TCNE and TCNQ ligands as efficient bridges in mixed-valence complexes containing iron-cyclopentadienyl and other organometallic systems

Article abstract of DOI:10.1016/S0277-5387(99)00331-9

Reaction between the π-accepting tetracyanoethene (TCNE) or tetracyano-p-quinodimethane (TCNQ) ligands and the π-electron-rich organometallic species Cp(dppe)Fe⁺ results in the formation of the complexes [{Cp(dppe)Fe}(n){η(n)-TCNX}](PF₆)(n) (n = 1, 2 and 4). Infrared spectroscopy, magnetic moment measurements, electron-spin resonance (ESR) and cyclic voltammetry data indicate a net transfer of one π electron to the TCNX acceptor ligand. The polynuclear complexes have an intense intervalence electron transfer absorption band in the near-infrared region. The values for the intervalence parameters α=0.02-0.06 and V(ab)=200-600 cm^-1 indicate that these complexes can be classified as class II according to Robin-Day. Estimation of the rate of electron transfer affords values in the order of 10⁸ s^-1, which are compared and discussed with values estimated for other polynuclear systems containing TCNE and TCNQ ligands as bridges. The solvent effect on the intervalence transition follows Hush's prediction for highly polar solvents, thereby permitting evaluation of the reorganizational energy. For the related series [{Cp(dppe)Fe}(n)-(μ-X)](PF₆)(n), the reorganization energy appears to depend mainly on the inner reorganization energy through changes to vibrational modes of the bridge. Reorganizational energy is not as important in determining the electron transfer rate as the activation energy, E(a). A linear relationship is found between the rate constant ln(k(et)) and the activation energy. These results can be discussed within the Marcus theory of electron transfer. The fast and efficient electron transfer between the metal fragments is discussed using molecular orbital arguments. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0277-5387(99)00331-9

www.elsevier.nl/locate/poly
Polyhedron 19 (2000) 137–145
TCNE and TCNQ ligands as efficient bridges in mixed-valence
complexes containing iron–cyclopentadienyl and other organometallic
systems
b
Carlos Diaz a, , Alejandra Arancibia
*
a Departamento de Quimica, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile
b Facultad de Quimica, P. Universidad Catolica de Chile, Casilla 306, Santiago, Chile
Received 10 May 1999; accepted 4 October 1999
Abstract
Reaction between the p-accepting tetracyanoethene (TCNE) or tetracyano-p-quinodimethane (TCNQ) ligands and the p-electron-rich
organometallic species Cp(dppe)Feq results in the formation of the complexes [{Cp(dppe)Fe}n{hn-TCNX}](PF6)n (ns1, 2 and 4).
Infrared spectroscopy, magnetic moment measurements, electron-spin resonance (ESR) and cyclic voltammetry data indicate a net transfer
of one p electron to the TCNX acceptor ligand. The polynuclear complexes have an intense intervalence electron transfer absorption band in
the near-infrared region. The values for the intervalence parameters as0.02–0.06 and Vabs200–600 cmy1 indicate that these complexes can
be classified as class II according to Robin-Day. Estimation of the rate of electron transfer affords values in the order of 108 sy1, which are
compared and discussed with values estimated for other polynuclear systems containing TCNE and TCNQ ligands as bridges. The solvent
effect on the intervalence transition follows Hush’s prediction for highly polar solvents, thereby permitting evaluation of the reorganizational
energy. For the related series [{Cp(dppe)Fe}n–(m-X)](PF6)n, the reorganization energy appears todepend mainly ontheinnerreorganization
energy through changes to vibrational modes of the bridge. Reorganizational energy is not as important in determining the electron transfer
rate as the activation energy, Ea. A linear relationship is found between the rate constant ln(ket) and the activation energy. These results can
be discussed within the Marcus theory of electron transfer. The fast and efficient electron transfer between the metal fragments is discussed
using molecular orbital arguments.
q2000 Elsevier Science Ltd All rights reserved.
Keywords: Hush theory; Iron complexes; Mixed valence; TCNE complexes; TCNQ complexes
1. Introduction
Tetracyanoethene (TCNE) and the related tetracyano-p-
quinodimethane (TCNQ) are molecules that present inter-
esting acceptor properties, which are associated with their
unusual electrical and magnetic behaviour [1–4]. TCNE and
TCNQ can act as s ligands (via the nitrile-N lone pair) as a
p-type arrangement (via the C_C olefinic bond in TCNE)
or ‘free’ (ion pair) mode II (Fig. 1) [5].
Many coordination compounds with considerable metal
ligand delocalization have been reported [6–10]. Several
special features make such compounds interesting for prac-
tical purposes: (1) intense charge-transfer absorption bands
in the visible and near-infrared regions can make such sys-
tems useful as dyes for information storage using optic fibre
and diode laser [11,12]; (2) the communication between
metal-connected ligands and/or ligand-connected metals has
Fig. 1. Coordination modes of TCNE and TCNQ ligands.
received attention in the area of low-dimensional polymers
with electron-propagating capabilities and within the concept
of ‘molecular electronics’ [13]; (3) the existence of several
close-lying and easily accessible redox transitions renders
* Corresponding author. Tel./fax: q56-2-271-3888
0277-5387/00/$ - see front matter q2000 Elsevier Science Ltd All rights reserved.
PII S0277-5387(99)00331-9
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

138
C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
these compounds suitable as ‘electron reservoirs’, especially
because small structural changes during the redox transitions
of extensively delocalized systems favour rapid electron
transfer due to the small reorganization barriers [14–16].
Non-reduced TCNE and TCNQ are rather poor ligands for
metal centres; however, on electron uptake by back-donation
or full reduction with one or two electrons, they can bind
metal fragments via modes I–IV.
In connection with points (2) and (3), we report here the
preparation of the complexes [{Cp(dppe)Fe}n{hn-TCNX}]-
(PF6)n and the study of the intermetallic electron transfer in
the following polynuclear systems:
respectively, to give dark black–green solids whose ele-
mental analysis and spectroscopic properties agree with the
mononuclear formulation [Cp(dppe)Fe(TCNX)]PF6. As
observed for another d5 paramagnetic mononuclear complex
[(C5R5)Mn(CO)2(h1-TCNX)] and others [8–10], mag-
netic resonance investigations of complexes 1 and 3 have
been hampered by the paramagnetism of the complexes. The
dark-black colour is typical of complexes of the type
Cp(dppe)FeIII [19–21]. Additional evidence for the reduc-
tion of the TCNX ligand is provided by from IR spectra [5]
(see Table 1). As for other monodentate TCNX complexes,
three nCN bands were observed [26] as expected for the non-
symmetrically bound ligands [5,27–33]. The nCN IR bands
are quite intense and shifted to lowerwavenumberscompared
to those of the free, non-reduced ligands. This indicates
nitrile-N coordination and sizeable back-donation of p elec-
tron density from the metal fragments to the TCNX ligand
[5,27–33]. The nCN wavenumber of the complexes lies
between those of the free anion radicals and the values found
for the dianions.
[Cp(dppe)Fe(TCNE)]PF6
(1)
(2)
(3)
(4)
[{Cp(dppe)Fe}2(TCNE](PF6)2
[Cp(dppe)Fe(TCNQ)]PF6
[{Cp(dppe)Fe}4(TCNQ](PF6)4
While several di- and polynuclear complexes with exten-
sive p delocalization including the metal centre have been
reported [6–10], no detailed quantitative studies on the elec-
tron transfer via TCNE and TCNQ bridges have been
reported.
Thus, the mononuclear complexes are distinguished from
polynuclear complexes and from the free ligands as the high-
est intensity is displayed by the low-energy band for the
polynuclear systems [6–10]. In the IR spectra of complexes
1 and 3 the highest intensity corresponds to the high-energy
band. IR bands for the Cp(dppe)Feq fragment appear nor-
mally [34]; the main values are displayed in Table 1.
We have selected the fragment Cp(dppe)Feq because its
electron richness permits efficient electronic delocalization
[11–21], it binds strongly to nitrogen cyanide containing
ligands [22–25] and the iron in the fragment is electrochem-
ically active [11–21].
2. Results and discussion
2.1.1. Magnetic measurements
Magnetic moments for complexes 1 and 3 are typical val-
ues for [Cp(dppe)FeX]q complexes [19,20], correspond-
ing to one unpaired electron in low spin FeIII. The unpaired
electron arising from the radical TCNXy could not be
observed due to spin coupling between them in solid state.
2.1. Mononuclear complexes
Cp(dppe)FeI reacts with an excess of TCNE or TCNQ in
the presence of NH4PF6 or TlPF6 in CH3OH or CH2Cl2,
Table 1
IR data for complexes 1–4 a
Complex
n(C5H5) b
n(dppe) c
n(PF6)
n(CN)
Solid
Solution
TCNE
2257
2220
2171
2059
1995
2251
2210
2200
2136
2065
1994
2197(sh)
2173
2223
2185
2129
[Cp(dppe)Fe(TCNE)]PF6 (1)
1098
696
843
[{Cp(dppe)Fe}2(TCNE)](PF6)2 (2)
1106
1094
1096
706
694
697
847
847
838
2197(sh)
2173
2223
2173
2128
2067
2099
2032
TCNQ
[Cp(dppe)Fe(TCNQ)]PF6 (3)
[{Cp(dppe)Fe}4(TCNQ)](PF6)4 (4)
2114
2035
a In KBr pellets, unless other indicated; data given in cmy1
b d(C–H) bending in-plane vibration.
.
c d(C–H) bending out-of-plane vibration.
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
139
Table 2
Long-wavelength absorption maxima l (nm), magnetic and ESR data for TCNX complexes
Complex
lmax a (CH2Cl2)
m (eff) c
ESR c: g1, g2, g3
1
2
3
4
545 (1187), 619 (1245)
441 b (2262), 512 (1769) b, 1073 (969)
498 (1189)
1.8
0.7
1.9
2.5
2.0189; 2.1912; 2400
1.9880; 2.2086; 2.420
1.9883; 2.0095; 2.2439
1.9940; 2.0393; 2.2417
498 (5741), 1008 (14300)
a Molar extinction coefficients (My1) are given in parentheses.
b Shoulder.
c In solid state.
characteristic of Cp(dppe)FeIIIX complexes [35–37]. Three
well-separated features corresponding to the three compo-
nents of the g tensor — expected for low symmetry pseudo-
octahedral piano–stoolcomplexes—wereobserved(seeFig.
2). Data are shown in Table 2. This indicates that the signal
observed in both cases may correspond to the unpaired elec-
tron centred on the metal. The signal corresponding to the
TCNEy and TCNQy radical ligands was not observed, prob-
ably due to intramolecular spin coupling between the
TCNXy molecules in the solid state. Consistent with this, in
CH2Cl2 solution the ESR spectra of 1 and 3 exhibit several
complex signals with various features (see Fig. 2), which
presumably include the signal of Cp(dppe)FeIIIL as well as
the signals of TCNXy. As found for other LnM(h1-
TCNXy) complexes [38,39], owing to the lowered sym-
metry, 16 (four different nitrogen) or 1296 (four different
nitrogen and four different 1H) lines for TCNE and TCNQ,
respectively, could be observed.
2.1.3. Cyclic voltammetry
Mononuclear systems are reduced at more negative poten-
tials than TCNX ligands (Table 3). This is a consequence of
metal-to-ligand electron transfer in the ground state, which
leaves either the single reduced ligand or the oxidized metal
fragment available for reduction at relatively negative
potentials.
Fig. 2. ESR spectra of complex 1 in CH2Cl2 solution (a) and in solid state
(b) at room temperature.
The ESR data discussed below are also in agreement with
this argument. Values are shown in Table 2.
Oxidation of the complexes is almost always irreversible;
it may thus be safely assumed that these are metal-based.
Thus peak potentials of about 0.7–0.8 V versus SCE
2.1.2. ESR spectra
The ESR spectra of mononuclear complexes 1 and 3 at
room temperature in solid state exhibit one single signal,
Table 3
Electrochemical data a for TCNX ligands and their complexes
ox
1/2
Compound
E
(metal)
Er1ed (0/1y) d
Er2ed (y/2y) d
TCNE
0.18
y0.61
y0.53
0.25
y0.49
y0.65
y0.85
y0.95
y0.88
y0.31
y0.70
y0.80
[Cp(dppe)Fe(TCNE)]PF6 (1)
[{Cp(dppe)Fe}2(TCNE)](PF6)2 (2)
TCNQ
[Cp(dppe)Fe(TCNQ)]PF6 (3)
[{Cp(dppe)Fe}4(TCNQ)](PF6)4 (4)
0.83 b
0.89, 0.70 c
0.78 b
0.93, 0.75 c
a In CH2Cl2–0.1 M Bu4NPF6; potential (in V) vs. SCE.
b Anodic peak potential are given for irreversible oxidation processes.
c Two very near quasi-reversible waves were observed.
d Quasi-reversible waves.
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

140
C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
(saturated calomel electrode) for complexes [Cp(dppe)-
Fe(nitrile)]PF6 have been observed [22–25].
having a 1,2-trans geometry: [(PPh3)2(CO)Ir]2TCNE
[29], (2176(m), 2097 vs); [L9L2Cu(m-TCNE)CuL2]n
[30], (2190, 2147); [Rh2(O2CCF3)4]2TCNE [31], (2230,
2210); [Tl(18-crown-6)TCNQ] [32], (2181, 2154);
[Ru(PPh3)2]2TCNQ [33] (2170, 2140).
2.1.4. Electronic spectroscopy
Complexes 1 and 3 exhibit intense bands at about 545 and
619 nm, which are typical of the absorption behaviour of
[Cp(dppe)FeIIIL](PF6)2 complexes [17–21] (Fig. 3).
Table 2 summarizes the data obtained. No absorption was
observed above 700 nm. This constitutes strong evidence for
the FeIII–(TCNXy) state because FeII–NCR complexes
exhibit only one band at about 450 nm [40].
In agreement with this, the binuclear complex 2 exhibits
two bands (one as a shoulder), suggesting the 1,2-trans con-
figuration for this complex. This is the favouredconfiguration
for bulky fragments [5] such as CpFe(dppe). As expected
for tetranuclear complexes (with a slightly perturbed D2h
symmetry), two bands are observed for complex 4. Thus
complexes [Ru(NH3)5–(m4-TCNX)] and [C5Me5–
Mn(CO)2]4–(m4-TCNX) exhibit bands at: 2163, 2121
cmy1 for the Ru–TCNE complex; 2153, 2096 cmy1 for the
Ru–TCNQ complex; 2160, 2110 cmy1 for the Mn–TCNE
complex; and 2170, 2105 cmy1 for the Mn–TCNQ complex.
As observed for similar (MLn)2TCNXnq and (MLn)4-
TCNXnq complexes the highest intensity is displayed by the
low-energy band, behaviour that is characteristic of polynu-
clear systems [6–10].
2.2. Polynuclear complexes
TCNE or TCNQs react with an excess of Cp(dppe)FeI in
the presence of NH4PF6 or TlPF6 in CH3OH or CH2Cl2,
respectively, to give black–green solids of composition
{[Cp(dppe)Fe]2TCNE}(PF6)2 and {[Cp(dppe)Fe]4-
TCNQ}(PF6)4. Similarly to complexes 1 and 3, NMRstudies
are precluded due to the paramagnetism of the complexes. In
agreement with this, the related complex [Ru(NH3)]4–
(m-TCNX)]10q is also paramagnetic [7]. No suitable crys-
tals were obtained for X-ray structural determination of the
complexes.
2.2.2. Magnetic measurements
The low magnetic moment for the binuclear complex 2
could be due to the delocalization of the unpaired electron
between the two iron atoms, which couple antiferromagnet-
ically with the unpaired electron of the TCNEy ligand. On
the other hand the tetranuclear complex 4 exhibits the highest
magnetic moment, which could be due to the complex inter-
action of several paramagnetic FeIII centres. Similar mag-
netic moments have been observed for tetranuclear
{[CpUMn(CO)2]4–(m-TCNE)} complexes [6–10].
2.2.1. Infrared spectra
As pointed out by Kaim [8] and others [28], for [MLn]n–
(TCNX) complexes some information about the structure of
the complexes can be obtained from IR spectroscopy. The
number and the pattern of bands observed by IR depend on
the symmetry of the complexes; for the binuclear systems
[MLn]2–(m-TCNX) three isomers can be expected: 1,2-
trans (C2h), 1,2-cis (C2v) or 1,19 (C2v). According to their
symmetry, two or four bands could be observed for the iso-
mers 1,2-trans, 1,1-cis or 1,2-cis, respectively. This is con-
sistent with IR data for [MLn]n–(m-TCNX) complexes
Fig. 3. Electronic absorption spectra of complexes 1 and 2 in CH2Cl2
solution.
Fig. 4. ESR spectra of compound 2 in CH2Cl2 solution (a) and in solid state
(b) at room temperature.
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
141
Table 4
Intervalence band parameters and derived electronic coupling parameter for the mixed-valence complexes a
nmax (cmy1
)
´ (My1 cmy1
)
Dn1/2 (cmy1
)
a
d (A)
Vab (cmy1
)
Ea (cmy1
)
Em (cmy1
)
Kth (segy1
)
˚
Complex
2
4
9320
9978
968
5960
1635
3159
0.027
0.063
9.9
14.2
253
625
2330
2480
9320
9921
2.18=108
1.22=108
a In CH2Cl2 solution.
2.2.3. ESR spectra
(Cp(dppe)FeII)qqTCNX™(Cp(dppe)FeIII–[TCNX])xy
(1)
The ESR spectrum of compound 2 in solid state shows a
broad line exhibiting recognizable anisotropic features (Fig.
4), behaviour typical of delocalized mixed-valence systems
[35–37]. The spectrum of compound 4 displays a similar
pattern with the g factors summarized in Table 3. Asobserved
for other multinuclear Ru–TCNX compounds [7], the ESR
signal for the radical TCNXy ligand was not observed. In
CH2Cl2 solution, similarly to the mononuclear complexes,
the signal of the iron centre as well as the complex signal of
the organic TCNXy radical were observed (see Fig. 4).
In the oxidation region of the metal fragment, the cyclic
voltammogram exhibits two non-well-defined waves which
probably involve the oxidation of the metal centres. The dif-
ference in peak potential indicates a considerable coupling
between the metal centres. From the wave separation, the
comproportionation constant, Kc, relative to the following
equilibrium:
FeII–TCNX–FeIIqFeIII–TCNX–FeIIIl2FeII–TCNX–FeIII
(2)
2.2.4. UV–Vis spectra
The UV–Vis spectra of complexes 2 and 4 revealed, in
addition to absorptions typical for Cp(dppe)FeIIIL [17–21]
and Cp(dppe)FeIIL [40] chromophores, a new absorption
band at about 1100 nm. The electronic spectrumofcompound
2 in CH2Cl2 is shown in Fig. 3.
These observations coupled with data on similar
Cp(dppe)FeIII–L–FeII(dppe)Cp systems [11–21] allow us
to assign this near-IR band to an intervalence transfer tran-
sition. Thus, on reducing the mixed valence complexes 2 and
4 with Red Al, sodium bis(2-methoxyethoxy)aluminium
hydride, the intervalence absorption disappears.
can be computed as Kcs1626 and 1202 for complexes 2 and
4, respectively. The relatively high Kc values indicate a con-
siderable electronic interaction between the iron atoms.
2.2.6. Electronic interaction in the mixed-valence species
Information about the degree of metal–metal interactionin
the mixed-valence complexes can be obtained from the inter-
valence absorption band using the Hush formalism [11–
16,41–45]. According to this, the degree of delocalization
parameter is given by the formula:
(4.2=10y4 )´maxDn1/2
as
(3)
nmax d 2
2.2.5. Cyclic voltammetry
where Dn1/2 is the bandwidth at half-maximum (cmy1), ´max
is the maximum extinction coefficient, nmax is the band posi-
The reduction potential of the polynuclear complexes with
TCNE and TCNQ (Table 3) are more negative than those of
the corresponding free ligands; this is consistent with the
electron transfer involved in the formation of the complexes,
tion (cmy1) and d is the metal–metal distance in A. Because
˚
no X-ray determination is available for complexes 2 and 4,
the metal–metal distance was estimated from structural data
for similar complexes [46–48]. The a values shown in Table
4 indicate a class II Robin-Day system.
On the other hand the electronic coupling term, Vab, can
be determined by the following formula [41–45]:
leaving the ligands as TCNXy or TCNX2y
.
Table 5
Electronic interaction parameter and electron transfer rate constants for 2, 4
and other polynuclear mixed-valence complexes containing TCNE and
TCNQ bridges
Vabsa1/2 nmax
(4)
Complexes
Vab (cmy1
)
ket (segy1
)
The calculated values, displayedinTable4, indicateamore
or less strong metal–metal coupling, although somewhat less
than for the cyanide ligand. From intervalence transfer tran-
sition data for other organometallic systems, the interaction
parameter Vab was estimated using Eq. (4). Values shown in
Table 5 for other complexes indicate that metallic fragments
such as (NH3)5Ru permit a stronger interaction than the iron
fragment, which can be due to steric hindrance of the
Cp(dppe)Feq arising from the bulky dppe [46]. From data
[{CpU(CO)2Mn}2(m-TCNE)] a
[{(NH3)5Ru}4(m-TCNE)](PF6)8
1302
2980
1256
1860
253
1.27=1010
2.7=108
1.26=109
2.1=109
2.2=108
6.2=108
a
[{(NH3)5Ru}4(m-NCC(t-Bu)CN)] a
a
[{(NH3)5Ru}4(m-TCNQ)](PF6)8
[{Cp(dppe)Fe}2(m-TCNE)](PF6)2
[{Cp(dppe)Fe}4(m-TCNQ)](PF6)4
b
b
625
a Determined using Eq. (4) and literature data [11–16,27–33] for the inter-
valence transition.
b This work.
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

142
C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
The solvent effect on the intervalence transition gives
information about the reorganization energy arising from
rearrangements of the inner and outer coordination spheres
[49,50]. With the assumption of a dielectric continuum
model, the dependence of Eop on solvent polarization is given
by:
EopsEiqEo
(6)
1
1
1
1
1
EopsEiqe2
q
y
=
y
ž2a / žD /
2a2
d
Ds
1
op
The intercept of a plot of Eop versus (1/Dopy1/Ds) for
several solvents affords the Ei values (Fig. 5 and Table 6).
From Eq. (6) the outer-sphere reorganizational parameters
Eo can be estimated. Values for complexes 2 and 4 and for
other complexes containing the CpFe(dppe) moiety are
shown in Table 6. Some comments can be made from this
data: comparing the E values for complexes 2 and 4, mostly
larger variations on the outer-reorganizational parameters
than for the inner-reorganization values are observed (dif-
ferences of 21% on Ei and of 43% on Eo are observed). This
can be explained by the similar metal–ligand vibrationmodes
involved in the electron-transfer process. Attention must be
drawn to the high inner-reorganization energies involved in
electron-transfer events in the complex {[Cp(dppe)]2–(m-
S–C5H5N)}(PF6)2. This may be due to an enhanced change
on vibrational modes of the thiopyridine ligand upon electron
transfer.
Fig. 5. Plot of intervalence absorption energy for Eop vs. (1/Dopy1/Ds).
in Table 5, it can also be concluded that the interaction
between the metal centres is highly dependent on the nature
of the organometallic fragment.
An approximate estimation of the rate of electron transfer
between the metal centres in complexes 2 and 4 can be made
using the expression [41–45]:
1/2
2p(Vab )2
p
Ea
kths
e y
(5)
ž /
≥ ¥≥ ¥
h
KTEm
RT
≥ ¥
where EmsEop and Eas[(Em)2/4Em]. Em, the total reor-
ganization energy, has two components: the outer-reorgani-
zation energy Eo, and the inner-reorganization energy Ei;thus
EmsEiqEo.
Although these equations differ in absolute values from
those measured experimentally, comparison for a related
series of compounds is valid [18]. The values about 108 sy1
calculated for compounds 2 and 4 indicate a fast electron
transfer between the iron atoms. Similar rates are obtained
using Eq. (5) for other organometallic systems, as can be
seen in Table 5. The similar ket values, notwithstanding the
different organometallic fragment, mean that the nature of
the bridge is most important in determining theelectrontrans-
fer rate. It is interesting to compare the TCNX ligands as
bridges with the cyanide bridging ligand for which electron-
transfer rates of 104 sy1 were estimated in the mixed-
valence complexes [Cp(dppe)Fe–CN–MLn]PF6 (MLns
Cp(dppe)Fe, Mn(CO)2(dppm)P(OPh)3) [18].
It is interesting to note that, although the complexes dis-
played in Table 7 all have the same fragment CpFe(dppe),
the reorganization energy is very different. For instance,
compare Ems12 697 cmy1 for [CpFe(dppe)]2–(m-S–
C5H5N)](PF6)2 [21] with Ems6038 cmy1 for {[CpFe-
(dppe)]2–m-CN]}PF6 [18].
This means that, for this series, the reorganizationalenergy
is highly dependent on some properties of the bridge ligand.
Because the data in Table 7indicatethatthemaincontribution
to the reorganizational energy arises from the inner-reorgan-
ization energy, the observed variations Em might stem from
changes in the vibrational mode of the bridging ligand.
Recent studies of electron transfer in the mixed-valence
systems [(NC)5MII–CN–RuIII(NH3)5]y (MsFe or Ru)
Table 6
Solvent dependence of the intervalence transition in complexes 2 and 4
Complex
Solvent
Eop (cmy1
)
(1/Dopy1/Ds)
Ei (cmy1
)
Eo (cmy1
)
Em
2
2
2
2
4
4
4
4
4
Dimethylsulfoxide (DMSO)
Acetone
MeOH
CH2Cl2
DMSO
9128
9238
9208
9320
9501
9872
9799
9780
9978
0.6551
0.6886
0.7234
0.5899
0.6550
0.7234
0.6722
0.6886
0.5899
6621
2587
9128
8394
1448
9501
MeOH
Ethyl methyl ketone
Acetone
CH2Cl2
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
143
Table 7
Solvent reorganization energies a for complexes 2, 4 and other related complexes containing the moiety CpFe(dppe)q
Complex
Ei (cmy1
)
Eo (cmy1
)
Em (cmy1
)
[{CpFe(dppe)}2–(m-TCNE)](PF6)2
[{CpFe(dppe)}4–(m-TCNQ)](PF6)4
[{CpFe(dppe)}2–(m-S–C5H4N)](PF6)2
[{CpFe(dppe)}2–(m-CN)]PF6
8394
6621
12470
2096
1448
2587
227
9842
9208
12697
6038
3942
a Values of Em, Ei and Eo were determined from Eop vs. (1/Dopy1/Ds) plots. Values for complexes other than complexes 2 and 4 were taken from [11–21].
have shown that a large part of the reorganization energy
comes from modes assigned to the bridging ligand [51,52].
Comparing the reorganizational values with electron-
transfer rates it can be observed that there does not exist a
correlation between them. According to semiclassical elec-
tron-transfer theory [41–45], electron-transfer reaction rates
(ket) are governed by three parameters (Eq. (5)): (1) the
electronic coupling (Vab) between reactants and products,
(2) the free-energy change for the reaction (DG8) and (3)
the total reorganization energy.
Bruker ECS 106 spectrometer. FTIR spectra were recorded
on an FT–IR Perkin Elmer 2000 spectrophotometer. Elec-
tronic absorption spectra were recorded on a Shimadzu UV
160. Cyclic voltammetry was performed on a three-electrode
configuration (glassy-carbon working electrode, satured cal-
omel as reference) with 0.1 M solution of NH4PF6 as elec-
trolyte support in CH2Cl2. Cp(dppe)FeI was prepared as
previously reported [22–25].
From Tables 5 and 7 it can be seen that no correlationexists
between ket and the electron coupling or the reorganization
energy. However, the electron-transfer rate appears to be
related to DG8 through the relationEas(Eop)2/(EopyDG8).
In fact, a linear correlation between ln(kth) and Ea holds, as
shown in Fig. 6. Furthermore, this expression has the shape
of the most general and simple relationship between ket and
Ea, the classical Marcus equation ketsAe(yEa) [41–45].
We believe that this is a remarkable example where the gen-
eral electron rate theory of Marcus works well in mixed-
valence complexes.
2.3. Nature of the electron transfer in binuclear complexes
2 and 4
Fig. 6. Estimated electron-transfer rates vs. activation energy Ea for
complexes 2, 4, {[CpFe(dppe)]2–(m-S–C5H4N)}(PF6)2 (5) and
{[CpFe(dppe)]2–(m-CN)} PF6 (6).
The efficient and fast electron transfer between the metal
centres in complexes 2 and 4 can be explained by considering
the molecularorbitalsinvolvedintheFe–Ninteraction.These
can be adequately established from an orbital interaction dia-
gram built from the molecular orbital (MO) fragments of
CpFe(dppe)q and nitriles [53–56]. As shown in Fig. 7, the
HOMO of complexes involves a p-delocalized molecular
orbital throughout the Fe–TNCX–Fe moiety.
3. Experimental
All operations were conducted under a pure dinitrogen or
argon atmosphere using standard Schlenk and gloveboxtech-
niques. Solvents were dried according to established proto-
cols and degassed prior to use. Unless otherwise specified,
reagents were obtained from commercial supplies and used
as received. Magnetic measurements were carried out by
Gouy’s method at room temperature using a JohnsonMatthey
magnetic susceptibility balance with tetrathiocyanate cobal-
tate(II) as calibrant balance. ESR spectra were recorded on
Fig. 7. Proposed HOMO-LUMO of complexes 2 and 4.
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

144
C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
3.1. [Cp(dppe)Fe–(TCNE)]PF6 (1)
3.5.1. [Cp(dppe)Fe–(TCNE)]PF6
A mixture of 0.15 g, 0.25 mmol of Cp(dppe)FeI, 0.04 g,
0.31 mmol of TCNE and 0.16 g, 0.45 mmol of TlPF6 (cau-
tion: thallous salts are very poisonous and should be handled
with care) in 30 ml ofCH2Cl2 was stirredatroomtemperature
for 3 days. The solution was then filtered through Celite and
the filtrate was removed in vacuo to about 15 ml. The black
solid precipitated was isolated by decantation and washed
twice with n-hexane–ether, and dried in vacuo to give 0.12
g, 60% yield.
A solution of Cp(dppe)FeI (70 mg, 0.10 mmol) in 20 ml
of CH3OH was added dropwise over 0.03 g, 0.23 mmol of
C2(CN)4 and 0.03 g, 0.18 mmol of NH4PF6. The mixture
was stirred at room temperature for24h. Thereactionmixture
was then evaporated to dryness, and the solid residue redis-
solved in 30 ml and the solution filtered through Celite and
evaporated to about 15 ml. After this a black solid was
formed, which was decanted and washed twice with hexane–
ether and dried in vacuo. Yield: 28 mg (29%). Anal. Calc.
for C41H29F6N4P3FePCH2Cl2: C, 52.08; H, 3.54; N, 6.38.
Found: C, 51.73; H, 4.02; N, 5.90%.
Acknowledgements
3.2. [{Cp(dppe)Fe}2–(m-TCNE)](PF6)2 (2)
Financial support of this work from D.I.D., Universidad
de Chile (Project 98-018), is gratefully acknowledged.
To a solution of Cp(dppe)FeI, 0.15 g, 0.23 mmol in 10 ml
of CH3OH, was added dropwise slowly over 0.015 g, 0.11
mmol of C2(CN)4 in 20 ml of CH3OH in the presence of
NH4PF6 (0.03 g, 0.18 mmol). The mixture was stirred at
room temperature for 24 h and then concentrated to dryness
under reduced pressure. The solid was dissolved in CH2Cl2
(30 ml) and filtered through Celite. The solution was con-
centrated in vacuum to about 15 ml and the black solid pre-
cipitated was washed twice with hexane–ether and dried
in vacuo. Yield: 0.045 g (28%). Anal. Calc. for
C72H58F12N8P2Fe2: C, 62.26; H, 4.4; N, 5.0. Found: C, 61.7;
H, 5.0; N, 5.0%.
References
[1] K.R. Dunbar, Angew. Chem., Int. Ed. Engl. 35 (1996) 1659.
[2] J.S. Miller, A.J. Epstein, Angew. Chem. 106 (1994) 399.
[3] J.S. Miller, A.J. Epstein, Angew. Chem., Int. Ed. Engl. 33 (1994)
385.
[4] J.S. Miller, Adv. Mater. 6 (1994) 322.
[5] W. Kaim, M. Moscherosch, Coord. Chem. Rev. 129 (1994) 157.
[6] E. Waldhor, W. Kaim, M. Lawson, J. Jordanov, Inorg. Chem. 36
(1997) 3248.
[7] M. Moscherosch, E. Waldhor, H. Binder, W. Kaim, J. Fiedler, Inorg.
Chem. 34 (1995) 4326.
[8] R. Gross-Lannert, W. Kaim, B. Olbrich-Deussner, Inorg. Chem. 29
(1990) 5046.
3.3. [Cp(dppe)Fe–(TCNQ)]PF6 (3)
[9] R. Gross-Lannert, W. Kaim, Angew. Chem., Int. Ed. Engl. 26 (1987)
251.
A solution of Cp(dppe)FeI, 0.15 g, 0.23 mmol in 10 ml
of CH3OH, was added dropwise over 0.07 g, 0.34 mmol of
TCNQ and in the presence of NH4PF6 (0.06 g, 0.36 mmol).
Following the procedure outlined for separation and purifi-
cation of compound 1, 0.2 g of black solid was obtained.
Yield: 68%. Anal. Calc. for C43H33F6N4PFePCH2Cl2: C,
55.43; H, 3.6; N, 5.8. Found: C, 54.3; H, 3.0; N, 4.8%.
[10] W. Kaim, T. Roth, B. Olbrich-Deussner, R. Gross-Lannert, J.
Jordanov, E.K.H. Roth, J. Am. Chem. Soc. 114 (1992) 5693.
[11] J. Fabian, K. Zahradnik, Angew. Chem. 101 (1989) 693.
[12] J. Fabian, K. Zahradnik, Angew. Chem., Int. Ed. Engl. 28 (1989)
677.
[13] D. Astruc, Acc. Chem. Res. 30 (1997) 383.
[14] D. Astruc, Electron Transfer and Radical Processes in Transition-
Metal Chemistry, VCH, New York, 1995.
[15] D. Astruc, Acc. Chem. Res. 12 (1986) 377.
[16] M.H. Desbois, D. Astruc, J. Guillin, F. Varret, A.X. Trautwein, G.
Villaneuve, J. Am. Chem. Soc. 111 (1989) 5800.
3.4. [{Cp(dppe)Fe}4–(m-TCNQ)](PF6)4 (4)
´
[17] G. Barrado, G.A. Carriedo, C. Dıaz, V. Riera, Inorg. Chem. 30 (1991)
To a solution of Cp(dppe)FeI, 0.15 g, 0.23 mmol in 20 ml
of CH3OH, was added dropwise slowly over 0.03 g, 0.14
mmol of TCNQ in 20 ml of CH3OH in the presence of
NH4PF6 (0.04 g, 0.25 mmol). The mixture was stirred at
room temperature for 24 h. Separation and purification pro-
cedures similar to that for compound 3 afforded a black pow-
der. Yield: 0.2 g (48%). Anal. Calc. for C136H120F24N4P4-
Fe4P2CH2Cl2: C, 57.09; H, 4.2; N, 1.96. Found: C, 55.39; H,
4.26; N, 2.6%.
4416.
´
´
´
´
[18] C. Dıaz, A. Arancibia, Inorg. Chim. Acta 269 (1998) 246.
[19] C. Dıaz, C. Leal, N. Yutronic, J. Organomet. Chem. 516 (1996) 59.
[20] C. Dıaz, Polyhedron 16 (1997) 999.
[21] C. Dıaz, A. Arancibia, unpublished results.
´
[22] C. Dıaz, N. Yutronic, Synth. React. Inorg. Met.-Org. Chem. 27(1997)
119.
´
´
[23] C. Dıaz, R. Latorre, Bol. Soc. Chil. Quim. 37 (1992) 211.
[24] C. Dıaz, A. Arancibia, Bol. Soc. Chil. Quim. 43 (1998) 303.
[25] G.A. Carriedo, A. Arancibia, C. Diaz, N. Yutronic, E.P. Carreno, S.G.
Granda, J. Organomet. Chem. 508 (1996) 23.
˜
[26] Strictly four bands must be observed for a Cs local symmetry of the
h1-nitrogen linked ligand. However, three bands are normally
observed, the fourth being observed as a shoulder, or masked for the
most intense absorption; see [5–10] and references cited in [27].
[27] B. Olbrich-Deubner, R. Gross, W. Kaim, J. Organomet. Chem. 366
(1989) 155.
3.5. Preparation of complexes 1–4 using TlPF6 as halide
abstractor and CH2Cl2 as solvent
Similar compounds were obtained using the following
typical procedure.
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288

C. Diaz, A. Arancibia / Polyhedron 19 (2000) 137–145
145
[28] J.L. Wesemann, M.H. Chisholm, Inorg. Chem. 36 (1997) 3258.
[29] G.T. Yee, J.C. Calabrese, C. Vazquez, J.S. Miller, Inorg. Chem. 32
(1993) 377.
[43]
[44]
[45]
[46]
N.S. Hush, Coord. Chem. Rev. 64 (1985) 135.
R.A. Marcus, N. Sutin, Biochim. Biophys. Acta 811 (1985) 265.
Y. Dong, J.T. Hupp, Inorg. Chem. 31 (1992) 3170.
P.E. Riley, C.E. Capshew, R. Pettit, R.E. Davis, Inorg. Chem. 17
(1978) 408.
´
[30] M.M. Olmstead, G. Speier, L. Szabo, J. Chem. Soc., Chem. Commun.
(1994) 541.
[31] F.A. Cotton, Y. Kim, J. Am. Chem. Soc. 221 (1993) 1.
[32] M.C. Grossel, S.C. Wester, J. Chem. Soc., Chem. Commun. (1992)
[47]
[48]
L. Ballester, M. Carmen Barral, A. Gutierrez, R. Jimenez-Aparicio, J.
˜
Martinez-Muyo, M.F. Perspinan, M.A. Monge, C. Ruiz-Valero, J.
1510.
Chem. Soc., Chem. Commun. (1991) 1395.
D. Ewan, A.J. Blake, T.A. Stephenson, M. Schroder, L.J. Tellouless,
J. Chem. Soc., Chem. Commun. (1988) 1533.
M.D. Newton, N. Sutin, Annu. Rev. Phys. Chem. 35 (1984) 437.
N. Sutin, B.S. Brunschrirg, C. Creutz, J.R. Winkler, Pure Appl. Chem.
60 (1998) 1817.
P. Forlano, F.D.C. Kiernik, O. Poizat, J. Olabe, J. Chem. Soc., Dalton
Trans. (1997) 1595.
K.S. Door, P.O. Stoutland, R.B. Dyer, H.W. Woodruff, Am. Chem.
Soc. 114 (1992) 3133.
[33] I. Ito, S. Kitazume, A. Yamamoto, S. Ikeda, J. Am. Chem. Soc. 92
(1970) 3011.
´
´
[34] M.M. Campos, C. Dıaz, K. Figueroa, L. Padilla, N. Lara, G. Dıaz,
Vibration Spectrosc. 7 (1994) 61.
[49]
[50]
[35] P. Hamon, I.R. Hamon, C. Lapinte, J. Chem. Soc., Chem. Commun.
(1992) 1602.
[36] P. Hamon, L. Toupet, J.R. Hamon, C. Lapinte, J. Chem. Soc., Chem.
Commun. (1994) 931.
[37] V. Mathias, S. Cron, L. Toupet, C. Lapinte, Organometallics 15
(1996) 5393.
[38] T. Roth, W. Kaim, Inorg. Chem. 31 (1992) 1930.
[39] S.E. Bell, J.S. Field, R.J. Haines, M. Moscherosh, W. Matheis, W.
Kaim, Inorg. Chem. 31 (1992) 3269.
[51]
[52]
[53]
For frontier OM of fragment Cp(dppe)Feq see: J.F. McGrady, T.
Lowell, R. Stranger, M.G. Humphrey, Organometallics 16 (1997)
4004.
´
[54]
[55]
See also our studies in [40].
N.M. Kostic, R.F. Fenske, Organometallics 1 (1992) 1979.
[40] C. Dıaz, B. Weiss, Polyhedron 12 (1993) 1403.
[41] N.S. Hush, Prog. Inorg. Chem. 8 (1967) 391.
[42] N.S. Hush, Electrochim. Acta 13 (1968) 1005.
[56] For frontier OM for TCNX, see [6,8].
Tuesday Feb 01 09:29 AM
StyleTag -- Journal: POLY (Polyhedron) Article: 3288