(Allenylidene)ruthenium complexes with redox-active substituents and ligands

Article abstract of DOI:10.1002/ejic.200390117

We describe the allenylidene complexes [TpL₂Ru=C=C= CPhR]⁺ SbF₆^- [Tp = HB(pz)₃^-, L₂ = 2 PPh₃ or 1,1′-bis(diphenylphosphanyl)ferrocene (dppf), R = Ph or ferrocenyl] and their spectroscopic and electrochemical characteristics. Three of these compounds possess redox-active, ferrocene-based substituents or ligands - that are oxidized at lower potentials than the ruthenium center itself - attached either to the terminal carbon atom of the allenylidene ligand or to the ruthenium atom. The Fc/Ph-substituted complexes 3a and 3b provide unique examples of hindered rotation of the allenylidene substituent around the M=C bond. For 3a (L₂ = 2 PPh₃), two isomers differing in the orientation of the vertically aligned, unsymmetrically substituted allenylidene ligand are discernible even at 388 K. The dppf -substituted congener 3b has the allenylidene ligand in a horizontal orientation and exhibits a rotational barrier, as determined by dynamic ³¹P NMR spectroscopy, of ΔG^≠ = 47 kJ/mol at Tc = 238 K. The changes in the spectroscopic and electrochemical properties upon replacement of the PPh₃ by a dppf ligand and the Ph by an Fc moiety can be explained in terms of the bonding within these systems. These effects are attenuated when the ferrocene-based redox tags are oxidized, as shown by IR and UV/Vis spectroelectrochemistry. Thus, infrared spectroelectrochemistry reveals a blue shift of the allenylidene stretch upon oxidation of the dppf ligand, while oxidation of the ferrocene covalently linked to the unsaturated C₃ ligand has the opposite effect. Oxidation of the ruthenium atom influences the bonding within the unsaturated ligand more profoundly. Results from IR spectroelectrochemistry point to a vinylidene structure in the Ru^III state. Reduction enhances the contribution of alkynyl-type resonance forms to the overall bonding description, as also shown by IR spectroelectrochemistry. For the ferrocenyl-substituted allenylidene complexes, oxidation and reduction result in bleaching of the intense optical low-energy absorption band attributed to a ferrocenyl-to-allenylidene charge-transfer process. The EPR spectra of the paramagnetic reduced forms of these complexes also indicate spin delocalization into the aryl substituents attached to the allenylidene ligand. The complexes Tp(dppf)RuCl and [Tp(dppf)Ru=C=C=CPh₂]⁺ SbF₆^- were also characterized by X-ray crystallography. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2003.

Full text of DOI:10.1002/ejic.200390117

FULL PAPER
(Allenylidene)ruthenium Complexes with Redox-Active Substituents and
Ligands
Stephan Hartmann,[a] Rainer F. Winter,*[a] Birgit M. Brunner,[a] Biprajit Sarkar,[a]
Axel Knödler,[a][ꢀ] and Ingo Hartenbach[a][ꢀ]
Keywords: Electrochemistry / Sandwich complexes / Ruthenium / Spectroelectrochemistry
We describe the allenylidene complexes [TpL2Ru=C=C=
CPhR]+ SbF6− [Tp = HB(pz)3−, L2 = 2 PPh3 or 1,1Ј-bis(diphen-
UV/Vis spectroelectrochemistry. Thus, infrared spectroelec-
trochemistry reveals a blue shift of the allenylidene stretch
ylphosphanyl)ferrocene (dppf), R = Ph or ferrocenyl] and upon oxidation of the dppf ligand, while oxidation of the
their spectroscopic and electrochemical characteristics.
Three of these compounds possess redox-active, ferrocene-
ferrocene covalently linked to the unsaturated C3 ligand has
the opposite effect. Oxidation of the ruthenium atom influ-
based substituents or ligands − that are oxidized at lower po- ences the bonding within the unsaturated ligand more pro-
tentials than the ruthenium center itself − attached either to foundly. Results from IR spectroelectrochemistry point to a
the terminal carbon atom of the allenylidene ligand or to the
ruthenium atom. The Fc/Ph-substituted complexes 3a and 3b
provide unique examples of hindered rotation of the allenyli-
vinylidene structure in the RuIII state. Reduction enhances
the contribution of alkynyl-type resonance forms to the over-
all bonding description, as also shown by IR spectroelectro-
dene substituent around the M=C bond. For 3a (L2 = 2 PPh3), chemistry. For the ferrocenyl-substituted allenylidene com-
two isomers differing in the orientation of the vertically
aligned, unsymmetrically substituted allenylidene ligand are
discernible even at 388 K. The dppf-substituted congener 3b
has the allenylidene ligand in a horizontal orientation and
exhibits a rotational barrier, as determined by dynamic 31P
NMR spectroscopy, of ∆G϶ = 47 kJ/mol at TC = 238 K. The
changes in the spectroscopic and electrochemical properties
upon replacement of the PPh3 by a dppf ligand and the Ph
by an Fc moiety can be explained in terms of the bonding
within these systems. These effects are attenuated when the
ferrocene-based redox tags are oxidized, as shown by IR and
plexes, oxidation and reduction result in bleaching of the in-
tense optical low-energy absorption band attributed to a
ferrocenyl-to-allenylidene charge-transfer process. The EPR
spectra of the paramagnetic reduced forms of these com-
plexes also indicate spin delocalization into the aryl substitu-
ents attached to the allenylidene ligand. The complexes
Tp(dppf)RuCl and [Tp(dppf)Ru=C=C=CPh2]+ SbF6 were
−
also characterized by X-ray crystallography.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
(CH3)]ϩ [5,6] and the related alkoxy-substituted iron derivat-
ive [Cp*(dppe)FeϭCϭCϭC(OCH3)(CH3)]ϩ.[7] Character-
istic attributes of allenylidene complexes include: (i) the
asymmetric valence band known as the ‘‘allenylidene
stretch’’, giving rise to highly intense absorption bands in
their infrared spectra, (ii) low-field signals for the carbon
atoms of the unsaturated ligands in their 13C NMR spec-
tra,[4,8] and (iii) their optical spectra, which feature two
bands in the visible region. The band at lower energy is due
to the forbidden HOMO Ǟ LUMO transition and so is
only weak, while that at higher energy is much more intense
and represents the allowed transition from the lower lying
second-highest occupied molecular orbital HOMO-1. Sys-
tematic studies on the complexes trans-[Cl(L2)2RuϭCϭCϭ
C(ERn)(RЈ)]ϩ revealed that their spectroscopic and electro-
chemical properties critically depend on the substituent
ERn (ERn ϭ NR2, SR, SeR, aryl) and the diphosphane li-
gands L2. Replacement of amino substituents NR2 by the
less potent π-donors SR, SeR, or aryl groups induces a sys-
tematic red shift of the electronic and infrared absorption
bands and a low-field shift of the resonance signals of the
In connection with our interest in the reactivity of prim-
ary butatrienylidene intermediates trans-[Cl(L2)2RuϭCϭ
CϭCϭCH2]ϩ (L2 ϭ chelating diphosphane ligand) we have
established new routes to heteroatom-substituted allenylid-
ene complexes trans-[Cl(L2)2RuϭCϭCϭC(ERn)(RЈ)]ϩ, in
part with unprecedented (ERn ϭ SR, SeR) substitution pat-
terns.[1,2] These rely either on the regioselective addition of
protic nucleophiles to the terminal CϭC double bond of
the unsaturated ligand or on the regioselective addition of
aprotic, allyl-substituted nucleophiles to carbon atom C-γ.
In this latter case, the initial addition is followed by a Cope-
type rearrangement of the resulting 3-heterohexa-1,5-diene
moiety.[1Ϫ4] The former approach had also been employed
by Bruce and others in the synthesis of amino-substituted
allenylidene complexes [Cp(PPh3)2RuϭCϭCϭC(NR2)-
[a]
Institut für Anorganische Chemie der Universität Stuttgart,
Pfaffenwaldring 55, 70569 Stuttgart, Germany
E-mail: winter@iac.uni-stuttgart.de
[‡]
X-ray structure determination
876
 2003 WILEY-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim 1434Ϫ1948/03/0305Ϫ0876 $ 20.00ϩ.50/0 Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
metal-bound carbon atom C-α and the carbon atom C-β maintaining the same overall structure. We compare our re-
adjacent to it.[1,2] Electrochemical investigations by cyclic sults for these systems with those for the corresponding
voltammetry revealed a cathodic displacement of the half- ‘‘parent’’ diphenylallenylidene complex [Tp(PPh3)2RuϭCϭ
wave potentials of the reduction and, albeit only to a minor CϭCPh2]ϩ, which lacks any additional redox-active moiety,
degree, of the oxidation potentials upon substitution of as a reference system.[13]
NR2 by SR, SeR, or aryl moieties. Broadly similar effects
are observed when the π-acidity of the diphosphane chelate
ligands is decreased. This mirrors the push-pull character-
Results and Discussion
istics of these complexes, with the ClRu(L2)2ϩ moiety acting
as the donor and the allenylidene unit as the acceptor. One
remarkable difference between the effects concomitant with
ligand and substituent replacement is their effect on the
redox potentials: replacement of the π-acidic ligands dppe
and dppm by the more basic depe facilitates oxidation by
more than 300 mV but leaves the reduction potential essen-
tially unchanged. This has been taken as an indication of a
ruthenium-based oxidation but a largely allenylidene-based
Synthesis, Molecular Structures, and Spectroscopic
Properties
The synthesis of allenylidene complexes containing the
[TpRuL2] moiety is straightforward and follows the method
elaborated by Hill and co-workers.[13] Treatment of a
CH2Cl2 solution of [TpRuL2Cl] (L2 ϭ 2 PPh3: 1a; L2 ϭ
dppf: 1b) with the appropriate propargylic alcohol
PhR(OH)CCϵCH [R ϭ Ph, Fc; Fc ϭ (η5-C5H4)Fe(η5-
reduction. Spectroelectrochemical experiments, monitoring
the spectroscopic changes associated with bulk electrolysis
by IR, UV/Vis, and EPR techniques, and results from
C5H5)] in the presence of AgSbF6 gave the corresponding
cationic allenylidene complexes [TpL2RuϭCϭCϭCRPh]ϩ
(2aϪ3b) in good to reasonable yields and as intensely
quantum mechanical calculations strongly support this
view.[1,4,9Ϫ11]
purple-colored, microcrystalline solids (Scheme 1). In the
course of these syntheses we obtained single crystals of
[Tp(dppf)RuCl] and the corresponding diphenyl-substituted
allenylidene complex [Tp(dppf)RuϭCϭCϭCPh2]ϩ (2b).
Their molecular structures in the solid state were estab-
lished by X-ray crystallography and are presented in
Figures 1Ϫ3, while selected bond parameters are provided
in Tables 1 and 2.
We have now turned our attention to allenylidene com-
plexes with either an additional redox-active moiety at-
tached to the C3 ligand or a redox-active diphosphane li-
gand. These were chosen such that they would undergo re-
versible electron transfer at lower potentials than the parent
allenylidene complex itself. This was to allow examination
of the effect of variation of the electron density at each of
these sites within the same basic system, without the need
for additional ligand or substituent replacement. Since
structural perturbations of the metallabutatriene chromo-
phore associated with electron transfer at a peripheral site
would be expected to be reasonably limited, the results from
such studies should reflect the effects of changes in the elec-
tron density at these sites more truly than any substitution.
In this respect, ferrocenyl-based systems are ideal candid-
ates, as they are known to undergo facile and reversible ox-
idation at substantially lower potentials than the RuII/III
couple. We have already reported on the ferrocenyl-substi-
tuted allenylidene complexes trans-[Cl(dppm)2RuϭCϭCϭ
C(NMe2)(C2H4Fc)]ϩ [Fc ϭ (η5-C5H4)Fe(η5-C5H5)][12] and
trans-[Cl(dppm)2RuϭCϭCϭC(SeFc)(C2H4CHϭCH2)]ϩ.[2]
In the amino-substituted complex the insulating ethylene
spacer separating the allenylidene chromophore from the
redox-active ferrocenyl unit prevents effective conjugation
between these entities. The spectroscopic changes associated
with ferrocene oxidation are therefore small. In the seleno-
substituted allenylidene complex, the redox-active group is
directly attached to the heteroatom. Electron transfer from
this site induces more substantial spectroscopic changes,
with a red shift of the IR frequency by 10 cmϪ1. Here we
present our results on a different class of allenylidene com-
plexes based on TpRuL2 units, as pioneered by Puerta and
Hill [Tp ϭ HB(pz)3Ϫ, tris(pyrazolyl)borate].[13Ϫ15] These
complexes offer the opportunity to introduce ferrocene-
based redox tags either into the allenylidene ligand or into
the diphosphane chelate ligand Ϫ or into both Ϫ while Scheme 1
Eur. J. Inorg. Chem. 2003, 876Ϫ891
877

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
FULL PAPER
Figure 1. Structure of [Tp(dppf)RuCl] in the solid state
Figure 3. Structure of [Tp(dppf)RuϭCϭCϭCPh2]ϩ in the solid
state
˚
Table 1. Selected bond lengths [A] and bond angles [°] for [Tp-
(dppf)RuCl] (1b)
RuϪN(2)
RuϪN(4)
RuϪN(6)
RuϪP(1)
RuϪP(2)
RuϪCl(1)
FeϪC(1)
FeϪC(2)
FeϪC(3)
FeϪC(4)
FeϪC(5)
FeϪC(6)
FeϪC(7)
FeϪC(8)
FeϪC(9)
FeϪC(10)
P(1)ϪC(6)
P(1)ϪC(11)
P(2)ϪC(1)
P(2)ϪC(31)
P(2)ϪC(41)
2.144(5)
2.144(5)
2.099(5)
2.3503(16)
2.3362(16)
2.4200(16)
2.001(6)
2.018(7)
2.063(7)
2.084(7)
2.052(7)
2.009(6)
2.013(7)
2.061(7)
2.061(7)
2.032(7)
1.835(6)
1.850(7)
1.807(6)
1.823(6)
1.862(6)
N(4)ϪRuϪN(2)
N(6)ϪRuϪN(2)
N(6)ϪRuϪN(4)
N(2)ϪRuϪP(1)
N(6)ϪRuϪP(1)
N(4)ϪRuϪP(1)
N(4)ϪRuϪP(2)
N(6)ϪRuϪP(2)
N(2)ϪRuϪP(2)
P(2)ϪRuϪP(1)
P(1)ϪRuϪCl
81.86(19)
87.57(19)
87.60(19)
89.70(14)
89.34(15)
171.13(14)
89.27(14)
93.59(14)
171.00(14)
99.23(6)
93.83(6)
93.28(6)
84.98(14)
88.15(14)
171.87(14)
99.23(6)
P(2)ϪRuϪCl
N(2)ϪRuϪCl
N(4)ϪRuϪCl
N(6)ϪRuϪCl
P(2)ϪRuϪP(1)
P(1)ϪRuϪCl
93.83(6)
93.28(6)
P(2)ϪRuϪCl
˚
Table 2. Selected bond lengths [A] and bond angles [°] for
[Tp(dppf)Ru(ϭCϭCϭCPh2)] SbF6 (2b)
Figure 2. Intermolecular contacts between the chloride ligand and
the solvate molecules of [Tp(dppf)RuCl]
RuϪP(1)
RuϪP(2)
RuϪN(2)
2.416(9)
2.376(11)
2.16(3)
2.18(3)
2.20(3)
1.88(4)
1.32(5)
1.26(5)
1.46(5)
1.52(5)
P(1)ϪRuϪP(2)
P(1)ϪRuϪC(60)
P(2)ϪRuϪC(60)
N(2)ϪRuϪN(4)
N(2)ϪRuϪN(6)
N(4)ϪRuϪN(6)
P(1)ϪRuϪN(2)
P(1)ϪRuϪN(4)
P(1)ϪRuϪN(6)
P(2)ϪRuϪN(2)
P(2)ϪRuϪN(4)
P(2)ϪRuϪN(6)
RuϪC(60)ϪC(61)
C(60)ϪC61)ϪC(62)
99.1(3)
94.7(11)
92.1(11)
84.8(10)
85.5(10)
81.8(11)
89.3(7)
169.2(8)
88.8(8)
95.2(7)
90.5(8)
172.2(8)
168(3)
[Tp(dppf)RuCl] (1b) crystallizes as a bis(solvate) from
chloroform, and a plot of its structure is presented as Fig- RuϪN(4)
RuϪN(6)
ure 1. Complex 1b displays bond parameters (Table 1) very
RuϪC(60)
similar to those of the PPh3-substituted congener.[17] The
C(60)ϪC(61)
P(1)ϪRuϪP(2) bite angle of the chelating dppf ligand is
C(61)ϪC(62)
only slightly smaller than the PϪRuϪP angle in the other
complex [99.23(6) against 101.9(1)°] and induces a compres-
sion of the opposite angle N(2)ϪRuϪN(4) to 81.9(2)°. All
other angles between cis-disposed ligands are in the
85.0Ϫ93.8° range, while mutually trans-disposed ligands
make angles of approximately 171° at the metal atom.
C(61)ϪC(71)
C(61)ϪC(81)
168(4)
878
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
The close similarity between [Tp(dppf)RuCl] and A plot of the structure is provided as Figure 3, while the
[Tp(PPh3)2RuCl] is also evident from the essentially ident- most pertinent bond parameters are collected in Table 2.
ical RuϪP and RuϪCl bond lengths in these two complexes
Complexes 2a and 2b and their Fc-substituted congeners
[2.336(2), 2.350(2) and 2.420(2) against 2.332(3), 2.349(3) 3a and 3b have very similar spectroscopic characteristics in
˚
and 2.409(3) A]. As is often observed for complexes of the their IR and NMR spectra. These are compared in Table 3.
Tp ligand,[13,18,19] the nitrogen atom trans to the chloride Samples (as KBr pellets) display the CϭCϭC stretch of the
ligand is closer to the metal center than those opposite to unsaturated ligand as a very intense band at 1946 cmϪ1 for
the π-accepting phosphane ligands. Both the cyclo- 3a and at 1941 cmϪ1 for 3b, which are some 5 cmϪ1 higher
pentadienyl (Cp) rings of the dppf ligand are essentially in energy than in the diphenyl-substituted complexes 2a and
planar, and the phosphorus atoms are situated in the plane 2b (see also Table 3). Replacement of the dppf by the less
of the ring to which they are attached. The ring centroids basic PPh3 ligands induces a blue shift of similar size. The
and the Fe atom make an angle of 178.2°, while the two observed effects, albeit small, match previous observations
rings adopt a staggered conformation and are rotated by that the presence of a weaker π-donor at the terminal car-
38° with respect to one another. The packing in the crystal bon atom C-γ (phenyl versus ferrocenyl) or of a more basic
lattice is stabilized by several inter- and intramolecular hy- metal center at the opposite end of the metallabutatriene
drogen bonds involving the solvate molecules, the chloride chain (the dppf versus two PPh3 ligands) shifts the IR ab-
ligand, some nitrogen atoms of the pyrazolyl rings, and one sorption of the allenylidene ligand to lower energies. This
hydrogen atom of each phenyl ring, as well as one of the has been explained in terms of a larger contribution from
Cp-hydrogen atoms of the dppf ligand. Intramolecular hy- the alkynyl-type resonance form II as opposed to the genu-
˚
drogen bonds of 2.528 and 2.663 A are observed between ine cumulenylidene-type resonance form I (Scheme 2) when
the ortho-hydrogen atom attached to C(26) and the nitrogen the metal center is more electron-rich and the substituents
atoms N(1) and N(2). A similar contact is found for H(32a) at C-γ are less able to support a positive charge adjacent to
Ϫ the hydrogen atom at C(32) Ϫ and N(4) (2.677 A). The them.[1,2,4,20,21] In the light of the stability of the ferrocenyl-
˚
chloride ligand forms a pair of contacts to one of the ortho- methylene carbenium ion[22] the ferrocenyl substituent may
hydrogen atoms of each of the adjacent phenyl rings, with be expected to stabilize the alternative alkynyl resonance
˚
a fairly short bond of 2.590 A to H(42) and a weaker con- form III, as indicated in Scheme 2, in line with the results
˚
tact of 2.871 A to H(16) on the other phenyl ring. The from IR spectroscopy. We note here that Le Bozec et al.
chloride ligand also forms strong intermolecular contacts arrived at a similar conclusion in their study of [(C6Me6)-
to the protons of the two cocrystallized CHCl3 solvate mo- Cl(PMe3)RuϭCϭCϭC(Ph)(Fc)]ϩ PF6Ϫ, closely related to
lecules, as evidenced by H···Cl distances of 2.393 and 2.423 3a and 3b.[23] The 13C NMR spectra also bear witness to
˚
A. The solute molecules are additionally fixed through this notion. Comparison of the relevant data for the pairs
weaker interactions to hydrogen atoms on the aromatic sub- of complexes 2a/3a and 2b/3b reveals that replacement of a
stituents. Thus, one chloroform molecule makes a contact phenyl by the more strongly electron-donating ferrocenyl
˚
of 2.834 A to one of the phenyl protons, while the other is substituent shifts the resonance signals of the carbon atoms
˚
2.792 A away from H(8a) at the 3-position of one of the Cp of the allenylidene ligand to higher field by ca. 30, 25, and
rings of the dppf ligand. Additional weaker interactions are 8 ppm. The largest shift is observed for the metal-coordin-
observed between the other chlorine atoms of the cocrystal- ated carbon atom C-α, while the terminal carbon atom C-
˚
lized solvent molecules and hydrogen atoms H(13) [3.292 A γ Ϫ where the actual substitution occurs Ϫ is influenced
˚
˚
to Cl(5), 3.249 A to Cl(7)], H(14) [3.247 A to Cl(2), 3.289 least. The sign and relative magnitude of the observed ef-
˚
˚
A to Cl(4)], and H(43) [3.274 A to Cl(3)] on the phenyl fects agree well with results from previous studies on the
˚
rings as well as to H(9) [3.175 A to Cl(6)] on the Cp ring. allenylidene
complexes
trans-[Cl(L2)2RuϭCϭCϭ
The H-bridges resulting from the stronger intermolecular C(ERn)(RЈ)]ϩ (ERn ϭ NR2, SR, SeR, aryl, RЈ ϭ al-
H···Cl contacts are shown in Figure 2.
kyl).[1,4,24]
The diphenylallenylidene complexes 2a and 2b, including
The triphenylphosphane-substituted complex 3a was
the solid-state structure of compound 2a, have already been found to exist as two isomers, as shown by the two sets of
reported by Hill.[13] We were able to obtain single crystals resonance signals for the carbon atoms of the cumulenylid-
of the dppf-substituted congener 2b, but these gave a data ene ligand. These are observed at δ ϭ 287.0, 184.3, and
set of only poor quality, preventing meaningful discussion 168.2 ppm for the major isomer and δ ϭ 282.8, 180.5 and
of bond lengths and angles within the allenylidene chain. 169.1 ppm for the minor counterpart, and point to hindered
We only note that, as in the chloro precursors 1a and 1b, rotation of the allenylidene ligand around the RuϭC bond.
the RuϪP bond lengths and the PϪRuϪP angles are nearly Each isomer exhibits a singlet resonance signal in the 31P
identical for both allenylidene complexes [2.376(11) and NMR spectrum. We therefore conclude that the plane of
˚
2.416(9) against 2.3779(7) and 2.4327(11) A as well as the terminal CR2 unit of the allenylidene ligand runs paral-
99.1(3) against 101.11 (3)°]. We also note that the terminal lel to the molecular symmetry plane and bisects the
CPh2 group of the allenylidene ligand lies approximately in PϪRuϪP angle, an orientation that has been termed ver-
the mirror plane of the molecule [the torsional angle tical.[25] Scheme 3 offers a Newman-type projection along
N(2)ϪRuϪC(62)ϪC(81) is 18.5°] and roughly bisects the the RuϭCϭCϭC axis. The isomers differ in which of the
PϪRuϪP angle. This orientation has been termed vertical. Ph or the Fc substituents points toward the Tp and the
Eur. J. Inorg. Chem. 2003, 876Ϫ891
879

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
Table 3. Selected spectroscopic data for complexes 2aϪ3b
FULL PAPER
Compd.
31P NMR:
δ [ppm]
13C NMR:
δ [ppm]
C-α
IR (KBr):
UV/Vis:
ν˜(CϭCϭC) [cmϪ1
]
λmax [nm] (log ε)[a]
HOMO-1 Ǟ LUMO
C-β
C-γ
HOMO Ǟ LUMO
2a
2b
3a
36.6
316.6
208.9
160.9
1941
1936
1946
527 (4.217)
522 (4.204)
538 (4.061)
907 (2.820)
902 (2.763)
Ϫ[d]
30.2
316.2
208.3
161.3
40.8[b]
40.5[c]
29.9
287.0[b]
282.8[c]
283.7
184.3[b]
180.5[c]
180.6
168.2[b]
169.1[c]
167.9
3b
1941
540 (4.190)
Ϫ[d]
[a]
[b]
[d]
In CH2Cl2. Data for major isomer.[c] Data for minor isomer.
Obscured by an intense ILCT band (see text).
Scheme 2
phosphane ligands. We propose that the major isomer has ates the metal ion through its Cp ring and the two pendant
the less bulky phenyl substituent oriented toward the PPh3 phosphane side-arms (∆G϶ ഠ 33 kJ/mol for secondary vi-
ligands. By integration of the 31P NMR resonance signals
the isomeric distribution was established as 4:1. Unfortu-
nately, we have not yet succeeded in growing X-ray quality
crystals of this complex to support this view. We note, how-
nylidene ligands and 44 kJ/mol for tertiary ones),[29] the di-
nuclear bis(vinylidene) manganese complex [(η5-
C5H4Me)(dmpe)MnϭCϭC(Ph)C(Ph)ϭCϭMn(dmpe)(η5-
C5H4Me)]2ϩ (∆G϶ ϭ 56.7 kJ/mol),[30] and the cyclohepta-
trienylmolybdenum complex [(η7-C7H7)(dppe)MoϭCϭ
CH2]ϩ (dppe ϭ Ph2PC2H4PPh2; ∆G϶ ϭ 51.9 kJ/mol).[31]
The highest activation energy for vinylidene rotation of
which we are aware has been observed for the rhenium com-
plexes [(η5-C5H5)(NO)(PPh3)ReϭCϭCPhMe]ϩ and [(η5-
C5H5)(NO)(PPh3)ReϭCϭCH2]ϩ (∆G϶ ϭ 77.9 kJ/mol).[32]
The butyl/methyl-substituted vinylidene complex [(η7-
C7H7)(dppe)MoϭCϭC(C4H9)(CH3)]ϩ displays separate
resonance signals for the two different rotamers at 348 K.[31]
The rotational barriers are usually thought of as being more
steric than electronic in origin. Calculations on (vinylidene)-
metal complexes indicate that the intrinsic (electronic) en-
ergy difference between the vertical and horizontal rotamers
amounts to no more than 10Ϫ20 kJ/mol.[25,33] The energy
barrier interconnecting them is thus expected to be fairly
low. In view of the further remoteness of the terminal
methylene group from the metalϪligand array it is therefore
not surprising that allenylidene complexes present even
lower rotational barriers than their vinylidene counterparts,
and so the observation of a truly significant rotational bar-
rier in an allenylidene ligand is quite unusual. This
prompted us to study the dynamic behavior of 3a in 1,2-
D2C2Cl4 solution by 31P NMR spectroscopy. Upon warm-
ing there was a very slight shift of the two singlet resonance
signals, but no broadening was observed up to 388 K, at
which point 3a decomposed.
Ϫ
ever, that both [Tp(PPh3)2RuϭCϭCϭCPh2]ϩ PF6 (2a)
and its dppf analogue 2b (vide supra) crystallize with the
allenylidene ligand in just this orientation.[13]
Scheme 3
To the best of our knowledge, complex 3a represents the
first instance of an allenylidene complex in which the energy
barrier to rotation of the allenylidene moiety is so high as
to allow observation of two different rotamers at ambient
temperature. Hindered rotations around MϭC double
bonds have, however, been noted in the related vinylidene
complexes {M}ϭCϭCR2.[26Ϫ28] Experimentally measured
energy barriers are usually of the order of 30Ϫ57 kJ/mol
and depend on the substituents on the vinylidene ligand
and the coordination environment at the metal center. More
recent examples include ruthenium complexes of the tri-
The sterically less hindered complex 3b exhibits one
podal CH3C(η5-C5H4)(CH2PPh2)2 ligand, which coordin- sharp singlet in its room-temperature 31P NMR spectrum.
880
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
When a solution in CD2Cl2 is cooled, this signal starts to Transfer (MLCT) character of this band.[4] This assignment
broaden at 258 K. At 233 K, splitting into two equally in- also seems to be valid here, since a much weaker absorption
tense, broad signals is observed, and at even lower temper-
atures the four resonance lines of a well-resolved AB spec-
trum are seen. The coupling constant Ϫ 30.65 Hz at 213 K
Ϫ is typical of cis-disposed phosphane ligands (Figure 4).
These findings would place the free energy of activation for
rotation of the allenylidene ligand at 47 kJ/mol at the co-
alescence temperature of 238 K. They also indicate a hori-
zontal orientation of the allenylidene ligand with the sub-
stituents on the terminal carbon atom orthogonal to the
symmetry plane of the [TpRuL2] moiety. This is in sharp
contrast to the vertical arrangement proposed for the bi-
s(triphenylphosphane)-substituted counterpart 3a. In the
absence of X-ray data we can only speculate that in the
vertical orientation the substituents on the terminal carbon
atom would each approach the ferrocenyl backbone of the
dppf ligand such that the horizontal orientation would be
preferred on steric grounds (see Scheme 3).
at considerably lower energies (907 nm in 2a and at 902 nm
in 2b) was observed for the forbidden HOMO Ǟ LUMO
transition. With reference to the trans-[Cl(dppm)2Ru]-de-
rived complexes, the HOMO should also be dominated by
the metal unit and have appreciable MLCT character as
well. As found for the HOMO-1 Ǟ LUMO transition, it
exhibits negative solvatochromism of a magnitude compar-
able to that of the main absorption band. We note that re-
placement of the trans-[Cl(L2)2Ru] moiety by [TpRuL2] ap-
parently results in a somewhat smaller energy gap between
the two highest occupied levels and the LUMO. This is in-
ferred by comparison of the data for 2a and 2b with those
for the related complex trans-[Cl(dppm)2RuϭCϭCϭ
CPh2]ϩ, which has its absorption maxima at 506 and
817 nm. The data also point to a slightly larger energy gap
between the two highest occupied levels for the [TpRuL2]-
derived complexes. The energy difference between the
HOMO-1 Ǟ LUMO and the HOMO Ǟ LUMO transitions
thus amounts to 1.01 eV for 2a and 0.98 eV for 2b, as com-
pared to 0.93 eV for trans-[Cl(dppm)2RuϭCϭCϭCPh2]ϩ.
In the ferrocenyl-substituted complexes 3a and 3b the in-
tense HOMO-1 Ǟ LUMO transition band is shifted to
somewhat lower energies, as in the diphenyl-substituted
analogues (λmax ϭ 538 nm for 3a, 540 nm for 3b, see
Table 3). At even lower energies another very intense, broad
band is found at 723 (3a) and 713 (3b) nm, tailing far into
the near IR region and burying the HOMO Ǟ LUMO
band. Again, a slight negative solvatochromism is noted
when spectra in CH2Cl2 and CH3CN are compared. This
band has no counterpart in 2a and 2b and is consequently
attributed to a ferrocenyl-to-ligand MLCT charge transfer
absorption. If the ferrocenyl substituent and the cumulenic
carbon atoms are regarded as belonging to the same chrom-
ophore, this band may also be termed an Intraligand
Charge Tranfer (ILCT) band, the charge density being
shifted between two different sites within the same basic
ligand. Irrespective of the exact interpretation, the presence
of this additional band points to effective conjugation be-
tween the ferrocenyl substituent and the allenylidene chro-
mophore. As discussed later, the above assignment is
strongly supported by our results from UV/Vis spectroelec-
trochemical investigations on the ferrocenyl-based one-elec-
tron oxidation.
Figure 4. Dynamic 31P NMR spectroscopy of compound 3b in
CD2Cl2 solution
Spectra of the dppf-substituted complexes 2b and 3b are
compared in Figure 5. The ferrocene-based absorption of
the dppf ligand could not be located, since it is obscured
by the much more intense, broad HOMO-1 Ǟ LUMO
transition. All complexes presented here display another in-
tense band at 350Ϫ356 nm (log ε ഠ 4). This feature seems
to be characteristic of aryl-substituted allenylidene com-
plexes and so is tentatively assigned to ILCT excitation
from the arene substituent(s) to the metallacumulene chro-
mophore. Absorptions at even higher energies in the UV
region arise from the n Ǟ π* and π Ǟ π* transitions of the
aryl-substituted phosphane ligands.[34]
In the visible range the electronic spectra of the intensely
colored complexes 2a and 3a are each dominated by a
strong absorption band, at 527 (2a) or 522 nm (2b) in
CH2Cl2 solution. This band is only very slightly solvato-
chromic, shifting to somewhat higher energies in the more
polar solvent CH3CN. For related allenylidene complexes
of the trans-[Cl(L2)2Ru] moiety, this absorption band has
been assigned to an allowed transition from the metal-
based second highest molecular orbital HOMO-1 to the
LUMO, which is delocalized over the allenylidene ligand.
This accounts for the strong Metal-to-Ligand Charge-
Eur. J. Inorg. Chem. 2003, 876Ϫ891
881

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
FULL PAPER
ible even at low sweep rates and allowed us to determine its
half-wave potential as ϩ1.05 V against the internal ferro-
cene/ferrocenium standard (Table 4). As with the dppm-
substituted complex 4, two consecutive one-electron reduc-
tions are observed. While the first reduction at E1/2
ϭ
Ϫ1.00 V is chemically reversible under ambient conditions,
the second reduction is not, and requires cooling to allow
detection of the associated anodic return peak. Even so,
only partial chemical reversibility could be achieved. It was,
however, still possible to determine the half-wave potential
of this redox couple as Ϫ2.05 V. We note that these poten-
tials are very similar to those of trans-[Cl(dppe)2RuϭCϭ
CϭCPh2]ϩ.[35] This further corroborates our previous find-
ing that Ϫ given the same allenylidene ligand Ϫ the reduc-
tion potentials are virtually independent of the electron
density on the metal, a further indication of the ligand-cent-
ered nature of this process.[4]
The analogous compound 2b contains an electroactive
dppf ligand, which would be expected to oxidize at consid-
erably lower potential than the LnRuII entity. Thus, 2b dis-
plays a reversible one-electron wave at ϩ0.485 V, more than
550 mV more negative than the RuII/III couple of the PPh3-
substituted congener 2a. No further anodic processes could
be observed within the accessible potential window. During
our studies on the ferrocenylethyl-substituted aminoal-
lenylidene complex trans-[Cl(dppm)2RuϭCϭCϭC(NMe2)-
(C2H4Fc)]ϩ we noticed that the RuII/III couple is shifted an-
odically by ca. 60 mV relative to closely related complexes
upon oxidation of the electroactive ferrocenyl moiety.[12] In
that particular example, the observed shift largely reflects
the increased electrostatic repulsion between the positively
charged centers. A much larger effect is to be expected in
2b, in which the electroactive ferrocenyl moiety is directly
linked to the metal ion. Given the proximity of the RuII/III
process to the anodic discharge limit of the solvent in com-
plex 2a, it is thus not surprising that it lies outside the ac-
cessible potential range in the case of 2b. The first reduction
of this complex is again a fully reversible one-electron pro-
cess, occurring at almost the same potential as in 2a (Ϫ1.01
V), but no further reduction was observed in this solvent.
Complex 3a has the electroactive ferrocenyl substituent at-
tached to the terminal carbon atom of the allenylidene li-
gand and gives rise to a reversible one-electron wave at
Figure 5. Optical spectra of complexes 2b and 3b in CH2Cl2
Electrochemistry
Electrochemical studies on the allenylidene complexes
trans-[Cl(L)2RuϭCϭCϭC(ERn)(RЈ)]ϩ have shown that
each of these complexes undergoes a one-electron oxidation
and a one-electron reduction.[1,4] By combining electro-
chemical and spectroscopic methods we have been able to
characterize the oxidation as a primarily metal-based pro-
cess (i.e., the RuII/III couple) and the reduction as ligand-
based, the latter assignment being confirmed by EPR and
IR spectroscopy on their reduced forms. The reduction of
the diphenyl-substituted complex trans-[Cl(dppe)2RuϭCϭ
CϭCPh2]ϩ (4), related to 2a, has meanwhile been investig-
ated by Touchard and co-workers. These authors arrived at
essentially the same conclusions and also corroborated the
reduction site by exploring the chemical reactivity of the
corresponding neutral radical.[35] In that particular com-
pound, a second, irreversible one-electron reduction was
observed at a potential Ϫ1.10 V lower than the first revers-
ible reduction process. The PPh3-substituted diphenylallen-
ylidene complex 2a exhibits essentially identical electro-
chemical behavior (Figure 6, bottom trace). The one-elec-
tron oxidation appears to be only partially reversible under
ambient conditions but a meaningful analysis of the degree
of chemical reversibility is precluded by its proximity to the
anodic discharge limit of the electrolyte solution. Cooling
to Ϫ78 °C renders the voltammetric wave chemically revers-
E1/2 ϭ ϩ0.36 V. As was the case for 2a, the ruthenium-based
oxidation is again shifted outside the potential window of
the electrolyte solution. While this precludes us from quan-
tifying the potential shift of the RuII/III couple induced by
ferrocene oxidation, it nevertheless provides further evid-
ence for effective conjugation between the ferrocene sub-
stituent and the cumulenic system, as already inferred from
the optical data. Complexes 2a and 3a also offer the oppor-
tunity to assess the effect on the reduction potentials of
replacement of a phenyl by a more strongly donating ferro-
cene substituent on the terminal carbon atom of the allenyl-
idene ligand. Given the allenylidene-centered nature of the
reduction processes, one would expect a cathodic shift of
the (ϩ/0) and the (0/Ϫ) couples. This is indeed the case.
The two reductions for 3a are observed at potentials 160
Figure 6. Cyclic voltammograms of complexes 2a (CH2Cl2, 215 K,
v ϭ 0.1 V/s, bottom trace) and 3b (CH2Cl2, 298 K, v ϭ 0.1 V/s,
top trace)
882
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
Table 4. Electrochemical data for complexes 2aϪ3b
E1/2 [RuII/III
]
E1/2 (dppf)[a]
E1/2 (Fc/Fcϩ)[a]
E1/2 (ϩ/0)[a]
E1/2 (0/Ϫ)[a]
[a]
Complex
Solvent
2a
2b
3a
3b
CH2Cl2
DMF
CH2Cl2
DMF
CH2Cl2
DMF
CH2Cl2
DMF
ϩ 1.05 (158)
n. d.
n. d.
n. d.
n. d.
n. d.
n. d.
n. d.
n. a.
n. d.
ϩ0.485 (142)
n. d.
n. a.
n. a.
ϩ0.51 (144)
n. d.
n. a.
n. a.
n. a.
n. a.
ϩ0.36 (138)
n. d.
ϩ0.37 (142)
n. d.
Ϫ1.00 (138)
Ϫ0.83 (78)
Ϫ1.01 (134)
Ϫ0.83 (72)
Ϫ1.16 (146)
Ϫ0.99 (80)
Ϫ1.165 (136)
Ϫ1.00 (78)
Ϫ2.05 (192)
Ϫ2.02[b]
n. d.
Ϫ2.00[b]
Ϫ2.375[b]
Ϫ2.16[b]
n. d.
Ϫ2.13[b]
[a]
[b]
E1/2 values (in V) are calibrated against the internal Fc/Fcϩ standard, ∆Ep values (in mV) are given in brackets.
Peak potential of
an irreversible process. n. d. ϭ not detected; n. a. ϭ not applicable.
and ca. 320 mV more negative than in its congener 2a. Des- torted responses in the more conducting DMF/NBu4PF6
pite all efforts, no return peak for the second reduction of electrolyte.
3a was observed. The reported shift of the (0/Ϫ) potential
Spectroelectrochemistry
is therefore based on the respective positions of the forward
peaks and provides only a qualitative measure.[36]
a) Oxidation
Lastly, complex 3b contains both secondary redox sites
employed in this study, the dppf ligand on the metal and
We now address the question of how electron transfer
the ferrocenyl substituent on the terminal carbon atom C- from the secondary electroactive moieties attached either
γ. Consequently, two reversible one-electron waves at poten- to the ruthenium or to the terminal carbon atom of the
tials considerably more negative than those of the RuII/III allenylidene ligand affects the spectroscopic properties of
couple of 2a are observed in the cyclic voltammogram of these complexes. As may be inferred from the above discus-
this complex (Figure 6, top trace). The half-wave potentials sion, a combination of bulk electrolysis with infrared spec-
were determined as ϩ0.37 and ϩ0.51 V relative to the ferro- troscopy in an optically transparent thin-layer electrolysis
cene/ferrocenium standard. Comparisons with those of 2b (OTTLE) cell provides the most sensitive tool with which
and 3a argue for an assignment of the less anodic wave to to examine bonding changes accompanying oxidation
the oxidation of the allenylidene-bound ferrocene moiety within the metallacumulene chromophore. The PPh3-ligated
and the other one to that of the dppf ligand on the metal diphenylallenylidene complex 2a does not contain any addi-
ion. Further evidence for this comes from spectroelectroch- tional redox-active moiety and thus serves as a benchmark
emical experiments, as detailed below. Paralleling the situ- system for elucidating the spectral changes concomitant
ation in compounds 2a and 3a, the reversible reduction of with oxidation of the [{Ru}ϭCϭCϭC(R)(Ph)]ϩ system
3b is shifted cathodically by 155 mV relative to the di- itself. It is thus discussed first.
phenyl-substituted 2b, which again reflects the higher elec-
Upon oxidation with IR monitoring, the strong CϭCϭ
tron density on the allenylidene ligand. No second reduc- C stretch of the allenylidene ligand vanishes and is replaced
tion was observed for 3b in CH2Cl2/Bu4NPF6 solution. In by a weaker absorption at 1587 cmϪ1. This new band is
order to assess the potential shifts induced on the second best accommodated by the vinylidene resonance forms IV
reduction by the presence of the ferrocenyl substituent, we and V (Scheme 4). Both place the unpaired spin on the in-
also recorded voltammograms in DMF/Bu4NPF6 solution. ternal carbon atom C-β of the allenylidene ligand. At a first
Thanks to its superior cathodic window we were now able approximation, one may expect that the electron lost upon
to observe this process for all compounds, albeit as an irre- oxidation should originate from the HOMO orbital. Ac-
versible event. The data in Table 5 show that in this solvent cording to recent DFT calculations on the parent cumulen-
the effect of substitution of one phenyl by a ferrocenyl sub- ylidene complexes of the Cr(CO)5 entity,[4,37]and also our
stituent is probably very similar for both reductions, al- own work on the amino-substituted model complex trans-
though the irreversible nature of the second process only [Cl(PH3)4RuϭCϭCϭC{N(CH3)2}(CH3)]ϩ,[4] this orbital is
allows for a qualitative statement.[36] No useful data, how- mainly composed of the ligand trans to the allenylidene li-
ever, could be obtained for any of the oxidations in the gand, the metal atom, and carbon atom C-β. Spin density
DMF solvent.
calculations performed on trans-[Cl(PH3)4RuϭCϭCϭ
The voltammetric data for all compounds under study C(ERn)(CH3)]2ϩ [ERn ϭ O(CH3), S(CH3), Se(CH3), CH3]
are listed in Table 4. The peak potential splittings signific- show that the orbital composition of the HOMO orbital
antly exceed the predicted theoretical values of 59 mV at faithfully translates into spin densities in the oxidized form.
298 K and 39 mV at 195 K.[36] This effect is largely due The mesomeric structures IV and V agree with these con-
to uncompensated resistance of the electrolyte solution, as siderations.
shown by the nearly identical peak-to-peak separations of
the internal ferrocene standard, which is thought to repres- a fairly intense IR band at around 1690Ϫ1580 cm
ent an ideal Nernstian redox system, and the much less dis- literature shows a clear trend toward lower energy absorp-
Vinylidene complexes of ruthenium are characterized by
.
Ϫ1 [38] The
Eur. J. Inorg. Chem. 2003, 876Ϫ891
883

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
Table 5. Spectroscopic data for complexes 2aϪ3b in different oxidation states
FULL PAPER
Compound[a]
IR:
UV/Vis:
λ [nm] (log ε)
ν˜(BϪH), (CϭC) or (CϭCϭC) [cmϪ1
]
2a [RuII]
2489 (w) (BH), 1945 (vs)
2521 (w) (BH), 1587 (s)
2469 (w), 1969 (m), 1959 (m)
354 (4.225), 525 (4.284), 909 (2.833)
411 (4.282), 491 (4.288), 830 (3.200)
350 (4.342), 400 (3.879), 480 (3.865), 534 (3.865),
812 (2.531)
2aϩ [RuIII
]
2aϪ
2b [FeII (dppf), RuII]
2b؉ [FeIII (dppf), RuII]
2b2ϩ [FeIII (dppf), RuIII
2bϪ
2487 (w),1939 (vs), 1573 (m), 1588 (m) 348 (3.995), 361 (sh), 526 (4.201), 901 (2.852)
2513 (w), 1957 (vs), 1948 (vs)
not detected
351 (4.060), 369 (4.211), 530 (4.044), 850 (2.903)
387 (4.090), 513 (3.949), 728 (3.083)
351 (4.156), 405 (3.760), 452 (3.680), 530 (5.735),
630 (sh, 3.057), 847 (2.813)
356 (3.903), 538 (4.061), 723 (3.756)
298 (4.097), 353 (3.863), 544 (4.000)
]
2473 (w), 1966 (m), 1956 (m)
3a [FeII (Fc), RuII]
2488 (w), 1948 (vs)
2506 (w), 1924 (s)
2514 (w), 1507 (s)
2463 (w), 1969 (s)
2488 (w), 1945 (vs)
3aϩ [FeIII (Fc), RuII]
3a2ϩ [FeIII (Fc), RuIII
]
not detected
3aϪ
340 (3.929), 410(sh), 451 (3.813), 544 (3.744), 693 (2.653)
353 (4.279), 538 (4.290), 713 (4.112)
353 (4.250), 540 (4.228), 701 (3.919)
315 (4.228), 353 (4.243), 475 (sh), 548 (4.225), 685 (3.829)
368 (4.188), 468 (4.215), 557 (3.914), 702 (4.047)
343 (4.270), 413 (3.941), 456 (4.072), 549 (4.005), 865 (2.771)
3b [FeII (Fc), FeII (dppf), RuII]
3bϩ [FeIII (Fc), FeII (dppf), RuII] 2506 (w), 1916 (m)
3b2ϩ [FeIII (Fc), FeIII (dppf), RuII] 2488 (w), 1945 (s), 1912 (m)
3b3ϩ [FeIII (Fc), FeIII (dppf), RuIII] 2519 (w), 1524 (s)
3bϪ
2470 (w), 1963 (m)
[a]
All data in C2H4Cl2/0.25  Bu4NPF6 at room temperature.
Scheme 4
Figure 7. IR spectroelectrochemistry of complex 3a (1,2-Cl2C2H4,
298 K); oxidation of the ferrocenyl substituent attached to the al-
lenylidene ligand
tions for acceptor-substituted vinylidene ligands. This is
demonstrated by the complexes [Cp(PPh3)2RuϭCϭ in Figure 7, electrolysis at a potential positive to the anodic
C(Ph)(R)]ϩ [ν˜(CϭC) ϭ 1675 cmϪ1 for R ϭ C7H7, 1655 wave of the ferrocenyl substituent induces a red shift of the
cmϪ1 for R ϭ CH2Ph, 1638 cmϪ1 for R ϭ I or Br, and allenylidene band by ca. 20 cmϪ1 to a value of 1924 cmϪ1
,
1585 cmϪ1 for R ϭ NNPh).[39Ϫ42] The IR band observed together with an appreciable decrease in absorptivity. Sub-
for the oxidized form of complex 2a is at the low-energy sequent reduction restores the original spectrum in nearly
border of this regime. In contrast to all other (vinylid- quantitative optical yield, showing that the smaller ab-
ene)ruthenium complexes so far reported it has the metal sorptivity is not an artifact of partial decomposition. In the
atom in a formal ϩIII oxidation state, which may account dipositive, monooxidized form, the unsaturated C3 ligand
for its fairly low energy. In any case, it is clear that ruth- bridges two positively charged entities. This should reduce
enium oxidation induces major bonding changes within the the change in dipole moment during this vibration and de-
metallacumulene chromophore.
crease the absorptivity of this band. The same reasoning
In complex 3a, oxidation involves the ferrocenyl moiety has been employed in explaining the considerably lower in-
at the terminal carbon atom of the allenylidene ligand. This tensity of the allenylidene stretch in neutral allenylidene
allows us to examine the effect of decreased electron density complexes as opposed to cationic ones, a situation qualitat-
at this site on the position of the IR band directly. As shown ively equivalent to the case here.[43]
884
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
The dppf-substituted complex 2b has its additional elec-
troactive moiety at exactly the opposite end of the metallac-
umulene chromophore. It provides a study case for the ef-
fect on the position of the IR band of diminished electron
density on the metal. Here, oxidation results in the disap-
pearance of the original IR band and the simultaneous
growth of a pair of closely spaced, equally intense bands at
1957 and 1948 cmϪ1, a blue shift relative to the RuII/FeII
precursor of an average of 15 cmϪ1 (Figure 8). The observa-
tion of a pair of bands instead of just one is most probably
the result of Fermi coupling of the asymmetric CϭCϭC
valence stretch with an overtone or a combination mode of
other vibrational mode(s) of closely related energies. Al-
Figure 9. IR spectroelectrochemistry of complex 3b (1,2-Cl2C2H4,
298 K); spectral changes concomitant with ruthenium oxidation
though too weak to be detected by themselves, they gain
intensity through coupling to the intense CϭCϭC stretch
of the unsaturated ligand, which then causes the additional
band. Similar observations have been made for the reduced less likely. No splitting of this band is observed for com-
forms of complexes 2a and 2b (vide infra). Notably, such plexes 3a or 3b, in which such rotamers are known to exist
apparent splitting of the valence stretch is observed only in and, from NMR spectroscopy (vide supra), not to intercon-
certain oxidation states. Generally speaking, Fermi coup- vert on the IR timescale. We also note that oxidation at the
ling is only encountered when there is a good energy match [Tp(dppf)Ru] site results in a net increase of absorptivity.
between the two underlying fundamentals. It is reasonable This can be explained well by the above reasoning, since
to assume that electron transfer alters the energy of at least the charge difference between the two sites bridged by the
one of the vibrational modes involved in this coupling, cumulenic chain, and hence the change in dipole moment
which would explain why such an effect may be observable associated with the asymmetric CϭCϭC stretch, is now en-
in one oxidation state but not in another. Paul, Lapinte, hanced.
and co-workers have very recently reported on such oxida-
Complex 3b, which features both types of secondary
tion-state-dependent Fermi coupling between CϵC and, redox centers, constitutes the most demanding case under
presumably, CϪCϵ modes[44] for alkynyl complexes study, since it poses the question of which ferrocenyl entity
[Cp*(dppe)FeϪCϵCϪC6H4R-4]nϩ (n ϭ 0, 1, FeII,III).[45,46] is oxidized first. We have already pointed out that the meas-
Similar coupling between the CϵN stretch and the ured redox potentials favor the first oxidation as involving
CϵNϪC bending modes of the isocyanide ligands in the the ferrocenyl substituent attached to the allenylidene li-
triruthenium clusters [Ru3(NCR)3(µ3-O)(µ-CH3COO)6]nϪ gand. IR spectroelectrochemistry provides compelling evid-
(n ϭ 0, 1) has recently been described by Kubiak et al.[47] ence for this assignment. During the first oxidation the
Since the band energy of the fundamental coupling with the CϭCϭC valence stretch at 1945 cmϪ1 is replaced by a sig-
CϭCϭC valence stretch is unknown, we adopt the center nificantly less intense absorption band at 1916 cmϪ1. The
of the Fermi doublet as the energy of the unperturbed second oxidation occurs at the dppf ligand and shifts the
CϭCϭC stretch (and also of the other mode), but this is IR band back to exactly its original position. This shows
only an approximation. The alternative explanation of there that Ϫ at least in terms of IR spectroscopy Ϫ the effects of
being two distinct conformers with appreciably different one-electron oxidation of the electroactive moieties posi-
ν˜(CϭCϭC) values cannot be discounted, but seems much tioned at opposite ends of the allenylidene ligand exactly
cancel each other. Both oxidized species display a high
degree of chemical stability even on the longer timescales of
spectroelectrochemical investigations. After bulk electrolysis
following the first and second oxidation the starting
material was recovered in ca. 95 and 92% optical yields,
respectively.
Despite being obscured by the rising background current
of the supporting electrolyte in cyclic voltammetry experi-
ments, it still proved possible to examine the ruthenium-
centered oxidation of complexes 3a and 3b under thin layer
spectroelectrochemistry conditions. The results are very
similar to those obtained for complex 2a: the IR bands in
the region typical of the CϭCϭC stretch of the cumulenic
ligand vanish and are replaced by new absorptions at 1507
(3a2ϩ) or 1524 cmϪ1 (3b3ϩ), again pointing to a vinylidene
Figure 8. IR spectroelectrochemistry of complex 2b (1,2-Cl2C2H4,
structure (see Figure 9). We also note a significant red shift
298 K); effect of oxidation of the ferrocene moiety of the dppf li-
gand
of the vinylidene band relative to the oxidized RuIII form
Eur. J. Inorg. Chem. 2003, 876Ϫ891
885

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
FULL PAPER
of 2a as the electron density of these complexes decreases secondary redox centers (Table 5). This reconfirms our pre-
further. vious assignment that it is the ferrocene moiety at the allen-
Electrolysis under UV/Vis monitoring conditions re- ylidene ligand that is oxidized first. The second oxidation
vealed the effects on the optical spectra of complexes 2b, of this compound causes only a slight blue shift of the lower
3a, and 3b of electron transfer from the secondary redox energy band, along with a further decrease in absorptivity.
sites. Spectral changes induced by oxidation of the dppf li-
For complexes 2a, 2b, and 3b we were even able to mon-
gand in 2b are rather small, the most prominent effects be- itor the spectral changes accompanying oxidation at the
ing a blue shift of the HOMO Ǟ LUMO band by 775 cmϪ1 ruthenium center. For 2a and 2b the effects are very similar.
and the development of a former shoulder at 361 nm into As an example, Figure 11 displays the spectroscopic traces
a distinct band located at 369 nm, while the main absorp- during the second oxidation of 2b. The HOMO Ǟ LUMO
tion band at 526 nm remains virtually unchanged. The most and HOMO-1 Ǟ LUMO bands experience shifts to higher
likely explanation is provided by reference to our earlier energies, by 1050 and 1300 cmϪ1 (2a) or 625 cmϪ1 for the
work on amino-substituted allenylidene complexes of the HOMO-1 Ǟ LUMO transition in 2b. In this latter com-
trans-[Cl(dppm)2Ru] fragment.[48] Here we found that the pound there is no distinct band for the excitation from the
electron density on the phosphane ligands was quite small HOMO level but only a broad absorption at ca. 700 nm.
for both occupied frontier levels, and this may also hold for As the oxidation proceeds, an intense band develops at
the mixed Tp/diphosphane complexes described here. Our 411 nm (2a) or 351 nm (2b). Since only one band in the
experimental observations can then be accommodated by 350Ϫ600-nm region is observed for the RuIII forms of allen-
the assumption that the HOMO level has some contribu- ylidene complexes of the ClRu(dppm)2 entity, we assume
tion from the phosphane ligands while there is only a negli- that the additional band (probably the one at higher energy)
gible one to the lower-lying HOMO-1 orbital. For com- arises from ligand-to-metal charge transfer from the Tp li-
plexes 3a and 3b, oxidation of the ferrocene moiety would gand to RuIII. The fully oxidized form of complex 3b differs
be expected to have a much more profound effect on the from the other two in that its spectrum, in terms of band
UV/Vis spectra, since the ferrocene moiety now constitutes positions, is surprisingly similar to that of the RuII/FeII(Fc)/
an integral part of the chromophore. Bulk electrolysis of 3a FeII(dppf) starting material. The intensities of the band
at a potential more positive than the first anodic wave in- near 700 and the one near 550 nm, however, are just the
deed causes a bleaching of the intense low-energy absorp- opposite of what they were in the beginning, the band at
tion band at 723 nm (Figure 10), this finding lending fur- lower energy now being very strong and the one at higher
ther support to our assignment of this band as resulting energy being rather weak. At present we have no plausible
from charge transfer from the ferrocene donor to the alleny- explanation for this pattern, but note that it seems to be
lidene acceptor moieties within this chromophore. As the associated with the ferrocenium moiety at the allenylidene
ferrocene moiety is oxidized, no such transition is possible ligand. This follows from comparison of the spectra of 2b2ϩ
any longer. Further changes are a slight red shift of the and 3b3ϩ. Another intense band at 468 nm may again arise
prominent absorption band by some 200 cmϪ1 and an ab- from a LMCT transition from the Tp ligand to RuIII
sorptivity decrease of the third prominent feature near
.
355 nm, accompanied by a slight shift to higher energies.
The direction of the shift of the former band is as would
be anticipated on the basis of our results on heteroatom-
substituted allenylidene complexes trans-[Cl(dppm)2Ruϭ
CϭCϭC(ERn)(RЈ)]ϩ, but its magnitude is again fairly
small. Nearly identical observations were made during the
first anodic step of compound 3b, featuring both kinds of
Figure 11. UV/Vis spectroelectrochemistry of complex 2b (1,2-
Cl2C2H4, 298 K); effect of oxidation of the ferrocene moiety of the
dppf ligand
b) Reduction
It has already been mentioned that the reduction of allen-
ylidene complexes mainly involves the allenylidene ligand.
From the shift of the IR band upon reduction of the SR
and SeR substituted complexes trans-[Cl(dppm)2RuϭCϭ
CϭC(ER)(RЈ)]ϩ and the coupling pattern of the unpaired
Figure 10. UV/Vis spectroelectrochemistry of complex 3a (1,2-
Cl2C2H4, 298 K); oxidation of the ferrocenyl substituent attached
to the allenylidene ligand
886
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
electron in their reduced forms, we have concluded that the
unpaired spin mainly resides on the terminal carbon
atom.[1,2] Touchard and co-workers have recently presented
convincing evidence that this is also true for the reduced
form of the diphenyl-substituted trans-[Cl(dppe)2RuϭCϭ
CϭCPh2]ϩ, closely related to the complexes discussed
here.[35] DFT calculations on variously substituted model
complexes
trans-[Cl(PH3)4RuϭCϭCϭC(ERn)(CH3)]·
[ERn ϭ N(CH3)2, O(CH3), S(CH3), Se(CH3), CH3], which
we will detail elsewhere, have broadly confirmed these con-
clusions but also indicate considerable spin density on the
metal-bound carbon atom C-α. This in turn fully agrees
with the orbital composition of the LUMO as established
by quantum mechanical calculations[4,9Ϫ11,37] and the well
established regiochemistry of nucleophilic addition to the
allenylidene ligand.[8,49] The reversible nature of the first re-
duction for each of the complexes presented here provided
us with the opportunity also to examine the spectral
changes accompanying this process. Again, we expected the
most revealing results to come from IR spectroelectrochem-
istry. Since none of the appended secondary redox sites can
be reduced, the results are very similar for all complexes.
Figure 12 displays the spectral traces recorded during the
reduction of complex 2b as a representative example. In
each case we observe a blue shift of the CϭCϭC band by
some 20 wavenumbers, along with an appreciable decrease
in intensity. We again confirmed that this was not caused
by decomposition of the reduced species, back-electrolysis
in each case restoring the original spectrum in optical yields
of more than 95%. We note that two IR bands spaced by
ca. 10 cmϪ1 were observed for the reduced forms of com-
plexes 2a and 2b, while there was only one for complexes
3a and 3b. The average shift, however, is almost constant
within the entire series and amounts to ca. 20 cmϪ1. While
all the mesomeric structures depicted in Scheme 5 place the
unpaired spin either on the carbon atom C-α or on C-γ, we
note that VI accounts for the high-energy shift of the IR
band. Relevant data are listed in Table 5.
Scheme 5
Figure 13. UV/Vis spectroelectrochemistry of complex 2b (1,2-
Cl2C2H4, 298 K); spectral changes accompanying reduction
In UV/Vis spectroelectrochemical experiments, reduction
was observed to change the intensities of the original ab-
sorption bands rather than their positions. Representative
Figure 14. UV/Vis spectroelectrochemistry of complex 3a (1,2-
Cl2C2H4, 298 K); spectral changes accompanying reduction
examples are provided in Figures 13 and 14, while the cor- responding data are listed in Table 5. This somewhat unex-
pected result points to a nearly parallel energy shift of all
frontier orbitals involved in the electronic transitions upon
reduction. The lower absorptivities may originate from par-
tial loss of the charge-transfer character of these bands as
the electron-accepting propensity of the allenylidene ligand
is diminished. We note that the main absorption band ori-
ginating from the HOMO-1 Ǟ LUMO transition seemingly
splits into two individual bands, which remain unresolved
for complex 2b but are clearly resolved for all other com-
plexes, the energy difference increasing in the order 2b Ͻ
2a Ͻ 3a ഠ 3b. Another striking observation is the complete
loss of the ferrocenyl-to-ligand (or intraligand) charge-
transfer absorption for the ferrocenyl-substituted complexes
3a and 3b, while a weak band identical to that seen for
the HOMO Ǟ LUMO transition in the reduced forms of
complexes 2a and 2b is now observed for 3b (Table 5). Des-
Figure 12. IR spectroelectrochemistry of complex 2b (1,2-Cl2C2H4,
298 K); spectral changes accompanying reduction
Eur. J. Inorg. Chem. 2003, 876Ϫ891
887

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
FULL PAPER
pite their loss of intensity, the transition bands in the visible
region are still quite strong and obscure the expected ferro-
cenyl-based absorptions. A new band at 400Ϫ410 nm was
observed for all complexes and cannot conclusively be as-
signed at present. Another feature common to all reduced
forms is an intense absorption at ca. 350 nm. Bands of this
type are familiar for aryl-substituted methyl radicals of the
type ArnCR3Ϫn·, which are characterized by sharp, intense
absorptions at 325Ϫ360 nm.[50Ϫ53] This argues for efficient
incorporation of the aryl substituents into the RuϭCϭCϭ
C(R)(RЈ) chromophore and the relevance of resonance
forms VI and VII, which both place the unpaired spin on
the carbon atom adjacent to the aromatic rings.
More evidence for spin delocalization into the aromatic
substituents was provided by EPR spectroscopy. The re-
duced forms of complexes 2a and 2b, frozen in a 1,2-
Cl2C2H4 glass at 4 K, display slightly broadened isotropic
signals at g-values typical of organic paramagnetic species
(i.e., close to the free electron value of 2.0023; Table 6). For
3a and 3b, however, the spectra were of the rhombic type,
with small but clearly detectable g-value anisotropies indic-
ating some transition metal contribution, most probably
from the 57Fe nucleus of the Fc substituent. Such coupling
also prevails in liquid solution, as is evident from a satellite
doublet with a signal splitting of 22.3 G (Figure 15; 57Fe
has a nuclear spin of 1/2 and a natural abundance of
2.15%). No further resolution of the EPR signal was ob-
tained, however. This is in sharp contrast to observations
made by Rigaut, Dixneuf, and co-workers and in our own
laboratory on the reduced forms of allenylidene complexes
bearing the trans-[Cl(L2)2Ru] fragment (L2 ϭ dppm, dppe).
These display resolved couplings to the hydrogen atoms on
the substituents bonded to the allenylidene ligand and the
Figure 15. EPR spectra of the reduced form of complex 3a in (a)
a 1,2-Cl2C2H4 glass at 4 K (top trace) or (b) in 1,2-Cl2C2H4 solu-
tion at 300 K (bottom trace)
c) The IR Frequency of the B؊H Stretch
The position of the BϪH stretch in infrared spectra has
phosphane P atoms coordinated to the metal ion.[1Ϫ3,5] In been recognized as an indicator for the coordination mode
the mixed Tp/phosphane complexes described here, coup- of the Tp ligand (i.e., κ3 versus κ2 coordination).[54,55] It
ling to the arene protons, the two magnetically different sets has, however, only very recently been realized that the fre-
of nitrogen atoms of the Tp ligand, and the phosphane P quency of this band also corresponds to the electron density
atoms would therefore be expected. It is possible that the on the metal ion to which it is coordinated and, more spe-
larger number of magnetically independent nuclei together cifically, to its oxidation state. Thus, O’Hare and co-workers
with some participation by a metal atom give rise to have just published a series of mixed-sandwich complexes
strongly overlapping resonance lines that cannot be re- containing both the Cp and the Tp ligand or the 3,5-di-
solved. It is quite revealing to note that the resonance sig- methyl derivative thereof, several of them in the adjacent
nals are significantly broadened relative to the analogous MII and MIII oxidation states (M ϭ V, Cr, Co, Ni).[55] In
complexes of the trans-[Cl(dppm)2Ru] entity.
all these cases oxidation of the metal atom was found to
induce a shift of the BϪH band to higher energies, though
of a strongly varying magnitude (7Ϫ72 cmϪ1). We have
made the same observations with the complexes reported
here. As the data for complexes 2a and 3b show (see
Table 5), oxidation at the ruthenium atom increases the en-
ergy of the BϪH stretch by about 30 cmϪ1 while the de-
creasing electron density on the ruthenium atom by oxida-
tion at the dppf ligand has a somewhat smaller effect of
some 25 cmϪ1 (2b). Ferrocene oxidation at the allenylidene
ligand also shifts the BϪH band to higher energy (3a).
Likewise, the reduced forms of these allenylidene complexes
display their BϪH band at some 15Ϫ25 cmϪ1 lower energy,
although this process mainly involves the organic ligand.
Thus, variation of the BϪH stretch with the oxidation state
Table 6. EPR data for the reduced forms of complexes 2aϪ3b
4 K
298 K
giso or gxx, gyy, gzz (gav)[a]
giso (A{57Fe})
2a
2b
3a
3b
2.0027
2.0027
2.0018 (Ϫ)
2.0030 (Ϫ)
2.0108 (22.3 G)
2.0110 (22.3 G)
2.0228, 2.0094, 2.0011 (2.0111)
2.0215, 2.0111, 2.0016 (2.0114)
[a]
2
2
gav ϭ {1/3 (gxx ϩ gyy ϩ gzz2)}1/2
.
888
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
of the metal atom or that of any redox-active ancillary li- the latter presented here.[13] For the dppf-substituted 3b,
gand bonded to it seems to be a quite general result and we however, splitting of the singlet 31P NMR resonance into a
expect additional examples of such behavior as more redox well-resolved AB quadruplet is observed upon cooling in
pairs of this type are characterized.
CD2Cl2, pointing to a horizontal alignment of the allenylid-
ene ligand. The energy barrier for allenylidene rotation is
determined as 47 kJ/mol at the coalescence temperature of
238 K.
Conclusions
We have investigated allenylidene complexes [TpL2Ruϭ
CϭCϭC(Ph)(R)]ϩ. Substitution of dppf by less strongly
donating PPh3 ligands or of the phenyl by a ferrocenyl sub-
stituent shifts the IR and electronic absorption bands to
lower energies, in line with previous results on complexes
based on the trans-[Cl(dppm)2Ru] moiety. Oxidation of
ferrocene-based redox sites attached either to the metal
atom (L2 ϭ dppf: 2b, 3b) or to the allenylidene ligand (R ϭ
Fc: 3a and 3b) strongly attenuates these effects. The same
holds for reduction, which mainly involves a ligand-cent-
ered, delocalized orbital. These studies also show that aro-
matic substituents at the allenylidene ligand are efficiently
incorporated into the chromophoric system. Thus, the Fc-
substituted complexes 3a and 3b display an intense, low-
energy absorption band attributable to charge transfer from
the ferrocenyl substituent to the allenylidene ligand. Oxida-
tion of the Fc moiety and reduction both cause bleaching
of this feature. EPR studies on the reduced forms of 3a and
3b reveal coupling of the unpaired spin with the 57Fe nuc-
leus of the ferrocenyl substituent. In their IR spectra some
of the oxidized or reduced forms display an apparent split-
ting of the allenylidene band. This is interpreted in terms
of Fermi coupling between the CϭCϭC fundamental and
an overtone or combination of (an)other vibration(s).
Cyclic voltammetry studies reveal that both types of
ferrocene-based redox systems are oxidized at much lower
potential than the ruthenium center. Although the RuII/III
process lies outside the anodic window of the supporting
electrolyte solution in voltammetric studies, it was still pos-
sible to perform this oxidation under thin layer conditions
in an OTTLE cell. This latter oxidation induces more sub-
stantial bonding changes within the cumulenylidene system.
Results from IR spectroelectrochemistry point to a vinylid-
ene resonance structure for the RuIII forms of these com-
plexes. The frequency of the BϪH stretch of the Tp ligand
is also sensitive to the overall oxidation state of the molec-
ule. Oxidation of any of the ferrocene entities or the ruth-
enium shifts this band to higher energies, while a red shift
is observed upon reduction.
Experimental Section
General: All manipulations were performed by standard Schlenk
techniques under argon. CH2Cl2 and 1,2-Cl2C2H4 were distilled
from CaH2 and stored over molecular sieves (4 A). All solvents
˚
were deoxygenated by at least three freeze-pump-thaw cycles and
saturation with argon prior to use. Benzoylferrocene,
[Tp(dppf)RuCl], [TpRu(PPh3)2Cl], and [TpL2RuϭCϭCϭCPh2]ϩ
(L2 ϭ 2 PPh3, dppf) were obtained by literature methods.[13,18,56,57]
Electrochemistry and spectroelectrochemistry were performed as
described in an earlier publication.[4] All experiments were per-
formed in the presence of 0.2 m Bu4NPF6 as the supporting elec-
trolyte and the given potentials are referenced to the ferrocene/
ferrocenium couple, set as 0.00 V. This scale can be related to satur-
ated calomel electrode (SCE) conditions by adding ϩ0.46 V for
CH2Cl2 and ϩ0.45 V for DMF solutions.[58] NMR: Bruker AC
250. IR: Elmer Paragon 1000 PC. UV/Vis: Omega 10 from Bruins
Instruments or J&M TIDAS diode array UV/Vis/NIR spectro-
meter equipped with a fiber-optic cable and operated with the
SCAN 2017 and SPECTRALYS software packages. Electrochem-
istry: EG&G potentiostat 273 A, BAS glassy carbon working elec-
trode (Ø, 3 mm), Pt wire as counter-electrode and Ag wire as
pseudo-reference electrode, potential calibration by addition of
ferrocene or decamethylferrocene as internal standard. EPR:
Bruker ESP 300, HP frequency counter 5350 B, Bruker NMR
gaussmeter ER 035 M, continuous flow cryostat ESR 900 from
Oxford Instruments for low-temperature work.
1-Ferrocenyl-1-phenylprop-2-yn-1-ol Trimethylsilyl Ether: nBuLi
(4.6 mL of a 2.5  solution in n-hexane) was added to trimethylsilyl-
acetylene (1.6 mL, 1.11 g, 11.2 mmol) in degassed THF (250 mL).
The clear, colorless solution was allowed to warm to room temper-
ature, and benzoylferrocene (2.32 g, 8 mmol) was added. The re-
sulting solution was stirred for 20 h, whereupon the color visibly
brightened to orange. It was hydrolyzed with 5 mL of distilled
water and the organic phase was taken to dryness. The brownish
residue was taken up in dichloromethane, washed with several por-
tions of a 5% solution of Na2CO3, and dried with MgSO4, and the
solvent was removed in vacuo. The product was purified by column
chromatography on neutral alumina with n-hexane/Et2O (3:1, v/v)
as the eluent. Slow evaporation of the solvents deposited orange
crystals (1.67 g, 4.3 mmol, 54%).
1-Ferrocenyl-1-phenylprop-2-yn-1-ol: Finely ground KOH (1.68 g,
30 mmol) was dissolved in MeOH, and the silyl ether (1.165 g,
3 mmol) was added. The orange solution was kept under a gentle
reflux for 18 h and the solvent was then distilled off. The residue
was hydrolyzed in water and carefully acidified with dilute HCl,
and the product was extracted with Et2O. The organic phase was
washed with water and dried with MgSO4, and the solvent was
removed in vacuo, which deposited the pure product as orange
crystals (0.81 g, 2.6 mmol, 85%).
The unsymmetrically substituted allenylidene complexes
[TpL2RuϭCϭCϭC(Ph)(Fc)]ϩ (3a and 3b) are dynamic in
solution, due to an energy barrier to rotation around the
RuϭC bond. Two isomeric forms differing in the relative
positions of the Fc and the Ph substituents are discernible
for the PPh3-substituted congener 3a up to 388 K in 1,2-
D2C2Cl4, at which point this compound decomposes. The
observation of just one singlet resonance for each isomer
by 31P NMR spectroscopy argues for a vertical orientation
of the allenylidene ligand. This is in agreement with the
[Tp(PPh3)2Ru؍C؍C؍CPh2]؉ SbF6Ϫ (2a): UV/Vis (CH3CN): λmax
(ε) ϭ 359 (10500), 527 (16500), 901 nm (710 Ϫ1 cmϪ1); (CH2Cl2):
crystallographically determined structures for 2a and 2b, λmax (ε) ϭ 356 (13000), 527 (16000), 907 nm (660 Ϫ1 cmϪ1).
Eur. J. Inorg. Chem. 2003, 876Ϫ891
889

S. Hartmann, R. F. Winter, B. M. Brunner, B. Sarkar, A. Knödler, I. Hartenbach
FULL PAPER
[Tp(dppf)Ru؍C؍C؍CPh2]؉ SbF6 (2b): UV/Vis (CH3CN): λmax Ph), 699 (s, Ph), 658 cmϪ1 (SbF6). UV/Vis (CH3CN): λmax (ε) ϭ
Ϫ
(ε) ϭ 346 (11000), 520 (13000), 901 nm (350 Ϫ1 cmϪ1); (CH2Cl2):
λmax (ε) ϭ 356 (16500), 522 (19600), 902 nm (580 Ϫ1 cmϪ1).
355 (3600), 534 (5200), 719 (2650 Ϫ1 cmϪ1); 825 nm (sh);
(CH2Cl2): λmax (ε) ϭ 356 (8000), 538 (11500), 724 nm (5700 Ϫ1
cmϪ1). C64H54BF6FeN6P2RuSb (1372.59): calcd. C 56.00, H 3.97,
N 6.12; found C 55.78, H 4.20, N 5.61.
Ϫ
[Tp(PPh3)2Ru؍C؍C؍CFcPh]؉ SbF6
(250 mg, 0.286 mmol),
(3a): Compound 1a
1-ferrocenyl-1-phenylprop-2-yn-1-ol
(260 mg, 0.82 mmol), and AgSbF6 (190 mg, 0.55 mmol) were mixed
in THF (10 mL) in a Schlenk flask and stirred in the dark at ambi-
ent temperature for 40 min, whereupon an intense purple color de-
veloped. THF was removed in vacuo and the solid residue was
extracted with CH2Cl2 (15 mL). Filtration through a fine frit co-
vered with Celite (10 cm) and silica (5 cm), with use of small por-
tions of CH2Cl2, removed the AgCl. The solid residue obtained
after the solvent had been pumped off, was subjected to column
chromatography on silica (Merck 60) with CH2Cl2/CH3CN (10:1,v/
v) as the eluent. The first, orange band contained the excess propar-
gylic alcohol, while the second, purple-black zone was slightly im-
pure product. Further purification was performed by three success-
ive reprecipitations from CH2Cl2/hexanes. After drying in vacuo,
spectroscopically pure product (251 mg, 0.183 mmol, 64%) was ob-
[Tp(dppf)Ru؍C؍C؍CFcPh]؉ SbF6 (3b): This compound was
Ϫ
prepared analogously to the method described for the synthesis and
purification of compound 3a. The product (289 mg, 74.7%) was
obtained from 1b (250 mg, 0.276 mmol), 1-ferrocenyl-1-
phenylprop-2-yn-1-ol (0.26 g, 0.82 mmol), and AgSbF6 (190 mg,
0.55 mmol). 1H NMR (CD2Cl2): δ ϭ 4.22 (s, 5 H, C5H5), 4.67, 4.76
3
[each br., 4 H, 2,3-H(dppf)], 5.05, 5.12 (each t, JH,H ϭ 2.1 Hz, 2
H, Cpsubst), 5.39 [t, 3JH,H ϭ 1.9 Hz, 2 H, 4-H(pz)], 5.93 [t, 3JH,H ϭ
3
2.2 Hz, 1 H, 4-H(pz)], 6.33 [d, JH,H ϭ 2.2 Hz, 1 H, 3-H(pz)], 6.39
[d, JH,H ϭ 1.9 Hz, 2 H, 3-H(pz)], 6.42Ϫ8.20 (several m, 26 H, 2
3
3
PPh2, Ph, HB), 7.64 [d, JH,H ϭ 2.2 Hz, 1 H, 5-H(pz)], 7.81 [d,
3JH,H ϭ 1.9 Hz, 2 H, 5-H(pz)] ppm. 13C{1H} NMR (62.9 MHz,
CD2Cl2): δ ϭ 70.6 (Cipso, Fc), 73.3 [t, 2JP,C ϭ 3.6 Hz, C-2/5(dppf)],
3
1
73.59 [Cipso(dppf)], 73.63 (C5H5), 73.9 (Cpsub), 74.6 [t, JP,C
2.9 Hz, C-3/4(dppf), 79.0 (Cpsub), 106.0, 106.5 [C-4(pz)],
128.0Ϫ133.5 (several m, 2 PPh2, Ph), 135.4, 137.3 [br., C-5(pz)],
144.0 (t, JP,C ϭ 0.7 Hz, Cipso, Ph), 144.8 [t, JP,C ϭ 1.1 Hz, C-
3(pz)], 147.3 (t, JP,C ϭ 1.5 Hz), 167.9 (t, JP,C ϭ 1.5 Hz, C-γ),
180.6 (br., C-β), 283.7 [t, JPC ϭ 18.2 Hz, C-α] ppm. 31P{1H} NMR
(101.3 MHz, CD2Cl2): δ ϭ 29.9 (s) ppm. IR (KBr): ν˜ ϭ 1941 (vs,
CϭCϭC), 1457 (s, Fc), 1435 (s, Ph), 1395 (m, Ph), 1382 (m, Ph),
1309 (m, Ph), 1049 (m, Ph), 749 (m, Ph), 697 (s, Ph), 658 (SbF6)
cmϪ1. UV/Vis (CH3CN): λmax (ε) ϭ 348 (14000), 534 (15500), 709
(10000), 1030 nm (800 Ϫ1 cmϪ1); (CH2Cl2): λmax (ε) ϭ 350
(16500), 540 (19500), 712 (13000), 1030 nm (2000 Ϫ1 cmϪ1).
ϭ
tained. Data for the major isomer: H NMR (CD2Cl2): δ ϭ 4.28
3
(s, 5 H, C5H5), 4.81, 5.05 (each t, JH,H ϭ 2.0 Hz, 2 H, Cpsubst),
3
3
5.58 [t, JH,H ϭ 2.3 Hz, 1 H, 4-H(pz)], 5.69 [d, JH,H ϭ 2.3 Hz, 1
H, 3-H(pz)], 5.72 [t, 3JH,H ϭ 2.3 Hz, 2 H, 4-H(pz)], 6.31 [d, 3JH,H ϭ
2.3 Hz, 2 H, 3-H(pz)], 6.85Ϫ8.03 (several m, 36 H, 2 PPh3, Ph,
5
3
3
4
3
3
HB), 7.66 [d, JH,H ϭ 2.3 Hz, 2 H, 5-H(pz)], 7.85 [d, JH,H
ϭ
2.3 Hz, 1 H, 5-H(pz)] ppm. 13C{1H} NMR (62.9 MHz, CD2Cl2):
δ ϭ 71.6 (Cipso, Fc), 73.2 (C5H5), 74.5, 78.3 (C-2Ϫ5, Cpsub), 105.7,
106.5 [br., C-4(pz)], 125.5Ϫ134.5 (several m, 2 PPh3, Ph), 136.5,
137.0 (br., C-5(pz)], 144.9 (Cipso, Ph), 145.3, 147.0 (C-3(pz)], 168.2
4
3
(t, JP,C ϭ 1.6 Hz, C-γ), 184.3 (t, JP,C ϭ 1.5 Hz, C-β), 287.0 [t,
JPC ϭ 18.2 Hz, C-α] ppm. 31P{1H} NMR (101.3 MHz, CD2Cl2):
δ ϭ 40.8 (s) ppm. IR (KBr): ν˜ ϭ 1946 (vs, CϭCϭC), 1458 (s, Fc), C62H52BF6Fe2N6P2RuSb (1402.40): calcd. C 53.10, H 3.73, N 5.99;
1436 (s, Ph), 1382 (m, Ph), 1309 (m, Ph), 1049 (m, Ph), 749 (m, found C 52.32, H 3.85, N 5.54.
Ϫ
Table 7. Crystallographic data for [TpRu(dppf)Cl]·2CHCl3 and [Tp(dppf)RuC3Ph2]ϩ SbF6
Ϫ
[TpRu(dppf)Cl]·2CHCl3
[Tp(dppf)RuC3Ph2]ϩ SbF6
Empirical formula
Formula mass
C45H40BCl7FeN6P2Ru
1142.65
173(2)
0.71073
C58H48BF6FeN6P2RuSb
1294.44
293(2)
0.71073
Temperature [K]
˚
Wavelength [A]
Crystal system
Space group
monoclinic
P21/n
13.977(3)
17.287(4)
orthorhombic
Pbca
21.9546(14)
27.959(3)
˚
a [A]
˚
b [A]
˚
c [A]
19.645(4)
22.152(2)
β [°]
91.85(3)
4744.1(16)
90
13597(2)
3
˚
V [A ]
Z
4
8
Density (calculated) [Mg/m3]
Absorption coefficient [mmϪ1
F(000)
1.600
1.127
2304
1.265
0.924
5184
]
Crystal size [mm]
θ range for data collection [°]
Index ranges for data collection
No. of reflections measured
No. of independent reflections
Refinement method
0.3 ϫ 0.25 ϫ 0.15
1.57 to 29.03.
0 Յ h Յ 19, Ϫ14 Յ k Յ 23, Ϫ22 Յ l Յ 26
12760
8069 [R(int) ϭ 0.0526]
full-matrix least-squares on F2
12373/0/573
0.25 ϫ 0.07 ϫ 0.07
1.96 to 19.04
Ϫ18 Յ h Յ 20, Ϫ25 Յ k Յ 25, Ϫ18 Յ l Յ 19
15729
4616
full-matrix least-squares on F2
4616/0/360
Data/restraints/parameters
Goodness-of-fit on F2
1.288
1.522
Final R indices [I Ͼ 4σ(I)]
R indices (all data)
Largest diff. peak/hole [e·A
R1 ϭ 0.0757, wR2 ϭ 0.1829
R1 ϭ 0.1269, wR2 ϭ 0.2218
2.10/Ϫ2.34
R1 ϭ 0.156, wR2 ϭ 0.4112
R1 ϭ 0.2473, wR2 ϭ 0.4528
1.648/Ϫ1.863
Ϫ3
˚
]
890
Eur. J. Inorg. Chem. 2003, 876Ϫ891

(Allenylidene)ruthenium Complexes with Redox-Active Substituents and Ligands
FULL PAPER
[23]
D. Pilette, K. Ouzzine, H. Le Bozec, P. H. Dixneuf, C. E. F.
Rickard, W. R. Roper, Organometallics 1992, 11, 809Ϫ817.
D. Touchard, N. Pirio, P. H. Dixneuf, Organometallics 1995,
14, 4920Ϫ4928.
N. M. Kostic, R. F. Fenske, Organometallics 1982, 1, 974Ϫ982.
G. Consiglio, F. Bangerter, C. Darpin, Organometallics 1984,
3, 1448Ϫ1450.
G. Consiglio, F. Morandini, Inorg. Chim. Acta 1987, 127,
79Ϫ85.
M. P. Gamasa, J. Gimeno, E. Lastra, B. M. Martın, A. Anillo,
A. Tiripicchio, Organometallics 1992, 11, 1373Ϫ1381.
K. Urtel, A. Frick, G. Huttner, L. Zsolnai, P. Kircher, P.
Rutsch, E. Kaifer, A. Jacobi, Eur. J. Inorg. Chem. 2000, 33Ϫ50.
D. Unseld, V. V. Krivykh, K. Heinze, F. Wild, G. Artus, H.
Schmalle, H. Berke, Organometallics 1999, 18, 1525Ϫ1541.
R. L. Beddoes, C. Bitchon, R. W. Grime, A. Ricalton, M. W.
Whiteley, J. Chem. Soc., Dalton Trans. 1995, 2873Ϫ2883.
A. Wong, J. A. Gladysz, J. Am. Chem. Soc. 1982, 104,
4948Ϫ4950.
B. E. R. Schilling, R. Hoffmann, D. L. Lichtenberger, J. Am.
Chem. Soc. 1979, 101, 585.
H. Goetz, F. Nerdel, K.-H. Wiechel, Justus Liebigs Ann. Chem.
1963, 665, 1Ϫ14.
X-ray Crystallography: Crystals of complex 1b were grown by slow
concentration of a saturated solution in CHCl3 while those of 2b
were obtained by cooling a solution in a THF/Et2O mixture to Ϫ15
°C. Experimental details regarding the X-ray structure solutions
are provided in Table 7. The structures were solved by direct
methods, least-squares refinement, and Fourier techniques. Struc-
ture solution and refinement were performed with SHELXTL.[59]
CCDC-192092 (2b) and -192093 (1b) contain the supplementary
crystallographic data for this paper. These data can be obtained
free of charge at www.ccdc.cam.ac.uk/conts/retrieving.html [or
from the Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK; Fax: (internat.) ϩ 44-1223/336-
033; E-mail: deposit@ccdc.cam.ac.uk].
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
´
Acknowledgments
We are grateful to the Volkswagenstiftung for financial support
within their program on intra- and intermolecular electron transfer.
We also wish to acknowledge Johnson & Matthey Company for a
generous loan of RuCl3 hydrate.
S. Rigaut, O. Maury, D. Touchard, P. H. Dixneuf, Chem. Com-
mun. 2001, 373Ϫ374.
[1]
R. F. Winter, Eur. J. Inorg. Chem. 1999, 2121Ϫ2126.
S. Hartmann, R. F. Winter, T. Scheiring, M. Wanner, J. Or-
[2]
R. S. Nicholson, I. Shain, Anal. Chem. 1964, 36, 706Ϫ723.
[37] [37a]
ganomet. Chem. 2001, 637Ϫ639, 240.
N. Re, A. Sgamellotti, C. Floriani, Organometallics 2000,
[3]
19, 1115Ϫ1122. [37b] A. Marrone, N. Re, Organometallics 2002,
21, 3562Ϫ3571.
R. F. Winter, Organometallics 1997, 16, 4248Ϫ4250.
[4]
´ ˇ
R. F. Winter, K.-W. Klinkhammer, S. Zalis, Organometallics
[38]
[39]
2001, 20, 1317Ϫ1333.
M. I. Bruce, P. Hinterding, P. J. Low, B. W. Skelton, A. H.
White, Chem. Commun. 1996, 1009Ϫ1010.
M. I. Bruce, P. Hinterding, P. J. Low, B. W. Skelton, A. H.
White, J. Chem. Soc., Dalton Trans. 1998, 467Ϫ473.
V. Guillaume, P. Thominot, F. Coat, A. Mari, C. Lapinte, J.
Organomet. Chem. 1998, 565, 75Ϫ80.
M. I. Bruce, Chem. Rev. 1998, 98, 2797Ϫ2858.
V. Cadierno, M. P. Gamasa, J. Gimeno, M. Gonzalez-Cueva,
E. Lastra, J. Borge, S. Garcıa-Granda, E. Perez-Carren˜o, Or-
ganometallics 1996, 15, 2137Ϫ2147.
M. A. Esteruelas, A. V. Gomez, A. M. Lopez, J. Modrego, E.
M. I. Bruce, Chem. Rev. 1991, 91, 107Ϫ257.
M. I. Bruce, M. G. Humphrey, G. A. Koutsantonis, M. J. Lid-
dell, J. Organomet. Chem. 1987, 326, 247Ϫ256.
M. I. Bruce, G. A. Koutsantonis, M. J. Liddell, B. K. Nichol-
son, J. Organomet. Chem. 1987, 320, 217Ϫ227.
M. I. Bruce, M. G. Humphrey, M. J. Liddell, J. Organomet.
Chem. 1987, 321, 91Ϫ102.
M. I. Bruce, M. G. Humphrey, Aust. J. Chem. 1989, 42,
1067Ϫ1075.
K. J. Harlow, A. F. Hill, J. D. E. T. Wilton-Ely, J. Chem. Soc.,
Dalton Trans. 1999, 285Ϫ291.
R. M. Badger, J. Chem. Phys. 1937, 5, 178.
R. Denis, L. Toupet, F. Paul, C. Lapinte, Organometallics 2000,
19, 4240Ϫ4251.
F. Paul, J.-Y. Mevellec, C. Lapinte, J. Chem. Soc., Dalton Trans.
2002, 1783Ϫ1790.
I. S. Zavarine, C. P. Kubiak, T. Yamaguchi, K.-i. Ota, T. Mat-
sui, T. Ito, Inorg. Chem. 2000, 39, 2696Ϫ2698.
R. F. Winter, S. Hartmann, S. Zalis, K.-W. Klinkhammer, to
be submitted to J. Chem. Soc., Dalton Trans.
V. Cadierno, M. P. Gamasa, J. Gimeno, Eur. J. Inorg. Chem.
2001, 571Ϫ591.
A. Bromberg, D. Meisel, J. Phys. Chem. 1985, 89, 2507Ϫ2513.
J. L. Faria, S. Steenken, J. Phys. Chem. 1993, 97, 1924Ϫ1930.
M. B. Jones, V. M. Maloney, M. S. Platz, J. Am. Chem. Soc.
1992, 114, 2163Ϫ2169.
[5]
[6]
[7]
[40]
[41]
[42]
[43]
[8]
[9]
´
´
´
[10]
[11]
[44]
[45]
´
´
On˜ate, Organometallics 1997, 16, 5826Ϫ5835.
´
M. Baya, P. Crochet, M. A. Esteruelas, E. Gutierrez-Puebla,
[46]
[47]
[48]
[49]
´
A. M. Lopez, J. Modrego, E. On˜ate, N. Vela, Organometallics
2000, 19, 2585Ϫ2596.
[12]
[13]
R. F. Winter, Chem. Commun. 1998, 2209Ϫ2210.
B. Buriez, I. D. Burns, A. F. Hill, A. J. P. White, D. J. Williams,
J. D. E. T. Wilton-Ely, Organometallics 1999, 18, 1504Ϫ1516.
B. Buriez, D. J. Cook, K. J. Harlow, A. F. Hill, T. Welton, A.
J. P. White, D. J. Williams, J. D. E. T. Wilton-Ely, J. Organomet.
Chem. 1999, 264Ϫ267.
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[50]
[51]
[52]
K. J. Harlow, A. F. Hill, T. Welton, J. Chem. Soc., Dalton
Trans. 1999, 1911Ϫ1912.
M. A. J. Tenorio, M. J. Tenorio, M. C. Puerta, P. Valerga, Or-
ganometallic 1997, 16, 5528Ϫ5535.
[53]
[54]
N. W. Alcock, I. D. Burns, K. S. Claire, A. F. Hill, Inorg. Chem.
1992, 31, 2906Ϫ2908.
D. Weir, J. Phys. Chem. 1990, 94, 5870Ϫ5875.
M. Akita, K. Ohta, Y. Takahashi, S. Hikichi, Y. Moro-Oka,
Organometallics 1997, 16, 4121Ϫ2128.
T. J. Bunker, A. R. Cowley, D. OЈHare, Organometallics 2002,
21, 3123Ϫ3138.
J. Tirouflet, R. Dabard, E. Laviron, Bull. Chem. Soc. Fr.
1963, 1655.
G. W. Gokel, I. K. Ugi, J. Chem. Educ. 1972, 49, 294.
N. G. Connelly, W. E. Geiger, Chem. Rev. 1996, 96, 877Ϫ910.
G. M. Sheldrick, Institut für Anorganische Chemie der Univer-
sität, Tammanstraße 4: Göttingen, 1997.
C. Slugovc, V. N. Sapunov, P. Wiede, K. Mereiter, R. Schmid,
K. Kirchner, J. Chem. Soc., Dalton Trans. 1997, 4209Ϫ4216.
S. Pavlik, C. Gemel, C. Slugovc, K. Mereiter, R. Schmid, K.
Kirchner, J. Organomet. Chem. 2001, 617Ϫ618, 301Ϫ310.
M. Tamm, T. Jentzsch, W. Werncke, Organometallics 1997,
16, 1418Ϫ1422.
M. Tamm, A. Grzegorzewski, I. Brüdgam, H. Hartl, J. Chem.
Soc., Dalton Trans. 1998, 3523Ϫ3528.
A. Z. Kreindlin, F. M. Dolgushin, A. I. Yanovsky, Z. A. Kerz-
ina, P. V. Petrovskii, M. I. Rybinskaya, J. Organomet. Chem.
2000, 616, 106Ϫ111.
[55]
[56]
[57]
[58]
[59]
Received August 26, 2002
[I02487]
Eur. J. Inorg. Chem. 2003, 876Ϫ891
891