand. The coordination of
CO stabilizes the 16e− cation 9a, and compound 10 can
be obtained as analytically pure material.
In contrast to 2a, the cation 9a is fairly inert. Reac-
tions with donor ligands like NCCH3 or P(OMe)3 are
too slow to allow for an isolation of possible reaction
products from by-products and intermediates. As we
have seen in the case of the triflate ion, the free
coordination site is effectively shielded by the sterically
demanding dppep ligand, and only small ligands like
CO or Cl− are able to coordinate to the tungsten center
in 9a.
Scheme 5.
2.4. Bonding characteristics of [W(CR)(CO)L(PR3)2]+
complexes
To gain a deeper understanding of the complex for-
mation chemistry of cation 2a, calculations were per-
formed on model compounds derived from the
transition metal fragment W(CH)(CO)2(PH3)2, [W]. The
computational work parallels the reactivity studies un-
dertaken with the novel tungsten cation, and it is
sought to identify characteristic bonding properties,
which might enable one to adjust and tune the course
of conceivable catalytic conversions. Besides the tung-
sten carbyne chloride [W]–Cl (I), the 16e− cation [W]+
(II+), and the phosphonic ester complex [W]–
OPH(OMe)2 (III+), a variety of donor adducts {[W]–
L}+ have been analyzed (L=PH3 (IIa+); PMe3 (IIb+);
P(OH)3 (IIc+); P(OMe)3 (IId+); CO (IIe+); NCMe
(IIf+); THF (IIg+); CH2Cl2 (IIh+); OPH(OH)2 (IIi+))
(Scheme 5).
The computational study is further extended to in-
clude the bimetallic model compound {[W]–Cl–[W]}+
(IV+).
2.4.1. Structure of selected model complexes
Two tungsten carbyne chloride complexes containing
the dppe ligand are described structurally in the litera-
ture, namely trans-W(C(CCHR))Cl(CO)2(dppe) (R=
)
[7] and trans-W(C(p-C6H4NH2))Cl(CO)2(dppe)
Fig. 3. Comparison of optimized model geometries I, II+ and III+
with experimental data of related compounds (in parentheses; aver-
aged values). For I, values are taken from Refs. [7,19]. Distances are
given in pm, angles in degrees.
,
[19]. In Fig. 3, we compare the crystallographic data of
these compounds with the optimized structure of I. The
coordination geometry around the tungsten center is
well reproduced in the simplified model system. The
deviations between measured and calculated W–L
bond lengths are around 2 pm.
seen, which are split into doublets of a doublet with
2JPP values smaller than 20 Hz, as can be expected for
The structure of the [W]+ fragment II+, a model for
the cation 2a, is of square pyramidal coordination
1
P-donor ligands in cis position [11]. One of the JPW

104
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
take place in the bonding between a PR3 donor with a
16e−–ML5 transition metal fragment [22].
geometry; the most significant change in the structure,
when comparing to I, is a shortening of the WꢀCH
bond by 5 pm. We note further a slight contraction of
the calculated d(CO) bond length, which is consistent
with the high-field shift of the w˜(CO) stretching fre-
quencies observed for 2a. Also shown in Fig. 3 is a
comparison of the solid state structure of 6a with the
optimized geometry of III+. The calculated values for
d(W–CR), d(W–CO) and d(W–P) are in fair agree-
ment with the experimental data, whereas the W–O
distance in III+ is significantly longer by 7 pm com-
pared to the cation in 6a. The PꢁO double bond of the
phosphonate ester elongates under coordination to the
metal fragment, the distance d(P–O) for optimized free
ligand amounts to 150 pm. The experimentally deter-
mined value for d(P–O) is slightly larger than that of
free phosphonates (d(P–O)=146 [20] and 145 pm [21]).
Selected geometric parameters for various {[W]–L}+
complexes are collected in Table 1.
Inspection of the data reveals the fact that the coor-
dination geometry of the phosphorus and the carbonyl
framework is essentially the same for all complexes
investigated. The W–CH distances vary between 181
and 184 pm. Both strong s donors as well as p accep-
tors trans to the carbyne unit cause an elongation of the
WꢀCH bond, since both electrostatic effects as well as
p-backdonation determine the tungsten–carbon triple
bond strength. The phosphorus systems IIa+ –IId+ all
The coordination of the ligand Lt trans to the car-
byne group is of special interest. All systems, in which
Lt is a phosphorus donor, show a prolonged W–P
distance, a manifestation of the strong trans influence
of the carbyne unit [3]. In complex IId+, the weak
bond of the COt group is indicated by the long W–
(CO) distance, and the value of d(CꢀO), which is
identical to that optimized for the free ligand, implies
that p-backbonding is of only minor importance. In the
next section, we investigate the trends in bonding,
which are suggested by the optimized geometries, in
more detail.
2.4.2. Bond analysis for selected model compounds
We analyze the energy changes associated with the
bond forming reaction as shown in Scheme 3, in terms
of steric contributions DE0 and orbital interactions
DEint [22,23]. The latter describe s-donation and p-
backdonation, whereas the former contain terms due to
Pauli repulsion, as well as an electrostatic component
DEelstat. The two terms can be combined to give the
ligand association energy DELA, which is related to the
bond snapping energy BEsnap,: DELA=DE0+DEint
=
−BEsnap. This value does not directly correlate with the
reaction enthalpy at 0 K, since — besides zero point
energy corrections —, a preparation energy term has
to be taken into account. It is however related to
metal–ligand bond enthalpy terms, which provide a
better description for bond strengths than bond dissoci-
ation enthalpies do [24]. The bond energy BE is ob-
show deviations from linearity in the angles Ú(CCH
–
W–L) and Ú(H–CCH–W), which is most prominent
for the trialkyl phosphine. This effect can be rational-
ized on the basis of particular orbital interactions that
Table 1
Selected optimized bond distances (pm) and bond angles (°) bond energy terms (kJ mol−1) and Hirshfeld charges (a.u.) for {[W]–L}+ systems
([W]=[W(CH)(CO)2(PH3)2])
IIa+
IIb+
IIc+
IId+
IIe+ a
IIf+
IIg+
IIh+
IIi+
III+
L
PH3
Pme3
P(OH)3
P(OMe)3
CO
NCMe
THF
CH2Cl2
OPH(OH)2 OPH(OMe)2
b
d(W–L)
d(W–CCH
d(W–CCO
274
266
184
203
252
168
171
−341
30
−241
211
182
266
183
205
253
175
178
−251
18
−178
160
140
267
183
204
253
172
174
−283
35
−207
172
140
228
184
205
254
179
179
−207
74
227
183
205
253
179
180
−245
−1
−151
152
145
236
182
205
254
176
179
−183
−9
−118
127
116
276
181
205
254
179
180
−94
32
234
181
205
254
179
179
−180
−25
−120
145
b
)
)
183
205
253
b
b
d(W–P)
Ú(Ct–W–L) 175
179
179
−202
−31
−134
165
Ú(H–C–W)
DEelstat
DE0
DEint
BEsnap
BE
179
−188
35
−152
117
−172
98
−101
69
107
88
58
130
147
qH([W])
qH(L)
0.87
0.13
0.77
0.23
0.89
0.11
0.83
0.17
1.01
−0.01
0.90
0.10
0.86
0.14
0.88
0.12
0.87
0.13
0.85
0.15
a d(C–O)CO=115 pm, d(C–O)L=114 pm.
b See Fig. 3.

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
105
and III+, the orbital interaction terms are comparable,
and it is again the influence of the electrostatic interac-
tion term that in most of the cases, except for IIh+,
causes a bonding DE0 contribution, and a fairly stable
metal to ligand bond. The weak W–L bond of the
non-coordinating solvent ligand is caused not by a
comparatively small orbital interaction term, but is due
to the steric repulsion. In the OPH(OR)2 systems, the
substitution of hydrogen by methyl groups leads to
Fig. 4. Calculated structure of the cation [{W(CH)(CO)2(PH3)2}2(m-
Cl)]+ (IV+). Distances and angles are reported in pm and degrees,
respectively.
increase in W–L bond strength of about 20 kJ mol−1
.
The fact that, in terms of BE, the W–L linkage is
somewhat stronger in III+ than in IId+, might con-
tribute to the driving force for the Arbuzov dealkyla-
tion as observed in the experiment.
tained when BEsnap is corrected by the preparation
energy, BE=BEsnap−DEprep. It is further instructive to
analyze the charge distribution on the interacting frag-
ments. The Hirshfeld charges [25] qH provide a measure
for the charge flow as it occurs during bond formation.
The results of our analysis are collected in Table 1.
Among the systems where L is a phosphorus donor,
IIa+ –IId+, the trialkyl substituted donor undergoes by
far the strongest interaction with the metal center, both
in terms of electrostatic as well as orbital interactions.
This can also be seen in the charge distribution which
show the largest amount of charge transfer for IIb+. Of
two phosphite ligands, the P(OMe)3 molecule posses
the higher W–L bond strength, mainly caused by an
increase in the orbital interaction term. However, due
to the difference in −DEprep, the bond energy terms BE
are similar for both P(OMe)3 and P(OH)3.
2.4.3. The {[W]–Cl–[W]}+ cation
In Fig. 4, the optimized structure of the bimetallic
cation IV+ is displayed. The geometry of the two [W]
fragments is similar to those found for the complexes
described before. The molecule possesses C26 symmetry;
calculations in which this symmetry constrained is lifted
did not result in structures with significantly varying
WꢀC or W–CO bond lengths, as observed in the
crystal structure. The large differences in the distances
W(1)–C(1) and W(1)–C(2) or W(1)–C(5) and W(2)–
C(6) observed for 7 (compare Fig. 2), are probably due
to the poor quality of the crystal used for the structure
determination (see experimental section).
For the geometry around the bridging m-Cl ligand,
we find W–Cl distances of 268 pm, which is 6 pm
longer than in the crystal structure, and an angle
Ú(Cl–W–Cl) of 134°, which is opened up by 8° com-
pared to 7. We have found similar discrepancies be-
tween calculated and observed M–Cl distances in
cationic {[Zr]–Cl–[Zr]}+ systems [26]. A theoretical
analysis revealed that the potential surface for internal
coordinates involving the m–Cl group is very shallow,
and that external effects like the presence of counteri-
ons may have a profound influence on the bridging
geometry.
When we compare the two system with Lt ligands
able of p-backbonding, we find in the case of CO a
weak W–L bond, although the orbital interaction term
is fairly large, and comparable to that of some phos-
phorus donors. The large destabilizing DE0 contribu-
tion substantially weakens the metal carbonyl bond.
We further observe that no net charge low takes place
in the formation of IIe+. The situation is different for
the acetonitrile ligand. The electrostatic contribution to
the W–L bond removes the antibonding character from
the steric term DE0, and as a consequence leads to the
formation of a stable W–(NCMe) bond. The p contri-
butions for both ligands amount to about 35% of the
term DEint, and do not represent the dominant orbital
For model compound IV+, we also find very soft
W–Cl stretching, and W–Cl–W bending modes. A
geometric arrangement with differences in W–Cl bond
lengths of 2 pm (d(W–Cl)=267, 269 pm) is less than 1
kJ mol−1 higher in energy. The distortion of the opti-
mized structure to adapt a m-Cl geometry as found in
the crystal requires only 6 kJ mol−1. Introduction of a
negative point charge q−, being 700 pm away from one
tungsten center, and located in the W–Cl–W mirror
plane, the W–Cl and W-q− vectors being perpendicu-
lar, reduces the symmetry of the molecular system to
Cs, and causes an elongation of the adjacent W–Cl
bond by 2 pm, but a contraction of the distant W–Cl
bond by 6 pm. The asymmetric m-Cl coordination of
the cation 7+ as deduced from the NMR spectra might
therefore be caused by the presence of outer ion pairs in
solution.
interaction.
t
–
The weak W (CO ) bond provides a rationale for the
fact that in 3 a phosphorus donor occupies the coordi-
nation site trans to the carbyne group. We have calcu-
lated that the isomerization mer-IId+ “fac-IId+ is
energetically favorable by 34 kJ mol−1. The d(CO)
values in fac-IId+ are 207 and 209 pm; about 2–4 pm
longer than the average calculated d(CO) distance
(compare the entries in Table 1). Both changes in DEint
and DE0 contribute to the stabilization energy, the
changes in the steric term being the more important
ones.
Lastly, we analyze the complexes containing ligand
capable of p-donation. For all complexes IIg+ –IIi+

106
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
The analysis for the bond forming reaction [W]–
the remanining silver salts were filtered off through
Celite. The clear, yellowish–red solution was evapo-
rated in vacuo, the yellow residue dissolved in CH2Cl2,
and precipitated with hexane. After several precipita-
tions, 2a (1.13 g, 1.13 mmol, 91%) was obtained as
yellow powder. IR (KBr, cm−1): 2000 (s), 1931 (s).
1H-NMR (300 MHz, CD2Cl2, ppm): 7.8–7.55 (m, 8H,
Ph), 7.40–7.10 (m, 12H, Ph), 6.83 (s, 2H, Mes), 3.10–
2.80 (m, 4H, PCH2), 2.13 (s, 6H, CH3-Mes), 2.11 (s,
3H, CH3-Mes). 13C{1H}-NMR (75.4 MHz, CD2Cl2,
Cl+[W]+“{[W]–Cl–[W]}+ reveals a strong W–Cl
bond with a value for BEsnap of 164 kJ mol−1. The
bond energy BE reduces to 147 kJ mol−1, but is still
comparable to that of systems with phosphorus donors.
The Hirshfeld charges of qH([W])=0.81 and qH(L)=
0.19 a.u. indicate that charge delocalization is a stabiliz-
ing factor in bond formation. This fact is also reflected
in the qH values for the tungsten atoms in I, II+ and
IV+, which amount to 0.10, 0.23, and 0.11 a.u.,
respectively.
2
ppm): 217.5 (dd, JCP=36.3×5.0 Hz, CO), 25.8–25.0
(m, CH2–dppe), 21.7 (s, CH3–Mes), 20.5 (s, 2CH3–
Mes). Due to decompostion of 2a, the remaining car-
bon atoms could not be assigned with certainty.
13C{1H}-NMR (75.4 MHz, d8-THF, ppm): 300 (m,
3. Conclusion
2
In this work, we have described a synthetic route to
the tungsten carbyne cations [W(CMes)(CO)2(dppe)]+
and [W(CMes)(CO)(dppep)]+. Preliminary reactivity
studies have been performed in reactions with donor
ligands. A theoretical analysis rationalizes the trends
observed in theses experiments, and reveals the impor-
tance of steric effects, especially of electrostatic nature,
for the metal to ligand bonding interaction.
CMes), 219.3 (dd, JCP=37.4×7.9 Hz, CO). 31P{1H}-
1
NMR (121.5 MHz, CD2Cl2, ppm): 50.7 (s, JPW=237
Hz). MS (FAB-Neg, LM=CH2Cl2, M=NBOH, e/z,
%): 769 (M+, 100), 741 (M+−CO, 20), 713 (M+−2
CO, 18). Anal. for C38H35F6O2P2SbW (1005.23). Calc.
C (45.40), H (3.51); Found: C (45.69), H (4.15)%.
4.1.2. W(CMes)(CO)2(dppe)(OTf ) (2b)
A solution of 1 (1.12g, 1.39 mmol) in CH2Cl2 (60 ml),
was treated with a solution AgOTf (0.54 g, 2.09 mmol)
in CH2Cl2 (30 ml). Immediate precipitation of AgCl
occured. After stirring for 3 h under the exclusion of
light, the reaction mixture was evaporated in vacuo,
and the resulting yellowish–brown solid was suspended
in a small amount of benzene, followed by filtration
through celite and removal of the solvent. Recrystal-
lization from CH2Cl2/hexane afforded 2b (1.14 g, 1.24
mmol) as yellow powder. IR (KBr, cm−1): 1998 (s),
1932 (s). 1H-NMR (300 MHz, CD2Cl2, ppm): 7.85–
7.75 (m, 4H, Ph), 7.70–7.60 (m, 4H, Ph), 7.50–7.20 (m,
12H, Ph), 6.55 (s, 2H, Mes), 3.12–2.92 (m, 2H, PCH2),
2.90–2.70 (m, 2H, PCH2), 2.15 (s, 3H, CH3–Mes), 1.82
(s, 6H, CH3–Mes). 13C{1H}-NMR (75.4 MHz, CD2Cl2,
4. Experimental
4.1. Preparati6e procedures
All manipulations were performed under a nitrogen
atmosphere using standard Schlenk techniques. Sol-
vents were dried according to standard procedures, and
distilled under N2 prior to use. The complexes trans-
W(CMes)Cl(CO)2(dppe)
(1)
and
trans-
W(CMes)Cl(CO)(dppep) (8) were prepared as described
elsewhere [3]. PMe3 [27] and PEt3 [28] were synthesized
following slightly modified literature procedures [29],
whereas all other reagents were purchased and used
without further purification. The following instruments
were used for spectroscopic and physical characteriza-
tion: Varian Gemini 200 and Gemini 300 NMR spec-
ppm): 289.6 (t, 2JCP=7.9 Hz, CMes), 216.4 (dd, 2JCP
=
40.4×7.2 Hz, CO), 143.6 (s, ipso-Mes), 139.4 (s, o-
Mes), 134.3 (s, p-Mes), 128.3 (s, m-Mes), 133.4, 133.2,
132.2, 132.1, 131.7, 131.4, 130.7, 129.4, 129.3, 129.2
(C6H5), 118.7 (q, 1JCF=319.5 Hz, CF3), 25.8 (dd,
1
trometers. H and 13C chemical shifts are referred to
Me4Si and referenced through the residual proton or
13C resonances of the deuterated solvent. 31P chemical
shifts were externally referenced to 85% H3PO4. Bio-
Rad FTS-45 IR spectrometer. Finnigan/MAT 8320
mass spectrometer. LECO CHNS-932 for elemental
analyses.
2
1JCP=27.4 Hz, JCP=11.7 Hz, PCH2), 21.5 (s, CH3–
Mes), 20.2 (s, 2CH3–Mes). 19F-NMR (282.3 MHz,
CD2Cl2, ppm): −78.0 (s). 31P{1H}-NMR (121.5 MHz,
1
CD2Cl2, ppm): 46.8 (s, JPW=239 Hz). MS (FAB-Neg,
LM=CH2Cl2, M=NBOH, e/z, %): 920 (M+, 1), 769
(M+−SO3CF3, 100), 741 (M+−SO3CF3−CO, 30),
712 (M+−SO3CF3−2 CO, 10). Anal. for
C39H35F3O5P2SW (918.11). Calc. C (50.97), H (3.84);
Found: C (51.23), H (3.91)%.
4.1.1. [W(CMes)(CO)2(dppe)][SbF6] (2a)
To a solution of 1 (1g, 1.24 mmol) in CH2Cl2 (30 ml),
AgSbF6 (0.64 g, 1.86 mmol), dissolved in CH2Cl2 (30
ml), was added at room temperature (r.t.). AgCl precip-
itated immediately. It was stirred for 10 min under the
exclusion of light, and the solvent was removed in
vacuo. The solid was suspended in benzene (60 ml), and
4.1.3. [W(CMes)(CO)3(dppe)][SbF6] (3)
To a solution of 1 (2.01g, 2.50 mmol) in CH2Cl2 (60
ml), AgSbF6 (1.28 g, 3.75 mmol), dissolved CH2Cl2 (30

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
107
ml), was added at r.t. under an atmosphere of carbon
monoxide. Stirring for 30 min under the exclusion of
light was followed by evaporation of the solvent. The
residue was suspended in benzene and filtered through
Celite. The clear, yellow filtrate was evaporated in
vacuo, and the solid washed with the hexane. Com-
pound 3 (2.46 g, 2.38 mmol, 95%) was isolated as
yellow powder. IR (KBr, cm−1): 2074 (w), 1993 (s),
1922 (s). 1H-NMR (300 MHz, CD2Cl2, ppm): 7.55–
7.40 (m, 20H, Ph), 6.80 (s, 2H, Mes), 3.10–2.80 (m, 4H,
PCH2), 2.22 (s, 3H, CH3–Mes), 2.20 (s, 6H, CH3–
2.19 (s, 3H, CH3–Mes), 1.90 (s, 6H, CH3–Mes), 1.45
(s, 3H, CH3CN). 13C{1H}-NMR (75.4 MHz, CD2Cl2,
ppm): 296.0 (t, 2JCP=8.4 Hz, CMes), 214.9 (dd, 2JCP
=
32.6×7.4 Hz, CO), 143.2 (s, ipso-Mes), 142.0 (s, o-
Mes), 141.2 (s, p-Mes), 128.6 (s, m-Mes), 133.5, 133.2,
131.9, 131.6, 131.4, 131.2, 130.0, 129.8, 129.6, 129.4,
1
2
(C6H5, CH3CN), 26.1 (dd, JCP=29.4 Hz, JCP=10.9
Hz, PCH2), 21.7 (s, CH3-Mes), 20.5 (s, 2CH3-Mes), 2.1
(s, CH3CN). 31P{1H}-NMR (121.5 MHz, C6D6, ppm):
1
46.3 (s, JPW=232 Hz). MS (FAB-Neg, LM=CH2Cl2,
M=NBOH, e/z, %): 810 (M+, 1), 770 (M+−CH3CN,
100), 741 (M+−CH3CN−CO, 15), 713 (M+−
CH3CN−2CO, 8). Anal. for C40H38NF6O2P2SbW·
CH2Cl2 (1046.28): Calc. C (43.53), H (3.56), N (1.24);
Found: C (43.63), H (4.17), N (0.87)%.
Mes). 13C{1H}-NMR (75.4 MHz, CD2Cl2, ppm): 308.8
2
(dd, 2JCP=20.1×7.8 Hz, CMes), 205.0 (dd, JCP
=
21.9×8.3 Hz, CO), 201.5 (t, 2JCP=7.3 Hz, 2CO), 145.1
(s, ipso-Mes), 144.1 (s, o-Mes), 133.7 (s, p-Mes), 129.6
(s, m-Mes), 132.3, 132.2, 132.1, 131.9, 130.6, 130.4,
1
2
130.1, 129.9 (C6H5), 29.7 (dd, JCP=30.1 Hz, JCP
=
4.1.6. [W(CMes)(CO)2(PMe3)(dppe)][SbF6] (5a)
1
2
13.3 Hz, PCH2), 26.6 (dd, JCP=26.1 Hz, JCP=9.8
Hz, PCH2), 22.2 (s, CH3–Mes), 21.0 (s, 2CH3–Mes).
31P{1H}-NMR (121.5 MHz, CD2Cl2, ppm): 37.9 (s, 1P,
At −20°C, PMe3 (83 ml, 0.76 mmol) was added to a
solution of 2a ( 0.63 g, 0.63 mmol) in toluene (40 ml).
The reaction mixture was allowed to warmed up to r.t.
for 1 h. After stirring for 30 min, the solvent was
removed in vacuo, and the resulting solid washed with
hexane (5 ml). 5a (0.51 g, 0.47 mmol, 75%) is isolated as
dark-yellow powder by precipitation with hexane from
a solution of the raw product in CH2Cl2. Repeated
precipitation yielded analytically pure material. IR
1JPW=237 Hz), 23.9 (s, 1P, JPW=91 Hz). MS (FAB-
1
Neg, LM=CH2Cl2, M=NBOH, e/z, %): 797 (M+,
50), 769 (M+−CO, 60), 741 (M+−2CO, 20), 713
(M+−3CO, 13). Anal. for C39H35F6O3P2SbW·2C6H6
(1033.25). Calc. C (51.50), H (3.98); Found: C (51.34),
H (4.12)%.
1
(KBr, cm−1): 1992 (s), 1921 (s). H-NMR (300 MHz,
4.1.4. [W(CMes)(OC4D8)(CO)2(dppe)][SbF6] (4a)
C6D6, ppm): 7.90–7.80 (m, 4H, Ph), 7.75–7.65 (m, 4H,
Ph), 7.30–7.20 (m, 4H, Ph), 7.15–6.90 (m, 4H, Ph),
6.85–6.75 (m, 4H, Ph), 6.36 (m, 2H, Mes), 3.5–3.25 (m,
2H, PCH2), 3.05–2.80 (m, 2H, PCH2), 1.98 (s, 3H,
CH3–Mes), 1.82 (s, 6H, CH3–Mes), 0.29 (d, 9H,
PMe3). 13C{1H}-NMR (50.3 MHz, CD2Cl2, ppm):
In an NMR tube, 2a (82 mg, 81.8 mmol) was dis-
solved in d8-THF (0.5 ml). After a few minutes at r.t.,
a stable distribution of compounds 2a, 3 and 4a was
obtained. The ratio of these complexes as determined
by 31P{1H}-NMR spectroscopy was 3:1:2. 4a could not
be isolated by precipitation, crystallization, or evapora-
tion of the solvent, and was therefore characterized by
2
300.8 (m, CMes), 214.6 (dt, JCP=26.1×7.3 Hz, CO),
143.9 (s br, ipso-Mes), 142.6 (s, o-Mes), 135.1 (s br,
p-Mes), 129.1 (s, m-Mes), 133.7, 133.2, 132.3, 132.0,
131.8, 131.4, 130.6, 130.0, 129.5, 129.2 (C6H5), 25.4 (dd,
1
NMR spectroscopy of the reaction mixture. H-NMR
(300 MHz, d8-THF, ppm): 6.53 (s, 2H, Mes), 2.05 (s,
3H, CH3-Mes), 1.77 (s, 6H, CH3–Mes). 13C{1H}-NMR
2
1JCP=30.0 Hz, JCP=11.9 Hz, PCH2), 21.9 (s, CH3–
2
(75.4 MHz, d8-THF, ppm): 218.5 (dd, JCP=38.2×7.2
Mes), 20.4 (s, 2CH3–Mes), 17.1 (m br, PMe3). 31P{1H}-
NMR (121.5 MHz, C6D6, ppm): 38.6 (d, 2P, 2JPP=22.3
Hz, 1JPW=227 Hz), −58.9 (t, 1P, 2JPP=22.3 Hz,
1JPW=83 Hz). MS (FAB-Neg, LM=CH2Cl2, M=
NBOH, e/z, %): 845 (M+, 20), 818 (M+−CO, 10), 795
(M+−2CO, 60), 769 (M+−PMe3, 80), 742 (M+−
PMe3−CO), 30), 714 (M+−PMe3−2CO, 20).
C41H44F6O2P3SbW (1081.32): Ber. C (45.54), H (4.10);
Gef. C (45.56), H (4.12)%.
Hz, CO), 70.2 (quint, OCD2CD2), 26.4 (quint,
OCD2CD2).31P{1H}-NMR (121.5 MHz, d8-THF, ppm):
1
47.1 (s, JPW=235 Hz).
4.1.5. [W(CMes)(CH3CN)(CO)2(dppe)][SbF6] (4b)
A solution of 2a (0.5 g, 0.62 mmol) in CH2Cl2 (15 ml)
and CH3CN (5 ml) was treated with a solution of
AgSbF6 (0.28 g, 0.81 mmol) in CH2Cl2 (15 ml). After
stirring for 2 h under exclusion of light, the resulting
yellow suspension was evaporated in vacuo. The dark-
yellow solid was suspended in benzene (35 ml) and
filtered through Celite. The solvent was removed, and
the raw product recrystallized from CH2Cl2/hexane. 4b
(0.61 g, 0.58 mmol, 94%) was obtained as yellow–or-
ange powder. IR (KBr, cm−1): 2005 (s), 1935 (s).
1H-NMR (300 MHz, CD2Cl2, ppm): 7.90–7.20 (m,
20H, Ph), 6.63 (s, 2H, Mes), 3.30–2.80 (m, 4H, PCH2),
4.1.7. [W(CMes)(CO)2(PEt3)(dppe)][SbF6] (5b)
As described for 5a, 2a (0.82 mg, 0.82 mmol) was
reacted with PEt3 (181 ml, 1.23 mmol) in toluene (60 ml)
at −20°C. 5b was obtained in 68% yield as dark-yellow
1
powder. IR (KBr, cm−1): 1990 (s), 1920 (s). H-NMR
(300 MHz, C6D6, ppm): 7.92–7.62 (m, 8H, Ph), 7.32–
7.22 (m, 4H, Ph), 7.15–6.95 (m, 4H, Ph), 6.85–6.73 (m,

108
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
1
4H, Ph), 6.30 (m, 2H, Mes), 1.97 (s, 3H, CH3–Mes),
1.89 (m, 12H, PCH2CH3), 1.86 (s, 6H, CH3–Mes), 0.73
(m, 18H, PCH2CH3). 13C{1H}-NMR (75.4 MHz,
Ph), 6.65 (s, 0.5H, OPH(OEt)2, JHP=748 Hz), 6.38 (s,
1
2H, Mes), 4.16 (s, 0.5H, OPH(OEt)2, JHP=748 Hz),
3.25–3.13 (m, 4H, OPH(OCH2CH3)2), 2.75–2.50 (m br,
2H, PCH2), 2.38–2.20 (m br, 2H, PCH2), 2.10 (s, 3H,
CH3–Mes), 1.93 (s, 6H, CH3–Mes), 0.79 (t, 6H,
2
CD2Cl2, ppm): 301.6 (dt, JCP=21.3×8.8 Hz, CMes),
215.0 (dt, 2JCP=26.8×7.9 Hz, CO), 143.0 (s, ipso-
Mes), 138.2 (s, o-Mes), 135.0 (s, p-Mes), 128.5 (s,
m-Mes), 133.7, 133.2, 132.3, 132.0, 131.8, 131.4, 130.6,
130.0, 129.5, 129.2 (C6H5), 25.9 (dd, 1JCP=28.3 Hz,
2JCP=11.9 Hz, PCH2), 21.8 (s, CH3–Mes), 20.9 (s,
2CH3–Mes), 21.0 (m, PCH2CH3), 8.0 (s br,
PCH2CH3). 31P{1H}-NMR (121.5 MHz, C6D6, ppm):
OPH(OCH2CH3)2). 13C{1H}-NMR (75.4 MHz, d8-
2
Toluol, ppm): 289.2 (m, CMes), 218.1 (dd, JCP
=
40.3×7.2 Hz, CO), 143.9 (s, ipso-Mes), 141.6 (s,
o-Mes), 139.3 (s, p-Mes), 133.6, 133.4, 132.6, 132.4,
130.8, 130.1, 129.3, 129.0, 128.9, 128.7, 128.5 (C6H5),
128.7 (s, m-Mes), 64.5 (d, 2JCP=5.8 Hz,
2
1
2
38.1 (d, 2P, JPP=21.5 Hz, JPW=226 Hz), −26.2 (t,
OPH(OCH2CH3)2), 24.9 (dd, 1JCP=28.6 Hz, JCP
=
2
1
1P, JPP=21.5 Hz, JPW=84 Hz).
11.4 Hz, PCH2), 21.8 (s, 1CH3–Mes), 20.7 (s, 2CH3–
Mes) 15.6 (d, 3JCP=7.3 Hz, OPH(OCH2CH3)2).
31P{1H}-NMR (121.5 MHz, C6D6, ppm): 48.7 (d, 2P,
4.1.8. [W(CMes)(CO)2(OPH(OMe)2)(dppe)][SbF6] (6a)
Following the procedure for 5a, 2a (0.56 g, 0.56
mmol) was reacted with P(OMe)3 (79 ml, 0.67 mmol) in
toluene (20 ml). Recrystallization from CH2Cl2/hexane
afforded 6a (0.57 g, 0.48 mmol, 86%) as yellow, crys-
talline product. IR (KBr, cm−1): 2000 (s), 1929 (s),
1050 (m, PO). 1H-NMR (300 MHz, d8-THF, ppm):
2JPP=6.3 Hz, JPW=236 Hz, dppe), 15.4 (t, 1P, JPP
=
1
2
6.3 Hz, OPH(OEt)2).
4.1.10. [{W(CMes)(CO)2(dppe)}2Cl][SbF6] (7)
AgSbF6 (0.26 g, 0.75 mmol) was added to a solution
of 2a (1.2 g, 1.49 mmol) in CH2Cl2. AgCl started to
precipitate, and after 5min, the solvent was removed in
vacuo. The resulting yellowish–grey solid was sus-
pendend in benzene (20 ml), and filtered through Celite.
Precipitation with hexane afforded a yellow solid; a
mixture of the compounds 1, 2a, 3 and 7. Compound 7
crystallizes in large, colorless crystals from a benzene
solution of the product mixture. It was not possible to
obtain a pure solution of 5. The crystals are extremely
air sensitive, and decompose even under N2 in the
absence of benzene. IR (C6H6, cm−1): 2007 (s), 1940
7.90–7.80 (m, 4H, Ph), 7.82–7.73 (m, 4H, Ph), 7.51–
1
7.20 (m, 12H, Ph), 6.65 (s, 0.5H, OPH(OMe)2, JHP
=
752 Hz), 6.60 (s, 2H, Mes), 4.13 (s, 0.5H,
1
2
OPH(OCH3)2, JHP=752 Hz), 3.18 (d, JHP=12.6 Hz,
6H, OPH(OMe)2), 3.10–2.90 (m br, 2H, PCH2), 2.38–
2.20 (m br, 2H, PCH2), 2.14 (s, 3H, CH3–Mes), 1.86 (s,
6H, CH3–Mes). 13C{1H}-NMR (75.4 MHz, CD2Cl2,
ppm): 272.1 (t, 2JCP=9.2 Hz, CMes), 216.1 (dd, 2JCP
=
43.2×6.6 Hz, CO), 144.1 (s, ipso-Mes), 138.2 (s, o-
Mes), 134.7 (s, p-Mes), 133.6, 133.4, 132.6, 132.4, 130.8,
130.1, 129.3, 129.0, 128.9, 128.7, 128,5 (C6H5), 128.2 (s,
m-Mes), 54.3, 53.8, 53.3 (s, OPH(OMe)2), 25.9 (dd,
1
(s). H-NMR (300 MHz, CD2Cl2, ppm): 6.47 (s, 2H,
Mes), 2.11 (s, 6H, CH3–Mes), 1.79 (s, 3H, CH3–Mes).
13C{1H}-NMR (75 MHz, CD2Cl2, ppm): 216.7 (dd,
2JCP=38.9×6.9 Hz, CO) The resonances of the re-
maining carbon atoms could not be assigned with cer-
tainty. 31P{1H}-NMR (121.5 MHz, d8-Toluol, 45°C,
2
1JCP=27.1 Hz, JCP=12.4 Hz, PCH2), 21.5 (s, 1CH3-
Mes), 20.3 (s, 2CH3–Mes). 31P{1H}-NMR (121.5 MHz,
2
1
C6D6, ppm): 49.0 (d, 2P, JPP=5.6 Hz, JPW=236 Hz,
dppe), 19.1 (t, 1P, 2JPP=5.6 Hz, OPH(OMe)2). 31P-
NMR (121.5 MHz, CD2Cl2, ppm): 49.2 (m br, 2P,
ppm): 42.5 (s br, JPW=230 Hz). 31P{1H}-NMR (121.5
1
1
1
dppe), 17.6 (dm, JPH=752 Hz, 1P, OPH(OMe)2). MS
MHz, d8-Toluol, −75°C, ppm): 45.3 (s br, 2P, JPW
=
1
(FAB-Neg, LM=CH2Cl2, M=NBOH, e/z, %): 879
(M+, 37), 769 (M+−OPH(OMe)2, 100), 741 (M+−
OPH(OMe)2−CO, 80), 712 (M+−OPH(OMe)2−
2CO, 25). Due to varying amounts of CH2Cl2
inclusions, it was not possible to obtain a reproducible
elemental analysis.
232 Hz), 40.8 (s br, 2P, JPW=228 Hz). 31P{1H}-NMR
1
(121.5 MHz, CD2Cl2, 25°C, ppm): 42.1 (s, JPW=231
Hz).
4.1.11. [W(CMes)(CO)({Ph2PCH2CH2}2PPh)][SbF6]
(9a)
AgSbF6 (0.22 g, 0.63 mmol) was added to a solution
of 8 (0.48 g, 0.53 mmol) in CH2Cl2 (60 ml). Immedi-
ately, AgCl started to precipitate. After 10 min, the
reaction mixture was filtered through Celite. Addition
of hexane to the clear filtrate yielded 9a (0.48 g, 0.43
mmol, 82%) as yellow powder. IR (CH2Cl2, cm−1):
4.1.9. W(CMes)(CO)2(OPH(OEt)2)(dppe)][SbF6] (6b)
At −40°C, P(OEt)3 (125 ml, 0.72 mmol) was added
to a solution of 3 (0.62 g, 0.60 mmol) in CH2Cl2 (30
ml). The reaction mixture was allowed to warmed up to
r.t. for 1 h, and was stirred over night. Evaporation in
vacuo afforded a yellow oil, which was extracted with
ether. 6b (0.58 g (0.47 mmol, 78%) was isolated as
yellow oil. IR (KBr, cm−1): 2001 (s), 1933 (s), 1023 (s,
PO).1H-NMR (300 MHz, C6D6, ppm): 7.90–7.80 (m,
4H, Ph), 7.75–7.68 (m, 4H, Ph), 7.40–6.80 (m, 12H,
1
1992 (s). H-NMR (300 MHz, C6D6, ppm): 7.85–6.65
(m, 25H, Ph), 6.35 (s, 2H, Mes), 3.60–3.40 (m, 4H,
PCH2), 2.60–2.35 (m, 4H, PCH2), 2.20 (s, 3H, CH3–
Mes), 1.81 (s, 6H, CH3–Mes). 13C{1H}-NMR (75.4
2
MHz, CD2Cl2, ppm): 316 (m, CMes), 218.8 (td, JCP
=

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
109
4.8×31.8 Hz, CO). 31P{1H}-NMR (121.5 MHz, C6D6,
from dichloromethane with one solvent molecule per
asymmetric unit. The CH2Cl2 molecule and one
methoxy oxygen atom appear to be disordered, and the
respective Cl and O atoms were refined with the PART
instructions in SHELXL-97. Positions of H atoms were
calculated after each cycle of refinement (riding model)
except for the hydrogen at P3, which has been located
by difference electron density calculations. The posi-
tional parameters of this atom could not be refined and
hence were fixed during the refinement. Crystal data:
C41H44Cl2F6NO5P3SbW; M=1200.17; triclinic, a=
2
1
ppm): 75.3 (t, 1P, JPP=13.9 Hz, JPW=195 Hz), 38.1
2
1
(d, 2P, JPP=13.9 Hz, JPW=260 Hz).
4.1.12. [W(CMes)(CO)({Ph2PCH2CH2}2PPh)][OTf ]
(9b)
Following the procedure for 9a, the reaction of 8
(0.32 g, 0.35 mmol) with AgOTf (0.11 g, 0.42 mmol)
afforded 9b (0.33 g, 0.32 mmol, 91%) as yellow powder.
IR (CH2Cl2, cm−1): 1993 (s). 1H-NMR (300 MHz,
C6D6, ppm): 7.85–6.65 (m, 25H, Ph), 6.33 (s, 2H, Mes),
3.60–3.40 (m, 4H, PCH2), 2.60–2.35 (m, 4H, PCH2),
2.22 (s, 3H, CH3–Mes), 1.81 (s, 6H, CH3–Mes).
13C{1H}-NMR (75.4 MHz, CD2Cl2, ppm): 316 (m,
,
8.728(2), b=13.658(3), c=20.176(3) A, h=72.6(5)°,
3
,
i=79.2(5)°, k=84.41(2)°; V=2253.2(8) A ; T=
−1
(
173(1) K, space group P1, Z=2, m=3.442 mm
,
2
CMes), 218.8 (td, JCP=4.8×31.8 Hz, CO). 19F-NMR
10476 reflections measured, 9817 unique (Rint=0.0488),
semi-emperical absorption correction; completeness to
q=27°: 99.8%; refined against F2, final R indices (I\
2s(I)): R1=0.0671, wR2=0.1500; final R indices (all
data): R1=0.1086, wR2=0.1744.
(282.3 MHz, CD2Cl2, ppm): −78.7 (s). 31P{1H}-NMR
2
(121.5 MHz, C6D6, ppm): 76.0 (t, 1P, JPP=14.0 Hz,
2
1
1JPW=195 Hz), 37.8 (d, 2P, JPP=14.0 Hz, JPW=260
Hz). MS (FAB-Neg, LM=THF, M=NBOH, e/z, %):
878 (M+, 18), 850 (M+−CO, 40).
4.2.2. Complex 7
4.1.13. [W(CMes)(CO)2({Ph2PCH2CH2}2PPh)][SbF6]
(10)
The only suitable crystal for the X-ray experiment
was of low quality with inclusion material. Therefore,
the R-values are relatively high. However, the confor-
mation of the tungsten moieties, which are bridged by a
Cl atom, is well established. The counterion SbF6−
appears to be disordered. Furthermore, five and a half
benzene molecules per dimeric unit were found in dif-
ference Fourier maps. For some atoms, non-positive-
definite anisotropic displacement parameters resulted in
the refinement; which were then treated isotropically.
For two benzene ring systems, distance restraints had
to be applied. All positions of the H atoms were
calculated after each cycle of refinement (riding model).
Crystal data: C109H103ClF6O4P4SbW2; M=2239.69;
8 was reacted with 1.2 equivalents of AgSbF6 in
CH2Cl2 under 1 atm. of CO, yielding compound 10.
Analytically pure material was obtained by repeated
precipitation with hexane from a CH2Cl2 solution. IR
1
(KBr, cm−1): 1994 (s), 1928 (s). H-NMR (300 MHz,
CD2Cl2, ppm): 7.95–6.90 (m, 25H, Ph), 6.61 (s, 2H,
Mes), 3.40–2.10 (m, 8H, PCH2), 2.19 (s, 3H, CH3–
Mes), 1.90 (s, 6H, CH3–Mes). 13C{1H}-NMR (75.4
MHz, CD2Cl2, ppm): 304.7 (m, CMes), 219.8 (td,
2
2JCP=8.3×15.4 Hz, CO), 209.8 (td, JCP=7.3×25.7
Hz). 31P{1H}-NMR (121.5 MHz, C6D6, ppm): 68.7 (dd,
2
1
1P, JPP=4.5×1.6 Hz, JPW=226 Hz), 41.0 (dd, 1P,
2JPP=17.4×4.5 Hz, 1JPW=223 Hz), 21.0 (dd, 1P,
2JPP=17.4×1.6 Hz, 1JPW=75 Hz). Anal. for
C46H44F6O2P3SbW (1141.37): Calc. C (48.41), H (3.89);
Found: C (48.30), H (3.85)%.
monoclinic, a=25.072(9), b=16.749(7), c=24.482(6)
3
,
,
A, i=105.680(10)°; V=9898(6) A ; T=293(2) K,
space group P21/c, Z=4, m=2.744 mm−1, 12115
reflections measured, 11708 unique (Rint=0.1101), nu-
merical absorption correction; completeness to q=24°:
72.5%; refined against F2, final R indices (I\2s(I)):
R1=0.0987, wR2=0.2510; final R indices (all data):
R1=0.1857, wR2=0.3076.
4.2. Crystal structure determination
Measurements were performed with a Siemens (Nico-
let) R3m/V2000 diffractometer, using Mo–Ka radiation
,
(u=0.71073 A). For cell refinement and data reduction
4.3. Computational details
the program package SHELXTL-PC [30] was used, struc-
ture solution and structure refinement were accom-
plished using SHELXS [31] and SHELXL [32] software,
respectively. Molecualr graphics have been prepared
with PLATON [33].
The ADF program package [34], release 1999.02, was
used throughout this work, except for point-charge
calculations, which were done as described elsewhere
[26]. Self consistent GGA calculations have been per-
formed, based on the LDA description by Vosko et al.
[35] with gradient corrections due to Becke [36] and
Perdew [37] (BP86). Use was made of the frozen core
approximation. All elements except hydrogen have been
described by a triple z-STO basis augmented by one d
polarization function for the main group elements, or
4.2.1. Complex 5a
A yellow plate (0.20×0.20×0.08) was selected for
the structure determination. Due to the relatively small
size, and due to impurities in the crystalline material,
the R-values are higher than expected. 5a crystallizes

110
E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
[6] (a) K.R. Birdwhistell, S.J.N. Burgmayer, J.L. Templeton,
J. Am. Chem. Soc. 105 (1983) 7789. (b) K.R. Birdwhistell, T.L.
Tonker, J.L. Templeton, J. Am. Chem. Soc. 107 (1985) 4474.
[7] L. Zhang, M.P. Gamasa, J. Gimeno, R.J. Carbajo, F. Lopez-Or-
tiz, M. Lanfranchi, A. Tiripiccho, Organometallics 15 (1996)
4274.
[8] A.S. Goldman, K. Krogh-Jespersen, J. Am. Chem. Soc. 118
(1996) 12159.
[9] H. Jacobsen, H. Berke, T. Brackemeyer, T. Eisenbla¨tter, G.
Erker, R. Fro¨hlich, O. Meyer, K. Bergander, Helv. Chim. Acta
81 (1998) 1692.
by one (n+1)p function for the transition metal (ADF
database IV). H-atoms were treated with a double z
plus p basis set (ADF database III). Relativistic effects
were included using a quasi-relativistic approach [38].
This is scheme has been applied successfully before to
complexes of tungsten [39]. Cs or C26 symmetry has
been employed, except for model complexes IIg+, IIi+
and III+ (no symmetry constraints).
[10] G.L. Hillhouse, B.L. Haymore, Inorg. Chem. 26 (1987) 1876.
[11] J.G. Verkade, L.D. Quin, Phosphorous-31 NMR Spectroscopy
in Stereochemical Analysis, vol. 8, VCH, Weinheim, 1987.
[12] (a) R.V. Honeychuck, W.H. Hersh, Inorg. Chem. 28
(1989) 2869. (b) W.H. Hersh, J. Am. Chem. Soc. 107 (1985)
4599.
5. Supplementary material
Crystallographic data for the structure analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC no. 143797 and CCDC no.
143798. Copies of this information may be obtained
free of charge from The Director, CCDC, 12 Union
Road, Cambridge CB2 1EZ UK (Fax: +44-1223-
336033; e-mail: deposit@ccdc.cam.ac.uk or www: http:/
/www.ccdc.cam.ac.uk).
[13] K. Su¨nkel, U. Nagel, W. Beck, J. Organomet. Chem. 251 (1983)
227.
[14] D.H. Johnston, D.F. Shriver, Inorg. Chem. 32 (1993) 1045.
[15] A.G. Orpen, L. Brammer, F.H. Allen, O. Kennard, D.G. Wat-
son, R. Taylor, J. Chem. Soc. Dalton Trans. (1989) S1.
[16] M.M. Crutchfield, C.H. Dungan, J.H. Letcher, V. Mark, J.R.V.
Wazer, P31 Nuclear Magnetic Resonance, vol. 5, Interscience,
New York, 1967.
[17] B.A. Arbuzov, Pure Appl. Chem. 9 (1964) 307.
[18] (a) D. Veghini, H. Berke, Inorg. Chem. 35 (1996) 4770. (b) G.
Weiler, G. Huttner, L. Zsolnai, H.Z. Berke, Naturforsch. Teil B
42 (1987) 203.
[19] M.P.Y. Yu, A. Mayr, K.-K. Cheung, J. Chem. Soc. Dalton
Trans. (1998) 475.
Acknowledgements
Financial support from the Swiss National Science
Foundation (SNSF) is gratefully acknowledged.
[20] W. Saenger, M. Mikolajczyk, Chem. Ber. 106 (1973) 3519.
[21] E. Mu¨ller, H.-B. Bu¨rgi, Helv. Chim. Acta 70 (1987) 1063.
[22] (a) H.-B. Kraatz, H. Jacobsen, T. Ziegler, P.M. Boorman,
Organometallics 12 (1993) 76. (b) H. Jacobsen, T. Ziegler, Com-
ments Inorg. Chem. 17 (1995) 301.
[23] F.M. Bickelhaupt, E.J. Baerends, Rev. Comp. Ch. 15 (2000) 1.
[24] J.A.M. Martinho Simo˜es, J.L. Beauchamp, Chem. Rev. 90
(1990) 629.
References
[1] (a) H. Fischer, P. Hofmann, F.R. Kreissl, R.R. Schrock, U.
Schubert, K. Weiss, Carbyne Complexes, VCH, Weinheim, 1988.
(b) A. Mayr, H. Hoffmeister, Adv. Organomet. Chem. 32 (1991)
227.
[25] F.L. Hirshfeld, Theor. Chim. Acta 44 (1977) 129.
[26] H. Jacobsen, T. Brackemeyer, H. Berke, G. Erker, R. Fro¨hlich,
Eur. J. Inorg. Chem. (2000) 1423.
[27] W. Wolfsberger, H. Schmidbaur, Syn. React. Inorg. Met.-Org.
Chem. 4 (1974) 149.
[2] F.W. Lee, M.C.W. Chan, K.K. Cheung, C.M. Che, J.
Organomet. Chem. 552 (1998) 255.
[3] E. Bannwart, H. Jacobsen, F. Furno, H. Berke, Organometallics
19 (2000) 3605.
[4] (a) F. Furno, E. Bannwart, H. Berke, Chimia 53 (1999) 350. (b)
F. Furno, T. Fox, H.W. Schmalle, H. Berke, Organometallics 19
(2000) 3620.
[28] M.L. Luetkens, A.P. Sattelberger, H.H. Murray, J.D. Basil, J.P.
Fackler, Inorg. Synth. 29 (1990) 7.
[29] E. Bannwart, PhD Thesis, Universita¨t Zu¨rich, 1998.
[30] SHELXTL-PC, Siemens Analytical X-ray Instruments, Inc., 1990.
[31] G.M. Sheldrick, Acta Crystallogr. A46 (1990) 467.
[32] G.M. Sheldrick, SHELXL-97, Universita¨t Go¨ttingen, Germany,
1997.
[5] (a) W. Koch, M.C. Holthausen, A Chemist’s Guide to Density
Functional Theory, Wiley–VCH, Weinheim, 2000. (b) J.F. Dob-
son, G. Vignale, M.P. Das (Eds.), Electronic Density Functional
Theory: Recent Progress and New Directions, Plenum Press,
New York, 1998. (c) H. Eschrig, The Fundamentals of Density
Functional Theory, Teubner, Stuttgart, 1996. (d) J.M. Seminario
(Ed.), Recent Developments and Applications of Modern Den-
sity Functional Theory, Elsevier, Amsterdam, 1996. (e) B.B.
Laird, R.B. Ross, T. Ziegler (Eds.), Chemical Applications of
Density Functional Theory, ACS Symposium Series 629, Wash-
ington, DC, 1996. (f) E.K.U. Gross, R.M. Dreizler (Eds.),
Density Functional Theory, Plenum, New York, 1995. (g) N.H.
March, Electron Density Theory of Atoms and Molecules, Aca-
demic Press, New York, 1992. (h) J. Labanowski, J. Andzelm
(Eds.), Density Functional Methods in Chemistry, Springer,
Heidelberg, 1991. (i) R.G. Parr, W. Yang, Density Functional
Theory of Atoms and Molecules, Oxford University Press, Ox-
ford, 1989.
[33] A.L. Spek, Acta Crystallogr. A46 (1990) C34.
[34] (a) E.J. Baerends, A. Be´rces, C. Bo, P.M. Boerrigter, L. Cavallo,
L. Deng, R.M. Dickson, D.E. Ellis, L. Fan, T.H. Fischer, C.
Fonseca Guerra, S.J.A. van Gisbergen, J.A. Groeneveld, O.V.
Gritsenko, F.E. Harris, P. van den Hoek, H. Jacobsen, G. van
Kessel, F. Kootstra, E. van Lenthe, V.P. Osinga, P.H.T.
Philipsen, D. Post, C.C. Pye, W. Ravenek, P. Ros, P.R.T.
Schipper, G. Schreckenbach, J.G. Snijders, M. Sola, D. Swer-
hone, G. teVelde, P. Vernooijs, L. Versluis, O. Visser, E. van
Wezenbeek, G. Wiesenekker, S.K. Wolff, T.K. Woo, T. Ziegler,
ADF Release 1999.01/02, SCM, Amsterdam, 1999. (b) C.F.
Guerra, J.G. Snijders, G. teVelde, E.J. Baerends, Theor. Chem.
Acc. 99 (1998) 391.

E. Bannwart et al. / Journal of Organometallic Chemistry 622 (2001) 97–111
111
[35] S.H. Vosko, L. Wilk, M. Nussair, Can. J. Phys. 58 (1980) 1200.
[36] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
[37] J.P. Perdew, Phys. Rev. B 33 (1986) 8822.
Ravenek, J. Phys. Chem. 93 (1989) 3050.
[39] (a) J. Li, G. Schreckenbach, T. Ziegler, J. Phys. Chem. 98 (1994)
4838. (b) H. Jacobsen, G. Schreckenbach, T. Ziegler, J. Phys.
Chem. 98 (1994) 11406.
[38] T. Ziegler, V. Tschinke, E.J. Baerends, J.G. Snijders, W.
.