Electron delocalization in vinyl ruthenium substituted cyclophanes: Assessment of the through-space and the through-bond pathways

Article abstract of DOI:10.1016/j.jorganchem.2011.06.028

Pseudo-para[2.2]paracyclophane- and [2.1]orthocyclophane-bridged diruthenium complexes 2 and 3 with two interlinked electroactive styryl ruthenium moieties have been prepared and investigated. Both complexes undergo two reversible consecutive one-electron oxidation processes which are separated by 270 or 105 mV. Stepwise electrolysis of the neutral complexes to first the mixed-valent radical cations and then the dioxidized dications under IR monitoring reveal incremental shifts of the charge-sensitive Ru(CO) bands and allow for an assignment of their radical cations as moderately or very weakly coupled mixed-valent systems of class II according to Robin and Day. Ground-state delocalization in the mixed-valent forms of these complexes as based on the CO band shifts is considerably larger for the "closed" paracyclophane as for the "half-open" orthocyclophane. Experimental findings are backed by the calculated IR band patterns and spin density distributions for radical cations of slightly simplified model complexes 2 ^Me+ and 3^Me+ with the PⁱPr₃ ligands replaced by PMe₃. Radical cations 2⁺ and 3⁺ feature a characteristic NIR band that is neither present in their neutral or fully oxidized forms nor in the radical cation of the monoruthenium [2.2]paracyclophane complex 1 with just one vinyl ruthenium moiety. These bands are thus assigned as intervalence charge-transfer (IVCT) transitions. Our results indicate that, for the radical cations, electronic coupling "through-space" via the stacked styrene decks is significantly more efficient than the "through-bond" pathway.

Full text of DOI:10.1016/j.jorganchem.2011.06.028

Journal of Organometallic Chemistry 696 (2011) 3186e3197
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Electron delocalization in vinyl ruthenium substituted cyclophanes: Assessment
of the through-space and the through-bond pathways
,
1
,
ꢀe
Philipp Mücke a, Manfred Zabel b, Ruth Edge c, David Collison c, Sébastien Clément d , Stanislav Zális
,
*
,
Rainer F. Winter a
**
a Fachbereich Chemie der Universität Konstanz, Universitätsstraße 31, D-78457 Konstanz, Germany
b Institut für Anorganische Chemie, Fachbereich Chemie und Pharmazie der Universität Regensburg, Universitätsstraße 31, D-93040 Regensburg, Germany
c EPSRC National EPR Service, University of Manchester, School of Chemistry, Oxford Road, Manchester, M13 9PL, United Kingdom
d Institut UTINAM UMR CNRS 6213, Université de Franche Comté, 25030 Besançon, France
e J. Heyrovský Institute of Physical Chemistry, v.v.i, Academy of Sciences of the Czech Republic, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history:
Pseudo-para[2.2]paracyclophane- and [2.1]orthocyclophane-bridged diruthenium complexes 2 and 3
with two interlinked electroactive styryl ruthenium moieties have been prepared and investigated. Both
complexes undergo two reversible consecutive one-electron oxidation processes which are separated by
270 or 105 mV. Stepwise electrolysis of the neutral complexes to first the mixed-valent radical cations
and then the dioxidized dications under IR monitoring reveal incremental shifts of the charge-sensitive
Ru(CO) bands and allow for an assignment of their radical cations as moderately or very weakly coupled
mixed-valent systems of class II according to Robin and Day. Ground-state delocalization in the mixed-
valent forms of these complexes as based on the CO band shifts is considerably larger for the “closed”
paracyclophane as for the “half-open” orthocyclophane. Experimental findings are backed by the
calculated IR band patterns and spin density distributions for radical cations of slightly simplified model
Received 4 May 2011
Received in revised form
23 June 2011
Accepted 24 June 2011
Keywords:
Vinyl complexes
Ruthenium
Cyclophanes
Electrochemistry
Spectroelectrochemistry
Mixed-valent
ꢀ
ꢀ
complexes 2Me D and 3Me D with the PiPr3 ligands replaced by PMe3. Radical cations 2ꢀD and 3ꢀD feature
a characteristic NIR band that is neither present in their neutral or fully oxidized forms nor in the radical
cation of the monoruthenium [2.2]paracyclophane complex 1 with just one vinyl ruthenium moiety.
These bands are thus assigned as intervalence charge-transfer (IVCT) transitions. Our results indicate
that, for the radical cations, electronic coupling “through-space” via the stacked styrene decks is
significantly more efficient than the “through-bond” pathway.
Ó 2011 Elsevier B.V. All rights reserved.
1. Introduction
DNA [29e32] or columnar mesophases made up of suitably func-
tionalized disk-shaped building blocks with extended aromatic
p
-Stacking has long been recognized as an important non-
cores [32e37]. Hupp et al. have beautifully demonstrated that
rectangular tetrametal box molecules with extended diimine p-
ligands as the long sides form, upon partial reduction, mixed-valent
radical anions that exhibit electronic coupling through-space
[38e40]. A clear dependency of the strength of the electronic
interaction on the stacking distance was observed.
covalent interaction governing the organization of matter [1e5].
Implications are numerous and include, inter alia, the photophysics
of luminophores [6e8] or of polymers having extended arene
substituents or backbones [9,10], the formation of self-assembled
coordination cages [3,5,11] and the exchange of magnetic [12e18]
or electronic information [8,16,19e28].
responsible for rapid charge migration over oligonucleotides and
p
-Stacking is also held
[n.n]Paracyclophanes have a particularly successful history as
testbeds for such interactions [41e44]. The groups of Neugebauer
[45,46] and, more recently, Grampp and Lambert [47] have
elegantly utilized the radical cations derived from electron-rich
methoxy or bis(triarylamine) substituted [n.n]paracyclophanes as
probes for electron delocalization on the EPR timescale and noted
that cyclophane bridges behave more like unsaturated and conju-
gated than as saturated bridges in terms of the electronic coupling
conveyed by them [42].
* Corresponding author.
** Corresponding author. Tel.: þ49 7531 885355; fax: þ49 7531 8883136.
ꢀ
E-mail addresses: stanislav.zalis@jh-inst.cas.cz (S. Zális), rainer.winter@uni-
konstanz.de (R.F. Winter).
1
Present address: Institut Charles Gerhardt e UMR5253, Equipe CMOS (Chimie
Moléculaire et Organisation du Solide), Université de Montpellier 2 e CC1701, Place
Eugène Bataillon, F-34095 Montpellier Cedex 05, France.
0022-328X/$ e see front matter Ó 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2011.06.028

P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
3187
We herein report [2.2]paracyclophanes where one or both arene
decks are elaborated into styryl ruthenium moieties RuCl(CH]
CHPh’)(CO)(PiPr3)2. The {RuCl(CO)(PiPr3)2} tags render the corre-
sponding styryl substituents electroactive at fairly low potential
and stabilize their associated radical cations to about the same
extent as dialkyl amino groups, thus making them amenable to
spectroscopic investigations. As an added benefit with respect to
classical organic electron donating groups they also provide the
characteristic Ru(CO) stretch as a reporter of the charge density at
the metal center. Its change in position upon oxidation is a conve-
nient spectroscopic probe of the loss of electron density from the
metal atom. Moreover, the CO band pattern and the relative CO band
shifts are indicative of the degree of intrinsic charge delocalization
on the short vibrational timescale of ca. 10ꢁ12 s in di- and oligonu-
clear complexes. Here we apply these Ru(CO) tags to measure
electron delocalization in a diruthenium pseudo-para-divinyl[2.2]
paracyclophane complex. The issue of the contribution of the
saturated alkylene straps to the overall electron delocalization is
addressed by comparison with the 3,7-divinyl-10,11-dihydro-5H-
dibenzo[a,d]cycloheptene-derived diruthenium complex 3 that
maintains the motif of two doubly alkyl linked styryl ruthenium
a platinum wire as counter electrode and a platinum working
electrode (Pt Ø ¼ 1.1 mm, BASi). Measurements were performed
under an argon atmosphere on 0.1 M solutions of NBu4PF6 in dried
CH2Cl2 and referenced to the Cp2Fe0/þ couple with ferrocene or
decamethylferrocene as an internal standard. The respective stan-
dard was added after all necessary data on the analyte had been
acquired. Representative scans were then repeated in the presence
of the standard. IR and spectroelectrochemical measurements were
recorded on a Nicolet iS10 instrument of ThermoFischer. UV/Vis/
NIR experiments were performed on an OMEGA 20 spectrometer of
Bruins Instruments or a TIDAS fiberorptic diode array spectrometer
(combined MCS UV/NIR and PGS NIR instrumentation) from j&m in
a home-built OTTLEecell following the basic design of F. Hartl et al.
[48]. EPR spectra were recorded at the EPSRC Multi-Frequency c.w.
EPR Service Centre at the University of Manchester on a Bruker
EMX spectrometer with different microwave bridges (X-band
(z9.4 GHz) and S-band (z3.8 GHz)). The temperature of the
sample was controlled by evaporation of liquid nitrogen and
a heater, using a digital temperature and gas flow control unit and
quartz cavity insert (X-band) or quartz immersion Dewar (S-band),
and was monitored with a thermocouple close to the sample
position. For EPR spectroscopic experiments, quartz tubes of 1 mm
inner diameter for X-band and of 3 mm inner diameter for S-band
measurements were used. In situ EPR spectroelectrochemical
experiments were carried out in dichloromethane with 0.1 M tet-
rabutylammonium hexafluorophosphate as the supporting elec-
trolyte using platinum working and auxiliary electrodes and a Ag
pseudo-reference electrode. The electrode wires were PTFE-coated
and all situated in the bulk solution inside the cavity of the spec-
trometer. The wires were all cut to different lengths and the PTFE
removed at the bottom of each wire to allow contact with the
solution, but so that no bare wires could touch each other [49]. The
potentiostat (Autolab, Type II) was controlled via a PC running
General Purpose Electrochemical System software, version 4.9 (Eco
Chemie BV, Utrecht). EPR simulations were performed using the
Bruker Win-EPR Simfonia software.
moieties but lacks the p-stacking of the individual styryl ruthenium
subunits as it is present in paracyclophanes (see Chart I).
2. Experimental section
2.1. General considerations
All reactions and manipulations were conducted using standard
Schlenk techniques. Solvents were dried over appropriate drying
agents, distilled under nitrogen and stored in airtight glass bulbs.
All NMR solvents were degassed by five “freezeepumpethaw”
cycles and stored in airtight Schlenk tubes over appropriate
molecular sieves. 1H, 13C and 31P NMR spectra were recorded on
a Bruker Avance 300 (300.13 MHz) or a Bruker Avance 400
(400.13 MHz) NMR spectrometer. Voltammetric measurements
were conducted on an EpsilonÔ potentiostat from BASi (Bio-
RuClH(CO)(PiPr3)2 was prepared according to Ref. [50] 4-Ethynyl
[2.2]paracyclophane
cyclophane were prepared according to Ref. [51]. The synthesis of
3,7-diethynyl-10,11-dihydro-5H-dibenzo[a,d]cycloheptene was
and
pseudo-para-diethynyl[2.2]para-
analytical Systems, Inc.) using
a cylindrical airtight one-
compartment cell with a spiral silver wire as reference electrode,
adapted from Lambert et al. [52] with some modifications of the
published procedure as given in the Supporting Information.
2.1.1. 4-(PiPr3)2(CO)ClRuꢁCH¼CHꢁ[2.2]paracyclophane, 1
73 mg (0.150 mmol) of RuClH(CO)(PiPr3)2 and 34 mg
(0.145 mmol) of 4-ethynyl[2.2]paracyclophane were dissolved in
dry CH2Cl2 under nitrogen. Soon after mixing, the solution color
intensified from orange-red to deep red. Stirring was continued for
1 h. After removal of the solvent, washing the residue with several
portions of hexane and drying under dynamic vacuum, slightly
impure 1 remained. This was purified from some unreacted
RuClH(CO)(PiPr3)2 by slowly dropping
a concentrated CH2Cl2
solution into vigorously stirred hexane and removing the super-
natant. After three repetitions and drying of the residue in oil pump
vacuum, 104 mg of the dichloromethane monosolvate 1 3 CH2Cl2
(89%) were obtained as a deep purple powder. 1H NMR (400.1 MHz,
3
3
CD2Cl2):
d
8.17 (dt, 1H, JHH ¼ 13.02 Hz, JPeH ¼ 1.01 Hz, H1), 6.55
3
4
3
(dd, 1H, JHH ¼ 7.96, JHH ¼ 1.84 Hz, H12), 6.53 (dd, 1H, JHH ¼ 7.96,
4JHH ¼ 1.84 Hz, H15), 6.48 (dd, 1H, JHH ¼ 7.76 Hz, JHH ¼ 1.76 Hz,
3
4
3
4
H16), 6.22 (dd, 1H, JHH ¼ 7.96 Hz, JHH ¼ 1.76 Hz, H13), 6.20 (d, 1H,
3JHH ¼ 7.56 Hz, H7), 6.18 (d, 1H, JHH ¼ 1.76 Hz, H4), 6.15 (dd, 1H,
3
3JHH ¼ 7.56 Hz, JHH ¼ 1.76 Hz, H6), 6.08 (dt, 1H, JHH ¼ 13.02 Hz,
4JPeH ¼ 2.14 Hz, H2), 5.32 (s, CH2Cl2), 3.37 (ddd, 1H, 3JHH ¼ 13.08 Hz,
3JHH ¼ 10.10 Hz, 1JHH ¼ 2.50 Hz, H9), 3.11e2.93 (m, 6H, H10, H17, H18),
2.86 (spt, 3H, 3JHH ¼ 7.90 Hz CH(CH3)2), 2.71 (m, 4H, CH(CH3)2, H9’),
4
3
Chart 1.

3188
P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
1.47e1.39 (m, 18H, CH(CH3)2), 1.28e1.15 (m, 18H, CH(CH3)2). 13C{1H}
2.1.3. {Ru(CO)Cl(PiPr3)2}2-3,7-divinyl-10,11-dihydro-5H-dibenzo
[a,d]cycloheptene, 3
2
NMR (150.9 MHz, CD2Cl2):
d
203.4 (t, JPeC ¼ 13.1 Hz, C19), 150.8 (t,
2JPeC ¼ 10.6 Hz, C1), 140.0 (s, C11), 139.9 (s, C14), 139.7 (s, C5), 139.5 (s,
C3), 134.9 (s, C7), 133.4 (s, C8), 133.3 (s, C16), 132.8 (s, C15), 132.5 (s,
C2), 132.4 (s, C13), 130.9 (s, C12), 130.7 (s, C4), 129.0 (s, C6), 35.8 (s,
C17), 35.7 (s, C18), 35.1 (s, C10), 33.5 (s, C9), 24.5 (vt, JPeC ¼ 15.0 Hz,
CH(CH3)2), 20.1, 20.0 (s, CH(CH3)2). 31P{1H} NMR (121.5 MHz,
138.8 mg (0.285 mmol) of RuClH(CO)(PiPr3)2 and 34.5 mg
(0.142 mmol) of 8 were dissolved in dry CH2Cl2 under nitrogen. The
solution color rapidly intensified from orange-red to purple-red.
After stirring at room temperature for 1 h the solvent was
removed under vacuum. The purple-red residue was washed with
several portions of hexane, then reprecipitated two times from
a CH2Cl2/hexane mixture and dried under vacuum. 171 mg (92%) of
the dichloromethane monosolvate were obtained. 1H NMR
3
3
CD2Cl2)
d
39.35 (d, JPP ¼ 411 Hz), 37.65 (d, JPP ¼ 411 Hz). Anal.
Calcd. for C38H61Cl3OP2Ru: C, 56.82; H, 7.65; Found: C, 57.21; H,
7.63.
(300.1 MHz, CD2Cl2):
d
8.43 (d, 2H, 3JHH ¼ 13.44 Hz, H1), 6.90 (d, 2H,
3JHH ¼ 7.68 Hz, H5), 6.82 (dd, 2H, 3JHH ¼ 7.68 Hz, 4JHH ¼ 1.71 Hz, H4),
4
3
6.79 (d, 2H, JHH ¼ 1.71 Hz, H8), 5.90 (dt, 2H, JHH ¼ 13.44 Hz,
4JPeH ¼ 2.05 Hz, H2), 3.91 (s, 2H, H10), 3.02 (s, 4H, H9), 2.80e2.65 (m,
12H, CH(CH3)2), 1.33e1.21 (m, 72H, CH(CH3)2). 13C{1H} NMR
2
(150.9 MHz, CD2Cl2):
d
203.4 (t, JP-C ¼ 13.2 Hz, C11), 149.0 (t,
2JPeC ¼ 11.1 Hz, C1), 139.5 (s, C7), 137.1 (s, C3), 135.2 (s, C6), 134.4 (s,
C2), 129.9 (s, C5), 125.0 (s, C8), 122.3 (s, C4), 41.5 (s, C10), 32.5 (s, C9),
24.8 (vt, JPeC ¼ 9.9 Hz, CH(CH3)2), 20.1, 19.9 (s, CH(CH3)2). 31P{1H}
NMR (121.5 MHz, CD2Cl2)
d
38.41 (s). Anal. Calcd. for
C58H102Cl4O2P4Ru2: C, 53.61; H, 7.91; Found: C, 53.14; H, 8.04.
2.1.2. Pseudo-para-{(PiPr3)2(CO)ClRuꢁCH]CH}2[2.2]
paracyclophane, 2
82 mg (0.17 mmol) of RuClH(CO)(PiPr3)2 and 21 mg (0.08 mmol)
of pseudo-para-diethynyl[2.2]paracyclophane were dissolved in dry
CH2Cl2 under nitrogen and stirred at room temperature for 1 h,
whereupon the orange-red color of the hydride complex changed
to deep purple. The solvent was removed and the residue was
washed with several portions of n-hexane and then dried under
vacuum to give 2 as a deep purple powder. Further purification was
achieved by repeated (two times) precipitation of 2 by slowly
adding a concentrated CH2Cl2 solution into vigorously stirred
hexanes. 99 mg (94%) of the dichloromethane monosolvate were
obtained after drying of the residue in vacuo. 1H NMR (300.1 MHz,
2.1.4. X-ray structure determination of complex 2
Geometry and intensity data were collected on an Oxford
Diffraction Gemini Ultra CCD diffractometer with multilayer optics
CD2Cl2):
d
8.12 (d, 2H, 3JHH ¼ 12.98 Hz, H1), 6.22 (s, 2H, H4), 6.21 (d,
2H, 3JHH ¼ 7.48 Hz, H6), 6.09 (dt, 2H, 3JHH ¼ 12.98 Hz, 4JHP ¼ 1.98 Hz,
H2), 5.98 (d, 2H, 3JHH ¼ 7.48 Hz, H7), 5.32 (s, CH2Cl2), 3.31 (ddd, 2H,
and with a Cu K
a
(
l
¼ 1.54184 Å, graphite monochromator) radia-
tion source. The data were collected at 123 K using an Oxford
Diffraction Cryojet Cooler. The structure was solved by direct
methods (SIR-97) and refined by full-matrix anisotropic least
squares (SHELXL97). The H-atoms were calculated geometrically
and a riding model was used during the refinement process.
3
2
3JHH ¼ 3.36 Hz, JHH ¼ 13.26 Hz, JHH ¼ 0.06 Hz, H9), 2.91e2.80 (m,
10H, H10, CH(CH3)2), 2.75e2.65 (m, 8H, H9’, CH(CH3)2), 1.44e1.38
(m, 36H, CH(CH3)2), 1.26e1.19 (m, 36H, CH(CH3)2). 13C{1H} NMR
2
(150.9 MHz, CD2Cl2):
d
203.5 (t, JPeC ¼ 13.13 Hz, C11), 149.9 (t,
1JPeC ¼ 10.6 Hz, C1), 139.9 (s, C5), 139.5 (s, C3), 133.9 (s, C7), 133.3 (s,
4
C8), 132.8 (t, JPeC ¼ 3.0 Hz, C2), 130.0 (s, C4), 127.1 (s, C6), 35.0 (s,
C10), 32.9 (s, C9), 24.6, 24.4 (vt, JPeC ¼ 9.8 Hz, CH(CH3)2), 20.1, 19.9 (s,
CH(CH3)2). 31P{1H} NMR (121.5 MHz, CD2Cl2): 39.29 (d,
3JPP ¼ 419 Hz), 37.58 (d, 3JPP ¼ 419 Hz). Anal. Calcd. for
C59H104Cl4O2P4Ru2: C, 53.96; H, 7.98; Found: C, 52.82; H, 8.22.
2.2. Quantum chemical calculations
The ground-state electronic structures were calculated by
density functional theory (DFT) methods using the Gaussian 09 [53]
and ADF [54,55] program packages. In order to reduce computa-
tional time to a reasonable limit, PiPr3 ligands were replaced by
PMe3. Models for experimental systems 1, 2 and 3 are denoted as
1Me, 2Me and 3Me. Quantum chemical studies were performed
without any symmetry constraints for all accessible oxidation
states. Open shell systems were calculated by the unrestricted
KohneSham approach (UKS). Geometry optimization followed by
vibrational analysis was made either in vacuum or in solvent media.
The quasirelativistic effective core pseudopotentials and the
corresponding optimized set of basis functions for Ru [56] and 6-
31G* polarized double-
z basis sets [57] for the remaining atoms
were employed together with the Perdew, Burke, Ernzerhof
exchange and correlation functional (PBE0) [58,59]. The Meta GGA
functional M05-2x containing larger admixture of HartreeeFock
(HF) exchange was used for comparison [60]. Solvent effects were

P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
3189
described by the polarizable conductor continuum model (CPCM)
[61].
The g-tensor (calculated by ADF) was obtained from a spin-
nonpolarized wave function after incorporating the spin-orbit
(SO) coupling by first-order perturbation theory from the ZORA
Hamiltonian in the presence of a time-independent magnetic field
[62,63]. Within ADF, Slater type orbital (STO) basis sets of triple-
x
quality with two polarization functions for Ru and one polarization
function for remaining atoms were employed. The inner shells
were represented by a frozen core approximation, viz. 1s for C, N, O,
1s-2p for P, Cl and 1s-3d for Ru were kept frozen. The calculations
were done with the functional including Becke’s gradient correc-
tion [64] to the local exchange expression in conjunction with
Perdew’s gradient correction [65] to the local correlation (ADF/BP).
The scalar relativistic (SR) zero order regular approximation (ZORA)
[66] was used within this study.
3. Results
3.1. [2.2]Paracyclophane complexes 1 and 2
Fig. 1. ORTEP showing the crystallographically determined structure of complex 2.
Ellipsoids are drawn at a 50% probability level.
The mono- and diruthenium [2.2]paracylophane complexes 1
and 2 of Chart 1 were prepared by hydroruthenation of 4-ethynyl
[2.2]paracyclophane or of pseudo-para-diethynyl[2.2]para-
cyclophane with RuClH(CO)(PiPr3)2 and accordingly characterized
by multinuclear NMR and IR spectroscopy. 1H and 13C NMR spectra
of the unsymmetrically substituted monoruthenium complex 1
feature separate resonances for every individual carbon and
hydrogen atom which could be assigned on the basis of one- and
two-dimensional NMR spectra and by application of HMBC and
HSQC pulse sequences. Of note is a slight high-field shift of the
bridgehead carbon atoms that link the benzene ring to the ethylene
straps for the more electron-rich, vinyl ruthenium substituted deck
when compared to the unsubstituted one. The vinyl ruthenium
moiety gives rise to 1H signals at 8.17 (RueCH) and 6.08 ppm
(RueCH]CH) which are both split into triplets of doublets by 3JHeH
(J ¼ 13.0 Hz) and JHeP or JHeP couplings of 1.0 or 2.1 Hz, respec-
tively, and 13C NMR signals at 150.8 (2JPeC ¼ 10.6 Hz, RueCH) and
132.5 ppm (RueCH]CH) for the vinyl group as well as the appro-
priate resonance signals of the PiPr3 ligands. The low symmetry of
the complexes (C1 for complex 1, C2h for complex 2) renders the two
PiPr3 ligands at each metal atom chemically and magnetically
inequivalent such that two closely spaced doublet signals with
coupling constants of 411 Hz (1) or 418 Hz (2) and a pronounced
between the CO and Cl ligands, which, however, could not be
resolved. Structural parameters of the central aromatic core
strongly resemble those of other [2.2]paracyclophane derivatives
and, particularly, those of 2,2-dibromovinyl-4-[2.2]paracyclophane
[51] and 4,12-bis(2,2-dibromovinyl)[2.2]paracyclophane [67]. Ring
strain imposed by the ethylene straps induces an opening of the
corresponding CeCeC angles at the aliphatic carbon atoms to
112.9(4) and 113.1(4)ꢂ and a boat like distortion of each phenyl
deck. This results in a smaller separation between the bridgehead
carbon atoms C(5)/C(90) of 2.773 Å as opposed to the distance
between the mean planes through the remaining carbon atoms of
3.065 Å. Interplanar angles between the planes through carbon
atoms C(3), C(5), C(6) or atoms (C4), C(9), C(10) forming the bow
and rear of the boat and the mean plane through the non-bridged
carbon atoms C(3), C(4), C(6) and C(10) amount to 12.1ꢂ. Compa-
rable values of 10.9e14.5ꢂ are observed for 2,2-dibromovinyl-4-
[2.2]paracyclophane [51] while the respective values for 4,12-
bis(2,2-dibromovinyl)[2.2]paracyclophane are 7.0 to 8.0ꢂ [67],
with a separation of 3.075 Å between the decks. In keeping with
similar five-coordinated vinyl ruthenium complexes [68e75] the
ruthenium atoms adopt a square pyramidal coordination geometry
3
4
roof effect are observed. The
observed as a strong resonance at 1910 cmꢁ1 while weaker bands
at 1553 and 1544 cmꢁ1 are probably of
(C]C) parentage. Simpler
n(CO) band of the Ru(CO) subunit is
n
Table 1
spectra are observed for dinuclear 2 in keeping with the presence of
an inversion center with almost identical NMR shifts for the 1H and
13C NMR vinyl resonances as were observed in 1. Thus, the vinyl
protons resonate at 8.12 and 6.09 ppm (3JHeH ¼ 13.0 Hz) while the
appropriate carbon resonances are observed at 149.9
(2JPeC ¼ 10.6 Hz) and at 132.8 ppm (3JPeC ¼ 3.0 Hz). Just like for 1,
the IR spectra feature diagnostic bands at 1909, 1555 and
1544 cmꢁ1. Again, two closely spaced doublets with JPeP ¼ 419 Hz
and a pronounced roof effect were observed in the 31P NMR
spectrum.
The DFT/PBE0 calculated symmetry averaged bond lengths (Å) and angles (ꢂ) for
model 2Me complex and the comparison with experimental structural parameters.
2Me [Å]
2 [Å]
RueC11
RueCl 1
RueC1
RueP1
RueP2
RueCl
C1eC2
C2eC3
C3eC4
C3eC5
C4eC9ʹ
C9e C8
C8eC7
C5eC9
Stacking distance
RueC1eC2
C1eC2eC3
1.815
2.423
1.990
2.354
2.354
2.423
1.348
1.472
1.405
1.416
1.394
1.508
1.593
2.782
3.072
134.0
126.9
1.828(5)
2.4432(11)
2.006(4)
2.4080(12)
2.4063(12)
2.4432(11)
1.322(6)
1.482(6)
1.413(6)
1.415(6)
1.380(7)
1.510(7)
1.592(7)
2.773
Deep red rod-shaped crystals of the CH2Cl2 bis(solvate) of
pseudo-para-divinyl[2.2]paracyclophane-bridged
diruthenium
complex 2 were obtained from concentrated solutions in CH2Cl2
and subjected to an X-ray diffraction study. The results are illus-
trated in Fig. 1 and Table 1. Details to the data collection and
structure refinement and a more complete listing of bond lengths
and angles can be found in Tables S4 and S5 of the Supporting
Information. The structure data suggest some degree of disorder
3.065
134.5(3)
125.5(4)

3190
P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
Table 2
with the vinyl substituent in the apical position and trans angles of
175.0(1) (C(11)eRueC(1) and 170.31(4)ꢂ (P(1)eRueP(2)) as well as
cis-angles in the range of 88.04(4) (Cl(1)eRueP(2) to 99.05(13)ꢂ
(P(2)eRueC(1)). This results from a 0.137 Å displacement of the
metal atom from the basal plane towards the apical vinyl ligand.
The vinyl group itself has a C]C bond length C(1)eC(2) of 1.322(6)
Å and suffers some opening of the Ru(1)eC(1)eC(2) angle to
134.5(3)ꢂ when compared to the C(1)eC(2)eC(3) angle of 125.5(4)ꢂ.
Torsional angles Ru(1)eC(1)eC(2)eC(3) of ꢁ170.0(4)ꢂ and C(1)e
C(2)eC(3)eC(4) of 12.6(7)ꢂ signal efficient conjugation throughout
the entire RueCH]CHePh moiety as it is routinely found in trans
configured vinyl ruthenium complexes bearing an aryl substituent.
This is in stark contrast to the aforementioned 4-(2,2-
dibromovinyl)[2.2]paracyclophane and 4,12-bis(2,2-dibromovinyl)
[2.2]paracyclophane where steric hindrance between the cis-
disposed bromine atom and the adjacent proton of the arene deck
Electrochemical data for complexes 1e3 (CH2Cl2/0.1 M NBu4PF6, r.t.).
0/þ
þ/2þ
Complex
E
1/2
[V] (ꢃ3 mV)
E
[V] (ꢃ3 mV)
D
E1/2 [mV]
Kcomp
1/2
1
2
3
0.165
0.125
0.184
0.92a
0.335
0.289
n. a.
n. a.
4100
64
0.210(ꢃ3)
0.105(ꢃ3)
a
Peak potential of an irreversible peak at v ¼ 0.1 V/s.
higher oxidation processes are shifted outside the accessible
potential window (Fig. 2). The appearance of two separate waves in
the 100e400 mV range instead of one suggests stepwise oxidation
of the redox-active electron-rich styryl ruthenium moieties. The
potential splitting and lowering of the first oxidation potential in 2
with respect to 1 provide first pieces of evidence for electronic
interaction between them but no quantitative measure for its
strength. Differences of redox potentials for stepwise electron
transfer in systems with two identical, interlinked redox sites are
subject to several contributions other than the “electronic
coupling”, in particular solvation energy changes and ion pairing
[76e80]. We nevertheless note that the half-wave potential sepa-
prevent a coplanar arrangement with torsional angles Carene
C
]
areneeCvinyl]Cvinyl of ca. 50ꢂ. Geometry optimization by density
functional theory (DFT, Gaussian 09 and ADF) on the simplified
model complex 2Me with the PiPr3 ligands replaced by PMe3 leads
to a structure close to the experimental one as is shown by the
comparison in Table 1. Bond lengths are reproduced within 0.02 Å,
and the calculated C5/C90 distance of 2.782 Å and the mean
distance of 3.072 Å between the C3eC4eC6eC10 and
C3’eC4’eC6’eC10’ planes are in good agreement with experi-
mental separations of 2.773 Å and 3.065 Å. A graphical represen-
tation of the calculated structure is provided in Figure S6 of the
Supporting Information.
ration in
2 is considerably larger as that found in Akita’s
Cp*Fe(dppe) capped diiron pseudo-meta-diethynyl[2.2]para-
cyclophane complex (Cp* ¼
h
5-C5Me5, dppe ¼ 1,2-Ph2PC2H4PPh2,
D
E1/2 ¼100 mV) [28] and in Connick’s pseudo-para-bis(picoli-
naldimino)-[2.2]paracyclophane-bridged dirhenium complexes
(D
E1/2 ¼ 80 mV) [20].
In search of more direct measures of the electronic coupling we
Individual molecules of 2 arrange into stepped stacks that run
along the diagonal of the ab-plane. The bulky PiPr3 ligands induce
a lateral shift of individual molecules residing in parallel planes and
generated radical cation 2ꢀD by electrolysis inside a transparent
thin-layer cell and compared its spectroscopic properties to those
of its mononuclear counterpart 1ꢀD. The latter serves as a bench-
mark system of a complex closely related to 2 without electronic
interactions with another styryl ruthenium subunit. Oxidation of
complex 1 under IR monitoring resulted in a 57 cmꢁ1 blue shift of
the Ru(CO) stretch from 1910 to 1967 cmꢁ1 and the appearance of
several new C]C bands in the 1640e1520 cmꢁ1 region which are
characteristic of oxidized styryl ruthenium complexes (Fig. 3) [81].
suppress any additional
p-stacking interactions between them.
Hydrogen bonding interactions of 2.726 and 2.818 Å are observed
between the ruthenium bonded Cl(1) atom and one hydrogen atom
of each cocrystallized CH2Cl2 solvate molecule which occupy voids
in between the stacks (see Figure S1 of the Supporting
Information).
Voltammetric studies on complex 1 reveal the presence of one
chemically and electrochemically reversible oxidation at 0.165 V
against the ferrocene/ferrocenium standard which is followed by
a chemically irreversible second oxidation process at 0.92 V (Fig. 2,
Table 2). These results are in full agreement with our observations
on the parent styryl complex RuCl(CH]CHPh)(CO)(PiPr3)2 [74]. In
dinuclear 2, however, two consecutive, reversible one-electron
waves are observed at 0.125 and 0.335 V while any additional
The rather modest magnitude of the n(CO) shift for the metal-
bonded carbonyl ligand is a consequence of the strong ligand
contribution to the oxidation of styryl-type complexes like 1 and
closely resembles our observations for the [RuCl(CH]
ꢀ
CHPh)(CO)(PiPr3)2]0/ þ pair [74]. Such behavior contrasts to metal
based redox processes where
n(CO) shifts in the range of
120e150 cmꢁ1 are expected [82e84]. Performing the same exper-
iment under UV/Vis/NIR monitoring induces the growth of new
bands with deconvoluted peak positions at 778, 706, 469 and
437 nm that are typical of the oxidized styryl ruthenium
Fig. 2. Cyclic voltammetry measurements on complexes 1 (top) and 2 (bottom) in
Fig. 3. Changes in the IR spectra upon the first oxidation of complex 1 (1,2-C2H4Cl2/
CH2Cl2/NBu4PF6 (0.1 M) at r.t. and at sweep rates of 50, 100, 200, and 500 mV/s.
NBu4PF6 (0.2 M) at r.t.).

P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
3191
spectroscopic regime between 3000 cmꢁ1 and the high-energy
limit of our detector (see insert of Fig. 5). Red-shifts of the (C]C)
absorptions of the vinyl and the arene groups are also observed.
Further oxidation to dicationic 22D restores the one-band
(CO)
n
n
pattern with the corresponding band peaking at 1969 cmꢁ1 and
causes complete bleaching of the broad feature of 2ꢀD at 5700 cmꢁ1
(see insert of Fig. 5). This latter conspicuous feature is also observed
to grow under UV/Vis/NIR monitoring of the first oxidation with
a better defined maximum at 1828 nm (5470 cmꢁ1). Other bands of
2ꢀþ resemble those of its mononuclear counterpart 1ꢀþ with a broad
structured feature near 700 nm whose low-energy shoulder
deconvolutes into a separate peak at 897 nm and a structured band
with peaks at 464, 426 and 397 nm. Full oxidation to 22D again
causes the disappearance of the characteristic 1828 nm band while
the other Vis/NIR absorptions of 2ꢀþ intensify and red-shift. Thus,
a new composite band with deconvoluted peaks at 1120, 903 and
730 nm is observed. The band near 450 nm also intensifies with
some loss of fine structure (Fig. 6). Nevertheless, separate maxima
at 457, 430 and 388 nm are identified upon deconvolution. Table 3
collects spectroscopic data for 1 and 2 in every accessible oxidation
state.
Fig. 4. Changes in the UV/Vis/NIR spectra upon the first oxidation of complex 1 (1,2-
C2H4Cl2/NBu4PF6 (0.2 M) at r.t.).
chromophore (Fig. 4) [81,85,86]. With reference to earlier work on
pyrimidinedione appended derivatives [73] we assign these
absorptions as a
p / p
* transition of the {Ru(CH]CHPh)}ꢀþ
subunit and metal / styrylꢀþ charge-transfer excitations, respec-
tively. All bands are red-shifted when compared to [RuCl(CH]
CHPh)(CO)(PiPr3)2]ꢀþ [74]. The maintenance of isosbestic points
throughout the measurements and the recovery of the spectra of
the parent neutral after a full oxidation/reduction cycle indicate
that 1ꢀD is stable on the longer electrolysis timescale.
Shifts of CO stretching frequencies were interpreted by quantum
chemistry. In vacuo DFT calculations with the PBE0 functional
strongly underestimate the splitting of CO frequencies in the radical
cation of diruthenium complex 2Me. This is in keeping with the
well-known fact that calculations with standard DFT functionals
frequently overestimate electron density delocalization. Recently it
was shown that larger admixture of HF exchange and/or the
inclusion of solvation effects leads to qualitatively correct descrip-
tions of electron delocalization in mixed-valent systems [86e89].
After inclusion of CPCM solvent correction the calculated sequence
of CO frequencies listed in Table 4 reproduces the experimental
As is expected on the basis of its voltammograms, oxidation of
complex 2 occurs as two separate steps with absorptions of the
intermediate mixed-valent radical cation 2ꢀD clearly distinct from
those of neutral 2 and fully oxidized 22D. In IR spectroelec-
trochemistry the single CO band of 2 at 1909 cmꢁ1 develops into
a two band pattern with separate absorptions at 1912 and
1962 cmꢁ1 (Fig. 5). This means that, in the radical cation state, the
two vinylRu(CO) subunits differ with respect to their intrinsic
electron densities. Of particular note is the growth of a broad
featureless band peaking at ca. 5700 cmꢁ1 that envelopes the entire
ꢀ
data reasonably well. In the case of radical cation 2Me D with PiPr3
Fig. 5. Changes in the IR spectra upon the first (top) and second (bottom) oxidation of
complex 2 (1,2-C2H4Cl2/NBu4PF6 (0.2 M) at r.t.). Insert: Range from 3000 to 8000 cmꢁ1
showing the growth and the collapse of the IVCT transition of the intermediate radical
Fig. 6. Changes in the UV/Vis/NIR spectra upon the first (top) and second (bottom)
oxidation of complex 2 (1,2-C2H4Cl2/NBu4PF6 (0.2 M) at r.t.).
cation 2ꢀD
.

3192
P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
Table 3
that the continuum model underlying the solvent correction is not
wholly adequate considering that the X-ray structures of all vinyl
ruthenium complexes that contain CH2Cl2 solvent molecules show
specific hydrogen bonding between the ruthenium bonded chlo-
ride ligand and the CH2Cl2 solvent molecules.
EPR spectra were recorded on electrochemically oxidized
samples of 1ꢀþ, 2ꢀþ and fully oxidized 22D at various temperatures
and concentrations in both the X- and the S-band (see Table 5 and
Figures S2 and S3 of the Supporting Information). At T ¼ 303 K,
monooxidized 1ꢀþ generated at a working potential of 0.6 V versus
an Ag pseudo-reference electrode shows a three-line resonance
signal at giso ¼ 2.035 with resolved 31P hyperfine splitting of 22.5 G.
Upon cooling, the spectra show anisotropy due to slow tumbling in
solution which results in general signal broadening. No further
couplings could be resolved even at lower modulation amplitudes.
At 120 K in a rigidly frozen glass 1ꢀþ exhibits an axial g-matrix with
IR and UV/Vis/NIR-spectroscopic data for 1, 1ꢀD, 2, 2ꢀD, 22D, 3, 3ꢀD and 32D in 1,2-
C2H4Cl2/NBu4PF6 (0.2 M).
ꢁ1
l
[nm] (ε [l molꢁ1 cmꢁ1])b
~
Complex
IR
n
[cm
]
1
1910(s), 1553(m), 1544(m),
1483(s), 1472(s), 1463(s),
1382(s), 1267(s), 1165(w),
1150(w), 1107(w), 1091(w),
1064(w), 1056(w), 1030(m)
1967(s), 1641(w), 1628(w),
1568(w), 1520(w), 1483(s),
1472(s), 1463(s), 1382(m),
1267(w), 1154(w), 1107(w),
1091(w), 1061(w), 1030(m)
1909, 1555(m), 1544(m),
1483(s), 1471(s), 1461(s),
1383(m), 1169(w), 1148(w),
1105(w), 1087(w), 1061(w)
5680,a 1912, 1962, 1571 (w),
1546(m), 1537(m), 1530(m),
1482(s), 1471(s), 1403(m),
1384(m), 1181(w), 1169(w),
1148(w), 1105(w), 1087(w),
1061(w), 1030(w)
515 (140), 353 (sh, 2100),
362 (2840), 263 (sh, 2700),
238 (sh, 5400), 218 (8600)
1ꢀD
735 (1070), 540 (sh, 640),
466 (3430), 434 (sh, 2650),
381 (1500), 274 (2170),
213 (8600)
2
519 (660), 328 (21,200),
243 (sh, 29,000)
2ꢀD
1828 (1100), 897 (sh, 2000),
700 (4000), 461 (12,700),
429 (13,000), 397 (12,000),
308 (17,000)
individual g-values of 2.064 (gt) and 2.035 (gıı) giving
<gav> ¼ 2.045. The proximity of giso and <gav> to the g-value of the
22D
1969, 1525(w), 1384(m),
1484(m), 1467(m), 1156(m),
1090 (w), 1061(m)
1120 (sh, 4700), 903 (9800),
730 (sh, 8200), 610 (4300),
450 (23,500), 388 (sh, 2560),
271 (18,000)
501 (500), 390 (3000),
310 (22,000), 228 (sh, 25,000)
3
1911,b 1600(w), 1576(m),
1549(m), 1383(s), 1170(m),
1152(m), 1107(w), 1061(w),
1029(w)
3ꢀD
1911,b 1972,b 1597(m),
1548(m), 1537(w), 1383(s),
1152(m), 1107(m), 1090(w),
1061(m), 1034(w),
917 (625), 674 (3500),
418 (sh, 10,000), 407 (10,500),
306 (15,000)
32D
1973,b 1582(m), 1169(sh),
1155(m), 1108(m), 1090(w),
1062(m), 1036(w), 1028(m)
676 (1760), 421 (sh, 6000),
405 (6400), 300 (16,600)
a
Electronic IVCT band of the respective radical cation.
Data from spectral deconvolution.
b
ligands modeled as PMe3, geometry optimization leads to a broken
symmetry solution (for details see Table S2 of the Supporting
Information) and to the concentration of spin density on one
styryl ruthenium subunit as it is depicted in Fig. 7 and listed in
Table S3 of the Supporting Information. The CO stretching
frequencꢁie1s calculated at 1911 cmꢁ1 for 2Me are shifted to 1915 and
ꢀ
1950 cm
for monooxidized 2Me D and finally to 1963 and
1964 cmꢁ1 for fully oxidized 2Me 2D irrespective of the overall spin
state (singlet diradical or triplet). It should be mentioned, though,
Table 4
The comparison of G09/PBE0/CPCM (1,2-C2H4Cl2) calculated CO stretching frequen-
cies for 1Me nD, 2Me nD and 3Me nD with experimental ones.
Calculateda
Experimental
n1 (CO)
n2 (CO)
n1 (CO)
n2 (CO)
1Me
1Me
2Me
2Me
1910
1957
1911
1915
1962b
1963c
1911
1915
1967b
1967c
e
1910
1967
1909
1912
1969
e
ꢀ
ꢀ
D
D
e
e
1911
1950
1963b
1964c
1912
1964
1968b
1968c
1909
1963
1969
2Me 2D
3Me
1911
1912
1972
1911
1972
1972
ꢀ
3Me D
3Me 2D
ꢀ
ꢀ
Fig. 7. DFT (PBE0/CPCM) calculated spin densities for 1Me D (top), 2Me D (middle), and
ꢀ
a
D
Calculated frequencies are scaled by a factor of 0.952.
Calculated for UKS singlet diradical state.
Calculated for UKS triplet state.
3Me
(bottom). Blue and green colors indicate positive and negative spin density,
b
c
respectively. (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)

P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
3193
Table 5
EPR data for radical cations 1ꢀD, 2ꢀD, 3ꢀD and dioxidized 22D and 32D
.
Complex
Experimental
Calculated
(T ¼ 273 K)
g׀׀, gt <gav> at 110 K; (
D
g)
giso
g1, g2, g3 (Dg)
1ꢀD
2ꢀD
22D
3ꢀD
32D
2.035; A(31P) ¼ 22.5 G
2.034
2.064, 2.035 <2.045>; (0.029)
2.062, 2.035 <2.044>; (0.027)
2.061, 2.035 <2.044>; (0.026)
2.071, 2.036 <2.048>; (0.035)
2.063, 2.037 <2.046>; (0.026)
2.026
2.029
2.037, 2.023, 2.019 (0.018)
2.041, 2.026, 2.019 (0.022)
2.034; A(31P) ¼ 22.7 G
2.040
2.037
2.052, 2.035, 2.024 (0.028)
2.038; A(31P) ¼ 24.0 G
free electron and the small g-matrix anisotropy
D
g of only 0.029 are
and 3.02 (C2H4) ppm as well as by one singlet signal at 38.4 ppm in
31P NMR spectroscopy. IR features include the characteristic Ru(CO)
stretch at 1911 cmꢁ1 and C]C bands at 1576 and 1549 cmꢁ1. Vol-
tammetric measurements show two chemically reversible one-
electron processes that are merged into a composite wave in
cyclic voltammetry but are resolved into individual peaks in
differential pulse and square wave voltammetric experiments
(Fig. 8, Table 2). Digital simulation of representative cyclic vol-
tammograms [93] or deconvolution of the square wave voltam-
both tokens of the dominant contribution of the arylvinyl ligand to
the SOMO of the radical cation as has been observed on previous
occasions [74,81,85,90e92]. Monooxidized 2ꢀþ gives an isotropic
signal at g ¼ 2.034. A broadening of the signal and some inflections
suggest underlying hyperfine splittings which, however, were not
resolved at any temperature in the 303e203 K range. Again we
observe anisotropic signal broadening as the temperature is low-
ered. At 120 K, the signal is axial with g-values of 2.062 (gt) and
2.035 (gıı) (<gav> ¼ 2.044). When the electrolysis was performed
for a longer time at an applied potential of 1.0 V (vs. Ag/Agþ), i.e. at
a potential sufficiently anodic of the 2ꢀD/2D wave, a strong isotropic
signal with well-resolved 31P hyperfine splitting of 22.7 G was
observed at giso ¼ 2.034 that we assign to dioxidized 22D. Cooling of
the solution caused signal broadening with the partial loss of
resolution until, in the frozen glass, an axial signal with gt ¼ 2.061,
gıı ¼ 2.035 and <gav> ¼ 2.044 was obtained. No half-field signal
was detected in either the fluid or the frozen solutions. Solutions of
2ꢀD generated by extensive electrolysis of 2 at an applied potential
of 0.6 V (against Ag/Agþ) did not show any better resolution in the
S-band than those recorded in the X-band (see Figure S5 of the
Supporting Information). Individual g-values and anisotropies for
mograms yield a half-wave potential difference
D
E1/2 of 105(ꢃ3)
mV and, according to Eq. (1), a comproportionation constant Kcomp
of 64(ꢃ8). This is to be compared with a value of 4.1(ꢃ9) ꢄ 103 for 2
(DE1/2 ¼ 210(ꢃ5)) mV where
p-stacking is operative.
nꢀ
ꢁ.
o
Kcomp ¼ exp n$F$
D
E1=2 ðR$TÞ
(1)
Spectroelectrochemical measurements were undertaken in
order to determine the electronic coupling in monooxidized 3ꢀD. IR
spectroelectrochemistry again revealed the presence of two
consecutive one-electron steps with the appearance of two Ru(CO)
bands during the first and their merging into a single one upon the
second oxidation (Fig. 9). This information is, however, hard to
glean from the experimental mid IR spectra pattern alone since
there is apparently no shift from the band of 3 to the low-energy
feature of 3ꢀD and from the high-energy band of 3ꢀD to the single
one in dioxidized 32D. There is, however, a broad high-energy
feature whose maximum lies just outside our detector range that
first grows in and then collapses as the oxidation from 3 to 3ꢀD and
ꢀ
D
2Me
were also calculated by DFT (see Table 5) and are in good
agreement with experimentally measured ones. UKS calculations
on doubly oxidized 2Me 2D were performed on diradical singlet and
triplet states. Both states are energetically close, the singlet being
slightly more stable by about 0.01 eV. In both states the spin density
is almost equally spread over the both vinyl ruthenium subunits
(Figure S7).
3.2. Half-open, ortho-connected 3: an analog of 2 without p-
stacking
There is a general consensus that through-space (i.e. p-stacking)
and through-bond pathways are both relevant for electron delo-
calization in odd-electron [n.n]paracyclophanes [22,42]. Literature
data are, however, somewhat inconclusive with respect to the
relative contributions of the two complementary pathways to
overall charge and spin delocalization. In an attempt to shed more
light on this issue we prepared the dinuclear complex 3 as a half-
open “[2.1]orthocyclophane” analog of 2 where the two styryl
decks are non-parallel and kept apart (Chart 1; for the DFT-
optimized structure see Figure S6 of the Supporting Information).
In order to render both available through-bond pathways roughly
equivalent and comparable to those in 2, the longer ethylene linker
is placed in direct conjugation with the styryl ruthenium subunits
(i.e. the para position) while the shorter methylene strap is in the
less favorable meta one.
Complex 3 was prepared from 3,7-diethynyl-10,11-dihydro-5H-
dibenzo[a,d]cycloheptene and the hydride complex RuClH(CO)(-
PiPr3)2. It is authenticated by the characteristic 1H and 13C NMR
resonances of the vinyl ruthenium subunit at 8.43 (RueCH]CH),
5.90 (RuCH]CH), 149.0 (RueCH) and 134.4 (RueCH]CH) ppm and
the expected signals of the bridging methylene groups at 3.91 (CH2)
Fig. 8. Electrochemical investigations on complex 3 in CH2Cl2/NBu4PF6 (0.1 M) at r. t.
Bottom: Cyclic voltammetry at sweep rates of 50, 100, 200, and 500 mV/s; top: Square
Wave Voltammetry; the peak on the right hand side corresponds to the internal
Cp*2Fe0/þ couple.

3194
P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
same observations as in the case of the corresponding para-
cyclophanes 2ꢀD and 22D. Thus, 3ꢀD gave a broadened isotropic
signal at g ¼ 2.040 without resolved 31P hyperfine splittings which
further broadened upon cooling and finally turned into an axial
pattern at T ¼ 120 K in the frozen matrix. As for 2ꢀD, the spectral
resolution and band widths of electrogenerated samples of 3ꢀD did
not change when they were recorded in the S-band. Prolonged
electrolysis at 1.0 V, a potential sufficiently positive of the 3ꢀD/2D
couple, produced a binomial 1:2:1 three-line signal at r.t. at
g ¼ 2.038 with A(31P) ¼ 24.0 G. Freezing the solution to 120 K
changed the spectrum to an axial pattern (gt ¼ 2.063, gıı ¼ 2.037,
<gav> ¼ 2.046), again with no half-field signal detected (Table 5
and Figure S4 of the Supporting Information). The somewhat
larger g-values and A(31P) coupling constants of 3nD when
compared to those of 2nD argue for an increased metal contribution
of the SOMO and are well reproduced by our quantum chemical
calculations (see Table
Information).
5 and Table S3 of the Supporting
4. Discussion
The main issue of this study was to experimentally probe for the
strength of the electronic coupling in the mixed-valent radical
cations 2ꢀD and 3ꢀD and for the extent to which the
p-stacking (or
through-space) pathway and the through-bond pathway via the
saturated alkylene straps contribute to it. While information as to
the extent of electron delocalization in mixed-valent systems can in
principle be obtained from several spectroscopic or even compu-
tational methods, the oxidation-induced IR band shifts provide the
most straightforward access to electron delocalization in the radical
cations 2ꢀD and 3ꢀD. Geiger et al. have defined a charge distribution
parameter Dr according to Eq. (2) which is based on the relative
shifts of the charge-sensitive CO stretches of metal-bonded
carbonyl ligands with respect to the bordering isovalent ones
[22,94]. In Eq. (2), nox and nred denote the band positions of the IR
Fig. 9. Changes in the IR spectra upon the first (top) and second (bottom) oxidation of
complex 3 (1,2-C2H4Cl2/NBu4PF6 (0.2 M) at r.t.). Insert: Range from 3000 to 8000 cmꢁ1
showing the IVCT transition of the intermediate radical cation 3ꢀD
.
then to 32D proceeds (see insert of Fig. 9). The spectrum with the
highest absorbance of this band thus defines the point at which the
concentration of monooxidized 3ꢀD reaches its maximum. This
low-energy band extending into the IR region is also seen in UV/
Vis/NIR spectroelectrochemical experiments (see insert of Fig. 10).
It has no equivalent in oxidized 1ꢀD but resembles that observed in
2ꢀD. It is thus therefore assigned as the IVCT transition of 3ꢀD. Band
label in the fully oxidized and the fully reduced forms while n ox
’
and
parameters as determined by spectral deconvolution are
n red are the band positions of the (formally) oxidized and (formally)
’
nmax ¼ 10,900 cmꢁ1
(917 nm),
εmax ¼ 625 l molꢁ1 cmꢁ1
and
~
reduced subunit of a mixed-valent system. According to that defi-
nition, Dr may assume values ranging from 0 to 0.5 where a value of
0 denotes the class I limit of a mixed-valent system with fully
localized valencies on two non-interacting redox sites while
Dr ¼ 0.5 heralds the class III limit of full charge delocalization
[22,94].
Dn1=2 ¼ 4250 cmꢁ1. The typical bands of the oxidized styryl ruthe-
nium chromophore are also observed as structured bands near
675 nm and 410 nm (Fig. 10); deconvoluted peak positions and
extinction coefficients are given in Table 3.
~
EPR studies on electrogenerated 3ꢀD (electrolysis at 0.6 V) and
fully oxidized 32D (prolonged electrolysis at 1.0 V) resulted in the
hꢀ
ꢁ
ꢀ
ꢁi.
Dr
¼
nox
ꢁ
n’ox
þ
n’red
ꢁ
nred
2ðnox
ꢁ
nred
Þ
(2)
ꢀ
Experimental data on the 20/ D/2D redox series (Table 3) yield
a
Dr value of 0.08 for 2ꢀD. According to that result 2ꢀþ is a moder-
ately coupled class II mixed-valent system. In keeping with that
assignment 2ꢀþ displays a characteristic low-energy electronic
absorption band at
l
¼ 1828 nm
(
nmax ¼ 5470 cmꢁ1
)
with
~
εmax ¼ 1100 l molꢁ1 cmꢁ1 and Dn1=2 ¼ 4870 cmꢁ1. This band has no
equivalent in oxidized 1ꢀþ (whose UV/Vis/NIR spectrum otherwise
closely resembles that of 2ꢀþ) with only one styryl ruthenium
moiety, or in reduced 2 and fully oxidized 22D, where the two styryl
ruthenium subunits are isovalent.
~
The combined effects of a very small magnitude of the oxida-
tion-induced Ru(CO) band shifts and of the only moderate com-
proportionation constant Kc pose a particular problem when
determining the CO band positions (and hence Dr) of 3ꢀD. A Kc value
of 64 dictates that, at the point of maximum concentration of the
intermediate radical cation, 80% of the total amount of compound 3
are in the 3ꢀD oxidation state whereas 10% of reduced 3 and fully
oxidized 32D each are present. Based on this consideration, the
hypothetical spectrum of 3ꢀD can be calculated from the
Fig. 10. Changes in the UV/Vis/NIR spectra upon the first oxidation of complex 3 (1,2-
C2H4Cl2/NBu4PF6 (0.2 M) at r. t.). Insert: Deconvolution of the bands at low-energy.
Experimental spectrum (black line), deconvoluted individual bands (green lines), and
superposition of individual bands (red line). (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)

P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
3195
experimental curves. Estimated
n
(CO) values are 1911 and
and 3ꢀD unfortunately does not allow us to determine the rates of
intramolecular electron transfer for these systems. Insufficient
signal resolution for detailed analysis might originate from a larger
number of unresolved hyperfine interactions with smaller coupling
constants when compared to 1ꢀD with just one oxidized styryl
ruthenium entity. A doubling of the number of hyperfine splittings
with simultaneous reduction of coupling constants by one half is
expected of a mixed-valent system with the spin delocalized over
two redox-active subunits and has been documented on several
occasions. Instructive examples are Kochi’s triphenylenes where
two redox-active 4-methyl-2,5-dimethoxyphenyl entities are
bridged by a central para-phenylene or -diphenylene bridge [95],
Gladysz’s butadiynediyl-bridged dirhenium complex [{Cp*Re(-
1972 cmꢁ1 as compared to 1911 cmꢁ1 for 3 and 1973 cmꢁ1 for 32D
.
From that result it is clear that the charge distribution parameter
for 3ꢀD is, at best, rather small and no larger than 0.01. Nevertheless,
3ꢀD also features an electronic band at fairly low-energy
nmax ¼ 10,900 cmꢁ1
,
lmax ¼ 917 nm) besides additional absorptions
~
(
that are characteristic of the oxidized styryl ruthenium chromo-
phore. This band has no equivalent in 1ꢀþ, 3 and 32D and is thus also
likely of IVCT origin. Taken together, 3ꢀD may be classified as
a weakly to very weakly coupled mixed-valent system of class II.
These conclusions based on experimental evidence are sup-
ported by our quantum chemical calculations on the PMe3 ligated
ꢀ
ꢀ
model complexes 2Me D and 3Me D. The calculated energy differ-
ence between the two CO stretches is considerably smaller for
NO)(PPh3)}2(
NO)(PPh3)(
m
-C4)]ꢀþ and its Pd/Re2 counterpart trans-[{Cp*Re(-
ꢀ
ꢀ
3Me D as for 2Me D. Calculated structures and spin density distri-
butions upon oxidation also point to a different degree of ground-
state charge and spin delocalization for the radical cations. For
m
-C4)}2Pd(PPh3)2]ꢀþ (Cp* ¼
h
5-C5Me5) [96] and the
bis(triarylamine)-appended paracyclophanes of Lambert and
Grampp (see Chart 2 for these systems) [47]. Another possible
source of signal broadening is homogeneous electron self-exchange
between the radical cations and the remaining neutrals in case of
incomplete oxidation of the neutral precursors to their radical
cations. Our present inability to produce pure samples 2ꢀD and 3ꢀD
by means of chemical oxidation (Agþ or acetylferrocenium) or by
quantitative coulometry under appropriate conditions does not
allow us to discriminate between these two possibilities. We
nevertheless note that similar [2.2]paracyclophane-bridged mixed-
valent radicals are all delocalized on the EPR timescale (see above)
and that the radical cation spectra of the triarylamine-substituted
paracyclophanes of Grampp’s and Lambert’s study were all
obtained in the presence of a ninefold excess of the corresponding
neutral [47]. We therefore consider the first explanation as the
more likely one. This view is also supported by our observation that
the X- and S-band EPR spectra of 2ꢀD and 3ꢀD are basically identical:
If the broadening was due to exchange the relative rates of that
exchange and of the experiment would be altered and a narrowing
of the spectra would be expected.
ꢀ
3Me D, removal of an electron structurally affects just one styryl
ruthenium site while there are hardly any changes on the other. In
contrast, both styryl ruthenium sites are affected upon oxidation of
ꢀ
2Me to 2Me D, albeit to a different degree. Moreover, the DFT (PBE0/
CPCM) calculated spin density distribution shows that the unpaired
spin resides on just one {Ru}-styryl subunit of the half-open
orthocyclophane 3MeꢀD whereas there is some degree of spin
delocalization between the two styryl ruthenium decks in the true
ꢀ
[2.2]paracyclophane radical cation 2Me D, (see Fig. 7 and Tables S2
and S3 of the Supporting Information).
The above conclusions pertain to the intrinsic electronic ground-
state delocalization on the timescale of molecular vibrations, that is
on timescales shorter than 10ꢁ11 s. They, however, provide not
much information about electron transfer dynamics. EPR is a highly
useful tool to address this issue since it operates on a slower
timescale of ca. 10ꢁ8 s. EPR investigations on the radical anions of
[2.2]paracyclophanes [41] or their dinitro derivatives [43] or on the
radical cations derived from electron-rich anisyl-derived or
triarylamine-decorated [2,2] and [3,3]paracyclophanes identified
them as intrinsically delocalized systems on the EPR timescale with
electron transfer rates of ca. 1$108 sꢁ1, irrespective of the strap
lengths [45e47]. Our failure to observe resolved hyperfine split-
tings in the X-band and S-band EPR spectra of radical cations 2ꢀD
Secondly, the absence of a half-field signal for dioxidized 22D
and 32D indicates, that the unpaired spins behave independently as
insulated spins, i.e, electronic coupling in the dioxidized state is, at
best, very weak. Our calculations on 2Me 2Dand 3Me 2D indeed place
diradical singlet and triplet ground states very close in energy with
Chart 2. An ¼ anisyl, 4-MeOC6H4.

3196
P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
the singlet diradical state as the slightly preferred one. For each
form the spin density is almost equally spread over both vinyl
ruthenium subunits.
Appendix. Supporting Information
Supplementary data associated with this article can be found in
the online version, at doi:10.1016/j.jorganchem.2011.06.028.
5. Conclusions
References
Complexes 2 and 3 feature two interlinked styryl-Ru(CO)
Cl(PiPr3)2 subunits, whose phenyl rings are integrated into a [2.2]
paracyclophane or a [2.1]orthocyclophane system. These two
architectures have virtually identical through-bond distances
between the individual styryl ruthenium chromophores and their
metal centers but strongly differ with respect to the mutual
[1] J.W. Steed, J.L. Atwood, Supramolecular Chemistry, Wiley, Chichester, 2000.
[2] Z. Chen, A. Lohr, C.R. Saha-Moeller, F. Würthner, Chem. Soc. Rev. 38 (2009)
564e584.
[3] H. Roesky, M. Andruh, Coord. Chem. Rev. 236 (2003) 91e119.
[4] C. Janiak, J. Chem. Soc., Dalton Trans. (2000) 3885e3896.
[5] P. Douglas, F.J. Stoddart, Angew. Chem. Int. Ed. Engl. 35 (1996) 1155e1196.
[6] N.W. Alcock, P.R. Barker, J.M. Haider, M.J. Hannon, C.L. Painting, Z. Pikramenou,
E.A. Plummer, K. Rissanen, P. Saarenketo, Dalton Trans. (2000) 1447e1462.
[7] A.S. Ionkin, W.J. Marshall, Y. Wang, Organometallics 24 (2005) 619e627.
[8] J.W. Hong, H.Y. Woo, B. Liu, G.C. Bazan, J. Am. Chem. Soc. 127 (2005)
7435e7443.
[9] S.R. Amrutha, M. Jayakannan, J. Phys. Chem. B 110 (2006) 4083e4091.
[10] P. Shen, G. Sang, J. Lu, B. Zhao, M. Wan, Y. Zou, Y. Li, S. Tan, Macromolecules 41
(2008) 5716e5722.
[11] Y. Yamauchi, M. Yoshizawa, M. Fujita, J. Am. Chem. Soc. 130 (2008)
5832e5833.
[12] L.-L. Li, K.-J. Lin, C.-J. Ho, C.-P. Sun, H.-D. Yang, Chem. Commun. (2006)
1286e1288.
[13] C. Elschenbroich, M. Wolf, O. Schiemann, K. Harms, O. Burghaus, J. Pebler,
Organometallics 21 (2002) 5810e5819.
[14] T. Cauchy, E. Ruiz, O. Jeannin, M. Nomura, M. Formigué, Chem. Eur. J. 13 (2007)
8858e8866.
[15] C. Elschenbroich, B. Kanellakopulos, F.H. Köhler, B. Metz, R. Lescouëzec,
N. Mitzel, W. Strauß, Chem. Eur. J. 13 (2007) 1191e1200.
[16] K.M. Knoblock, C.J. Silvestri, D.M. Collard, J. Am. Chem. Soc. 128 (2006)
13680e13681.
arrangements of the arene decks. Enforced p-stacking at distances
well beyond the sum of the van der Waals radii in the case of the
[2.2]paracyclophane-derived complex 2 contrasts with a consider-
ably larger separation and a strong tilt induced by the ortho
ethylene and methylene hinges in 3. Complexes 2 and 3 are thus
ideal probes for studying the mutual contributions of the
“through-space” and “through-bond” pathways for electron delo-
calization in such odd-electron phane architectures. The relative
shifts of the Ru(CO) bands upon stepwise oxidation to their radical
cations 2ꢀD and 3ꢀD and their dications 22D and 32D provide
a quantitative measure of ground-state delocalization through the
charge distribution parameter Dr. Based on the experimental Dr
values, 2ꢀD is a moderately and 3ꢀD a weakly coupled mixed-valent
system of class II. The direct comparison of 2ꢀD and 3ꢀD, which
only differ with respect to the mutual arrangement of the indi-
vidual styryl ruthenium subunits, suggests that the through-space
pathway dominates over the through-bond one in effecting
[17] C. Hou, J.-M. Shi, Y.-M. Sun, W. Shi, P. Cheng, L.-D. Liu, Dalton Trans. (2008)
5970e5976.
[18] D.L. Reger, A. Debreczeni, B. Reinecke, V. Rassolov, M.D. Smith, Inorg. Chem. 58
(2009) 8911e8924.
[19] I.V. Fedyanin, K.A. Lyssenko, Z.A. Starikova, M.Y. Antipin, Russ. Chem. Bulletin
53 (2004) 1153e1158.
[20] P.J. Ball, T.R. Shtoyko, J.A. Krause Brauer, W.J. Oldham, W.B. Connick, Inorg.
Chem. 43 (2004) 622e632.
[21] J. Casado, K. Takimiya, T. Otsubo, F.J. Ramírez, J.J. Quirante, R.P. Ortiz,
S.R. González, J.T. López Navarrete, J. Am. Chem. Soc. 130 (2008)
14028e14029.
[22] M.E. Stoll, S.R. Lovelace, W.E. Geiger, H. Schimanke, I. Hyla-Kryspin, R. Gleiter,
J. Am. Chem. Soc. 121 (1999) 9343e9351.
[23] S. Gath, R. Gleiter, F. Rominger, C. Bleiholder, Organometallics 26 (2007)
644e650.
[24] S.F. Nelsen, G. Li, K.P. Schultz, H.Q. Tran, I.A. Guzei, J. Am. Chem. Soc. 130
(2008) 11620e11622.
[25] D.-L. Sun, S.V. Rosokha, S.V. Lindeman, J.K. Kochi, J. Am. Chem. Soc. 125 (2003)
15950e15963.
[26] A. Fürstner, M. Alcarazo, H. Krause, C.W. Lehmann, J. Am. Chem. Soc. 129
(2007) 12676e12677.
[27] G.B. Bartholomew, G.C. Bazan, Acc. Chem. Res. 34 (2001) 30e39.
[28] Y. Tanaka, T. Ozawa, A. Inagaki, M. Akita, Dalton Trans. (2007) 928e933.
[29] M.R. Arkin, E.D.A. Stemp, R.E. Holmlin, J.K. Barton, A. Hörmann, E.J.C. Olson,
P.F. Barbara, Science 273 (1996) 475e480.
[30] S.O. Kelly, K.K. Barton, Chem. Biol. 5 (1998) 413e425.
[31] T.T. Tashica, J.K. Barton, in: M. Demeunynck, C. Bailly, W.D. Wilson (Eds.),
Small Molecule DNA and RNA Binders, Wiley-VCH, Weinheim, Germany,
2003, pp. 146e177.
delocalization
in
mixed-valent,
odd-electron
[2.2]para-
cyclophanes. This is under the assumption that delocalization via
the through-bond pathway is equal for both compounds. Our
results on odd-electron radical cations 2ꢀD and 3ꢀD contrast to
quantum chemical calculations on neutral closed-shell para-
cyclophanes, where the through-bond pathway was calculated to
be the more efficient one for the [2.2]paracyclophane architecture
[97].
Quantum chemical calculations on the simplified model
complexes 2Me nD and 3Me nD with the PiPr3 ligands replaced by
PMe3 reproduce the experimental structure of 2 and predict that
ꢀ
radical cation 2Me D assumes a completely delocalized electronic
structure in the gas phase. Inclusion of the CPCM solvent correction,
however, reproduced our experimental findings of partial or weak
ground-state delocalization in 2ꢀD and 3ꢀD as is indicated by the
structural differences between the individual styryl halves, the
good agreement of the calculated CO band splittings with the
experimental ones and by the calculated spin density distributions.
Radical cations 2ꢀD and 3ꢀD show a moderately intense elec-
tronic NIR band which, by comparison with the spectra of the
reduced or dioxidized forms and those of the monoruthenium [2.2]
paracyclophane radical cation 1ꢀD, is identified as an intervalence
charge-transfer (IVCT) transition.
[32] A. Tirosi, Mol. Simulat. 32 (2006) 707e716.
[33] A. van de Craats, N. Stutzmann, O. Bunk, M.N. Nielsen, M. Watson, K. Müllen,
H.D. Chanzy, H. Sirringhaus, R.H. Friend, Adv. Mater. 15 (2003) 495e499.
[34] L. Li, S.-W. Kang, J. Harden, Q. Sun, X. Zhu, L. Dai, A. Jakli, S. Kumar, Q. Li, Liquid
Crystals 35 (2008) 233e239.
[35] S. Laschat, A. Baro, N. Steinke, F. Giesselmann, C. Hägele, G. Scalia, R. Judele,
E. Kapatsina, S. Sauer, A. Schreivogel, M. Tosoni, Angew. Chem. Int. Ed. Engl. 46
(2007) 4832e4887.
[36] S. Sergeyev, W. Pisula, Y.H. Geerts, Chem. Soc. Rev. 36 (2007) 1902e1929.
[37] S. Rathgeber, D. Bastos de Toledo, E. Birckner, H. Hoppe, D.A.M. Egbe,
Macromolecules 43 (2010) 306e315.
[38] P.H. Dinolfo, M.E. Williams, C.L. Stern, J.T. Hupp, J. Am. Chem. Soc. 126 (2004)
12989e13001.
[39] P.H. Dinolfo, J.T. Hupp, J. Am. Chem. Soc. 126 (2004) 16814e16819.
[40] P.H. Dinolfo, S.J. Lee, V. Coropceanu, J.-L. Brédas, J.T. Hupp, Inorg. Chem. 44
(2005) 5789e5797.
[41] F. Gerson, Top. Curr. Chem. 115 (1983) 57e105.
[42] S. Amthor, C. Lambert, J. Phys. Chem. A 110 (2006) 1177e1189.
[43] S.F. Nelsen, A.E. Konradsson, J.P. Telo, J. Am. Chem. Soc. 127 (2005) 920e925.
[44] Y. Morisaki, Y. Chujo, Angew. Chem. Int. Ed. Engl. 45 (2006) 6430e6437.
Acknowledgment
We are grateful to Deutsche Forschungsgemeinschaft
(project Wi 1272/7e2) for financial support of this work. S. Z.
wishes to thank the Ministry of Education of the Czech Republic
(Grant COST LD11086) and the Grant Agency of the Academy of
Sciences of the Czech Republic (KAN100400702). We also wish
to thank Silvia Domingo-Köhler and Dr. Malte Drescher for the
recording of additional EPR spectra of complex 1þ. R. F. W. also
wishes to thank a referee for insightful comments and Prof. Dr.
Stephan Landgraf (Technical University of Graz) for helpful
discussions.

P. Mücke et al. / Journal of Organometallic Chemistry 696 (2011) 3186e3197
3197
[45] A. Wartini, J. Valenzuela, H.A. Staab, F.A. Neugebauer, Eur. J. Org. Chem. (1998)
139e148.
[71] N.W. Alcock, J. Cartwright, A.F. Hill, M. Marcellin, H.M. Rawles, J. Chem. Soc.,
Chem. Commun. (1995) 369e370.
[46] A. Wartini, H.A. Staab, F.A. Neugebauer, Eur. J. Org. Chem. (1998) 1161e1170.
[47] D.R. Kattnig, B. Mladenova, G. Grampp, C. Kaiser, A. Heckmann, C. Lambert,
J. Phys. Chem. C 113 (2009) 2983e2995.
[48] M. Krejcik, M. Danek, F. Hartl, J. Electroanal. Chem. 317 (1991) 179e187.
[49] P.R. Murray, L.J. Yellowlees, in: W. Kaim, A. Klein (Eds.), Spectroelec-
trochemistry, RSC Publishing, Cambridge, 2008, pp. 207e231.
[50] M.A. Esteruelas, H. Werner, J. Organomet. Chem. 303 (1986) 221e231.
[51] S. Clément, L. Guyard, M. Knorr, S. Dilsky, C. Strohmann, M. Arroyo,
J. Organomet. Chem. 692 (2007) 839e850.
[52] M. Holzapfel, C. Lambert, J. Phys. Chem. C 112 (2008) 1227e1243.
[53] M.J. Frisch, G. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Blonio, G. Zheng, J.L. Sonnenberg,
M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,
Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta,
F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov,
R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev,
A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma,
V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich,
A.D. Daniels, J.B. ÖFarkasForesman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussion
09, Revision A.02. Gaussian, Inc, Wallingford CT, 2009.
[72] Y. Maruyama, K. Yamamura, T. Sagawa, H. Katayama, F. Ozawa, Organome-
tallics 19 (2000) 1308e1318.
[73] M. Pichlmaier, R.F. Winter, M. Zabel, S. . Zális¸ , J. Am. Chem. Soc. 131 (2009)
4892e4903.
ꢀ
[74] J. Maurer, M. Linseis, B. Sarkar, B. Schwederski, M. Niemeyer, W. Kaim, S. Zális,
C. Anson, M. Zabel, R.F. Winter, J. Am. Chem. Soc. 130 (2008) 259e268.
ꢀ
[75] K. Kowalski, M. Linseis, R.F. Winter, M. Zabel, S. Zális, H. Kelm, H.-J. Krüger,
B. Sarkar, W. Kaim, Organometallics 28 (2009) 4196e4209.
[76] D.E. Richardson, H. Taube, Coord. Chem. Rev. 60 (1984) 107e129.
[77] R.J. Crutchley, Adv. Inorg. Chem. 41 (1994) 273e325.
[78] Y.-C. Lin, W.-T. Chen, J. Tai, D. Su, S.-Y. Huang, I. Lin, J.-L. Lin, M.M. Lee,
M.F. Chiou, Y.-H. Liu, K.-S. Kwan, Y.-J. Chen, Inorg. Chem. 48 (2009)
1857e1870.
[79] F. Barrière, N. Camire, W.E. Geiger, U.T. Mueller-Westerhoff, R. Sanders, J. Am.
Chem. Soc. 124 (2002) 7262e7263.
[80] D.M. D’Alessandro, F.R. Keene, Dalton Trans. (2004) 3950e3954.
ꢀ
[81] J. Maurer, R.F. Winter, B. Sarkar, J. Fiedler, S. Zális, Chem. Commun. (2004)
1900e1901.
ꢀ
[82] F. Hartl, D.J. Stufkens, A. Vlcek, Inorg. Chem. 31 (1992) 1687e1695.
[83] F. Hartl, A.K. Renfrew, F. Lafolet, T. Mahabiersing, M.J. Calhorda, S. Chardon-
Noblat, A. Deronzier, Inorg. Chem. 48 (2009) 8233e8244.
[84] S.J. Sherlock, D.C. Boyd, B. Moasser, W.L. Gladfelter, Inorg. Chem. 20 (1991)
3626e3632.
ꢀ
[54] G. te Velde, F.M. Bickelhaupt, S.J.A. van Gisbergen, C. Fonseca Guerra,
E.J. Baerends, J.G. Snijders, T. Ziegler, J. Comput. Chem. 22 (2001) 931e967.
[55] ADF 2010.01. in: V.U. SCM, Amsterdam (Ed.), edition 2005 ed., http://www.
scm.com, Amsterdam, The Netherlands.
[85] (a) J. Maurer, B. Sarkar, B. Schwederski, W. Kaim, R.F. Winter, S. Zális,
Organometallics 25 (2006) 3701e3712;
(b) X.H. Wu, S. Jin, J.H. Liang, Y.Y. Li, G.-ao Yu, S.H. Liu, Organometallics 28
(2009) 2450e2459.
[56] D. Andrae, U. Haeussermann, M. Dolg, H. Stoll, H. Preuss, Theor. Chim. Acta 77
(1990) 123.
[86] W.Y. Man, J.-L. Xia, N.J. Brown, J.D. Farmer, D.S. Yufit, J.A.K. Howard, S.H. Liu,
P.J. Low, Organometallics 30 (2011) 1852e1858.
[57] P.H. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213e222.
[58] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865e3868.
[59] C. Adamo, V. Barone, J. Chem. Phys. 110 (1999) 6158e6170.
[60] Y. Zhao, N.E. Schultz, D.G. Truhlar, J. Chem. Theory Comput. 2 (2006) 364e382.
[61] M. Cossi, N. Rega, G. Scalmani, V. Barone, J. Comput. Chem. 24 (2003)
669e681.
[87] M.A. Fox, R.L. Roberts, T.E. Baines, B. Le Guennic, J.-F. Halet, F. Hartl, D.S. Yufit,
D. Albesa-Jové, J.A.K. Howard, P.J. Low, J. Am. Chem. Soc. 130 (2008)
3566e3578.
[88] S. Zális, R.F. Winter, W. Kaim, Coord. Chem. Rev. 254 (2010) 1383e1396.
[89] M. Renz, K. Theilacker, C. Lambert, M. Kaupp, J. Am. Chem. Soc. 131 (2009)
ꢀ
16292e16302.
ꢀ
[62] E. van Lenthe, A. van der Avoird, P.E.S. Wormer, J. Chem. Phys. 107 (1997)
2488e2498.
[90] J. Maurer, B. Sarkar, W. Kaim, R.F. Winter, S. Zális, Chem. Eur. J. 13 (2007)
10257e10272.
[63] E. van Lenthe, A. van der Avoird, P.E.S. Wormer, J. Chem. Phys. 108 (1998)
4783e4796.
[64] A.D. Becke, Phys. Rev. A 38 (1988) 3098e3100.
[91] J. Maurer, B. Sarkar, S. Zalis, R.F. Winter, J. Solid State Electrochem. 9 (2005)
738e749.
ꢀ
[92] M. Linseis, R.F. Winter, B. Sarkar, W. Kaim, S. Zális, Organometallics 27 (2008)
[65] J.P. Perdew, Phys. Rev. B 33 (1986) 8822e8824.
3321e3324.
[66] E. van Lenthe, A.E. Ehlers, E.J. Baerends, J. Chem. Phys. 110 (1999) 8943e8953.
[67] S. Clément, L. Guyard, M. Knorr, C. Däschlein, C. Strohmann, Acta Cryst. E65
(2009) o528eu1861.
[68] H. Werner, M.A. Esteruelas, H. Otto, Organometallics 5 (1986) 2295.
[69] D. Huang, W.E. Streib, J.C. Bollinger, K.G. Caulton, R.F. Winter, T. Scheiring,
J. Am. Chem. Soc. 121 (1999) 8087e8097.
[93] M. Rudolph, S. Feldberg, DigiSim3, Version 3.03. Bioanalytical Systems, Inc,
1994.
[94] C.G. Atwood, W.E. Geiger, J. Am. Chem. Soc. 122 (2000) 5477e5485.
[95] S.V. Lindemann, S.V. Rosokha, J.K. Kochi, J. Am. Chem. Soc. 124 (2002)
843e855.
[96] W. Weng, T. Bartik, M. Brady, B. Bartik, J.A. Ramsden, A.M. Arif, J.A. Gladysz,
J. Am. Chem. Soc. 117 (1995) 11922e11931.
[70] S. Jung, K. Ilg, C.D. Brandt, J. Wolf, H. Werner, Eur. J. Inorg. Chem. (2004)
469e480.
[97] G.F. Caramori, S.E. Galembeck, J. Phys. Chem. A 111 (2007) 1705e1712.