Cleavage of polarized P-P bonds in N-Heterocyclic diphosphines in reactions with metal olefin complexes

Article abstract of DOI:10.1021/om801008s

A CC-saturated N-heterocyclic diphosphine was prepared and reacted with W(CO)₄(cod) in acetonitrile to give a mixture of two complexes arising from insertion of an acetonitrile molecule or a metal dicarbonyl fragment, respectively, into the P-P

Full text of DOI:10.1021/om801008s

Organometallics 2009, 28, 1447–1452
1447
Cleavage of Polarized P-P Bonds in N-Heterocyclic Diphosphines in
Reactions with Metal Olefin Complexes
Sebastian Burck,† Dietrich Gudat,*,† and Martin Nieger‡
Institut fu¨r Anorganische Chemie, UniVersita¨t Stuttgart, Pfaffenwaldring 55, 70550 Stuttgart, Germany,
and Laboratory of Inorganic Chemistry, UniVersity of Helsinki, A. I. Virtasen aukio 1, Helsinki, Finland
ReceiVed October 20, 2008
A CC-saturated N-heterocyclic diphosphine was prepared and reacted with W(CO)4(cod) in acetonitrile
to give a mixture of two complexes arising from insertion of an acetonitrile molecule or a metal dicarbonyl
fragment, respectively, into the P-P bond of the starting material. Cleavage of P-P bonds was likewise
observed upon treating a diphosphine with metal olefin complexes; these reactions proceeded either by
metal insertion to give phosphenium-phospholide complexes or via metathesis with formation of a
P-chloro N-heterocyclic phosphine (NHP) and transfer of a phospholyl unit to the metal. The newly
formed complexes were characterized by spectroscopic data and in several cases by single-crystal X-ray
diffraction studies, and mechanistic explanations for the observed product diversity are proposed.
Chart 1
Introduction
N-heterocyclic diphosphines such as 1a,b (Chart 1) exhibit
highly reactive P-P bonds that add under very mild conditions
to electron-poor double or triple bonds to give unsymmetrically
substituted bidentate 1,2-bisphosphines.1,2
As these transformations allow the simultaneous introduction
of two P-donor moieties into an organic backbone, they may
be regarded as complementary to known approaches to syn-
thesize bidentate O,O or N,N ligands via dihydroxylation3 or
diamination4 of olefins, respectively. Of special interest is also
a recent report on the transition-metal-induced addition of 1b
to a nitrile, which gave directly the chelate complex 22 with a
hybrid bidentate P-donor ligand of a type that has recently
gathered interest in catalysis;5 thermolysis of 2 proceeded further
via extrusion of the nitrile and produced the complex 3, which
features an intriguing combination of both nucleophilic (phos-
pholide) and electrophilic (phosphenium) ligands and promises
a rich, yet unexplored, reactivity (Scheme 1).2 In order to
establish if the facile synthetic approach to 2 and 3 can be used
to access a wider range of similar but structurally diverse
complexes, it appeared of interest to establish if insertion of a
metal atom into a P-P bond is also feasible for diphosphines
with P-P bonds less reactive than those in 1a,b or for
complexes with metal atoms other than tungsten. We report here
on the synthesis of products similar to 2 and 3 from the reaction
of a C-saturated N-heterocyclic diphosphine with (cod)W(CO)5
and acetonitrile, the preparation of a phosphenium-phospholide
complex of cobalt from Jonas’ reagent, CpCo(C2H4)2, and a
C-unsaturated N-heterocyclic diphosphine, and the reactions of
the latter with iron dichloride that occurs likewise with rupture
of the P-P bond. A discussion of the results permits us to draw
conclusions on some mechanistic aspects of the metal insertion
and the scope of this reaction.
Results and Discussion
The C-saturated N-heterocyclic diphosphine 4 was prepared
by analogy to the synthesis of 1b1 via metathesis of the
appropriate cyclic chlorophosphine with tetraethyl lithium
phospholide and characterized by spectroscopic and analytical
data and a single-crystal X-ray diffraction study. The molecular
structure (Figure 1) is characterized by the presence of a twisted
rather than planar conformation of the N-heterocyclic ring and
a P-P bond (2.346(1) Å) decidedly shorter than in 1b (2.484(1)
Å1). The formal CC double and single bonds in the phosphole
ring are 1 pm shorter and 2 pm longer, respectively, than in
1b, and although these differences are on the verge of being
significant in relation to the sum of standard deviations, they
are in accord with the assumption of a lower degree of P-P
bond polarization in 4 that is suggested by the observed change
in P-P bond lengths.1 The 31P NMR data of 4 are distinguished
* To whom correspondence should be addressed. Fax: +49 711
68564241. E-mail: gudat@iac.uni-stuttgart.de.
† Universita¨t Stuttgart.
‡ University of Helsinki.
(1) Burck, S.; Gudat, D.; Nieger, M. Angew. Chem. 2004, 116, 4905;
Angew. Chem., Int. Ed. 2004, 43, 4801.
1
by a noteworthy increase of JPP (262 Hz vs 188 Hz for 1b).
The observation of two anisochronic signals for the methylene
protons in the N-heterocyclic ring indicates that inversion of
the phosphorus atom in the ring is slow on the NMR time scale,
whereas the observation of a single set of resonances for the
nuclei in the o-CH3 groups of the N-mesityl substituents point
to unrestricted rotation of the aryl rings around the N-C bonds.
Studies of the interaction of the C-saturated N-heterocyclic
diphosphine 4 with (cod)W(CO)4 and acetonitrile revealed that
(2) Burck, S.; Gudat, D.; Nieger, M. Angew. Chem. 2007, 119, 2977;
Angew. Chem., Int. Ed. 2007, 46, 2919.
(3) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV.
1994, 94, 2483. Sharpless, K. B. Angew. Chem. 2002, 114, 2126; Angew.
Chem., Int. Ed. 2002, 41, 2024.
(4) Muniz, K. New J. Chem. 2005, 29, 1371.
(5) Selected references: Sakai, N.; Mano, S.; Nozaki, K.; Takaya, H.
J. Am. Chem. Soc. 1993, 115, 7033. Nozaki, K.; Sakai, N.; Nanno, T.;
Higashijima, T.; Mano, S.; Horiuchi, T.; Takaya, H. J. Am. Chem. Soc.
1997, 119, 4413. Yan, Y.; Zhang, X. J. Am. Chem. Soc. 2006, 128, 7198.
10.1021/om801008s CCC: $40.75
2009 American Chemical Society
Publication on Web 02/09/2009

1448 Organometallics, Vol. 28, No. 5, 2009
Burck et al.
Figure 1. Molecular structure of 4 (H atoms omitted for clarity;
50% probability thermal ellipsoids). Selected bond lengths (Å):
P(1)-N(5) ) 1.694(1), P(1)-N(2) ) 1.710(1), P(1)-P(2) )
2.346(1), N(2)-C(3) ) 1.478(2), C(3)-C(4) ) 1.509(2), C(4)-N(5)
) 1.469(2), N(5)-C(15) ) 1.426(2), C(6)-C(11) ) 1.403(2),
P(2)-C(27) ) 1.800(2), P(2)-C(24) ) 1.810(2), C(24)-C(25) )
1.365(2), C(25)-C(26) ) 1.462(2), C(26)-C(27) ) 1.362(2).
Scheme 1a
Figure 2. Molecular structure of 5 (H atoms omitted for clarity;
50% probability thermal ellipsoids). Selected bond lengths (Å):
W(1)-C(1B) ) 1.998(3), W(1)-C(1A) ) 2.006(3), W(1)-C(1D)
) 2.038(3), W(1)-C(1C) ) 2.041(3), W(1)-P(2) ) 2.466(1),
W(1)-P(1) ) 2.476(1), N(1)-C(1) ) 1.262(3), N(1)-P(1) )
1.755(2), C(1)-P(2) ) 1.882(2), P(1)-N(5) ) 1.664(2), P(1)-N(2)
) 1.676(2).
Scheme 2a
1
a
R ) 2,4,6-trimethylphenyl.
a reaction was only observed if all three components were
present and that the diphosphine 4 reacted, in a fashion similar
to that for 1b,2 neither with the metal complex nor with
acetonitrile alone. However, whereas the formation of 2 from
1b occurred already at ambient temperature, the reaction of 4
was only observed under more forcing conditions (50 °C) and
produced, according to in situ 31P NMR studies, a mixture of
the acetonitrile addition product 5 and the nitrile-free phospholyl
complex 6 (Scheme 2). Attempts to separate both complexes
failed and isolation of pure products on a preparative scale thus
remained unfeasible, but 5 was further characterized by an X-ray
diffraction study of a single crystal which was serendipitously
obtained by manual selection from the isolated product mixture.
As in the case of 2, cleavage of acetonitrile from 5 occurred
upon heating to 200 °C and allowed quantitative conversion of
the original mixture into 6, which was isolated by crystallization
and unambiguously characterized by analytical and spectro-
scopic data and a single-crystal X-ray diffraction study.
a
R1 ) 2,4,6-trimethylphenyl.
the N-heterocycles) not significant in relation to the estimated
standard deviations and will not be discussed further. The
spectroscopic data of 5 and 6 likewise closely resemble those
of the species 2 and 3 with C-unsaturated N,P-heterocycles. The
change from a formal bis-phosphine (5) to a phosphenium-
phospholide (6) coordination mode environment is reflected by
a strong increase of both δ(31P) and JWP for the phosphorus
1
The molecular structures of the complexes 5 and 6 are
displayed in Figures 2 and 3, respectively. The X-ray diffraction
study confirms the coordination of a cis-substituted tungsten
tetracarbonyl fragment by a chelating bis-phosphine for 5 and
a piano-stool like geometry with a η5-bound phospholyl moiety
for 6. The terminal phosphenium ligand in 5 displays a planar
coordination of the phosphorus atom. The N-heterocyclic rings
in 5 and 6 are not planar as in 2 and 3 but adopt a flat twist
conformation. Although the individual bond distances in 5 and
6 differ numerically from the corresponding values for 2 and
3,2 respectively, these deviations are (apart from the deviation
in N-C and C-C bonds involving sp2 and sp3 carbon atoms in
atom in the phosphenium moiety, which is indicative of partial
M-P double-bond character;6 at the same time, the switch from
η1 to η5 coordination of the phospholyl moiety induces a strong
shielding of the phosphorus resonance and a decrease of JWP
to a value of 0.
If one considers that 5 is stable in solution up to temperatures
of at least 100 °C and that its conversion into 6 occurs only
around 200 °C, it appears unreasonable to postulate 5 as an
intermediate in the formation of 6 during the reaction of 4 and
1
(6) Cowley, A. H.; Kemp, R. A. Chem. ReV. 1985, 85, 372. and literature
cited therein.

Metal-Promoted P-P Bond CleaVage
Organometallics, Vol. 28, No. 5, 2009 1449
Scheme 3a
1
a
R ) 2,6-Me2C6H4.
peratures in both cases are easily rationalized. The fact that a
phosphenium-phospholide complex (6) is formed in the reac-
tion with 4, but not with 1b, is rationalized by postulating that
the nitrile-containing intermediates (A or its successors) are
thermally labile and may easily lose both nitriles at elevated
temperature. Stabilization of the coordinatively unsaturated metal
atom may then occur by a switch from η1(P) to π coordination
of the phospholyl ligand, and it is conceivable that after
additional loss of one carbonyl the transient π-phospholyl
complexes B are produced. Conversion of these intermediates
into the final products 3 and 6 via recombination of the ionic
fragments and decarbonylation is a known reaction that has
precedence in the formation of a previously described phos-
phenium-cobalt complex.11 The mechanistic hypothesis is
further supported by the observation that reaction of 1b with
W(CO)3(MeCN)3 yields likewise a mixture of the nitrile addition
product 2 and the phosphenium-phospholide complex 3 unless
performed under strict temperature control (formation of 3 is
obviously facilitated by a lower number of stabilizing carbonyl
ligands and thus an increased thermal lability for W(CO)3-
(MeCN)3 as compared to W(CO)4(MeCN)2).12 Although inser-
tion into the P-P bond of symmetrical diphosphines has been
shown to follow a radical mechanism,13 we want to rule out
this alternative in the present case, as 1b (which reacts with
(cod)W(CO)4/nitrile to give the same products as 4) had
definitely previously been disproved to undergo similar radical
reactions,1 and even the reactions of alkynes with diphosphines
having P-P bonds less polar than in 4 have been considered to
proceed via heterolytic rather than homolytic bond cleavage.8
Figure 3. Molecular structure of one of the two crystallographically
independent molecules of 6 (H atoms omitted for clarity; 50%
probability thermal ellipsoids; only W and P atoms refined
anisotropically). Selected bond lengths (Å) (values in brackets
denote bond lengths for the second molecule): W(1)-C(1B) )
1.933(15) [1.922(15)], W(1)-C(1A) ) 1.937(14) [1.959(17)],
W(1)-P(1) ) 2.223(3) [2.223(4)], W(1)-P(2) ) 2.572(4) [2.566(4)],
W(1)-cent ) 2.021(6) [1.959(6)], P(1)-N(2) ) 1.647(12),
[1.665(12)], P(1)-N(5) ) 1.666(12) [1.643(12)].
Chart 2
W(CO)4(cod). We propose therefore an alternative mechanism
which explains the formation of both products via a common
precursor and allows us at the same time to rationalize the
previously reported behavior of 1.2 Following this model, we
suggest that the reaction is initiated by displacement of
cyclooctadiene by two nitrile ligands and subsequent attack of
a coordinated CN triple bond by the nucleophilic phospholyl
moiety of the diphosphine. A plausible structure of a resulting
key intermediate is formulated as A (Chart 2), which resembles
the betaines arising from interaction of electron-deficient alkynes
with tertiary phosphines7 or diphosphines.8 Since we had
previously established that interaction of the phosphinyl moiety
in 1 with electrophiles facilitates P-P bond dissociation,9 it is
reasonable to assume that the next reaction step involves
heterolytic P-P bond fission. The observed products 2 and 5
are finally formed via migration of the phosphenium cation
fragment to the imino nitrogen atom, extrusion of the remaining
nitrile ligand, and coordination of both phosphine donor moieties
to the metal atom.10 Since the increased P-P bond lengthening
in 1b as compared to the bonds in 4 must be considered to
facilitate the bond-breaking step,1,9 the different reaction tem-
In order to establish the feasibility of P-P bond activation
with metals other than tungsten, we studied the behavior of the
more reactive diphosphine 1c1 toward anhydrous iron dichloride
and CpCo(C2H4)2 (Jonas’ reagent), respectively. Analysis of 31
P
NMR spectra of reaction solutions of 1c and FeCl2 indicated
formation of a 1:1 mixture of the chlorophosphine 714 and the
diphosphaferrocene 815 (Scheme 3). Separation of the products
by selective extraction with hexane proved unfeasible, and
further attempts toward isolation of pure compounds were thus
(11) Burck, S.; Daniels, J.; Gans-Eichler, T.; Gudat, D.; Na¨ttinen, K.;
Nieger, M. Z. Anorg. Allg. Chem. 2005, 631, 1403.
(12) Burck, S. Ph.D. Thesis, Universita¨t Stuttgart, Stuttgart, Germany,
2006.
(13) Bezombes, J. P.; Hitchcock, P. B.; Lappert, M. F.; Nycz, J. E.
Dalton Trans. 2004, 499.
(7) Recent references: Duan, Z.; Zhang, J.; Tian, R.; Bai, L.; Mathey,
F. Organometallics 2008, 27, 5169. Liu, H.; Jiang, H. Tetrahedron 2008,
64, 2120.
(8) Dodds, D. L.; Haddow, M. F.; Orpen, A. G.; Pringle, P. G.;
Woodward, G. Organometallics 2006, 25, 5937.
(9) Gudat, D.; Burck, S.; Nieger, M.; Vindusˇ, D. Eur. J. Inorg. Chem.
2008, 704.
(10) The exact sequence of the individual steps and the nature of any
further intermediates formed must currently remain undecided.
(14) Burck, S.; Gudat, D.; Na¨ttinen, K.; Nieger, M.; Niemeyer, M.;
Schmid, D. Eur. J. Inorg. Chem. 2007, 5112.
(15) Sava, X.; Ricard, L.; Mathey, F.; Le Floch, P. Organometallics
2000, 19, 4899.

1450 Organometallics, Vol. 28, No. 5, 2009
Burck et al.
Chart 3
Scheme 4. Schematic Representation of the Proposed Reaction
Mechanism for the Reaction of N-Heterocyclic Diphosphines
with Metal Olefin Complexes
abandoned, but both products were unambiguously identified
by comparison of their spectroscopic data with those of authentic
samples.16
Reaction of 1c with CpCo(C2H4)2 proceeded with gas
evolution (C2H4) to produce a dark-colored solution from which
a black crystalline product was isolated. The crystals were
unsuitable for a single-crystal X-ray diffraction study, but
analytical and spectroscopic (NMR, MS) data allowed unam-
biguous identification of the product as cobalt complex 10. The
presence of a phospholyl, a cyclopentadienyl, and a N-
heterocyclic phosphenium moiety is clearly evident from the
1H and 13C NMR spectra, and the observation of two 31P NMR
signals connected by a homonuclear coupling (32 Hz) proves
that both phosphorus-containing ligands are attached to the metal
but are no longer directly bonded. The substantial broadening
of the signal of the N-substituted phosphorus atom by partially
unreasonable to assume that an anionic 20-VE [CpCo(phos-
pholyl)]- moiety should coordinate a further cationic diaza-
phospholenium moiety to give a complex with a formal valence
electron count of 22. Even though 10′′ seems to be better in
accord with the observed chemical shifts, a reliable distinction
between both models is unfeasible from the available spectro-
scopic data.
If one assumes that the previously described reactions are
likewise initiated by displacement of a coordinated alkene unit
by the phosphole moiety of 1c to give a betaine-like structure
similar to A, their course can be rationalized by a similar
mechanism involving subsequent P-P bond cleavage to give
the putative intermediate C (Scheme 4) and, finally, electrophilic
attack of the phosphenium cation at the counteranion to produce
the observed products. Whereas in the reaction with Jonas’
reagent the metal atom remains the only nucleophilic center,
intermediates derived from metal chlorides possess two nucleo-
philic sites in the metal and the chlorine atom. The observed
regioselectivitysattack at the chlorine atom to give metathesis
products rather than interaction with the metal atomsis in accord
with the previous finding22 that N-heterocyclic phosphenium
atoms seem to prefer charge-controlled reactions with harder
Lewis bases over orbital-controlled reactions with softer bases.
7
collapsed scalar coupling to the quadrupolar (I ) /2) 59Co
nucleus indicates that the coupling constant 1JCoPsalthough not
directly measurablesis of substantial magnitude and is a typical
characteristic of phosphenium complexes with a formal M-P
double bond and planar coordination at phosphorus.11 In
contrast, the narrow line width of the other signal reveals that
1JCoP is small and suggests that coordination of the phospholyl
ligand occurs via the π system or in an η1 mode which does
not involve the phosphorus lone pair.17 The C5H5 moiety gives
1
rise to a single H and 13C NMR signal, respectively, which
display the typical chemical shifts of a π-coordinated ligand. If
one assumes for the metal atom in 10 a valence electron (VE)
count of 18 and considers that a phosphenium-Co(I) fragment
contributes 10 VE, the molecular structure can be described by
either of the two models shown in Chart 3. In the first case, the
Cp- and phospholide ligands represent η3-bound 4-VE donors
and must be fluxional to match the observed NMR data, whereas
Conclusions
the second structure contains η5-Cp and η1-phospholyl moi-
Comparison of the reactivity of the N-heterocyclic diphos-
phines 1b and 4 toward acetonitrile/W(CO)4(cod) indicates that
the P-P bond strength in the starting material influences the
competition between a reaction under direct oxidative addition
of the P-P bond to the metal atom and the diphosphination of
the nitrile triple bond, respectively, and that high P-P bond
reactivity in the diphosphine is crucial for the observation of
the unusual nitrile activation. The reactivity of 1c toward further
low-valent metal olefin complexes reveals that P-P cleavage
reactions indeed have a larger scope and proceed either via
metathesis under transfer of only a phospholyl unit to the metal
or via insertion under transfer of both a phospholyl and a
phosphenium unit. The available results suggest that the
selection between both pathways depends on the ability of the
metal complex to act as ambident nucleophile, with metal halides
featuring a hard Lewis base center preferring the metathesis
reaction.
17,19
eties;18 both η1
and η3 binding modes20 of phospholyl
ligands have precedents in the literature. An alternative as-
sumption that 10 features both rings in an η5 binding mode as
in phosphacobaltocenes21 and phosphanickelocenes19 with
formal 19- and 20-VE counts is unlikely, as it appears
(16) Similarly, reaction of 1c with [(cod)RhCl]2 gave, according to a
31P NMR survey, a mixture of 7 and a complex assigned as (tetraethylphos-
pholyl)(cyclooctadiene)rhodium(I) on the basis of the similarity of the
1
observed NMR data (δ(31P)-4.1, JRhP ) 3.8 Hz) to those of the known
complex (2,5-di-tert-butylphospholyl)(cyclooctadiene)rhodium(I) (δ(31P)-
3.8, 1JRhP ) 15.2 Hz); see: Forissier, K.; Ricard, L.; Carmichael, D.; Mathey,
F. Chem. Commun. 1999, 1273.
(17) Mercier, F.; Ricard, L.; Mathey, F. Organometallics 1993, 12, 98.
Marinetti, A.; Le Floch, P.; Mathey, F. Organometallics 1991, 10, 1190.
Abel, E. W.; Towers, C. Dalton Trans. 1979, 814.
(18) A third alternative comprising a fluxional η1-Cp and a η5-phospholyl
ligand appears unlikely, in view of the observed δ(13C) values of the carbon
atoms in both rings.
(19) Burney, C.; Carmichael, D.; Forissier, K.; Green, J. C.; Mathey,
F.; Ricard, L. Chem. Eur. J. 2005, 11, 5381.
(20) Niecke, E.; Nieger, M.; Wenderoth, P. Angew. Chem. 1994, 106,
362; Angew. Chem., Int. Ed. Engl. 1994, 33, 353.
(21) Burney, C.; Carmichael, D.; Forissier, K.; Green, J. C.; Mathey,
F.; Ricard, L. Chem. Eur. J. 2005, 9, 2567. Cabon, Y.; Carmichael, D.;
Forissier, K.; Mathey, F.; Ricard, L.; Seeboth, N. Organometallics 2007,
26, 5468.
Experimental Section
All manipulations were carried out under an atmosphere of dry
argon using standard vacuum line techniques. Solvents were dried
(22) Burck, S.; Gudat, D. Inorg. Chem. 2008, 47, 315.

Metal-Promoted P-P Bond CleaVage
Organometallics, Vol. 28, No. 5, 2009 1451
Table 1. Crystallographic Data and Summary of Data Collection and Refinement for 4-6
4
5
6
formula
formula wt
C32H46N2P2
520.65
C38H49N3O4P2W
857.59
C34H46N2O2P2W
760.52
cryst size, mm
cryst syst; space group
0.40 × 0.30 × 0.20
0.30 × 0.20 × 0.20
0.45 × 0.20 × 0.10
monoclinic; P21/n
orthorhombic; Pbca
monoclinic; C2/c
unit cell dimens
a, Å
b, Å
c, Å
9.7825(2)
22.4622(4)
13.7718(3)
14.5680(1)
20.7786(1)
25.2384(2)
23.8525(5)
18.1101(4)
34.0334(6)
ꢀ, deg
90.014(2)
3026.17(11)
4; 1.14
0.166
98.180(1)
14551.9(5)
16; 1.39
3.292
V, Å3
7639.73(9)
8; 1.49
3.149
3472
Z; Fcalcd, Mg/m3
µ, mm-1
F(000)
1128
6144
θ range for data collecn, deg
completeness of data for θ ) 25°, %
limiting indices
3.0-27.5
3.0-27.5
2.9-25.0
99.8
99.7
96.7
-12 e h e 7, -29 e k e 22,
-18 e h e 18, -26 e k e 25,
-28 e h e 23, -20 e k e 21,
-27 e l e 40
29 175/12 438 (Rint ) 0.101)
semiempirical from equivalents
0.463 80, 0.304 45
12 438/0/371
1.015
0.080
0.217
4.579, -1.758
-17 e l e 17
-32 e l e 32
no. of collected/unique rflns
abs cor
21 121/6837 (Rint ) 0.044)
none
67 805/8682 (Rint ) 0.054)
semiempirical from equivalents.
0.535 73, 0.466 49
8682/0/440
max, min transmissn
no. of data/restraints/params
goodness of fit on F2
R1 (I > 2σ(I))
6837/0/331
1.026
0.040
0.107
0.354, -0.282
1.038
0.024
0.060
1.996, -1.600
wR2 (all data)
largest diff peak, hole, e Å-3
by standard procedures. NMR spectra were recorded on Bruker
Avance 400 (1H, 400.1 MHz; 13C, 100.5 MHz; 31P, 161.9 MHz)
and Avance 250 (1H, 250.1 MHz; 13C, 62.8 MHz; 31P, 101.2 MHz)
NMR spectrometers at 303 K; chemical shifts are referenced to
external TMS (1H, 13C) or 85% H3PO4 (ꢀ ) 40.480 747 MHz, 31P).
Coupling constants are given as absolute values; i, o, m, and p
denote the positions in phenyl and 2,6-dimethylphenyl (DMP,
denoted as C6H3) rings. EI-MS was carried out on a Varian MAT
711 instrument at 70 eV. Elemental analyses were determined on
a Perkin-Elmer 24000CHN/O analyzer. Melting points were
determined in sealed capillaries.
Reaction of 4 with (cod)W(CO)4. A solution of (cod)W(CO)4
(810 mg, 2 mmol) in CH3CN (30 mL) was added dropwise to a
solution of 4 (1.36 g, 2 mmol) in CH3CN (30 mL), and the resulting
mixture was stirred for 4 h at 50 °C. Formation of a mixture of 5
and 6 was established by 31P NMR. Attempts toward separation of
the reaction products by evaporation of volatiles and fractionating
crystallization from hexane failed and produced a precipitate
consisting of a mixture of crystals of both specimens. Manual
separation of the different sorts of crystals on a large scale was
unfeasible, but 5 could be further characterized by a single-crystal
X-ray diffraction study of a manually selected single crystal.
31P{1H} NMR (C6D6): δ 218.1 (d, 2JPP ) 10.2 Hz, 1JPW ) 718 Hz,
1,3-Dimesityl-2-(2′,3′,4′,5′-tetraethylphospholyl)tetrahydro-
1,3,2-diazaphosphole (4). Lithium (70 mg, 10 mmol) was added
to a solution of 1-chloro-2,3,4,5-tetraethylphosphole (1.15 g, 5
mmol) in THF (50 mL) and the mixture stirred for 3 h at room
temperature. Unreacted metal was filtered off, and the solution was
added dropwise to a cooled (-78 °C) solution of 1-chloro-1,3-
dimesityltetrahydro-1,3,2-diazaphosphole (1.81 g, 5 mmol) in THF
(100 mL). After the addition was finished, the mixture was warmed
to room temperature and stirred for 1 h more. Volatiles were
evaporated under vacuum, the residue was dissolved in hexane (100
mL), and this solution was filtered over Celite. The filtrate was
concentrated under reduced pressure to a total volume of 30 mL
and stored at -20 °C for crystallization. Yellow crystals separated
that were collected by filtration and dried under vacuum: yield
1.87 g (69%); mp 101 °C. 1H NMR (C6D6): δ 6.75 (s, 4 H, m-CH),
3.53 (m, 2 H, N-CH2), 2.82 (m, 2 H, N-CH2), 2.40 (s, 12 H,
3
1
N2P, 6), 158.2 (d, JPP ) 28.8 Hz, JPW ) 154 Hz, 5), 100.1 (d,
1
2
3JPP ) 28.8 Hz, JPW ) 119 Hz, 5), -16.1 (d, JPP ) 10.2 Hz, 6).
IR (KBr, Nujol): 2017 (5), 1920 (6), 1905 (5), 1881 (6), 1856 cm-1
(5).
Dicarbonyl-2-(1,3-dimesityltetrahydro-1,3,2-diazaphospho-
lyl)(η5-2′,3′,4′,5′-tetraethylphospholyl)tungsten(0) (6). The crude
mixture of 5 and 6 (1.72 g) was placed in a Schlenk tube and heated
briefly to 200 °C. The dark-colored melt was allowed to cool to
ambient temperature and dissolved in toluene (10 mL). Storage of
the solution at -20 °C produced a yellow crystalline precipitate,
which was collected by filtration and dried under vacuum: yield
1.40 g (92%); mp 199 °C. 1H NMR (C6D6): δ 6.93 (s, 4 H, m-CH),
3.77 (d, 3JPH ) 4.5 Hz, 4 H, N-CH2), 2.52 (m, 4 H, CH2), 2.37 (s,
12 H, o-CH3), 2.29 (s, 6 H, p-CH3), 2.29 (m, 4 H, CH2), 2.10 (dq,
2JHH ) 7.4 Hz, 3JHH ) 7.4 Hz, 2 H, CH2), 1.68 (dq, 2JHH ) 7.6 Hz,
3JHH ) 7.6 Hz, 2 H, CH2), 1.05 (t, 3JHH ) 7.6 Hz, 6 H, CH3), 0.79
3
o-CH3), 2.20 (q, JHH ) 7.3 Hz, 4 H, CH2), 2.10 (s, 6 H, p-CH3),
3
3
2.01 (m, 2 H, CH2), 1.9 (m, 2 H, CH2), 1.07 (t, JHH ) 7.5 Hz, 6
(t, JHH ) 7.4 Hz, 6 H, CH3). 31P{1H} NMR (C6D6): δ 220.1 (d,
H, CH3), 0.90 (t, 3JHH ) 7.3 Hz, 6 H, CH3). 13C{1H} NMR (C6D6):
1JPW ) 718 Hz, JPP ) 10.3 Hz, N2P), -16.0 (d, JPP ) 10.3 Hz,
P(phosphole)). IR (KBr, Nujol): 1920, 1881 cm-1. MS (EI, 70 eV,
430 K): m/e (%) 760.3 (100) [M]+, 732.0 (6) [M - CO]+, 704.0
(14) [M - 2CO]+, 325.2 (81) [C20H26N2P]+. Anal. Calcd for
C34H46N2O2P2W (760.55): C, 53.69; H, 6.20; N, 3.68. Found: C,
54.41; H, 6.46; N, 3.46.
2
2
δ 151.6 (d, 2JPC ) 5.2 Hz, i-C), 142.9 (dd, 1JPC ) 17.6 Hz, 2JPC
)
11.4 Hz, C(phosphole)), 139.9 (dd, JPC ) 12.7, 1.1 Hz, C(phosp-
4
5
hole)), 136.2 (broad s, o-C), 134.9 (dd, JPC ) 1.9 Hz, JPC ) 0.1
5
Hz, m-C), 130.3 (d, JPC ) 1.1 Hz, p-C), 54.0 (dd, JPC ) 7.9, 0.3
Hz, N-CH2), 21.7 (dd, JPC ) 20.7, 1.6 Hz, CH2), 21.3 (dd, 4JPC
)
5
6
1.5 Hz, JPC ) 0.6 Hz, o-CH3), 20.8 (d, JPC ) 0.6 Hz, m-CH3),
Reaction of 1c with FeCl2. 1c (49 mg, 0.1 mmol) and anhydrous
FeCl2 (10 mg, 0.05 mmol) were dissolved in C6D6. Quantitative
conversion of the starting materials into the products was established
by 31P and 1H NMR spectroscopy. The products were identified by
comparison of the data with those of authentic samples; no attempts
toward separation or isolation were made. 1H NMR (C6D6): δ 6.96
(broad, 6 H, m-/p-CH), 5.78 (broad, 2 H, N-CH), 2.45 (broad, 4
H, CH2), 2.35 (broad, 4 H, CH2), 2.19 (s, 12 H, o-CH3), 1.20 (broad,
20.6 (dd, 3JPC ) 8.3 Hz, 4JPC ) 2.3 Hz, CH2), 18.6 (dd, 3JPC ) 5.4
4
4
Hz, JPC ) 1.7 Hz, CH3), 16.2 (d, JPC ) 1.3 Hz, CH3). 31P{1H}
NMR: δ 143.1 (d, 1JPP ) 261.6 Hz, N2P), 9.1 (d, 1JPP ) 261,6 Hz,
P(phosphole)). MS (EI, 70 eV, 400 K): m/e (%) 520.3 (0.3) [M]+,
325.2 (100) [M - C12H20P]+, 196.1 (42) [M - C20H25N2P]+, 148.1
(54). Anal. Calcd for C32H46N2P2 (520.68): C, 73.82; H, 8.91; N,
5.38. Found: C, 73.43; H, 8.96; N, 5.32.

1452 Organometallics, Vol. 28, No. 5, 2009
Burck et al.
6 H, CH3), 1.09 (broad, 6 H, CH3). 31P{1H} NMR (C6D6): δ 147.8
(s, 7), -62.9 (broad s, 8).
1. Direct methods (SHELXS-97)23 were used for structure solution,
and refinement was carried out using SHELXL-97 (full-matrix least
squares on F2).23 Hydrogen atoms were refined using a riding
model. Due to the low quality of the diffraction data of 6, only the
W and P atoms were refined anisotropically. There are large voids
(approximately 400 Å3) in the structure. These voids probably
contain highly disordered solvent (toluene), which could not be
refined reasonably. Application of SQUEEZE suggested that these
voids should contain approximately 800 electrons, or 16 molecules
of toluene, but this did not help to improve the quality of the
structure.
(η1-Cyclopentadienyl)(η1-1,3-bis(2′,6′-dimethylphenyl)-2,3-di-
hydro-1,3,2-diazaphospholyl)(η5-2,3,4,5-tetraethylphospholyl)co-
balt(I) (9). A solution of 1 (343 mg, 0.7 mmol) and CpCo(C2H2)2
(126 mg, 0.7 mmol) in dry THF (10 mL) was stirred for 1 h.
Volatiles were then removed under reduced pressure, the residue
was dissolved in pentane (10 mL), and this solution was stored at
-20 °C. Black crystals formed, which were collected by filtration
and dried under vacuum: yield 342 mg (81%); mp 155 °C. 1H NMR
3
(C6D6): δ 6.85-7.00 (m, 6 H, m-/p-H), 5.64 (d, JPH ) 6.7 Hz, 2
H, N-CH), 3.90 (s, 5 H, Cp), 2.32-2.78 (m, 8 H, CH2), 2.33 (s,
Acknowledgment. We thank Dr. J. Opitz, J. Trinkner, and
K. Wohlbold (Institut fu¨r Organische Chemie, Universita¨t
Stuttgart) for recording the mass spectra. The Deutsche
Forschungsgemeinschaft is acknowledged for financial support.
3
3
12 H, o-CH3), 1.27 (t, JHH ) 7.5 Hz, 6 H, CH3), 1.22 (t, JHH
)
7.5 Hz, 6 H, CH3). 13C{1H} NMR (C6D6): δ 142.9 (d, JPC ) 6.1
1
Hz, PC(phosphole)), 139.4 (d, 2JPC ) 5.7 Hz, i-C), 136.3 (s, o-C),
2
4
128.9 (s, p-CH), 128.8 (s, m-CH), 122.4 (dd, JPC ) 2.3 Hz, JPC
2
2
) 1.0 Hz, N-CH), 82.3 (d, JPC ) 2.9 Hz, Cp), 23.2 (d, JPC
)
3
Supporting Information Available: CIF files giving X-ray
structural information on 4-6. This material is available free of
charge via the Internet at http://pubs.acs.org. In addition CCDC-
704594 (4), CCDC-704595 (5), and CCDC-704596 (6) contain
supplementary crystallographic data for this paper. These data can
be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
21.8 Hz, CH2), 22.3 (s, CH2), 18.7 (d, JPC ) 7.4 Hz, CH3), 18.6
(s, o-CH3), 16.8 (d, 3JPC ) 4.0 Hz, CH3). 31P{1H} NMR (C6D6): δ
2
157.5 (broad, N2P), -81.0 (d, JPP ) 32.4 Hz, PC4). MS (EI, 70
eV, 410 K): m/e (%) 614.3 (<0.1) [M]+, 419.1 (<0.1) [M - Et4-
phospholyl]+, 319.1 (100) [CpCo(Et4-phospholyl)]+, 295.0 (37)
[C18H20PN2]+. Anal. Calcd for C35H45N2P2Co (614.64): C, 68.40;
H, 7.38; N, 4.56. Found: C, 67.93; H, 7.36; N, 4.12.
Crystallography. The crystal structure determinations of 4-6
were performed on a Nonius KappaCCD diffractometer at 123(2)
K using Mo KR radiation (λ ) 0.710 73 Å). Crystal data, data
collection parameters, and results of the analyses are given in Table
OM801008S
(23) Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112.