Homoleptic diphosphacyclobutadiene complexes [M(η(4)-P2C2R2)2]x- (M = Fe, Co; x = 0, 1).

Abstract

The preparation and comprehensive characterization of a series of homoleptic sandwich complexes containing diphosphacyclobutadiene ligands are reported. Compounds [K([18]crown-6)(thf)(2)][Fe(η(4)-P(2)C(2)tBu(2))(2)] (K1), [K([18]crown-6)(thf)(2)][Co(η(4)-P(2)C(2)tBu(2))(2)] (K2), and [K([18]crown-6)(thf)(2)][Co(η(4)-P(2)C(2)Ad(2))(2)] (K3, Ad = adamantyl) were obtained from reactions of [K([18]crown-6)(thf)(2)][M(η(4)-C(14)H(10))(2)] (M = Fe, Co) with tBuC[triple bond]P (1, 2), or with AdC[triple bond]P (3). Neutral sandwiches [M(η(4)-P(2)C(2)tBu(2))(2)] (4: M = Fe 5: M = Co) were obtained by oxidizing 1 and 2 with [Cp(2)Fe]PF(6). Cyclic voltammetry and spectro-electrochemistry indicate that the two [M(η(4)-P(2)C(2)tBu(2))(2)](-)/[M(η(4)-P(2)C(2)tBu(2))(2)] moieties can be reversibly interconverted by one electron oxidation and reduction, respectively. Complexes 1-5 were characterized by multinuclear NMR, EPR (1 and 5), UV/Vis, and Mössbauer spectroscopies (1 and 4), mass spectrometry (4 and 5), and microanalysis (1-3). The molecular structures of 1-5 were determined by using X-ray crystallography. Essentially D(2d)-symmetric structures were found for all five complexes, which show the two 1,3-diphosphacyclobutadiene rings in a staggered orientation. Density functional theory calculations revealed the importance of covalent metal-ligand π bonding in 1-5. Possible oxidation state assignments for the metal ions are discussed.

Full text

Page 1

DOI: 10.1002/chem.201001913
Homoleptic Diphosphacyclobutadiene Complexes [M(h4-P2C2R2)2]x?
(M=Fe, Co; x=0, 1)
Robert Wolf,*[a, b]Andreas W. Ehlers,[b]Marat M. Khusniyarov,[c]Frantis ˇek Hartl,[d, e]
Bas de Bruin,[d]Gary J. Long,[f]Fernande Grandjean,[g]Falko M. Schappacher,[a]
Rainer Pçttgen,[a]J. Chris Slootweg,[b]Martin Lutz,[h]Anthony L. Spek,[h]and
Koop Lammertsma*[b]
Abstract: The preparation and compre-
hensive characterization of a series of
homoleptic sandwich complexes con-
tainingdiphosphacyclobutadiene
gandsarereported.
li-
Compounds
[K([18]crown-6)(thf)2][Fe(h4-
P2C2tBu2)2] (K1), [K([18]crown-6)-
(K2),
(thf)2][Co(h4-P2C2tBu2)2] and
[K([18]crown-6)(thf)2][Co(h4-
P2C2Ad2)2] (K3, Ad=adamantyl) were
obtained fromreactionsof
[K([18]crown-6)(thf)2][M(h4-C14H10)2]
(M=Fe, Co) with tBuC?P (1, 2), or
with AdC?P (3). Neutral sandwiches
[M(h4-P2C2tBu2)2] (4: M=Fe 5: M=
Co) were obtained by oxidizing 1 and 2
with [Cp2Fe]PF6. Cyclic voltammetry
and spectro-electrochemistry indicate
that the two [M(h4-P2C2tBu2)2]?/[M(h4-
P2C2tBu2)2] moieties can be reversibly
interconverted by one electron oxida-
tion and reduction, respectively. Com-
plexes 1–5 were characterized by multi-
nuclear NMR, EPR (1 and 5), UV/Vis,
and Mçssbauer spectroscopies (1 and
4), mass spectrometry (4 and 5), and
microanalysis(1–3).
structures of 1–5 were determined by
using X-ray crystallography. Essentially
D2d-symmetric structures were found
for all five complexes, which show the
two 1,3-diphosphacyclobutadiene rings
in a staggered orientation. Density
functional theory calculations revealed
the importance of covalent metal–
ligand p bonding in 1–5. Possible oxi-
dation state assignments for the metal
ions are discussed.
The molecular
Keywords: cobalt · iron · metalates ·
phosphorus · sandwich complexes
[a] Dr. R. Wolf, Dr. F. M. Schappacher, Prof. Dr. R. Pçttgen
Institute of Inorganic and Analytical Chemistry
University of M?nster
Corrensstrasse 30, 48149 M?nster (Germany)
Fax: (+ +49)251-833-6610
E-mail: r.wolf@uni-muenster.de
[b] Dr. R. Wolf, Dr. A. W. Ehlers, Dr. J. C. Slootweg,
Prof. Dr. K. Lammertsma
Department of Organic and Inorganic Chemistry
Faculty of Sciences, VU University Amsterdam
De Boelelaan 1083, 1081 HV, Amsterdam (The Netherlands)
Fax: (+ +31)20-5987488
E-mail: K.Lammertsma@few.vu.nl
[c] Dr. M. M. Khusniyarov
Department of Chemistry and Pharmacy
Friedrich-Alexander-Universit?t Erlangen-N?rnberg
Egerlandstrasse 1, 91058 Erlangen (Germany)
[d] Prof. Dr. F. Hartl, Dr. B. de Bruin
Homogeneous and Supramolecular Catalysis
Van?t Hoff Institute for Molecular Sciences
University of Amsterdam
Science Park 904, 1098 XH Amsterdam (The Netherlands)
[e] Prof. Dr. F. Hartl
Department of Chemistry, University of Reading
Whiteknights, Reading, RG6 6AD (UK)
[f] Prof. Dr. G. J. Long
Department of Chemistry
Missouri University of Science and Technology
University of Missouri, Rolla, MO 65409-0010 (USA)
[g] Prof. Dr. F. Grandjean
Department of Physics, B5, University of Li?ge
4000 Sart-Tilman (Belgium)
[h] Dr. M. Lutz, Prof. Dr. A. L. Spek
Bijvoet Center for Biomolecular Research
Crystal and Structural Chemistry, Utrecht University
Padualaan 8, 3584 CH, Utrecht (The Netherlands)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem201001913.
Chem. Eur. J. 2010, 00, 0–0 ? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
These are not the final page numbers!??
&1&
FULL PAPER

Page 2

Introduction
Since the discovery of ferrocene, [Cp2Fe], sandwich com-
plexes have fascinated the chemical community. Sandwiches
containing cyclopentadienyl and arene ligands are most
common. In particular, ferrocene derivatives have found nu-
merous applications.[1,2]The chemistry of complexes contain-
ing four-membered cyclobutadiene rings is comparatively
less developed.[3,4]In particular, the paucity of homoleptic
cyclobutadiene complexes is striking.[5]
Phosphaalkynes,suchastert-butylphosphaalkyne
(tBu?C?P), can undergo transition-metal-mediated cyclooli-
gomerization, which is a versatile method that can generate
complexes with phosphaorganic ligands.[6,7]Cyclodimeriza-
tion to heteroleptic 1,3-diphosphacyclobutadiene complexes
is frequently encountered. For example, the reaction of
[Fe(CO)5] with tBu?C?P results in the formation of com-
plex A, a phosphorus analogue of Petit?s complex [Fe(h4-
C4H4)(CO)3].[8]By using cyclopentadienyl half-sandwich
complexes, for example, [CpCo(h2-C2H2)2], heteroleptic
complexes, such as B, are formed.[9]Complex C, which has
been prepared from [Ni(h4-cod)2] and tBu?C?P, represents
the only homoleptic diphosphacyclobutadiene complex re-
ported prior to our work.[10]
Unique and structurally diverse products, for example,
complexes D–F, can be obtained by using “gaseous” metal
atoms that are accessible by means of metal vapor (MV)
synthesis.[11]Unfortunately, the applicability of metal vapor
synthesis is limited by the poor selectivity, low yields, and
the special experimental apparatus required. Product mix-
tures are often formed. For example, the reaction of “gas-
eous” cobalt atoms with tBu?C?P yielded three complexes
G–I, which contain diphosphacyclobutadiene, phosphacyclo-
pentadienyl ligands, and a protonated tetraphosphabarre-
lene.[12]
Considering the fascinating products that are obtained
from reactions of “gaseous” metal atoms with phosphaal-
kynes we reasoned that unprecedented anionic sandwich
complexes could be synthesized by using “naked” transition-
metal anions. Ellis and co-workers have recently shown that
reactive arene metalates may serve as efficient synthetic
equivalents for Mx?synthons.[13]The preparation of the de-
caphosphatitanocene dianion [Ti(h5-P5)2]2?by means of the
reaction of an anionic titanium naphthalene complex with
P4is a striking illustration of the potential of such com-
pounds.[14]Apart from this landmark example, arene metal-
ates have not been utilized for the preparation of organome-
tallic sandwich compounds. Therefore, we endeavored to
study reactions of homoleptic metalates with phosphaal-
kynes. Here we report a full account of our preliminary
studies.[15]First, we describe the synthesis of the anionic di-
phosphacyclobutadiene sandwich complexes [K([18]crown-
(K1),
6)(thf)2][Fe(h4-P2C2tBu2)2] [K([18]crown-6)(thf)2]
[Co(h4-P2C2tBu2)2] (K2), and [K([18]crown-6)(thf)2][Co(h4-
P2C2Ad2)2] (K3, Ad=adamantyl) the structures and spectro-
scopic properties of which are discussed. Second, we present
an analysis of the redox properties of anions 1 and 2. We de-
scribe the synthesis of the neutral oxidation products [Fe(h4-
P2C2tBu2)2] (4) and [Co(h4-P2C2tBu2)2] (5), which may be
consideredasrarephosphorusanaloguesofelusive
bis(cyclobutadiene) complexes [Fe(h4-C4R4)2] and [Co(h4-
C4R4)2]. Third, we report the Mçssbauer spectra of 1 and 4
and analyze DFT calculations for model compounds, which
provide, for the first time, detailed insight into the electronic
structures of homoleptic diphosphacyclobutadiene com-
plexes.
Results and Discussion
Syntheses, structures and spectroscopic properties of the
anionic sandwich complexes K1–K3: Orange potassium salts
[K([18]crown-6)(thf)2][Fe(h4-P2C2tBu2)2] (K1),
and
[K([18]crown-6)(thf)2][Co(h4-P2C2tBu2)2](K2),
[K([18]crown-6)(thf)2][Co(h4-P2C2Ad2)2] (K3) were obtained
in moderate to good yields by reacting four equivalents of
phosphaalkyne (tBu?C?P or Ad?C?P)[16,17]with the metal-
ates[Fe(h4-C14H10)2]?
(J)
Scheme 1).[18,19]Although highly sensitive to moisture and
oxygen, all three compounds are thermally very robust and
decompose at temperatures above 2308 8C.
The identity of K1–K3 was established by X-ray crystal-
lography, microanalyses, EPR (1), NMR, and UV/Vis spec-
troscopies. In the solid state, all three compounds show simi-
lar ion-separated structures that comprise homoleptic anions
[Fe(h4-P2C2tBu2)2]?(1), [Co(h4-P2C2tBu2)2]?(2), and [Co(h4-
P2C2Ad2)2]?
(3), aswellas
and[Co(h4-C14H10)2]?
(K,
a[K([18]crown-6)(thf)2]+ +
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim Chem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&2&

Page 3

countercation.[18,19]The molecular structure of adamantyl-
substituted K3 is shown as an example in Figure 1. The
anions feature two h4-coordinated diphosphacyclobutadiene
rings in a staggered orientation, owing to the steric repulsion
of the tBu (1, 2) or Ad (3) substituents. The P2C2rings of
the ligands display equal P?C bond lengths (Table 1). The
P?C bond lengths (1.7999(18)–1.8028(18) ?) are typical for
1,3-diphosphacyclobutadienes coordinated to first row tran-
sition metals.[6]Metal–phosphorus and metal–carbon bond
lengths are in the range reported for related, neutral com-
plexes, for example, A–C.[8–10,20]The Fe?P and Fe?C bond
lengths in the iron complex 1 are slightly larger than the
Co?P and Co?C bond lengths in the cobalt species 2 and 3,
due to the larger radius of iron.
[Fe(h4-P2C2tBu2)2]?(1) is a paramagnetic 17-electron spe-
cies. The magnetic moment, determined by the Evans
method in [D8]THF solution, of 1.77 mBindicates the pres-
ence of one unpaired electron per iron center. In accordance
with the paramagnetism of 1, only a broad signal in the
1H NMR spectrum (C6D6) was detected for the tBu groups
at d??1.9 ppm. The experimental EPR spectrum of a
frozen THF solution (T=40 K, Figure 2) revealed an axial,
g-tensor, without any resolved hyperfine couplings with the
31P nuclei. A satisfactory spectral simulation was obtained
with the g values g11=2.279, g22=2.026, and g33=2.026. The
observed pattern is typical for an axially symmetric d9spe-
cies with one unpaired electron and thus further confirms
the S=1/2 (doublet) ground state. The EPR properties were
also calculated with ADF (OPBE/TZ2P) by using the opti-
mized geometry of [Fe(h4-P2C2Me2)2]?(1?) as a smaller com-
putational model of the experimental system. The agree-
ments between the experimental EPR parameters and the
DFT-calculated parameters are very reasonable (Table 2).
The related heteroleptic 17-electron complexes [Fe(h6-
C7H8)(h4-P2C2tBu2)2]+ +(g11=2.53, g22=g33=2.01) and [Fe(h5-
P2C3tBu3)2(h4-P2C2tBu2)2] (g11=2.407, g22=g33=2.026), de-
scribed by Zenneck and co-workers,[21]display similar EPR
spectra as 1, with no observable31P hyperfine coupling. The
Scheme 1.
Figure 1. Solid-state structure of [K([18]crown-6)(thf)2][Co(h4-P2C2Ad2)2]
(K3), displacement ellipsoids at 50% probability level, H atoms and dis-
ordered solvent molecules are omitted for clarity. Symmetry equivalents
used to generate equivalent atoms: a) 1?x, 0.5?z; b) 1.5?x, 0.5?y, ?z.
Table 1. Selected bond lengths [?] and angles [8 8] of 1–5.
K1[15]
K2[15]
K34[15]
5
M?P
M?C
P?C
P-C-P
C-P-C
2.2969(5)–2.3024(5)
2.0939(16)–2.1027(16)
1.8000(17)–1.8048(17)
98.71(8)–99.07(8)
80.60(8)–80.94(8)
2.2537(6)–2.2598(6)
2.0649(19)–2.0724(19)
1.791(2)–1.798(2)
98.66(10)–99.02(10)
80.79(9)–81.02(9)
2.2601(5)–2.2620(4)
2.0603(17)–2.0671(18)
1.7999(18)–1.8028(18)
99.03(8),99.03(8)
80.61(8),80.67(8)
2.3044(10)–2.3081(9)
2.090(3)–2.100(3)
1.796(3)–1.808(3)
97.98(15)–98.61(15)
80.91(15)–81.41(15)
2.2811(7)–2.2830(7)
2.086(2)–2.087(2)
1.801(2)
98.66(11)–98.81(11)
80.83(11)–80.99(11)
Figure 2. Experimental and simulated X-band EPR spectra of anionic
complex 1. Experimental conditions: T=40 K, attenuation=30 dB, field
modulation amplitude=4 G, microwave frequency=9.380276 GHz.
Chem. Eur. J. 2010, 00, 0–0? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
These are not the final page numbers!??
&3&
FULL PAPER
Diphosphacyclobutadiene Complexes

Page 4

significant g anisotropies indicate that the unpaired electron
is metal-centered in all three cases. This is nicely confirmed
for 1 by our DFT calculations (see below). In contrast, the
19-electron anion [Fe(h6-C7H8)(h4-P2C2tBu2)2]?(g11=2.032,
g22=1.992, and g33=1.946) displays a much smaller g aniso-
tropy and thus seems to be a primarily ligand-centered radi-
cal.[21]
The cobaltate anions [Co(h4-P2C2tBu2)2]?(2) and [Co(h4-
P2C2Ad2)2]?(3) are diamagnetic 18-electron complexes.
Multinuclear NMR spectroscopy supports the structural for-
mulations obtained from the X-ray crystal structure analysis.
Both complexes gave rise to one singlet in the31P{1H} NMR
spectrum (K2: d= + +2.4 ppm, K3: d=?2.6 ppm).
13C{1H} NMR spectra of K2 showed a single tBu environ-
ment. Only one set of adamantyl signals was detected in the
1H and
13C{1H} NMR spectrum, confirming the highly symmetric
nature of K3. The UV/Vis spectrum of iron complex 1 in
THF showed two intense absorptions in the UV region at
l=296 and 383 nm. In addition, a very weak, broad absorp-
tion was detected at l=699 nm. The cobaltates K2 and K3
displayed very similar UV/Vis spectra in THF with intense
absorptions in the UV around l=290 and 330 nm. TD-DFT
calculations at the spin-unrestricted B3LYP level of theory
indicate that these absorptions mainly arise as a result of
charge transfer between occupied metal-centered orbitals
and empty ligand orbitals, that is metal-to-ligand charge
transfer (MLCT).
Spectro-electrochemistry of complexes 1 and 2: Redox activ-
ity is an important facet of organometallic sandwich com-
pounds. Anionic complexes 1 and 2 display identical 1,3-di-
phospha-2,4-tert-butylcyclobutadiene ligands, but differ in
the metal center (Fe vs. Co) and electron count (17e vs.
18e). Therefore, we decided to perform a comparative study
of their redox properties. First, we investigated the redox
behavior of 1 and 2 by cyclic voltammetry (CV) in THF.
Both complexes are reversibly oxidized at low potentials (1:
E1/2=?0.97 V, 2: E1/2=?0.73 V vs. Fc/Fc+ +, Figure 3, inset).
The iron-containing 17-electron anion 1 is oxidized even
more easily than the 18-electron cobalt complex 2.[22]The
electrode potential of the reversible Co redox couple 2/5
(18e/17e) shifts with the solvent polarity, the HOMO of 2
being apparently stabilized in more polar solvents: E1/2(V
vs. Fc/Fc+ +)=?0.62 in MeCN, ?0.68 in CD2Cl2, and ?0.73 in
THF. Furthermore, the CV spectra of 2, in CD2Cl2recorded
at 213 K, revealed a new, irreversible one-electron wave at
Ep,c=?2.70 V versus Fc/Fc+ +. This cathodic step might indi-
cate the formation of the 19-electron complex [Co(h4-
P2C2tBu2)2]2?that is highly unstable.
Characterization of oxidized complexes 4 and 5: Subsequent
synthetic investigations led to the access of the neutral spe-
cies 4 and 5 on a preparative scale by oxidizing anions 1 and
2 with [Cp2Fe]PF6(Scheme 2). The by-product ferrocene is
Table 2. Experimental and DFT-calculated EPR parameters of 1 and 5.
1
OPBE/
TZ2P[b]
5
Exp[a]
Exp[a]
OPBE/
TZ2P[c]
B3LYP/
TZVP)[d]
g11
g22
g33
AP
2.279
2.026
2.026
NR[e]
<30
NR[e]
<20
NR[e]
<20
2.346
2.028
2.024
4? ?24
2.642
2.038
2.038
NR[e]
2.317
2.063
2.061
4? ?31
2.295
2.071
2.048
4? ?19
11
AP
22
4? + +16 NR[e]
4? + +12 4? + +8
AP
33
4? + +24NR[e]
4? + +84? + +21
ACo
ACo
11 –
22 –
–
–
118
NR[e]
<20
NR[e]
<20
+ +476
+ +468
+ +445
?8
ACo
33 ––
?51
+ +48
[a] Parameters from spectral simulations (least squares ?best fit?). Hyper-
fine couplings in MHz. [b] DFT-calculated parameters for [Fe(h4-
P2C2Me2)2]?(1?, ADF, OPBE, TZ2P). [c] DFT-calculated parameters of 5
(optimized with Turbomole, BP86/SV(P), EPR parameters calculated
with ADF, OPBE/TZ2P). [d] DFT-calculated parameters of 5 (optimized
with Turbomole, B3LYP/TZVP, EPR parameters calculated with ORCA,
B3LYP/TZVP). [e] Not resolved (NR) in the experimental X-band spec-
tra.
Figure 3. UV/Vis monitoring of the electrochemical oxidation of 1 and 2 in THF/Bu4NPF6in an OTTLE cell, insets: cyclic voltammograms of 1 and 2 in
THF/Bu4NPF6.
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&4&
R. Wolf, K. Lammertsma et al.

Page 5

easily removed by vacuum sublimation. Complexes 4 and 5
were isolated in moderate yield as orange crystals that are
highly soluble in n-pentane. Their identity was established
by using mass spectrometry,1H NMR and EPR spectroscop-
ies, and single-crystal X-ray structure analyses.
Isostructural complexes 4 and 5 crystallize in the mono-
clinic space group C2/c with four molecules in the unit
cell.[15]The structure of 5 is depicted in Figure 4. The mole-
cules reside on crystallographic C2axes, and feature h4-coor-
dinated diphosphacyclobutadiene ligands in a staggered ori-
entation. The P?C bond lengths in the rhombic P2C2rings
are equal within experimental error (Table 1). The M?P and
M?C bond lengths of 4 are marginally longer (0.01–0.02 ?)
than in 5. Comparison of anionic and neutral sandwich
structures 1–3 reveals that removing one electron from the
anions only has negligible structural effects. In fact, the
bond lengths of neutral 16e iron complex 4 are mostly iden-
tical to those of the 17-electron anion 1 within experimental
error. Co?P and Co?C bond lengths of 17-electron 5 are
only slightly longer than those of 18-electron 1 by about
0.02 ?.
Compound 4 is a rare example of a 16-electron iron com-
plex with P-heterocyclic ligands.[23]The solution magnetic
moment of 4 (Evans method, C6D6solution) of 2.74 mBis
close to the expected spin-only value for two unpaired elec-
trons (2.82 mB). This is in accord with our DFT calculations,
which predict a triplet ground state (vide infra). Owing to
its paramagnetic nature, the
1H NMR spectrum of 4 in
[D6]benzene featured a broad signal at d=2.6 ppm for the
tBu groups, whereas the
plex 5 gave rise to a very broad tBu signal at d=?2.6 ppm
in C6D6. Its solution magnetic moment of 1.73 mB(deter-
mined in the same solvent) perfectly matches the expected
spin-only value for one unpaired electron in a low-spin com-
plex.
The UV/Vis spectrum of 4 in THF shows intense absorp-
tions in the UV region with a maximum at l=275 nm and a
shoulder at l=320 nm, and a broad medium-intensity ab-
sorption band at l=443 nm (emax=11000m?1cm?1) tailing
down to l=600 nm. The electronic absorption spectra of
the isoelectronic 17-electron species 1 and 5 feature a re-
markably similar pattern. Absorptions of 5 at l=313 and at
440 nm are bathochromically shifted compared to the bands
observed for 1 at l=296 and 383 nm (Figure 3).
The EPR spectrum of Co compound 5 revealed an axial g
tensor, again without any resolved hyperfine couplings with
the four
with
tral simulation was obtained with the following parameters:
g11=2.642, g22=g33=2.038, and ACo
1H NMR spectrum of cobalt com-
31P nuclei, but with resolved hyperfine coupling
59Co (I=7/2) along g11(Figure 5). A satisfactory spec-
11=118 MHz.
In an attempt to gain a more detailed understanding of
the electronic structure of 5, we performed DFT EPR prop-
erty calculations. We used both ADF and ORCA for this
purpose. ORCA allowed us to use the hybrid B3LYP func-
tional for these calculations. The EPR parameters of Co
complex 5 are not predicted very well (Table 2), neither by
ORCA nor by ADF. The DFT calculations in all cases pre-
dict much too low g11values and too large ACovalues. It is
also clear that slight variations in the structure (Me4model
versus tBu4 model) and geometry (optimized at OPBE,
Scheme 2.
Figure 4. Solid-state structure of [Co(h4-P2C2tBu2)2] (5). Displacement el-
lipsoids at 50% probability level, H atoms omitted for clarity.
Figure 5. Experimental and simulated X-band EPR spectra of neutral Co
complex 5 (inset: expanded area of the spectrum showing the hyperfine
coupling to the
50 K, microwave power 0.2 mW, field modulation amplitude=2 Gauss,
microwave frequency=9.382254 GHz. The simulated spectrum was ob-
tained with the parameters shown in Table 2.
59Co nucleus). Experimental conditions: Temperature=
Chem. Eur. J. 2010, 00, 0–0? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
These are not the final page numbers!??
&5&
FULL PAPER
Diphosphacyclobutadiene Complexes

Page 6

BP86 or B3LYP levels of theory) have a huge influence on
the DFT calculated EPR parameters. The poor performance
of both ADF and ORCA in predicting the EPR parameters
of complex 5 is perhaps not surprising. DFT EPR property
predictions are typically poor in cases with nearly degener-
ate metal centered orbitals, leading to large g11values (as is
the case for 5), for which small geometrical changes have a
large influence on the orbital splittings.[24]
The relatively large g anisotropy and the substantial hy-
perfine coupling observed for 5 is typical for a metal-cen-
tered radical. Related (tetrahedral) d9cobalt(0) complexes
display smaller g11values, but larger hyperfine couplings.[25]
Smaller g anisotropies than for 5 have been observed for
various 19-electron cobalt complexes.[26]The 19e formally
cobalt(0) species [CpCo(CO)2]?displays g values (g11=
2.018, g22=1.995, g33=2.041) close to the free electron value
geand thus seems to be a ligand radical complex.[27]
Mçssbauer spectra of complexes 1 and 4: The
bauer spectra of 1 have been measured between 4.2 and
265 K and some of the spectra are shown in Figure 6. Two
aspects of the spectra are rather unusual. First, the quadru-
pole doublet has very different component line widths and,
second, there is very little change in the absorption profile
with temperature. At this point it is important to note that
virtually identical spectra have been obtained for three sep-
57Fe Mçss-
arate absorbers made from two separate preparations of 1
in two different laboratories; hence, the spectra are reprodu-
cible.
Although one might expect more dramatic changes in the
line shape with changing temperature, the spectral profiles
shown in Figure 6 are indicative of the onset of slow para-
magnetic relaxation of the effective hyperfine field associat-
ed with the low-spin S=1=2ground state of 1. As a conse-
quence of the line shape, the spectra have been fit with a re-
laxation profile based on the formalism developed by Datta-
gupta and Blume, and an Arrhenius plot of the resulting re-
laxation rates is shown in Figure S1 of the Supporting
Information.[28,29]The results of these fits are shown as the
solid lines in Figure 6. The resulting parameters are given in
Table 3, a graphical representation of the temperature de-
pendencies is given in Figure S2. The relaxation frequency is
essentially the same upon warming from 4.2 to 85 K and in-
creases slightly between 85 and 265 K; the observed temper-
ature dependence is consistent with the presence of spin-
spin mediated relaxation in 1 as a result of intermolecular
iron-iron magnetic exchange interactions. The temperature
dependence of the isomer shift of 1 is well fit with the
Debye model[30]for the second-order Doppler shift with a
characteristic Mçssbauer temperature, VM, of 604(17) K, a
value that is reasonable for the environment of the low-spin
iron(III) ion, and the temperature dependence of the spec-
tral area yields a Debye temperature, VD, of 134(3) K (for a
graphical representation see Figure S2 in the Supporting In-
formation).[31]
To confirm that the relaxation observed in the Mçssbauer
spectra of 1 is mediated by spin–spin interactions, the spec-
trum of 1 has also been measured at 80 K in a frozen THF
solution. The resulting spectrum (see Figure S3 in the Sup-
porting Information) has been fit with the same relaxation
model as used above with a fixed line width of 0.30 mms?1
and a fixed hyperfine field of 11 T; a slightly better fit may
be obtained if the line width is increased to 0.385 mms?1.
The parameters derived from these fits are given in Table 3.
It is clear from these results that the relaxation rate is sub-
Table 3.
57Fe Mçssbauer spectroscopic parameters for 1 and 4.
ComplexT
[K]
d
[mms?1][a]
e2Qq/2
[mms?1]
n
[MHz]
Spectral area
[(%e)
(mms?1)][b]
1
265[c]
225[c]
155[c]
85[c]
4.2[c]
80[c]
80[d]
0.360
0.378
0.410
0.431
0.433
0.63(2)
0.61(2)
?1.144
?1.139
?1.145
?1.156
?1.163
?0.75(4)
?0.80(4)
245(20)
220(10)
205(5)
168(7)
171(5)
33(1)
36(1)
1.276
1.784
3.126
4.882
7.550
–
–
1 in frozen
solution
4
77[c]
0.309 1.30–1.689
[a] The isomer shifts are given relative to 295 K a-iron powder. [b] %e in-
dicates %effect. [c] Line width fixed at 0.30 mms?1. [d] Line width fixed
at 0.385 mms?1.
Figure 6. The57Fe Mçssbauer spectra of 1 obtained at the indicated tem-
peratures and fit with a relaxation profile.
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim Chem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&6&
R. Wolf, K. Lammertsma et al.

Page 7

stantially reduced in the frozen solution, as compared with
the bulk sample, as a result of the dramatically increased
iron–iron distance. This results in a reduction in the spin–
spin interactions and thus the spin–spin mediation of the re-
laxation. It is also apparent that in the frozen solution the
magnitude of the quadrupole interaction is somewhat re-
duced and, more surprisingly, the isomer shift is increased
more than might be expected.
In contrast to 1, the Mçssbauer spectrum of 4 obtained at
77 K is a simple symmetric quadrupole doublet (see Fig-
ure S4 in the Supporting Information) with a line width of
0.30 mms?1; the spectral hyperfine parameters are given in
Table 3. The isomer shifts of 1 are similar to those observed
for ferrocenium-like low-spin iron(III) complexes.[33,34]It is
noteworthy that the observed isomer shift for 4 is somewhat
smaller than that of 1, a decrease that agrees with and sup-
ports the calculated values (see Table 4). It is immediately
apparent from this table that there is good qualitative agree-
ment with the observed results given in Table 3. Further-
more, it is not surprising that the asymmetry parameter h is
zero for both complexes, owing to D2dsymmetry. In con-
trast, it is rather unexpected that a one-electron dz2-centered
oxidation results only in a minor increase of the isomer
shift. This can be nicely explained by our theoretical calcula-
tions, which show that the electron populations of the dxz
and dyzorbitals increase significantly after the loss of one
electron from the dz2 orbital on oxidation (vide infra), a loss
that decreases their s-electron shielding, increases the s-elec-
tron density at the
isomer shift of 4 relative to that of 1.
57Fe nucleus and, thus, decreases the
Sandwiches 1 and 4 may be regarded as iron(+ +III) and
iron(+ +IV) complexes. The isomer shift of 1 may indeed
argue for iron(+ +III). However, with respect to 4 it should
be noted that isomer shift values for comparable iron(IV)
complexes are unavailable. Mçssbauer spectroscopic data
for related low-oxidation state complexes are scarce, and a
firm correlation between isomer shifts and iron oxidation
states is not established for organoiron compounds. There-
fore, it remains difficult to assign the oxidation states to iron
in 1 and 4 on the basis of the Mçssbauer spectroscopic re-
sults alone.
Electronic structures of 1–5: To gain insight into the elec-
tronic structures of complexes 1–5 we performed a density
functional theory (DFT) study. For reasons of computational
efficiency, we decided to use truncated models [Fe(h4-
P2C2Me2)2]?(1?), [Co(h4-P2C2Me2)2]?(2?), [Fe(h4-P2C2Me2)2]
(4?), and [Co(h4-P2C2Me2)2] (5?) for which the bulky alkyl
substituents in 1–5 were replaced by methyl groups. The
model geometries (D2dsymmetry), optimized at the OPBE/
TZ2P level of theory by using ADF,[35–37]reproduce the
structural features of 1–5 very well. The calculations gave a
significant energetic preference for the doublet state of 17-
electron complex 1? over both the quartet (DE =30.4 kcal
mol?1) and the sextet (DE =64.0 kcalmol?1). Similar energy
differences were found for 16e 4? for which the triplet
ground state is favored over both the open-shell singlet and
the quintet by 20.0 and 37.6 kcalmol?1, respectively. Molecu-
lar orbital (MO) analysis of 1? and 5?, showed dominant
metal dx2?y2 contributions to the singly occupied molecular
orbital (SOMO, 1?: 84% dx2?y2, 5?: 55% dx2?y2). The two
singly occupied MOs of 4? also display the most significant
contributions from the iron orbitals (SOMO-1: 65% dz2,
SOMO: 65% dx2?y2).
We examined theground
state electronic structures of
models 1?, 2?, 4?, and 5? in great-
er detail at the spin-unrestricted
B3LYP[38]level of theory. At-
temptstofind
broken-symmetry
were unsuccessful. All energeti-
cally low-lying states feature
slightly increased (<9%) ex-
low-energy
states[39,40]
pectation values of the Sˆ2operator, pointing to the absence
of ligand radicals.[40]The values and the shape of the atomic
spin density distributions are in good agreement with metal-
based radicals for 1?, 4?, and 5? (see Figure S5 in the Sup-
porting Information).
Compositions of the quasi-restricted[41]frontier MOs with
substantial metal character are collected in Table 5. For sim-
plicity, we begin our discussion with the diamagnetic 18-
electron complex [Co(h4-P2C2Me2)2]?(2?). Three doubly oc-
cupied dxy(b2), dz2 (a1), and dx2?y2 (b1) orbitals can be clearly
identified within the frontier orbitals of 2? (Figure 7). In con-
trast to ferrocene, the dxyand dx2?y2 orbitals have different
symmetry and are thus not degenerate. The degenerate dxz
and dyzorbitals of e symmetry are highly delocalized over
several MOs and are discussed below.
To gain insight into the electronic spectrum of 2? we per-
formed an excited state calculation at the B3LYP-TD-DFT
level of theory.[38,42]The calculated UV/Vis spectrum closely
resembles the experimental ones of 2 and 3 (Figure 3 and
Figure S6 in the Supporting Information). The calculated
states at 20212 and 29466 cm?1mainly arise due to excita-
tions from dxy(b2), dz2 (a1) and dx2?y2 (b1) type orbitals to the
degenerate unoccupied 3e orbital. The latter orbital displays
substantial ligand character (Table 5). To a first approxima-
tion, these transitions may therefore be regarded as metal-
to-ligand charge transfer (MLCT). The state at 31452 cm?1
essentially arises from the 1e!3e transition and may be as-
signed to a delocalized p(ML)!p*(ML) transition. The 1e
and the 3e orbitals show contributions from both metal and
Table 4. Calculated Mçssbauer spectral parameters.[a]
Complex
1(0) [a.u.?3]
1(0) [??3]
dFe
[b][mms?1]Vzz
?15.92(3)?1021
+ +4.15(3) ?1021
?7.81(3) ?1021
[c][Vm?2]eQVzz/2[c][mms?1]
h
[Fe(P2C2Me2)2]2?
11816.43(2)
11816.83(2)
11816.87(2)
79730.9(1)
79733.5(1)
79733.9(1)
0.520(5)
0.375(5)
0.360(5)
+ +2.65(1)
?0.69(1)
+ +1.81(1)
0
0
0
[Fe(P2C2Me2)2]?, 1?
[Fe(P2C2Me2)2], 4?
[a] The parameters have been obtained from spin-unrestricted B3LYP-DFT calculations. [b] The isomer shifts
are given relative to a-iron, and calculated[53]from dFe,=?0.367(1(0)?11800)+ +6.55. [c] 1?1021Vm?2corre-
sponds to a (eQVzz)/2 of 0.166 mms?1for a nuclear quadrupole moment of 0.16(1)?10?28m2.
Chem. Eur. J. 2010, 00, 0–0? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
These are not the final page numbers!??
&7&
FULL PAPER
Diphosphacyclobutadiene Complexes

Page 8

ligand orbitals (Figure 7) with 1e being p bonding with re-
spect to the metal-ligand bond and 3e p antibonding.
Higher energy states at 35460 and 36419 cm?1are of a com-
plicated origin, with major contributions from the b2!a1
(MLCT)transition.Finally,
37974 cm?1may be assigned to transitions from the occu-
pied 1e and 2e orbitals to unoccupied a1and b2orbitals,
which display ligand p* character. We conclude that the visi-
ble part of the spectrum of 2? clearly originates from MLCT,
the statesat38489 and
whereas intraligand p!p* transitions play an important
role in the UV spectral region.
Upon 2?!5? one-electron oxidation, the dx2?y2 orbital in
[Co(h4-P2C2Me2)2] (5?) becomes singly occupied (see Fig-
ure S7 in the Supporting Information); the composition of
the frontier orbitals does not change. Neutral 17-electron
cobalt complex 5? is isoelectronic with the anionic iron com-
plex [Fe(h4-P2C2Me2)2]?(1?). Both complexes show a very
similar composition of the frontier orbitals (Figure S7).
One-electron oxidation of 1? produces 16-electron complex
[Fe(h4-P2C2Me2)2] (4?), which features singly occupied dx2?y2
and dz2 orbitals (Figure 8). Population analyses of the indi-
vidual d orbitals confirm that the total d electron density is
reduced upon oxidation of 1? and 2? (Table 6). However, this
depletion of the total d orbital charge amounts to far less
than one electron due to charge transfer. As a consequence
of p bonding, the occupations of the dxzand dyzorbitals ac-
tually increase upon oxidation of 1? and 2? (Table 6). This
Figure 7. Qualitative MO scheme for 2?.
Figure 8. Qualitative MO scheme for 4?.
Table 5. Composition of the frontier quasi-restricted molecular orbitals with significant metal d-character (spin-unrestricted B3LYP-DFT calculations,
Lçwdin populations).
[Fe(P2C2Me2)2]?(1?)[Co(P2C2Me2)2]?(2?)[Fe(P2C2Me2)2] (4?)[Co(P2C2Me2)2] (5?)
orbital metal
contr. [%]
compos.occup.metal
contr. [%]
compos.occup.metal
contr. [%]
compos.occup. metal
contr. [%]
compos. occup.
3e (LUMO)
b1
a1
2e
b2
1e
41
96
82
17
78
27
dxz+ +dyz
dx2?y2
dz2
dxz+ +dyz
dxy
dxz+ +dyz
0
1
2
2
2
2
35
88
78
12
76
34
dxz+ +dyz
dx2?y2
dz2
dxz+ +dyz
dxy
dxz+ +dyz
0
2
2
2
2
2
51
96
92
18
81
20
dxz+ +dyz
dx2?y2
dz2
dxz+ +dyz
dxy
dxz+ +dyz
0
1
1
2
2
2
30
96
77
9
69
45
dxz+ +dyz
dx2?y2
dz2
dxz+ +dyz
dxy
dxz+ +dyz
0
1
2
2
2
2
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&8&
R. Wolf, K. Lammertsma et al.

Page 9

finding correlates well with the observed trend in the Mçss-
bauer isomer shifts for complexes 1? and 4? (vide supra).
Our population analysis, Table 5, shows that b2(dxy), a1
(dz2), and b1 (dx2?y2) are essentially metal-centered. Inter-
mediate-spin d4(4?), low-spin d5(1? and 5?), and low-spin d6
(2?) configurations should be assigned to the metal if the dxz
and dyzorbitals remained unoccupied. Fully occupied dxz
and dyzorbitals would result in low-spin d8(4?), d9(1? and
5?), and d10(2?) electronic configurations for the metal ions.
However, owing to the mentioned interaction between the
dxzand dyzmetal orbitals (e symmetry) and ligand p orbitals,
it is difficult to determine the exact d occupations of the
metal atoms. The occupied 1e orbitals feature a pronounced
dxz/dyzcharacter, which is somewhat higher in the cobalt
compounds 2? and 5? (34 and 45%, Table 5) as compared to
the Fe complexes 1? and 4? (27 and 20%). The unoccupied
3e orbitals also feature substantial metal character (1?: 41%,
4?: 51%, 2?: 35%, 5?: 30%).
Analysis of reduced orbital charges and orbital spin densi-
ties yields insight into the electron (charge) and spin popula-
tions of each metal d orbital.[43,44]An electron population of
two and a spin population of zero are benchmarks for a
fully occupied d orbital. An electron population of one and
a spin population of one indicate a singly occupied orbital,
whereas the electron and spin populations are both zero for
unoccupied orbitals. In our experience, the actual popula-
tions can deviate in transition metal complexes from the
given ideal values, owing to charge transfer and spin-polari-
zation effects, but the trend is easy to follow.[43]
For example, the one-electron oxidation 1?!4? is accom-
panied by a large decrease of the d electron population
(1.93!1.10) and a simultaneous increase of the spin popula-
tion (0.02!0.90) of the dz2 orbital (Table 6). It immediately
follows that the 1?!4? oxidation therefore corresponds to a
metal-based (dz2)2!(dz2)1oxidation. Analogously, the one-
electron oxidation 2?!5? is a metal-based (dx2?y2)2!(dx2?y2)1
redox event. It is now revealing to analyze the “problemat-
ic” dxzand dyzorbitals in 1?, 2?, 4?, and 5? in the same
manner. The spin populations do not exceed 0.20 (Table 6).
This is a strong indication that these orbitals should be
either fully occupied or unoccupied. However, the electron
populations of 1.21–1.50 for dxz/dyzare much lower than the
1.7–2.0 typically found for doubly occupied d orbitals, but
much higher than expected for
unoccupied orbitals (eventually
0.0 for a totally ionic system).
Owing to the covalent nature
of the metal-ligand bonding a
“black or white” picture fails to
provide an apt description of
the electronic
complexes
1–5.
electronic situation must be de-
scribedasintermediate
tween two extreme cases, that
is,
for
1,
structures
Instead,
of
the
be-
[Fe?I(L0)2]?$[FeIII(L2?)2]?
[Co?I(L0)2]?$[CoIII-
(L2?)2]?
for
2,and
3, [Fe0(L0)2]$[FeIV(L2?)2]for
4,
[Co0(L0)2]$[CoIV(L2?)2] for 5, for which L0is the neutral di-
phosphacyclobutadiene ligand (P2C2R2) and L2?represents
the doubly reduced form (P2C2R2)2?. The bonding situation
may be understood in terms of a low-valent metal ion that is
stabilized by very strong p-acceptor ligands, L0, or, alterna-
tively, as a high-valent metal ion that interacts with two very
strong p donors, L2?. The two scenarios are essentially
equivalent. In both cases, significant electron transfer results
in considerable, though incomplete, population of the metal
dxz/dyzorbitals as shown by our calculations.
Conclusion
Reactions
C10H14)2]?(M=Fe, Co) yield homoleptic sandwich com-
plexes [M(h4-P2C2R2)2]?, which are formed by means of
head-to-tail cyclodimerization of the phosphaalkyne at the
iron or cobalt center. It should be noted that the related re-
action of [Fe(h4-C10H14)2]?with diphenylacetylene yielded
an arene complex through alkyne cyclotrimerization in-
stead.[45]
Complexes [Fe(h4-P2C2tBu2)2]?
P2C2tBu2)2]?(2), and [Co(h4-P2C2tBu2)2]?(3) have been fully
characterized spectroscopically. Electrochemical and prepa-
rative investigations show that, owing to their electron-rich
nature, anions 1 and 2 are readily and reversibly oxidized to
neutral sandwiches [M(h4-P2C2tBu2)2] (4: M=Fe, 5: M=
Co). Complexes 1–5 feature essentially D2d-symmetric sand-
wich structures with h4-coordinated 1,3-diphosphacyclobuta-
diene ligands. There is little structural change on oxidation
of anions 1 and 2. Density functional studies show similar
molecular orbital compositions for both neutral and monoa-
nionic sandwiches. The highest occupied molecular orbitals
display dominant metal character. This observation is nicely
corroborated by the EPR spectra of the 17-electron species
1 and 5, which are consistent with the dx2?y2 character of the
SOMOs. Theisomer shifts
P2C2tBu2)2]?(1) agree well with the shifts of ferrocenium-
of phosphaalkyneswith metalates[M(h4-
(1),[Co(h4-
of monoanion[Fe(h4-
like iron(III) complexes. However, the DFT calculations
clearly show that covalent metal–ligand p bonding is highly
significant in these complexes. It is therefore equally appro-
priate to assign the + +III or the ?I oxidation state to the
Table 6. Reduced orbital charges and orbital spin densities (spin-unrestricted B3LYP-DFT calculations,
Lçwdin populations).
[Fe(P2C2Me2)2]?(1?)[Co(P2C2Me2)2]?(2?) [Fe(P2C2Me2)2] (4?) [Co(P2C2Me2)2] (5?)
electron spin electronspin electronspinelectron spin
dz2
dxz
dyz
dx2?y2
dxy
Sdi
Mtotal
Ltotal
1.93
1.21
1.21
1.04
1.71
7.11
0.02
0.14
0.14
0.96
0.05
1.31
1.33
?0.33[c]
1.95
1.28
1.28
1.99
1.67
8.18
0
0
0
0
0
0
0
0
1.10
1.45
1.45
1.04
1.79
6.85
0.90
0.20
0.20
0.96
0.07
2.32
2.37
?0.37[c]
1.97
1.50
1.50
1.05
1.86
7.86
0.01
0.13
0.13
0.95
0.02
1.24
1.26
?0.26[c]
[a]
[b]
[a] The total spin density at metal ion including s and p orbitals. [b] The total spin density at the ligands.
[c] Due to spin polarization.
Chem. Eur. J. 2010, 00, 0–0? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
These are not the final page numbers!??
&9&
FULL PAPER
Diphosphacyclobutadiene Complexes

Page 10

metal centers in anions 1–3, and oxidation state + +IV or 0 to
the metal atoms in 4 and 5. In the former case, electroneu-
trality[46]of a high-valent metal ion will be maintained by p
donation from the ligand, whereas, in the latter case, exces-
sive charge on a low-valent metal center will be dissipated
by metal–ligand p backbonding.
With this work, the properties of a series of homoleptic
diphosphacyclobutadiene complexes have been established
for the first time. Syntheses, structures, spectroscopic charac-
teristics, and bonding of these complexes are now well-un-
derstood. In addition to the homoleptic species described
herein, related heteroleptic sandwich anions have very re-
cently become accessible as well.[47]Future challenges clear-
ly lie in the application of the diphosphacyclobutadiene
sandwiches, for example, as building blocks of more elabo-
rate supramolecular structures and multimetallic arrange-
ments. Investigations towards these goals are on-going.
Experimental Section
All experiments were performed under an atmosphere of dry argon, by
using standard Schlenk and glovebox techniques. Solvents were purified,
dried, and degassed by standard techniques. NMR spectra were recorded
(298 K) by using Bruker Advance 250 (31P; 85% H3PO4) and MSL 400
spectrometers (1H,
resonances. IR spectra were recorded by using a Shimadzu FTIR-84005
spectrophotometer. EI mass spectrometry was performed by using a
JEOL JMS SX/SX 102 A four-sector mass spectrometer, coupled to a
JEOL MS-MP9021D/UPD system program. Melting points were mea-
sured on samples in sealed capillaries and are uncorrected. Compounds
tBu?C?P,[16]
13C; SiMe4), internally referenced to residual solvent
Ad?C?P,[17]
[K([18]crown-6)(thf)2][Fe(h4-C14H10)2],[18]
and
[K([18]crown[6])(thf)2][Co(h4-C14H10)2][19]were prepared according to lit-
erature procedures. [Cp2Fe]PF6and [Cp2Fe] were purchased from Al-
drich and used as received.
EPR spectroscopy: Experimental X-band EPR spectra were recorded by
using a Bruker EMX spectrometer equipped with a He temperature con-
trol cryostat system (Oxford Instruments). The spectra were simulated by
iteration of the anisotropic g values, (super)hyperfine coupling constants,
and line widths by using the EPR simulation program W95EPR, which is
available upon request from Prof. Frank Neese, University of Bonn.
Mçssbauer spectroscopy: The Mçssbauer spectra of 1 have been mea-
sured between 4.2 and 265 K and the spectrum of 4 has been measured
at 77 K on constant-acceleration spectrometers, which utilized a rhodium
matrix cobalt-57 source and were calibrated at 295 K with a-iron powder.
The Mçssbauer spectral absorbers were prepared and mounted in the
cryostat under an atmosphere of dinitrogen. The relative statistical errors
are ?0.005 mms?1for the isomer shifts and quadrupole splittings and
?0.005 (%e)(mms?1) for the total spectral absorption area; double the
statistical error associated with the relaxation frequencies are given in
Table 3; the absolute errors are approximately twice the statistical errors
indicated here.
Cyclic voltammetry and spectroelectrochemistry: Cyclic voltammograms
were recorded by using an EG&G PAR model 283 potentiostat operated
with the PAR Power CV software. The single-compartment air-tight elec-
trochemical cell contained a 0.42 mm2Pt microdisc working electrode
polished between scans with a 0.25 mm diamond paste (Oberfl?chentech-
nologien Ziesmer, Kempen, Germany), a Pt wire auxiliary electrode and
an Ag wire pseudoreference electrode (combined with the ferrocene/fer-
rocenium (Fc/Fc+ +) redox couple used as an internal reference). The satu-
rated calomel electrode (SCE) shows a potential of ?0.45 V against Fc/
Fc+ +in dichloromethane.
Thin-layer UV/Vis spectroelectrochemistry was performed by using an
optically transparent electrochemical (OTTLE) cell equipped with Pt
minigrid working and auxiliary electrodes, an Ag microwire pseudorefer-
ence electrode and CaF2windows.[48]Thin-layer cyclic voltammograms
were recorded in the course of each OTTLE experiment for a precise po-
tential control achieved with a PA4 potentiostat (EKOM, Poln?, Czech
Republic), and for monitoring the progress of the electrolyses by decreas-
ing Faradaic current. The spectroelectrochemical samples were approxi-
mately 10?3m in the studied complex and 3?10?1m in the supporting elec-
trolyte. UV/Vis spectra of the electrolyzed solutions were obtained by
using a HP 8453 A diode array spectrophotometer.
Theoretical calculations: Geometry optimizations of 1?, 2?, 4?, and 5?
were performed in D2dsymmetry by using ADF2007.01.[35]The exchange-
correlation potential is based on the GGA exchange functional OPTX in
combination with the non-empirical PBE (OPBE) and a non-contracted
triple-zeta valence-plus-2-polarization STO (TZ2P) basis set was used for
all atoms.[37],[38]The inner core electrons of carbon (1s), phosphorus and
iron (1s, 2s, 2p) were kept frozen. The truncated X-ray structures of 1, 2,
4, and 5 were used as the starting geometries. All optimized structures
were verified as minima by frequency calculations (no imaginary fre-
quencies). For the electronic structures and Mçssbauer spectroscopic pa-
rameters, single-point calculations were then performed on the optimized
geometries using the program package ORCA 2.6 revision 35[42]and the
B3LYP functional.[38]Tight convergence criteria were used for the SCF
procedure (TIGHTSCF). Triple-z basis sets with one-set of polarization
functions[49](TZVP) were used for iron, cobalt, and the phosphorus
atoms, and double-z basis sets with one-set of polarization functions[50]
(SVP) were used for all other atoms. The resolution of the identity ap-
proximation (RIJONX) was employed[51,52]with matching auxiliary basis
sets.[52]For calculation of Mçssbauer spectroscopic parameters, the
“core” CP(PPP) basis set for iron[53]with enhanced integration accuracy
on the metal (SPECIALGRIDINTACC 7) was used. All reduced orbital
charges and spin densities[44]were calculated according to Lçwdin popu-
lation analysis.[54]The reduced orbital population is defined as a popula-
tion per angular momentum,[44]meaning the decomposition of the total
spin or charge population at the given atom into the population of s, pi,
di, and fiorbitals of the atom. Time-dependent DFT (TD-DFT) calcula-
tions were performed by using the B3LYP functional and the conductor-
like screening model (COSMO) with THF as a solvent.[55]The first 90
states were calculated, for which the maximum dimension of the expan-
sion space in the Davidson procedure (MAXDIM) was set to 900. Molec-
ular orbitals and spin densities were visualized with the program Mole-
kel.[56]
[K([18]crown-6)(thf)2][Fe(h4-P2C2tBu2)2] (K1): The phosphaalkyne tBu?
C?P (5.00 mL, 10.00 mmol, 2m solution in hexane) was added dropwise
to a deep brown solution of [K([18]crown-6)(thf)2][Fe(h4-C14H10)2]
(2.13 g, 2.48 mmol) in THF (?40 mL) at ?788 8C and the mixture was al-
lowed to warm to room temperature overnight. The dark orange solution
wasfiltered, concentratedto approximately 10 mLandtoluene
(?30 mL) was added. K1 was isolated as a yellow-orange crystalline solid
after storage at ?208 8C overnight, washed with toluene (3?15 mL) and
dried in vacuo. Crystals suitable for X-ray crystallography were grown by
layering a THF solution of K1 with n-pentane. Yield 1.41 g (75%); M.p.
slow decompostion >3008 8C;
(very br s; tBu), 1.73 (br s; THF), 3.72 ppm (br s; THF); elemental analy-
sis (%) calcd for C32H60O6P4FeK (M=759.66): C 50.59, H 7.96; found: C
49.74, H 7.95; magnetic susceptibility (Evans method, 258 8C, [D8]THF):
meff=1.77 mB; UV/Vis (THF): lmax(e)=296 (48100), 383 (10000), 699 nm
(310 mol?1Lcm?1).
1H NMR (250.13 MHz, [D8]THF): d=?1.9
[K([18]crown-6)(thf)2][Co(h4-P2C2tBu2)2] (K2): Compound K2 was pre-
pared by an analogous procedure to K1 by adding tBu?C?P (1.05 mL,
2.1 mmol, 2m solution in hexane) to a deep red-brown solution of
[K([18]crown[6])(thf)2][Co(h4-C14H10)2](0.46 g,0.53 mmol)in THF
(?8 mL). Yield 0.23 g (48%) of K2; m.p.: slow decomposition >2508 8C;
1H NMR (250.13 MHz, [D8]THF): d=1.04 (br s, 36H; tBu), 1.71 (br s,
8H, THF), ?3.5 (very br. s; [18]crown-6), 3.59 ppm (br s, 8H; THF);
13C{1H} NMR (100.62 MHz, [D8]THF): d=34.1 (s; C(CH3)3), 36.3 (s; C-
(CH3)3), 103.1 ppm (t,
1J(C,P)=52.3 Hz; C2P2tBu2); the signals for
[18]crown-6 and THF were not observed due to overlap with solvent sig-
nals.31P{1H} NMR (101.23 MHz, [D8]THF): d=2.4 ppm; elemental anal-
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&10&
R. Wolf, K. Lammertsma et al.

Page 11

ysis (%) calcd for C40H76O8P4CoK (K2–1THF): C 51.79, H 8.21; found: C
51.53, H 8.57; UV/Vis (THF): lmax(e)=264 (shoulder), 284 (35500), 329
(8900 mol?1Lcm?1), 503 nm (shoulder).
[K([18]crown-6)(thf)2][Co(h4-P2C2Ad2)2] (K3): Compound K3 was pre-
pared by an analogous procedure to K1 and K2 by adding Ad?C?P
(2.1 mL, 1.04 mmol, 0.5m solution in hexane) to a deep orange-brown so-
lution of [K([18]crown-6)(thf)2][Co(h4-C14H10)2] (0.46 g, 0.53 mmol) in
THF (?8 mL). Yield 0.24 g (76%) of K3; m.p.: slow decomposition
>2308 8C;
lapping s, THF), 1.51 (br overlapping s, THF), 1.5–2.0 (overlapping m,
15H; Ad), 3.21 (very br, [18]crown-6), 3.49 (overlapping br s; THF),
3.52 ppm (overlapping br s; THF);13C{1H} NMR (100.62 MHz, [D8]THF/
C6D61:1): d=30.3 (s; C-3,C-5,C-7 of Ad), 37.3 (s; C-1 of Ad), 38.0 (s; C-
1H NMR (250.13 MHz, [D8]THF/C6D61:1): d=1.46 (br over-
4, C-6, C-10 of Ad), 46.3 (s, C-2, C-8, C-9 of Ad), 103.8 ppm (t,1J(C,P)=
51.3 Hz; C2P2Ad2);31P{1H} NMR (101.23 MHz, [D8]THF): d=?2.6 ppm;
elemental analysis (%) calcd for C64H100O6P4CoK (M=759.66): C 63.04,
H 8.26; found: C 62.37, H 8.20; UV/Vis (THF): lmax(e)=286 (47400),
330 nm (10900 mol?1Lcm?1).
[Fe(h4-P2C2tBu2)2] (4): [Cp2Fe]PF6(0.050 g, 0.15 mmol) was added to a
dark-orange solution of K2 (0.106 g, 0.14 mmol) in THF (6 mL) and the
mixture was stirred at room temperature overnight. The solvent was re-
moved completely, the ferrocene byproduct removed by sublimation
(10?2Torr, 508 8C), and the dark residue was extracted into n-pentane
(10 mL). Concentrating the orange-red extract to about 1 mL yielded red
crystals of 3 after storage at ?208 8C for several days. Yield 0.023 g (36%).
M.p. 184–1868 8C (dark oil);
br. s; tBu); magnetic susceptibility (Evans method, 258 8C, C6D6): meff=
2.74 mB; UV/Vis (THF): lmax (e)=275 (65700), 320 (sh), 443 nm
(11000 mol?1Lcm?1). HRMS (EI): m/z (%): 456.1 (60); calcd for
C20H36P4Fe: 456.1100; found: 456.1117.
[Co(h4-P2C2tBu2)2] (5): Complex 5 was prepared by an analogous proce-
dure to 4 by oxidation of K2 (0.221 g, 0.29 mmol) with [Cp2Fe]PF6
(0.099 g, 0.30 mmol) in THF. Yield 0.041 g (31%). M.p. 193–1948 8C (dark
oil);
netic susceptibility (Evans method, 258 8C, C6D6): meff=1.73 mB; UV/Vis
(THF): lmax (e)=221 (30700), 284 (37000), 313 (43300 mol?1Lcm?1),
440 nm (sh); HRMS (EI): m/z (%): 459 (51); calcd for C20H36P4Co:
459.10995; found: 459.10973.
X-ray crystallography: Reflections were measured by using a Nonius
KappaCCD (K3) and a Bruker APEXII (5, Table 7) diffractometers with
rotating anodes and MoKaradiation (l=0.71073 ?). Absorption correc-
tions were performed by using SADABS.[57]The structures were solved
by direct methods and refined against F2.[58]The crystal of K3 (Table 7)
contained large voids (1324 ?3per unit cell) filled with disordered sol-
vent molecules. Their contribution to the structure factors was secured by
back-Fouriertransformation using
PLATON,[49]resulting in 74 electrons per unit cell. Geometry calculations
and check for higher symmetry was performed with PLATON.[59]Details
of the structural analyses of K1, K2 and 4 have been reported previous-
ly.[15]CCDC-771227 (K3) and CCDC-771228 (5) contain the supplemen-
tary 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.
1H NMR (250.13 MHz, C6D6): d=2.6 (very
1H NMR (200.13 MHz, C6D6): d=?2.6 ppm (very br s; tBu); mag-
the SQUEEZEroutineof
Acknowledgements
We are indebted to Dr. E. Bill (MPI M?lheim/Ruhr) for measuring the
solution Mçssbauer spectrum of K1 and valuable discussions. We thank
Dr. E.P.A. Couzijn (VU University Amsterdam) for experimental assis-
tance, Prof. Dr. D. Gudat (Universit?t Stuttgart) for help with microanal-
ysis, and H. Peeters (Universiteit van Amsterdam) for mass spectromet-
ric measurements. Funding from the DFG (IRTG1444 and WO1496/1–1
to 4–1), the Fonds der Chemischen Industrie (Liebig-fellowships to R.W.
and M.M.K) and PhoSciNet is gratefully acknowledged. M.M.K is grate-
ful to the Max Planck Society for a postdoctoral fellowship. R.W. thanks
Prof. Dr. W. Uhl for his generous support.
[1] a) T. J. Kealy, P. L. Pauson, Nature 1951, 168, 1039; b) G. Wilkinson,
M. Rosenblum, M. C. Whiting, R. B. Woodward, J. Am. Chem. Soc.
1952, 74, 2125; c) E. O. Fischer, W. Pfab, Z. Naturforsch. B 1952, 7,
377.
[2] Metallocenes (Eds.: A. Togni, R. L. Halterman), Wiley-VCH, Wein-
heim, 1998.
[3] C. Elschenbroich, Organometallchemie, 6th ed., Teubner-Verlag,
Wiesbaden, 2008, pp. 446–449.
[4] a) A. Efraty, Chem. Rev. 1977, 77, 691, b) L. Veiros, G. Dazinger, K.
Kirchner, M. C. Calhorda, R. Schmid, Chem. Eur. J. 2004, 10, 5860,
and references therein; c) D. Seyferth, Organometallics 2003, 22, 2;
d) U. H. F. Bunz, Top. Curr. Chem. 1999, 201, 131.
[5] Rare examples of homoleptic cyclobutadiene nickel complexes:
a) H. Hoberg, R. Krause-Gçing, R. Mynott, Angew. Chem. 1978, 90,
138; Angew. Chem. Int. Ed. Engl. 1978, 17, 123; b) H. Hoberg, C.
Froehlich, J. Organomet. Chem. 1981, 213, C49.
[6] a) M. Regitz, Chem. Rev. 1990, 90, 191; b) A. C. Gaumont, J. Denis,
Chem. Rev. 1994, 94, 1413; c) J. F. Nixon, Coord. Chem. Rev. 1995,
145, 201; d) Multiple Bonds and Low Coordination in Phosphorus
Chemistry (Eds.: M. Regitz, O. J. Scherer), Thieme, Stuttgart, 1990;
e) K. B. Dillon, F. Mathey, J. F. Nixon, Phosphorus: The Carbon
Copy, Wiley, Chichester, 1998; f) F. Mathey, Angew. Chem. 2003,
115, 1616; Angew. Chem. Int. Ed. 2003, 42, 1578.
[7] The catalytic trimerization of phosphaalkynes to triphosphabenzenes
has also been described: F. Tabellion, C. Peter, U. Fischbeck, M.
Regitz, F. Preuss, Chem. Eur. J. 2000, 6, 4558.
[8] a) P. Binger, B. Biedenbach, R. Schneider, M. Regitz, Synthesis
1989, 960; b) D. Himmel, M. Seitz, M. Scheer, Z. Anorg. Allg.
Chem. 2004, 630, 1220.
[9] a) P. Binger, R. Milczarek, R. Mynott, M. Regitz, W. Rçsch, Angew.
Chem. 1986, 98, 645; Angew. Chem. Int. Ed. Engl. 1986, 25, 644;
b) P. B. Hitchcock, M. J. Maah, J. F. Nixon, J. Chem. Soc. Chem.
Table 7. Crystal data and structure refinement of K3 and 5.
K35
formula [C20H40KO8]
[C44H60CoP4]
+ +disordered solvent
0.48?0.27?0.27
red
1219.35[a]
C2/c
14.5182(1)
24.7076(2)
20.8839(2)
101.6076(5)
7338.05(10)
4
150(2)
1.10[a]
47294/54.4
8360
6889
354
0.4[a]
0.0419
0.1294
1.08
?0.41/0.70
C20H36P4Co
cryst. size [mm]
color
formula weight
space group
a [?]
b [?]
c [?]
b [8 8]
V [?3]
Z
T [K]
1calcd[gcm?3]
reflns collected/2qmax[8 8]
unique reflections
refl. obs. [I>2s(I)]
no. of parameters
m [mm?1]
R1 [I>2s(I)]
wR2 (all data)
GOF on F2
residual density [e??3] (min/
max)
0.11?0.05?0.02
orange
459.30
C2/c
18.0153(17)
8.8666(9)
16.783(2)
117.0460(10)
2387.6(5)
4
150(2)
1.28
11356/54.88
2715
2328
186
1.0
0.0362
0.1026
1.04
-0.32/1.15
[a] Derived values do not contain the contribution of the disordered sol-
vent.
Chem. Eur. J. 2010, 00, 0–0? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
These are not the final page numbers!??
&11&
FULL PAPER
Diphosphacyclobutadiene Complexes

Page 12

Commun. 1986, 737; c) P. Binger, R. Milczarek, R. Mynott, C.
Kr?ger, Y.-H. Tsay, E. Raabe, M. Regitz, Chem. Ber. 1988, 121, 637.
[10] T. Wettling, G. Wolmersh?user, P. Binger, M. Regitz, J. Chem. Soc.
Chem. Commun. 1990, 1541.
[11] a) P. L. Arnold, F. G. N. Cloke, P. B. Hitchcock, J. F. Nixon, J. Am.
Chem. Soc. 1996, 118, 7630; b) F. G. N. Cloke, K. R. Flower, P. B.
Hitchcock, J. F. Nixon, J. Chem. Soc. Chem. Commun. 1994, 489;
c) A. G. Avent, F. G. N. Cloke, K. R. Flower, P. B. Hitchcock, J. F.
Nixon, D. M. Vickers, Angew. Chem. 1994, 106, 2406; Angew. Chem.
Int. Ed. Engl. 1994, 33, 2330.
[12] F. G. N. Cloke, P. B. Hitchcock, J. F. Nixon, D. M. Vickers, J. Organo-
met. Chem. 2001, 635, 212.
[13] J. E. Ellis, Inorg. Chem. 2006, 45, 3167.
[14] E. Urne ˙z ˇius, W. W. Brennessel, C. J. Cramer, J. E. Ellis, P. v. R.
Schleyer, Science 2002, 295, 832.
[15] Preliminary communications: a) R. Wolf, A. W. Ehlers, J. C. Sloot-
weg, M. Lutz, D. Gudat, M. Hunger, A. L. Spek, K. Lammertsma,
Angew. Chem. 2008, 120, 4660; Angew. Chem. Int. Ed. 2008, 47,
4584; b) R. Wolf, J. C. Slootweg, A. W. Ehlers, F. Hartl, B. de Bruin,
M. Lutz, A. L. Spek, K. Lammertsma, Angew. Chem. 2009, 121,
3150; Angew. Chem. Int. Ed. 2009, 48, 3104.
[16] a) G. Becker, G. Gresser, W. Uhl, Z. Naturforsch. B 1981, 36, 16;
b) W. Rçsch, T. Allspach, U. Bergstr?ßer, M. Regitz in Synthetic
Methods of Organometallic and Inorganic Chemistry, Vol. 3 (Ed.:
W. A. Herrmann), Thieme, Stuttgart, 1996, p. 11.
[17] T. Allspach, M. Regitz, G. Becker, W. Becker, Synthesis 1986, 31.
[18] W. W. Brennessel, R. E. Jilek, J. E. Ellis, Angew. Chem. 2007, 119,
6244; Angew. Chem. Int. Ed. 2007, 46, 6132.
[19] W. W. Brennessel, V. G. Young, Jr., J. E. Ellis, Angew. Chem. 2002,
114, 1259; Angew. Chem. Int. Ed. 2002, 41, 1211.
[20] a) D. Bçhm, F. Knoch, S. Kummer, U. Schmidt, U. Zenneck, Angew.
Chem. 1995, 107, 251; Angew. Chem. Int. Ed. Engl. 1995, 34, 198;
b) F. W. Heinemann, S. Kummer, U. Seiss-Brandl, U. Zenneck, Or-
ganometallics 1999, 18, 2021.
[21] M. Driess, D. Hu, H. Pritzkow, H. Sch?ufele, U. Zenneck, M.
Regitz, W. Rçsch, J. Organomet. Chem. 1987, 334, C35.
[22] The reduction of anion 1 to the hypothetical 18e dianion [Fe(h4-
P2C2tBu2)2]2?
was not observed in the experimental potential
window. In contrast, the neutral 18-electron complex [Fe(h6-
C7H8)(h4-P2C2tBu2)2] can be reduced reversibly at low temperatures
to the 19-electronradical anion
(?2.50 V vs. SCE in DME at ?608 8C) and also oxidized to the corre-
[Fe(h6-C7H8)(h4-P2C2tBu2)2]?
sponding 17-electron radical cation [Fe(h6-C7H8)(h4-P2C2tBu2)2]+ +
(+ +0.55 V vs. SCE in CH2Cl2).[21]The different, more negatively shift-
ed redox potentials of 1 compared to the neutral, heteroleptic spe-
cies [Fe(h6-C7H8)(h4-P2C2tBu2)2] are most likely a consequence of
the different charges and formal metal oxidation states of both com-
plexes. The frontier orbitals of [Fe(h6-C7H8)(h4-P2C2tBu2)2] are not
known in detail, but it is expected that the reduction is ligand-cen-
tered, whereas the oxidation is metal-centered.
[23] a) D. Hu, H. Sch?ufele, H. Pritzkow, U. Zenneck, Angew. Chem.
1989, 101, 929; Angew. Chem. Int. Ed. Engl. 1989, 28, 900; b) U.
Zenneck, Angew. Chem. 1990, 102, 171; Angew. Chem. Int. Ed.
Engl. 1990, 29, 126.
[24] a) M. Atanasov, C. A. Daul, M.-M. Rohmer, T. Venkatachalam,
Chem. Phys. Lett. 2006, 427, 449; b) M. Atanasov, E. J. Baerends, P.
Baettig, R. Bruyndonckx, C. Daul, C. Rauzy, M. Zbri, Chem. Phys.
Lett. 2004, 399, 433.
[25] S. Deblon, L. Liesum, J. Harmer, H. Schçnberg, A. Schweiger, H.
Gr?tzmacher, Chem. Eur. J. 2002, 8, 601, and references therein.
[26] a) H. van Willigen, W. E. Geiger, M. D. Rausch, Inorg. Chem. 1977,
16, 581; b) L. Vasquez, H. Pritzkow, U. Zenneck, Angew. Chem.
1988, 100, 705; Angew. Chem. Int. Ed. Engl. 1988, 27, 706; c) W. E.
Geiger, P. H. Rieger, C. Corbato, J. Edwin, E. Fonseca, G. A. Lane,
J. M. Mevs, J. Am. Chem. Soc. 1993, 115, 2314.
[27] N. G. Connelly, W. E. Geiger, G. A. Lane, S. J. Raven, P. H. Rieger,
J. Am. Chem. Soc. 1986, 108, 6219.
[28] In these fits the isomer shift, d, the quadrupole splitting, e2Qq/2, and
the relaxation rate, d, have been varied, whereas the line width has
been constrained to 0.30 mms?1, the approximate line width ob-
served for the higher velocity line at all temperatures. Further, the
effective hyperfine field has been constrained to 11 T, a field that is
consistent with the S=1=2ground state of 1. The relaxation of the ef-
fective hyperfine field cannot be parallel with the principal axis of
the electric field gradient at the iron site as this would yield a sym-
metric spectrum, but must be transverse to this axis. The value of
e2Qq/2 may be either positive or negative if the relaxation is normal
to this gradient and along the y or x axis, respectively, that is, in the
plane containing the singly occupied dx2?y2 orbital, and the latter
case has been chosen to be consistent with the calculated negative
quadrupole splitting, see Table 4; the asymmetry parameter, h, has
also been constrained to zero in agreement with the structure of 1
and the calculations.
[29] S. Dattagupta, M. Blume, Phys. Rev. B 1974, 10, 4540.
[30] G. K. Shenoy, F. E. Wagner, G. M. Kalvius in Mçssbauer Isomer
Shifts, (Eds.: G. K. Shenoy, F. E. Wagner), North-Holland, Amster-
dam, 1978, p. 49.
[31] It is well known[30]that the Mçssbauer and Debye temperatures, VM
and VD, obtained from the two temperature dependencies, are usu-
ally different because they depend on <v2> and <x2>, respective-
ly, for which <v2> is the root-mean-square vibrational velocity of
the57Fe nuclide and <x2> is the root-mean-square displacement of
the57Fe nuclide; unfortunately, there is no model independent rela-
tionship[30]between these mean-square values. However, the values
of these temperatures reported[32]for other iron complexes indicate
that VM, which is more sensitive to the high-frequency phonons, is
often up to four times larger than VD.
[32] T. Owen, F. Grandjean, G. J. Long, K. V. Domasevitch, N. Gerasim-
chuk, Inorg. Chem. 2008, 47, 8704, and the references therein.
[33] R. V. Parish in The Organic Chemistry of Iron, Vol. 1 (Eds.: E. A. K.
von Gustorf, F. Grevels, I. Fischler), Academic Press, New York,
1978, p. 192.
[34] R. Deschenaux, M. Schweissguth, M. T. Vilches, A. M. Levelut, D.
Hautot, G. J. Long, and D. Luneau, Organometallics 1999, 18, 5553.
[35] ADF2007: E. J. Baerends, J. Autschbach, A. B?rces, F. M. Bickel-
haupt, C. Bo, P. M. Boerrigter, L. Cavallo, D. P. Chong, L. Deng,
R. M. Dickson, D. E. Ellis, M. van Faassen, L. Fan, T. H. Fischer, C.
Fonseca Guerra, S. J. A. van Gisbergen, J. A. Groeneveld, O. V.
Gritsenko, M. Gr?ning, F. E. Harris, P. van den Hoek, C. R. Jacob,
H. Jacobsen, L. Jensen, G. van Kessel, F. Kootstra, E. van Lenthe,
D. A. McCormack, A. Michalak, J. Neugebauer, V. P. Nicu, V. P.
Osinga, S. Patchkovskii, P. H. T. Philipsen, D. Post, C. C. Pye, W.
Ravenek, P. Ros, P. R. T. Schipper, G. Schreckenbach, J. G. Snijders,
M. Sol?, M. Swart, D. Swerhone, G. te Velde, P. Vernooijs, L. Ver-
sluis, L. Visscher, O. Visser, F. Wang, T. A. Wesolowski, E. M. van
Wezenbeek, G. Wiesenekker, S. K. Wolff, T. K. Woo, A. L. Yakov-
lev, T. Ziegler, SCM, Theoretical Chemistry, Vrije Universiteit, Am-
sterdam, The Netherlands, 2007.
[36] N. C. Handy, A. J. Cohen, Mol. Phys. 2001, 99, 403.
[37] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.
[38] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648; b) C. T. Lee, W. T.
Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.
[39] a) C. Adamo, V. Barone, A. Bencini, F. Totti, I. Ciofini, Inorg.
Chem. 1999, 38, 1996; b) A. P. Ginsberg, J. Am. Chem. Soc. 1980,
102, 111; c) L. Noodleman, J. Chem. Phys. 1981, 74, 5737; d) L. Noo-
dleman, J. G. Norman, J. H. Osborne, A. Aizman, D. A. Case, J.
Am. Chem. Soc. 1985, 107, 3418; e) L. Noodleman, E. R. Davidson,
Chem. Phys. 1986, 109, 131; f) L. Noodleman, D. A. Case, A.
Aizman, J. Am. Chem. Soc. 1988, 110, 1001; g) L. Noodleman, C. Y.
Peng, D. A. Case, J. M. Mouesca, Coord. Chem. Rev. 1995, 144, 199;
h) A. A. Ovchinnikov, J. K. Labanowski, Phys. Rev. A 1996, 53,
3946; i) D. Herebian, K. E. Wieghardt, F. Neese, J. Am. Chem. Soc.
2003, 125, 10997.
[40] V. Bachler, G. Olbrich, F. Neese, K. Wieghardt, Inorg. Chem. 2002,
41, 4179.
[41] F. Neese, J. Am. Chem. Soc. 2006, 128, 10213.
[42] F. Neese, ORCA—an Ab Initio, Density Functional and Semiempir-
ical Program Package, version 2.6, revision 35; Institut f?r Physikali-
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim Chem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&12&
R. Wolf, K. Lammertsma et al.

Page 13

sche und Theoretische Chemie, Universit?t Bonn, Germany, July
2007.
[43] a) M. M. Khusniyarov, E. Bill, T. Weyherm?ller, E. Bothe, K.
Harms, J. Sundermeyer, K. Wieghardt, Chem. Eur. J. 2008, 14, 7608;
b) M. M. Khusniyarov, T. Weyherm?ller, E. Bill, K. Wieghardt, J.
Am. Chem. Soc. 2009, 131, 1208.
[44] ORCA 2.6 manual, February 2007, p. 296.
[45] R. Wolf, N. Ghavtadze, K. Weber, E.-M. Schnçckelborg, B. de
Bruin, A. W. Ehlers, K. Lammertsma, Dalton Trans. 2010, 39, 1453.
[46] L. Pauling, The Nature of the Chemical Bond, 3rd ed., Cornell Uni-
versity Press, Ithaca, New York, 1960, p. 172f.
[47] R. Wolf, E.-M. Schnçckelborg, Chem. Commun. 2010, 46, 2832.
[48] M. Krejc ˇ?k, M. Dane ˇk, F. Hartl, J. Electroanal. Chem. Interfacial
Electrochem. 1991, 317, 179.
[49] A. Sch?fer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829.
[50] A. Sch?fer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571.
[51] a) E. J. Baerends, D. E. Ellis, P. Ros, Chem. Phys. 1973, 2, 41; b) B. I.
Dunlap, J. W. D. Connolly, J. R. Sabin, J. Chem. Phys. 1979, 71,
3396; c) O. Vahtras, J. Almlof, M. W. Feyereisen, Chem. Phys. Lett.
1993, 213, 514.
[52] a) K. Eichkorn, O. Treutler, H. Ohm, M. Haser, R. Ahlrichs, Chem.
Phys. Lett. 1995, 242, 652; b) K. Eichkorn, F. Weigend, O. Treutler,
R. Ahlrichs, Theor. Chem. Acc. 1997, 97, 119.
[53] a) F. Neese, Inorg. Chim. Acta 2002, 337, 181; b) S. Sinnecker, L. D.
Slep, E. Bill, F. Neese, Inorg. Chem. 2005, 44, 2245.
[54] a) P. O. Lçwdin, J. Chem. Phys. 1950, 18, 365; b) P. O. Lçwdin, Adv.
Quantum Chem. 1970, 5, 185.
[55] A. Klamt, G. Schuurmann, J. Chem. Soc. Perkin Trans. 2 1993, 799.
[56] S. Portmann, Molekel, version 4.3.win32, CSCS/UNI Geneva, Swit-
zerland, November 2002.
[57] G. M. Sheldrick, SADABS/TWINABS: Area-Detector Absorption
Correction, 1999, Universit?t Gçttingen, Germany.
[58] G. M. Sheldrick, Acta Crystallogr. A 2008, 64, 112.
[59] A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7.
Received: July 7, 2010
Published online: && &&, 2010
Chem. Eur. J. 2010, 00, 0–0 ? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
These are not the final page numbers!??
&13&
FULL PAPER
Diphosphacyclobutadiene Complexes

Page 14

Sandwich Complexes
R. Wolf,* A. W. Ehlers,
M. M. Khusniyarov, F. Hartl,
B. de Bruin, G. J. Long, F. Grandjean,
F. M. Schappacher, R. Pçttgen,
J. C. Slootweg, M. Lutz, A. L. Spek,
K. Lammertsma* ............. &&&&—&&&&
Homoleptic Diphosphacyclobutadiene
Complexes [M(h4-P2C2R2)2]x?
(M=Fe, Co; x=0, 1)
Fancy a sandwich? Homoleptic sand-
wich anions that contain diphosphacy-
clobutadiene ligands result from the
reaction of anthracene metalates
[Fe(h4-C14H10)2]?and [Co(h4-C14H10)2]?
with phosphaalkynes. Electrochemical
and preparative investigations of
[M(h4-P2C2tBu2)2]?(M=Fe, Co)
revealed that these anions are readily
oxidized to neutral derivatives [M(h4-
P2C2tBu2)2]. The electronic structures
of the new sandwich complexes have
been elucidated by using spectroscopic
techniques and density functional
theory calculations.
www.chemeurj.org
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Eur. J. 0000, 00, 0–0
??These are not the final page numbers!
&14&
R. Wolf, K. Lammertsma et al.