Benzo[c]thiophene chromophores linked to cationic Fe and Ru derivatives for NLO materials: Synthesis characterization and quadratic hyperpolarizabilities

Article abstract of DOI:10.1002/ejic.201300048

η⁵-Monocyclopentadienyliron(II)/ruthenium(II) complexes of the general formula [M(η⁵-C₅H₅)(PP)(L1)] [PF₆] {M = Fe, PP = dppe; M = Ru, PP = dppe or 2PPh₃; L1 = 5-[3-(thiophen-2-yl)benzo[c]th

Full text of DOI:10.1002/ejic.201300048

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
FULL PAPER
Nonlinear Optical Materials
T. J. L. Silva, P. J. Mendes,*
M. H. Garcia, M. Paula Robalo,
J. P. Prates Ramalho,
A. J. Palace Carvalho, M. Büchert,
C. Wittenburg, J. Heck ................... 1–13
η5-Monocyclopentadienyliron(II)/ruthe-
nium(II) complexes with 5-[3-(thiophen-2-
yl)benzo[c]thiophenyl]thiophene-2-carbo-
nitrile as ligand have been synthesized.
Quadratic hyperpolarizabilities in the range
of (105–164)ϫ10–30 esu were evaluated by
hyper-Rayleigh scattering at 1500 nm. DFT
calculations were performed to explain
their non-linear optical properties.
Benzo[c]thiophene Chromophores Linked
to Cationic Fe and Ru Derivatives for NLO
Materials: Synthesis Characterization and
Quadratic Hyperpolarizabilities
Keywords: Ruthenium
/ Iron / Sulfur
heterocycles / Nonlinear optics / Hyper-
polarizabilities
0000, 0–0
1
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
DOI:10.1002/ejic.201300048
Benzo[c]thiophene Chromophores Linked to Cationic Fe
and Ru Derivatives for NLO Materials: Synthesis
Characterization and Quadratic Hyperpolarizabilities
Tiago J. L. Silva,[a] Paulo J. Mendes,*[b] M. Helena Garcia,[a]
M. Paula Robalo,[c,d] J. P. Prates Ramalho,[b]
A. J. Palace Carvalho,[b] Marina Büchert,[e] Christian Wittenburg,[e]
and Jürgen Heck[e]
Keywords: Ruthenium / Iron / Sulfur heterocycles / Nonlinear optics / Hyperpolarizabilities
η5-Monocyclopentadienyliron(II)/ruthenium(II) complexes of
the general formula [M(η5-C5H5)(PP)(L1)][PF6] {M = Fe, PP =
dppe; M = Ru, PP = dppe or 2PPh3; L1 = 5-[3-(thiophen-2-yl)-
benzo[c]thiophenyl]thiophene-2-carbonitrile} have been
synthesized and studied to evaluate their molecular qua-
dratic hyperpolarizabilities. The compounds were fully char-
acterized by NMR, FTIR and UV/Vis spectroscopy and their
electrochemical behaviour studied by cyclic voltammetry.
Quadratic hyperpolarizabilities (β) were determined by hy-
per-Rayleigh scattering measurements at a fundamental
wavelength of 1500 nm. Density functional theory calcula-
tions were employed to rationalize the second-order non-lin-
ear optical properties of these complexes.
gand charge transfer (MLCT), ligand-to-metal charge
transfer (LMCT) or intraligand charge transfer (ILCT) ex-
citations. Among the organometallic compounds presenting
this structural feature, our group and others have carried
out systematic studies on η5-monocyclopentadienylmetal
complexes with benzene or oligothiophene conjugated
chains coordinated to the metal centres through nitrile or
acetylide linkages.[8,12–18] In particular, very efficient NLO
responses were found when strong electron donors such as
iron and ruthenium organometallic moieties were coupled
with strong electron acceptors like NO2.
Although fundamental research on NLO properties has
mostly been devoted to the preparation of compounds with
large optical non-linearities, the use of these properties in
molecular switching has attracted considerable inter-
est.[19–29] Organotransition-metal compounds revealed en-
couraging results because the presence of a redox-active
metal centre within a conjugated system provides excellent
opportunities for reversible modulation of the second-order
non-linear optical (SONLO) properties. Our on-going work
in this field was motivated by benzo[c]thiophene-based
chromophores, the unique electronic behaviour of which,
originating from their low HOMO–LUMO energy
gaps,[30,31] can potentially yield interesting NLO effects. A
soluble form of polyisothianaphthene was found to origi-
nate a large third-order non-linear optical response[32] and
a ferrocenylethenyl-thienyl-2-thienylbenzo[c]thiophene or-
ganometallic complex exhibits moderate quadratic hyper-
polarizabilities.[33] However, the structure of this metallo-
cene derivative, in which the metal is placed orthogonally
Introduction
The exploitation of organometallic chemistry for the syn-
thesis of compounds with non-linear optical (NLO) proper-
ties has been mainly motivated by their use in optical de-
vices.[1] During the last two decades, significant work has
been published outlining the most important progress in
the field.[2–11] According to the overall results, the general
understanding is that second-order non-linearities are
strongly related to asymmetric push–pull systems. In the
case of organometallic compounds, the metal centre can be
bound to a highly polarizable conjugated backbone, thereby
acting as an electron-releasing or -withdrawing group. This
type of structural feature leads to large quadratic hyperpo-
larizabilities arising from intense low-energy metal-to-li-
[a] Centro de Ciências Materiais e Moleculares, Faculdade de
Ciências da Universidade de Lisboa,
Campo Grande, 1749-016 Lisboa, Portugal
[b] Centro de Química de Évora,
Rua Romão Ramalho 59, 7002-554 Évora, Portugal
Fax: +351-266745303
E-mail: pjgm@uevora.pt
Homepage: www.cqe.uevora.pt/
[c] Centro de Química Estrutural, Instituto Superior Técnico,
Universidade Técnica de Lisboa,
Av. Rovisco Pais, 1049-001 Lisboa, Portugal
[d] Área Departamental de Engenharia Química, Instituto
Superior de Engenharia de Lisboa,
Rua Conselheiro Emídio Navarro 1, 1959-007 Lisboa, Portugal
[e] Institut für Anorganische und Angewandte Chemie, Universität
Hamburg,
Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201300048.
Eur. J. Inorg. Chem. 0000, 0–0
2
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
to the extended π system, does not favour better coupling measured by hyper-Rayleigh scattering (HRS) at the funda-
between the organometallic fragment and the conjugated mental wavelength of 1500 nm. DFT calculations were per-
chromophore to achieve high quadratic hyperpolarizabili- formed to gain an understanding of the second-order NLO
ties. Comparing the metallocene systems that have been mechanism in the titled complexes in solvated media with
studied for NLO purposes, the main structural feature of the objective of obtaining quantitatively satisfactory results
our η5-monocyclopentadienylmetal complexes is the pres- for both the calculated electronic excitations and the qua-
ence of the metal centre in the same plane as the π system, dratic hyperpolarizabilities in comparison with experimen-
this allowing better coupling between the organometallic tal data.
fragment and the conjugated chromophores. Recently we
performed DFT calculations to investigate the quadratic
hyperpolarizabilities of η5-monocyclopentadienyliron(II)
and -ruthenium(II) model complexes with 5-[3-(thiophen-2-
yl)benzo[c]thiophenyl]thiophene-2-carbonitrile as a ligand
to evaluate their potential applications as SONLO
Results and Discussion
Synthesis and Spectroscopic Studies
The procedure used for the synthesis of the ligand 5-[3-
switches.[34] Based on previous reports that higher hyper- (thiophen-2-yl)benzo[c]thiophenyl]thiophene-2-carbonitrile
polarizabilities can be achieved upon oxidation and/or re- (L1) is depicted in Scheme 1. The 1,3-di(thiophen-2-yl)-
duction, our motivation in that work was to study mo- benzo[c]thiophene skeleton (3) was synthesized according
lecules with architectural features different to the tradi- to the method of Kiebooms et al.[35] in three steps. First,
tional D–π–A system. We adopted a more likely less studied the reaction of phthaloyl dichloride with 2-mercaptopyr-
D–π–DЈ feature (“off” form), from which a D–π–AЈ struc- idine gave the S-(2-pyridinyl) thioester compound (2),
ture (“on” form) can be achieved upon oxidation and hence which was treated with the freshly prepared Grignard rea-
an expected increase in the corresponding quadratic hyper- gent thiophen-2-ylmagnesium bromide. The resulting 1,2-
polarizability. The results showed that these complexes phenylenebis(thiophen-2-ylmethanone) (3) was treated with
could be good candidates for SONLO switches acting Lawesson’s reagent to give the desired 1,3-di(thiophen-2-yl)-
through a redox mechanism because an 8.3-fold increase in benzo[c]thiophene (4) in good yields. The 1,3-di(thiophen-
the quadratic hyperpolarizability was predicted upon oxi- 2-yl)benzo[c]thiophene was further functionalized by the
dation.
Wielsmeyer–Haack reaction to afford the corresponding al-
In this work we report the synthesis of a series of η5- dehyde 5. The reaction of 5 with hydroxylammonium chlor-
monocyclopentadienyliron(II)/ruthenium(II) complexes of ide in pyridine followed by dehydration with acetic anhy-
the general formula [M(η5-C5H5)(PP)(L)][PF6] {M = Fe, PP dride afforded L1 in good yield. The ligand was fully char-
= dppe; M = Ru, PP = dppe or 2PPh3, L = 5-[3-(thiophen- acterized by FTIR and 1H and 13C NMR spectroscopy. The
2-yl)benzo[c]thiophenyl]thiophene-2-carbonitrile} to evalu- solid-state IR spectra (KBr pellets) showed the characteris-
ate their molecular quadratic hyperpolarizabilities. tic stretching vibration of the nitrile functional group at
Attempts to obtain the oxidized MIII complexes to study 2203 cm–1.
the corresponding quadratic hyperpolarizabilities for their
Complexes of general formula [M(η5-C5H5)(PP)-
potential use as SONLO switches operating by a redox (L1)][PF6] (M = Ru, PP = dppe, 2PPh3; M = Fe, PP =
mechanism were unsuccessful due to the low stability of the dppe) were prepared by halide abstraction with TlPF6 from
oxidized species. The quadratic hyperpolarizabilities were the parent neutral complexes [M(η5-C5H5)(PP)X] (M =
Scheme 1. Synthesis of the ligand L1.
Eur. J. Inorg. Chem. 0000, 0–0
3
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
Scheme 2. Synthesis of benzo[c]thiophene-derived Ru/Fe complexes and the numbering scheme for the NMR spectral assignments.
1
Table 1. Selected IR and H and 13C NMR data for the RuII/FeII organometallic complexes.[a]
ν(NϵC) [cm–1]
δH [ppm]
δC [ppm]
C-3
3-H
4-H
C-1
C-2
C-4
C-5
1Fe
2196
(–7)
2211
(+8)
2213
(+10)
6.88
(–1.05)
6.93
(–1.00)
7.67
7.30
(–0.30)
7.33
(–0.27)
7.56
128.38
(13.82)
131.48
(16.92)
125.68
(11.12)
108.01
(–0.48)
107.16
(–1.33)
107.47
(–1.02)
139.81
(–0.12)
140.25
(0.32)
141.09
(1.16)
125.38
(–1.06)
125.52
(–0.92)
125.99
(–0.45)
143.47
(0.13)
144.29
(0.95)
144.78
(1.44)
1aRu
1bRu
(–0.26)
(–0.04)
1
[a] Differences in the IR ν(NϵC) (νcomplex – νligand) and NMR H and 13C resonances (δcomplex – δligand) of the ligand upon coordination
to the organometallic fragments are shown in parentheses.
FeII, X = I; M = RuII, X = Cl) in dichloromethane at room nitrile ligand is mainly observed at the 3-H and 4-H protons
temperature in the presence of a sufficient excess of the cor- (see Scheme 2 for numbering), the resonances of which are
responding ligand (Scheme 2). The complexes are fairly shifted upfield upon coordination to the organometallic
stable towards oxidation in air and to moisture both in the moieties. The higher shielding effect observed for 1Fe is
solid state and in solution. The compounds were charac- consistent with the higher degree of π-back-donation for
terized by analytical data, FTIR and 1H, 13C and 31P NMR this compound deduced from the FTIR data. The substitu-
spectroscopy, which supported the proposed formulations. tion of the dppe co-ligand by 2PPh3 in the ruthenium com-
Relevant spectroscopic data are presented in Table 1.
plex results in reduced shielding of the 3-H and 4-H protons
Characteristic IR bands confirm the presence of the cy- upon coordination. The 13C NMR spectroscopic data con-
–
1
clopentadienyl co-ligand (ca. 3000–3100 cm–1), the PF6
firm the trends found in the H NMR spectra. The major
counter anion (838 and 557 cm–1) and the coordinated ni- changes in the ligand carbon resonances upon coordination
trile (2196–2213 cm–1) in all the complexes. Comparison of were observed in the thienyl-nitrile moiety (carbons C-1 to
1
ν(NϵC) upon coordination of the ligand reveals a negative C-5), in agreement with the behaviour observed in the H
shift of 7 cm–1 for 1Fe and a positive shift of 8 and 10 cm–1 NMR spectra. In particular, significant deshielding of the
for 1aRu and 1bRu, respectively. The negative shift can be NC carbon (C-1) of the ligand was observed upon coordi-
attributed to the enhancement of the π-back-donation from nation, as expected. The 31P NMR spectra show a single
the metal d orbitals to the π* orbital of the NC group, sharp signal for the phosphane co-ligands, which were de-
which leads to a decrease in the bond order of NϵC. The shielded upon coordination, as expected, in accord with the
magnitude of this π-back-donation is thus higher for the σ-donor character of these ligands. The resonances are in
iron(II) compound, as expected considering the better π- the range usually observed for monocationic ruthenium(II)
donating ability of the iron(II) moiety. The benzo[c]thio- and iron(II) complexes. The same general behaviour of the
phene derivatives described here show more effective π- NMRresonanceshasbeenobservedpreviouslyinrelatedoligo-
back-donation than those possessing terthiophene units thiophene iron(II)[14] and ruthenium(II),[13] and 1,2-di(2-
with an acceptor nitro end-group in related oligothiophene thienyl)ethene iron(II) and ruthenium(II)[12] complexes.
iron(II)[14] and ruthenium(II)[13] complexes. Moreover, less
The optical absorption spectra of the complexes at con-
effective π-back-donation was found in comparison with centrations of 1.0ϫ10–5 m were recorded in chloroform and
parent 1,2-di(2-thienyl)ethene iron(II) and ruthenium(II)[12] DMF. The spectrum of 1Fe exemplifies the behaviour of
complexes, which possess one vinylidene unit between the the compounds studied in this work (Figure 1) and the op-
two thiophene rings.
tical data are summarized in Table 2. Figure 1 also includes
1H NMR resonances for the cyclopentadienyl ring are in the spectrum of L1 for comparison.
the range usually observed for monocationic ruthenium(II)
The spectrum of L1 is characterized by one absorption
and iron(II) complexes. The effect of coordination of the band in the UV region and one strong and broad absorp-
Eur. J. Inorg. Chem. 0000, 0–0
4
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
The energy and shape of the lowest-energy bands of the
complexes resemble that of the uncoordinated ligand, which
seems to indicate that they have high ILCT character. Ac-
cording to our recent TD-DFT calculations on model iron
and ruthenium complexes with the formula [M(η5-
C5H5)(H2PCH2CH2PH2)(L1)]+ (similar to 1Fe and 1aRu)
in the gas phase,[34] this band can be assigned to charge
transfer mainly within the ligand fragment (ILCT). How-
ever, it is well known that solvation interactions are, in some
cases, critical for obtaining quantitatively satisfactory re-
sults of the calculated electronic excitations in comparison
with experimental data. Thus, for a deeper knowledge of
the relevant electronic transitions involved, which can be
helpful for further discussions on our experimentally deter-
mined first hyperpolarizabilities, we performed theoretical
studies on model complexes taking into account solvation
effects. The polarizable continuum model (PCM) was used
to simulate the interaction between the complexes with
chloroform. Also, we extended the theoretical studies to the
synthesized ruthenium complexes (1aRu and 1bRu) to eval-
uate the use of model phosphanes to describe the behaviour
found with real molecules. The ligand L1 was also included
in these studies to enable a comparison with the results
found for FeII/RuII complexes. Figure 2 shows the struc-
tures and atom labelling of the complexes used in our calcu-
lations.
Figure 1. Electronic spectra of 1Fe and the free ligand L1 (dashed
line) in chloroform.
Table 2. Optical spectral data for the synthesized complexes [M(η5-
C5H5)(PP)(L1)][PF6] in chloroform and DMF.
λ [nm] (ε [104 m–1 cm–1])
Δλ [nm]
CHCl3
DMF
1Fe
478 (3.0)
≈290
464 (2.3)
≈291
–14
–2
1aRu
1bRu
475 (3.0)
≈280
473 (1.8)
≈288
473 (2.7)
≈280
453 (2.2)
≈291
–20
tion band in the visible region. The lowest-energy band is
within the range found for benzo[c]thiophene deriva-
tives[36,37] and can be assigned to a π–π* electronic transi-
tion. The main feature of the spectra of the complexes re-
corded in CHCl3 is the presence of an intense broad band
in the range of 473–478 nm with a band at higher energy
(in the range of ca. 280–290 nm) depending on the organo-
metallic fragment. To study the solvatochromic behaviour
of the lowest-energy band, absorption spectra were also re-
corded in DMF. For all complexes a hypsochromic shift
was observed upon increasing the solvent polarity, particu-
larly for 1Fe and 1bRu (–14 and –20 nm, respectively). The
negative solvatochromic behaviour exhibited by these com-
pounds is characteristic of electronic transitions in which
there is a decrease in the dipole moment upon excitation.
A blueshift was also observed in the solvatochromic study
Figure 2. Structures and atom labeling of the complexes used in
the DFT calculations.
The optimized structures and selected structural data can
of the ferrocenylethenyl-thienyl-2-thienylbenzo[c]thiophene be found in the Supporting Information. The angles and
organometallic complex.[33]
bond lengths are consistent with experimental crystal data
Table 3. Optical data for the studied compounds obtained by TD-DFT calculations in chloroform.
[a]
λeg [nm]
f[b]
Attribution[c]
Character of the CT[d]
L1
431 (452)
447 (478)
486 (475)
477 (473)
465 (475)
475 (473)
0.613
0.852
0.964
0.936
0.952
1.008
HǞL (97%)
HǞL (93%)
HǞL (96%)
HǞL (96%)
HǞL (96%)
HǞL (96%)
BcT (55), T2 (45)ǞNC-T1 (100)
1ЈFe
BcT-T2 (93), Fe (4), Cp (3)ǞNC-T1 (96), P (4)
BcT-T2 (92), Ru (4), Cp (4)ǞNC-T1 (96), P (4)
BcT-T2 (95), Ru (5)ǞNC-T1 (100)
BcT-T2 (90), Ru (5), P (5)ǞNC-T1 (100)
BcT-T2 (75), Ru (19), Cp (6)ǞNC-T1 (100)
1aЈRu
1bЈRu
1aRu
1bRu
[a] Absorption wavelength of the main transitions (experimental values are shown in parentheses). [b] Oscillator strength. [c] H = HOMO,
L = LUMO. [d] Based on the represented molecule fragments (overall % of the charge transfer is given in parentheses). Cp = cyclopen-
tadienyl, NC = nitrile group, T = thiophene rings (T1 = thiophene close to NC; T2 = terminal thiophene), BcT = benzo[c]thiophene,
P = phosphane co-ligands.
Eur. J. Inorg. Chem. 0000, 0–0
5
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
for the parent iron(II)[12] and ruthenium(II)[13] thiophene sults show that the character of this excitation resembles
derivatives and other η5-monocyclopentadienylmetal com- that of the free ligand, thus having mainly ILCT character.
plexes.[15,38,39] After the geometry optimizations, the elec- The corresponding electron density difference maps
tronic spectra were simulated by means of TD-DFT calcu- (EDDMs) clearly illustrate the transfer of the electron den-
lations. Two bands are predicted, in accordance with our sity from the terminal thiophene ring (T2) and benzo-
experimental spectra (a simulated spectrum for 1ЈFe is given [c]thiophene group of the ligand to the NC-T1 moiety (Fig-
in the Supporting Information). The results show that the ure 4). However, there is also some contribution from
lowest-energy band for all complexes, which is the key to MLCT. An enhanced contribution of the MLCT character
second-order non-linear optical properties, arises mainly is predicted when the real phosphanes dppe and PPh3 were
from a single electronic transition (HOMOǞLUMO). The used in the calculations instead of H2P(CH2)2PH2 and PH3.
wavelengths, oscillator strengths and composition of this In particular, substitution of the model phosphane PH3
electronic transition in terms of the contributions of groups (1bЈRu) by PPh3 (1bRu) results in an increase in the contri-
of atoms involved are given in Table 3.
bution of the MLCT (from 5 to 25%) to the overall exci-
Very good results for the calculated values of λmax were tation.
obtained in comparison with the experimental data by
using either model or real phosphanes in the calculations.
In particular, the values of λmax calculated for the ruthe-
nium complexes (465 nm for 1aRu and 475 nm for 1bRu)
are very close to the experimental values (475 and 473 nm,
respectively). The analysis of the electronic transition in
terms of the contributions of groups of atoms involved al-
lowed us to predict reliable assignments of the band ob-
served in the experimental spectra. The orbitals involved in
the electronic transition are depicted in Figure 3. The re-
Figure 4. Electron density difference maps (EDDMs) of the
HOMO–LUMO transition for the complexes 1ЈFe, 1aЈRu, 1bЈRu,
1aRu and 1bRu (red indicates an increase in electron density and
blue indicates a decrease, isovalue = 0.002).
Electrochemical Studies
The redox behaviour of the compounds was studied by
cyclic voltammetry in dichloromethane and acetonitrile, be-
tween the limits imposed by the solvents, to evaluate the
electron richness at the active redox centres and the revers-
ibility of the oxidation/reduction processes. The study of the
reversibility of the redox processes gives an insight into the
stability of the oxidized and/or reduced species. This is par-
ticularly important when dealing with NLO redox switching
properties. In fact, to be used as molecular switches, both
“on” and “off” forms must be stable and the switching pro-
cess must be reversible. Thus, electrochemical studies are
helpful to screen molecules with potential as molecular
switches. As an example, the electrochemical response of
1Fe in dichloromethane is shown in Figure 5 and the most
important data exhibited by all the complexes and the free
ligand in dichloromethane and acetonitrile are summarized
in Table 4.
Figure 3. Representation of the main orbitals involved in the elec-
tronic transitions in the complexes 1ЈFe, 1aЈRu, 1bЈRu, 1aRu and
1bRu.
Eur. J. Inorg. Chem. 0000, 0–0
6
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
5-(2-thiophen-2-ylvinyl)thiophene-2-carbonitrile[12]
and
2,2Ј:5Ј,2ЈЈ-terthiophene-5-carbonitrile[40,41]
compounds.
The main electrochemical behaviour of the complexes is
characterized by the presence of an oxidative process attrib-
uted to the MII/MIII redox pair in addition to the processes
at the positive and negative potentials already attributed to
the coordinated ligand L1. The MII/MIII redox process is
reversible for the iron complex 1Fe in both solvents (E1/2
=
0.73 V in dichloromethane, E1/2 = 0.68 V in acetonitrile).
However, for the ruthenium complexes, this redox process
was found to be irreversible (Epa = 1.28 V for 1aRu and Epa
= 1.29 V for 1bRu in dichloromethane; Epa ≈ 1.09 V for
1aRu and Epa = 1.23 V for 1bRu in acetonitrile). The poten-
tial values of the ligand-related oxidative and reductive pro-
cesses remained almost unchanged after coordination to the
metal centres in both solvents. For the ruthenium com-
plexes, all the processes are irreversible, and the same be-
haviour was found in dichloromethane for 1Fe, but the first
Figure 5. Cyclic voltammogram of complex 1Fe in CH2Cl2 at
200 mV/s showing the reversibility of the isolated oxidative pro-
cesses: (···) FeII/FeIII, (---) both L1/L1ox and FeII/FeIII
.
one becomes quasi-reversible in acetonitrile (E1/2
=
The electrochemistry of the free ligand L1 is charac-
terized by an irreversible oxidative process (Epa = 1.00 V in
dichloromethane, Epa = 0.98 V acetonitrile) and two re-
ductive processes at negative potentials. The reductive pro-
cesses are irreversible in dichloromethane, but the first one
becomes quasi-reversible in acetonitrile (E1/2 = –1.36 V).
Similar electrochemical behaviour was found for the related
–1.38 V). In general, the organometallic complexes exhibit
quite similar electrochemical behaviour to that observed for
iron(II) and ruthenium(II) complexes bearing the related
substituted thiophene-2-carbonitrile ligands.[12,13] In view of
their potential use as second-order non-linear optical
switches operating through redox mechanisms, and accord-
Table 4. Electrochemical data for complexes [M(η5-C5H5)(PP)(L1)]PF6] and free L1 in CH2Cl2 and MeCN.
Epa [V]
Epc [V]
E1/2 [V]
Epa – Epc [mV]
ipc/ipa
Dichloromethane
L1
1.00
–
–
1.07
0.77
–
–
1.28
0.98
–
–
–
–
1.03
0.73
–
–
–
–
–
–
–
80
90
–
–
–
–
–
–
–
0.9
0.9
–
–
–
–
–1.38
–1.53
0.99
0.68
–1.37
–1.53
–
[Fe(η5-C5H5)(dppe)(L1)][PF6] (1Fe)
[Ru(η5-C5H5)(dppe)(L1)][PF6] (1aRu)
[Ru(η5-C5H5)(PPh3)2(L1)][PF6] (1bRu)
–
–1.36
–1.53
–
1.29
1.01
–
–
–
–
–
–
–
–
–
–
–
–
–
–1.38
–1.55
Acetonitrile
L1
1.13
0.98
–1.32
–
–
–
–
–
70
–
0.89
–1.39
–1.71
–
0.4
–1.36
0.8[a]
[Fe(η5-C5H5)(dppe)(L1)][PF6] (1Fe)
1.17
1.09
–
–
–
–
–
–
–
0.72
–1.34
0.64
–1.42
–1.72
–
–1.38
–1.70
–
0.68
–1.38
90
80
1.0
0.7[a]
[Ru(η5-C5H5)(dppe)(L1)][PF6] (1aRu)
[Ru(η5-C5H5)(PPh3)2(L1)][PF6] (1bRu)
1.09[b]
–
–
1.23
1.10
0.97
–
–
–
–
–
–
–
–
–
–
–
–1.39
–1.70
[a] ipc/ipa values. [b] Very broad wave that incorporates three oxidative processes.
Eur. J. Inorg. Chem. 0000, 0–0
7
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
ing to the electrochemical behaviour of the complexes, it field, and only one component of the β tensor dominates
seems that only the iron complex may be suitable for this the NLO response (i.e., a unidirectional charge-transfer
purpose owing to the reversibility of the FeII/FeIII couple. transition). This model establishes a link between the mo-
However, attempts to synthesize and isolate the oxidized lecular hyperpolarizability and a low-lying energy charge-
FeIII complex have so far been unsuccessful due to its low transfer transition expressed by Equation (1)
stability. Further studies seeking to modify the molecular
structure to obtain stable oxidized forms are currently in
progress. Because it was not possible to achieve the oxidized
MIII forms of the complexes, it will not be possible to evalu-
(1)
ate the molecular switching based on the second-order
NLO properties by redox means. Thus, only the quadratic in which Δμeg is the difference between the dipole moments
hyperpolarizabilities of the MII complexes were studied.
of the ground (g) and excited (e) states, feg is the oscillator
strength and Eeg is the transition energy. An optimal combi-
nation of these factors will provide higher β values. The
DFT calculations of the molecular hyperpolarizabilities
show that all complexes are dominated by the βzzz tensor
component (along the charge-transfer axis) with a depolar-
ization ratio (ρ) of around 1:5 [see Eq. (6) in the Exp. Sect.
and the Supporting Information), which is expected for a
prototypical 1D dipolar molecule.[44] In addition, TD-DFT
calculations predicted a single dominant electronic transi-
tion in the visible region (HOMOǞLUMO) for these com-
plexes (Table 3). Thus, an analysis of the results according
to the TLM seems to be reasonable for these complexes.
The experimental quadratic hyperpolarizabilities are rela-
tively low, as expected considering the molecular nature of
these complexes, which can be viewed as having a structure
of the type D–π–DЈ in which the well-known electron-do-
nating character of the organometallic fragments is com-
bined with the thiophene-benzo[c]thiophene moiety that be-
haves mainly as a donor during the lowest-energy transition
(see above). The results show some resonant enhancement,
although the relative orderings were maintained with the
TLM corrected values: 1Fe Ͼ1aRuϾ1bRu. In spite of the
relatively similar magnitudes of the experimental quadratic
hyperpolarizabilities for all the complexes (the values for
1Fe and 1aRu are within the experimental error), it is inter-
esting to note that experimental quadratic hyperpolarizabil-
ities follow the same trend as observed for the relative π-
back-donation of the complexes deduced from the FTIR
and NMR spectroscopic data. The existence of this interac-
tion was found to play an important role in the second-
order NLO response of the η5-monocyclopentadienyliron,
-ruthenium, -nickel and -cobalt complexes with para-substi-
tuted benzonitrile chromophores, with the high values
found for the RuII and FeII complexes being attributed to π-
back-donation.[18] The fact that the experimental quadratic
hyperpolarizabilities for all the complexes are relatively sim-
ilar is not surprising considering the spectroscopic data dis-
cussed above. In fact, for the low-energy band, which is the
key to second-order non-linear optical properties, λmax is
very similar for all the complexes and no significant solva-
tochromic effect was observed. In addition, charge transfer
was predicted by theoretical calculations to be of the same
Quadratic Hyperpolarizabilities
Because the complexes exhibit optical transitions in the
visible range, close to the second harmonic at 532 nm, high-
resonance enhancements could be expected for β measure-
ments at 1064 nm on the basis of the two-level model
(TLM) of Oudar and Chemla.[42] Thus, experimental mo-
lecular quadratic hyperpolarizabilities (β) were measured by
hyper-Rayleigh scattering (HRS) at 1500 nm to avoid the
superposition of absorptions in the UV/Vis region and the
second-harmonic signal (750 nm), where more reasonable
results can be expected from the TLM analysis. The static
hyperpolarizabilities of 1aRu, 1bRu and the model com-
plexes 1ЈFe, 1aЈRu and 1bЈRu were computed by using DFT
calculations to give an insight into the applicability of the
TLM to the rationalization of the experimental NLO prop-
erties of the complexes studied. The results obtained are
summarized in Table 5 together with relevant spectroscopic
data.
Table 5. Quadratic hyperpolarizabilities and relevant spectroscopic
data for the complexes [M(η5-C5H5)(PP)(L1)][PF6].[a]
[b]
Δν(NC)
λexp [nm]
βHRS [10–30 esu]
[cm–1]
(ε [104 m–1 cm–1])
Exp.
TLM[c] DFT[d]
1Fe
–7
+8
478 (3.0)
475 (3.0)
164
147
87
80
20.06[e]
24.73
1aRu
24.72[f]
26.36
1bRu
+10
473 (2.7)
105
57
29.43[g]
[a] All measurements were performed in CHCl3 solution. The β
values were measured at 1500 nm (experimental error: 15%). [b]
Differences in the IR ν(NϵC) band upon ligand coordination:
νcomplex –νligand. [c] β corrected for resonance enhancement by using
the two-level model with βTLM = βexp[1–(2λmax/λHRS)/2][1–(λmax
/
λHRS)/2] (damping factors not included). [d] βHRS calculated by
DFT using Equations (3)–(5) (the orientational averages and β ten-
sor components are shown in the Supporting Information). The B
convention[43] was used for direct comparison with experimental
data. [e] Model complex 1ЈFe. [f] Model complex 1aЈRu. [g] Model
complex 1bЈRu.
It is well known that the two-level model (TLM) is a nature, mainly within the ligand (ILCT) with some MLCT
good approximation for estimating quadratic hyperpolariz- contribution (see above). The use of dppe and PPh3 instead
abilities in cases in which only one excited state is coupled of the models H2P(CH2)2PH2 and PH3 in the DFT calcula-
strongly enough to the ground state by the applied electric tions, however, can give an additional insight into the differ-
Eur. J. Inorg. Chem. 0000, 0–0
8
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
ences in the experimental quadratic hyperpolarizabilities of phosphanes, with the obvious gain in computational cost,
the ruthenium complexes. As discussed above, DFT calcula- proves to be a good compromise between the accuracy of
tions on 1bRu predicted an enhanced contribution of the the predicted excitation energies and computational effort.
MLCT to the overall excitation (ca. 25%). For 1aRu, the However, the results show that the use of real phosphanes
contribution of the MLCT is lower (ca. 10%). The ILCT in DFT calculations can be important to provide an ad-
arises from charge transfer from the terminal thienyl ring ditional understanding of the experimental values of the
and the benzo[c]thiophene moiety to the thienyl-nitrile unit quadratic hyperpolarizabilities. In comparison with the ex-
whereas the MLCT occurs from the organometallic moiety perimental TLM corrected data, DFT calculations under-
also to the thienyl-nitrile unit. Thus, ILCT and MLCT have estimate to some extent the βHRS values of the studied com-
opposite directions of charge transfer. This might result in plexes.
a lower charge-transfer efficiency during the electronic exci-
tation for 1bRu (which has a large MLCT contribution)
and an expected lowering effect on the experimental qua-
Experimental Section
dratic hyperpolarizability.
General Procedures: Syntheses were carried out under nitrogen by
using current Schlenk techniques and the solvents were dried by
standard methods.[46] Commercial reagents were used without fur-
ther purification. Organometallic starting materials were prepared
following methods described in the literature: [Fe(η5-C5H5)-
(dppe)I],[14] [Ru(η5-C5H5)(PPh3)2Cl] and [Ru(η5-C5H5)(dppe)-
Cl].[47] Compounds 1–3 were prepared following published pro-
cedures[35] with some modifications (see details below). 5-(3-Thi-
ophen-2-yl-benzo[c]thiophenyl)thiophene-2-carbaldehyde (5) was
prepared following a published procedure[48] but with a different
purification method (see details below). FTIR spectra were re-
corded with a Mattson Satellite FTIR spectrophotometer in dry
KBr pellets (only significant bands are cited). 1H, 13C and 31P
NMR spectra were recorded with a Bruker Avance 400 spectrome-
ter at probe temperature using CDCl3 or (CD3)2CO as solvent. The
1H and 13C chemical shifts are reported in parts per million (ppm)
downfield from internal Me4Si and the 31P chemical shifts are re-
ported in ppm downfield from 85% H3PO4 as external standard.
Coupling constants are reported in Hz and spectral assignments
follow the numbering scheme shown in Scheme 2 and were assigned
by using HMBC, HMQC and/or COSY NMR techniques. UV/Vis
spectra were recorded with a Jasco V-560 spectrophotometer in the
range of 200–900 nm in dried solvents. Elemental analyses were
obtained at our laboratories (Laboratório de Análises, Instituto Su-
perior Técnico) by using a Fisons Instruments EA1108 system.
Data acquisition, integration and handling were performed by
using a PC with the EAGER-200 software package (Carlo–Erba
Instruments).
DFT calculations predicted low βHRS values for the
model complexes, in the range of (20.06–29.43)ϫ10–30 esu.
The substitution of the model phosphanes by dppe and
PPh3 originates small differences (up to ca. 10%) in the cor-
responding calculated βHRS. The results show some undere-
stimation of the DFT-calculated βHRS values when com-
pared with the experimental two-level model corrected val-
ues for 1Fe, 1aRu and 1bRu. Different reasons for these
discrepancies can be conjectured. For example, the use of
averaged configurations of the complexes for the calcula-
tions instead of the minimum energy alone and the in-
clusion of solvated media with a more detailed model than
PCM could probably lead to more realistic results. These
issues have been discussed, for example, by Jensen and
van Duijnen for the case of the widely studied p-nitroanil-
ine[45] and further studies will be carried out taking them
into account.
Conclusions
η5-Monocyclopentadienyliron(II)/ruthenium(II)
com-
plexes bearing the 5-[3-(thiophen-2-yl)benzo[c]thiophenyl]-
thiophene-2-carbonitrile ligand have been synthesized and
fully characterized. Quadratic hyperpolarizabilities were de-
termined by HRS measurements at 1500 nm, showing reso-
nant values in the range (105–164)ϫ10–30 esu and corrected
values, by application of the two-level-model, in the range
(57–87)ϫ10–30 esu. Time-dependent density functional
theory calculations were employed to gain a deeper knowl-
edge of the relevant electronic transitions involved, which
was very helpful in the discussion on the experimental qua-
dratic hyperpolarizabilities. These properties were also com-
puted by using DFT. These results, together with TD-DFT
data, showed that TLM could be used to rationalize the
experimentally observed NLO properties of the studied
complexes. The observed quadratic hyperpolarizabilities,
similar for all complexes, can be explained by the very sim-
ilar energies and intensities of the lowest-energy electronic
transition, which are shown to have mainly ILCT character.
The use of real and model phosphanes to simulate the elec-
tronic spectra of the complexes by DFT calculations shows
no significant differences between the calculated energies
Synthesis
S,S-Di(pyridin-2-yl)benzene-1,2-bis(carbothioate) (2): Triethylamine
(5 mL) and 2-mercaptopyridine (3.3 g, 30 mmol) were added to
dried thf (50 mL). After stirring for 10 min, full dissolution of the
2-mercaptopyridine was observed, the mixture was placed in an ice-
cold bath and stirred at 0 °C for a further 15 min. A solution of
phthaloyl chloride (2.2 mL, 15 mmol) in dried thf (50 mL) was
added in one portion and with vigorous stirring. The reaction was
quenched immediately with 2% aq. HCl solution (200 mL) and ex-
tracted with chloroform. The combined organic fractions were
washed with 10% aq. NaOH and water until neutral, dried with
anhydrous MgSO4 and the solvent was evaporated in vacuo.
Crystallization from CHCl3/diethyl ether afforded the desired prod-
uct as white needles in 80% yield. 1H NMR (400 MHz, CDCl3,
25 °C): δ = 7.24 (dt, 2 H, Hpy), 7.59 (m, 2 H, Hbenz), 7.71 (m, 4 H,
Hpy), 7.82 (m, 2 H, Hbenz), 8.57 (d, 2 H, Hpy) ppm.
1,2-Phenylenebis(thiophen-2-ylmethanone) (3): 2-Bromothiophene
(4.5 mL, 46 mmol) was added to a Schlenk flask with thf (100 mL)
containing activated magnesium turnings (1.2 g, 46 mmol) and a
and the experimental data. Therefore the use of model catalytic amount of iodine. The mixture was heated at a gentle re-
Eur. J. Inorg. Chem. 0000, 0–0
9
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
flux for 3 h until complete reaction of the magnesium and then left
to cool at room temperature. The Grignard solution was added
dropwise through a Teflon cannula to a stirred solution of 2 (7.95 g,
23 mmol) in thf (150 mL) at 0 °C. The resulting mixture was stirred
for an additional 30 min and quenched with 10% aq. HCl (200 mL)
solution. After extraction with several portions of chloroform, the
combined organic extracts were washed with 10% aq. NaOH, fol-
lowed by aq. NaHCO3, and finally water. After drying with anhy-
drous MgSO4, the solvent was removed in vacuo. Crystallization
from CHCl3/petroleum ether (40–60) afforded a brownish orange
solid in 90% yield. 1H NMR [400 MHz, (CD3)2CO, 25 °C]: δ =
7.19 (dt, 2 H, Hth), 7.59 (dd, 2 H, Hth), 7.78 (m, 2 H, Hbenz), 7.83
(m, 2 H, Hbenz), 7.97 (dd, 2 H, Hth) ppm.
(C-6), 126.33 (C-11), 126.44 (C-4), 127.66 (C-17), 127.74 (C-8),
128.07 (C-15), 129.35 (C-16), 130.57 (C-13), 135.10 (C-14), 136.09
(C-12), 137.13 (C-7), 139.93 (C-3), 143.34 (C-5) ppm.
UV/Vis (CHCl3): λmax (ε)
=
452 nm (2.1ϫ104 m–1 cm–1).
C17H9NS3·0.1CH2Cl2 (323.46): calcd. C 61.87, H 2.79, N 4.22, S
28.98; found C 61.81, H 3.01, N 3.96, S 27.75.
General Procedure for the Synthesis of the [M(η5-C5H5)(LL)L1]PF6
Complexes: The complexes were synthesized by halogen abstraction
from the organometallic precursors in a Schlenk tube. A mixture
of [M(η5-C5H5)(LL)X] (0.5 mmol), TlPF6 (1.5 equiv.) and L1
(1.1 equiv.) in freshly dried dichloromethane (30 mL) was heated at
reflux overnight under nitrogen. After filtration to remove the TlX
formed during the reaction, the solvent was evaporated under vac-
uum and the solid residues were recrystallized twice from dichloro-
methane/hexane to afford the desired compounds as reddish solids.
1,3-Di(thiophen-2-yl)benzo[c]thiophene (4): Lawesson’s reagent
(8.13 g, 20.1 mmol) was added in two portions to a solution of
3 (6.0 g, 20.1 mmol) in dichloromethane (100 mL). After stirring
overnight, the solvent was removed under vacuum and ethanol
(100 mL) was added. This solution was stirred with heating for
30 min. The solvent was removed and the brown residue was added
to dichloromethane, washed with 10% aq. NaOH and several times
with water. After drying with anhydrous MgSO4 followed by fil-
tration, the solvent was evaporated in vacuo and the resultant
brownish orange solid was purified by chromatography on silica
gel. Elution with petroleum ether (40–60) afforded the desired
product as orange crystals in 55% yield. 1H NMR (400 MHz,
(CD3)2CO, 25 °C): δ = 7.26 (m, 2 H, Hth), 7.51 (dd, 2 H, Hth), 7.64
(dd, 2 H, Hth), 8.00 (m, 2 H, Hbenz) ppm.
[Fe(η5-C5H5)(dppe)L1]PF6 (1Fe): Dark red, yield 360 mg, 73%. IR
(KBr): ν = 2196 (NϵC), 838 and 557 (PF ) cm–1. 1H NMR
–
˜
6
[400 MHz, (CD3)2CO, 25 °C]: δ = 2.67 (m, 2 H, CH2), 2.84 (m, 2
3
H, CH2), 4.71 (s, 5 H, η5-C5H5), 6.88 (d, JH,H = 4.1 Hz, 1 H, 3-
3
H), 7.28 (dt, 1 H, 16-H), 7.30 (d, JH,H = 4.1 Hz, 1 H, 4-H), 7.34
(dt, 1 H, 8-H), 7.43 (dt, 1 H, 11-H), 7.55 (dt, 1 H, 17-H), 7.57 (m,
8 H, Ho-Ph), 7.68 (m, 8 H, Hm-Ph), 7.71 (dt, 1 H, 15-H), 7.81 (d,
3
3JH,H = 8.9 Hz, 1 H, 10-H), 8.04 (d, JH,H = 8.9 Hz, 1 H, 9-H),
8.12 (m, 4 H, Hp-Ph) ppm. 13C NMR [100 MHz, (CD3)2CO, 25 °C]:
1
δ = 28.51 (t, JC,P = 42.0 Hz, CH2-dppe), 81.12 (η5-C5H5), 108.01
(C-2), 121.56 (C-10), 122.83 (C-9), 123.77 (C-6), 125.38 (C-4),
126.52 (C-11), 127.89 (C-15), 127.92 (C-8), 128.31 (C-17), 128.38
(C-1), 129.43 (C-16), 130.06 and 130.27 (t, JC,P = 10.0 Hz, Cm-Ph),
131.04 (C-13), 131.41 and 131.85 (s, Cp-Ph), 132.41 and 134.03 (t,
JC,P = 10.0 Hz, Co-Ph), 134.92 (C-14), 136.07 (C-12), 137.03 (C-7),
137.61 (t, JC,P = 41.0 Hz, Ci-Ph), 139.81 (C-3), 143.47 (C-5) ppm.
31P NMR [161 MHz, (CD3)2CO, 25 °C]: δ = 96.98 (s, dppe),
5-[3-(Thiophen-2-yl)benzo[c]thiophenyl]thiophene-2-carbaldehyde
(5): A mixture of phosphorus oxychloride (2.42 mL, 26 mmol) and
DMF (2 mL, 26 mmol) in dichloromethane (10 mL) was added to
a solution of 4 (1.55 g, 5.2 mmol) in dichloromethane (50 mL) at
room temperature. After stirring overnight, a saturated solution of
sodium acetate (100 mL) was added and the mixture was stirred
for 1 h to complete the hydrolysis. The suspension formed was ex-
tracted with dichloromethane. The combined organic layers were
washed with water, dried with MgSO4, and the solvent removed in
vacuo. The red solid obtained was purified by chromatography on
silica gel. Elution with petroleum ether (40–60)/diethyl ether (7:3
to 1:1) afforded the desired product as red crystals in 80% yield.
1H NMR [400 MHz, (CD3)2CO, 25 °C]: δ = 7.26 (m, 2 H, 2-H),
7.51 (dd, 2 H, 3-H), 7.64 (dd, 2 H, 1-H), 8.00 (m, 2 H, 4-H, 5-
H) ppm.
–
–144.25 (sept, PF6 ) ppm. UV/Vis. (CHCl3): λmax (ε) = 478 nm
(3.0ϫ104 m–1 cm–1). C48H38F6FeNP3S3·1/3CH2Cl2 (987.77): calcd.
C 57.13, H 3.84, N 1.38, S 9.47; found C 57.25, H 3.66, N 1.42, S
9.34.
[Ru(η5-C5H5)(dppe)L1]PF6 (1aRu): Light red, yield 284 mg, 55%
yield. IR (KBr): ν = 2211 (NϵC), 838 and 557 (PF ) cm–1. 1H
–
˜
6
NMR (400 MHz, (CD3)2CO, 25 °C): δ = 1.89 (m, 2 H, CH2), 2.20
3
(m, 2 H, CH2), 5.06 (s, 5 H, η5-C5H5), 6.93 (d, JH,H = 4.0 Hz, 1
3
H, 3-H), 7.28 (dt, 1 H, 16-H), 7.33 (d, JH,H = 4.0 Hz, 1 H, 4-H),
7.43 (dt, 1 H, 8-H), 7.50 (1 H, 11-H), 7.53 (m, 8 H, Ho-Ph), 7.56
(dt, 1 H, 17-H), 7.66 (m, 8 H, Hm-Ph), 7.73 (dt, 1 H, 15-H), 7.83
5-[3-(Thiophen-2-yl)benzo[c]thiophenyl]thiophene-2-carbonitrile
(L1): A solution of H2NOHHCl (2.13 g, 30.7 mmol) in pyridine
(10 mL) was added to a stirred solution of 5 (2.0 g, 6.13 mmol) in
pyridine (30 mL) cooled to 0 °C. The reaction was allowed to warm
to room temperature. After stirring for 1 h, acetic anhydride (5 mL)
was added dropwise and the mixture was heated at reflux for 2 h.
After cooling to room temperature, the mixture was poured into a
mixture of ice and 10% aq. HCl. The resulting orange precipitate
was dissolved in dichloromethane, washed with water and dried
with MgSO4. The solvent was removed in vacuo and the crude
product was purified by column chromatography on silica gel. Elu-
tion with a solvent gradient from petroleum ether (40–60)/diethyl
3
(d, JH,H = 8.0 Hz, 1 H, 10-H), 8.05 (1 H, 9-H), 8.06 (m, 8 H,
H
p-Ph) ppm. 13C NMR [100 MHz, (CD3)2CO, 25 °C]: δ = 28.53 (t,
1JC,P = 46 Hz, CH2-dppe), 83.53 (η5-C5H5), 107.16 (C-2), 121.54
(C-10), 122.86 (C-9), 123.63 (C-6), 125.52 (C-4), 126.54 (C-11),
127.95 (C-15), 128.01 (C-8), 128.38 (C-17), 129.45 (C-16), 129.88
and 130.11 (t, JC,P = 9.00 Hz, Cm-Ph), 131.30 and 131.80 (s, Cp-Ph),
131.48 (C-1), 131.98 and 134.28 (t, JC,P = 22.0 Hz, Co-Ph), 134.90
(C-14), 136.10 (C-12), 137.12 (C-7), 140.25 (C-3), 144.29 (C-
5) ppm ppm. 31P NMR [161 MHz, (CD3)2CO, 25 °C]: δ = 79.07 (s,
–
dppe), –144.26 (sept, PF6 ) ppm. UV/Vis. (CHCl3): λmax (ε) = 475
(3.0ϫ104 m–1 cm–1). C48H38F6NP3RuS3·2CH2Cl2 (1033.00): calcd.
C 49.93, H 3.52, N 1.16, S 8.00; found C 50.40, H 3.60, N 1.20, S
8.00.
ether, 7:3 to 1:1, afforded the desired product as an orange solid in
1
72% yield. IR (KBr): ν = 2203 (NϵC) cm–1. H NMR [400 MHz,
˜
(CD3)2CO, 25 °C]: δ = 7.27 (dt, 1 H, 16-H), 7.30 (dt, 1 H, 11-H),
[Ru(η5-C5H5)(PPh3)2L1]PF6 (1bRu): Dark red, yield 361 mg, 62%.
3
–
7.37 (dt, 1 H, 8-H), 7.57 [dd, JH,H = 3.7, 1.0 Hz, 1 H, 17-H], 7.60
IR (KBr): ν = 2213 (NϵC), 838 and 557 (PF ) cm–1. 1H NMR
˜
6
3
3
(d, JH,H = 4.1 Hz, 1 H, 4-H), 7.71 (dd, JH,H = 5.3, 1.0 Hz, 1 H, [400 MHz, (CD3)2CO, 25 °C]: δ = 4.79 (s, 5 H, η5-C5H5), 7.27 (m,
3
3
15-H), 7.93 (d, JH,H = 4.1 Hz, 1 H, 3-H), 8.03 (t, JH,H = 9.0 Hz, 12 H, Ho-Ph), 7.29 (1 H, 16-H), 7.34 (dt, 1 H, 8-H), 7.39 (m, 12 H,
1 H, 9-H, 10-H) ppm. 13C NMR [100 MHz, (CD3)2CO, 25 °C]: δ
Hm-Ph), 7.44 (dt, 1 H, 11-H), 7.49 (m, 6 H, Hp-Ph), 7.56 (d, JH,H =
3
3
= 108.49 (C-2), 114.56 (C-1), 121.76 (C-10), 122.63 (C-9), 123.87 4.1 Hz, 1 H, 4-H), 7.59 (dd, JH,H = 4.0, 0.9 Hz, 1 H, 17-H), 7.67
Eur. J. Inorg. Chem. 0000, 0–0
10
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
3
3
(d, JH,H = 4.1 Hz, 1 H, 3-H), 7.73 (dd, JH,H = 4.0, 0.9 Hz, 1 H, tered signal, the orientational average over β is given by Equa-
15-H), 7.94 (d, 3JH,H = 8.0 Hz, 1 H, 10-H), 8.07 (d, 3JH,H = 8.0 Hz,
1 H, 9-H) ppm. 13C NMR [100 MHz, (CD3)2CO, 25 °C]: δ = 85.31
(η5-C5H5), 107.47 (C-2), 121.62 (C-10), 122.88 (C-9), 123.67 (C-6),
125.68 (C-1), 125.99 (C-4), 126.58 (C-11), 128.00 (C-15), 128.09 (C-
17), 128.41 (C-8), 129.44 (t, JCP = 9.0 Hz, Cm-Ph), 131.14 (s,
tion (3).[44]
(2)
(3)
C
p-Ph), 131.49 (C-13), 134.29 (t, JC,P = 11.0 Hz, Co-Ph), 134.91 (C-
14), 136.13 (C-12), 136.48 (t, JC,P = 44.0 Hz, Ci-Ph), 137.27 (C-7),
141.09 (C-3), 144.78 (C-5) ppm. 31P NMR [161 MHz, (CD3)2CO,
–
25 °C]: δ = 41.35 (s, dppe), –144.27 (sept, PF6 ) ppm. UV/Vis
(CHCl3): λmax (ε) = 473 (2.7ϫ104 m–1 cm–1). C58H44F6NP3RuS3·1/
4CH2Cl2 (1159.15): calcd. C 59.27, H 3.80, N 1.19, S 8.15; found
C 59.27, H 4.12, N 1.19, S 7.82.
The macroscopic averages in Equation (3) can be written as a func-
tion of the β-tensor components in the molecular frame (without
assuming Kleinman’s symmetry conditions) according to the
method developed by Bersohn et al.[53] For transparent and weak
optically dispersive materials, the Kleinman symmetry condi-
tions[54] can be applied to the β-tensor components. Thus, the ori-
entational averages involved in Equation (3) can be simplified to
Equations (4) and (5).[55]
Electrochemical Studies: Electrochemical experiments were per-
formed with an EG&G Princeton Applied Research Potentiostat/
Galvanostat Model 273A instrument equipped with Electrochemi-
cal PowerSuite v2.51 software for electrochemical analysis. Cyclic
voltammograms were recorded in CH2Cl2 and CH3CN with
[Bu4N][PF6] (0.1 m) as the supporting electrolyte. The electrochemi-
cal cell was a three-electrode configuration cell with a platinum disc
working electrode (1.0 mm) probed by a Luggin capillary con-
nected to a silver wire pseudo-reference electrode and a platinum
wire auxiliary electrode. All the experiments were performed under
argon at room temperature. All the reported potentials were mea-
sured against the ferrocene/ferrocenium redox couple as internal
standard and are normally quoted relative to SCE (using the ferro-
cenium/ferrocene redox couple E1/2 = 0.46 or 0.40 V vs. SCE for
dichloromethane or acetonitrile, respectively[49]). The electrochemi-
cal-grade electrolyte was purchased from Aldrich Chemical Co. and
dried under vacuum at 110 °C for 24 h. Reagent-grade solvents
were dried, purified by standard procedures,[46] and distilled under
nitrogen before use.
(4)
(5)
The associated depolarization ratio (ρ) is defined by Equation (6).
HRS Measurements of the First Hyperpolarizabilities: The first hy-
perpolarizabilities (β) were measured by using the harmonic light
scattering technique (also named hyper-Rayleigh) in chloroform.
The 10–3–10–5 m solutions of the complexes were placed in a 4 cm
long fluorimetric cell after being carefully filtered through a
0.2 mm filter to eliminate the white light noise resulting from
microburning any of the remaining dust particles by the incoming
laser beam. The measurements were performed at a fundamental
wavelength of 1500 nm as described by Stadler et al.[50] using a Q-
switched Nd:YAG laser operating in the 10 Hz repetition range.
The scattered second harmonic signal was collected at 90° with
respect to the direction of the incoming laser beam. The molecular
hyperpolarizabilities were derived from the HRS data by using the
external reference method. Disperse red 1 (DR1) dissolved in chlo-
roform was used as the external standard. The reference hyperpo-
larizability (β) of DR1 in CHCl3 was measured by comparison of
the slopes of I2w vs. the concentration plot of the standard in
CH2Cl2 and CHCl3.[44] By using the hyperpolarizability of DR1 in
dichloromethane [βDR1 (CH2Cl2) = 70ϫ10–30 esu],[51] the hyperpo-
(6)
As an example, for a prototypical one-dimensional extended di-
polar molecule with the single major diagonal tensor component
βzzz, the depolarization ratio is 1:5.[44]
DFT Calculations: All calculations were performed at the DFT
level of theory by using the Gaussian 09 package.[56] The solvation
effects were simulated by using the polarizable continuum model
(PCM) as implemented in Gaussian 09. The hybrid functional
CAM-B3LYP[57] (Coulomb-Attenuating Method applied to
B3LYP) was used for the calculations. As a compromise between
accuracy and computational effort we adopted the 6–31G* (for
geometry optimizations) and the 6–31+G* basis sets (for calcula-
tion of the hyperpolarizabilities) for C, H, N, O and H, and the
LANL2DZ effective core potential basis set for S, P, Fe and
Ru.[58,59] In the case of the calculations of the hyperpolarizabilities,
the LANL2DZ basis set was also augmented with a polarization
larizability of DR1 in CHCl3 was estimated to be 80ϫ10–30 esu, function (exponents of 0.496 and 0.364) and a diffuse function (ex-
which is very close to the published value of 74ϫ10–30 esu.[50] The
effect of the refractive indices of the solvents was corrected by using
a simple Lorentzian local field.[52] Assuming that the scattering
contribution from the solvent is negligibly small, this external refer-
ence method was then used to calculate the βHRS values of com-
ponents of 0.0347 and 0.0298) for elements S and P, respec-
tively.[60–62] Geometry optimizations were performed without any
symmetry constraints. In all cases, the Hessian was computed to
confirm the stationary points of the potential energy surfaces (PES)
as true minima. The static first hyperpolarizability tensor compo-
nents (βijk), for all compounds were calculated by means of the
analytic gradient methodology adopted in the Gaussian 09 pro-
gram package. These terms were used to calculate βHRS according
to Equation (3), after the summations given by Equations (4) and
(5). The reported βijk tensor components (presented in the Support-
2
plexes according to Equation (2)[44] in which ͗βHRS ͘ is the orienta-
tional average of the β tensor and S is the slope of the appropriate
“I2w versus concentration” plot (the subscripts c and ref refer to
the complexes and reference sample, respectively). For the classic
90° angle geometry HRS experiment and for an unpolarized scat-
Eur. J. Inorg. Chem. 0000, 0–0
11
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
[16] W. Wenseleers, E. Goovaerts, P. Hepp, M. H. Garcia, M. P. Ro-
balo, A. R. Dias, M. F. M. Piedade, M. T. Duarte, Chem. Phys.
Lett. 2003, 367, 390–397.
[17] M. H. Garcia, M. P. Robalo, A. R. Dias, M. T. Duarte, W.
Wenseleers, G. Aerts, E. Goovaerts, M. P. Cifuentes, S. Hurst,
M. G. Humphrey, M. Samoc, B. Luther-Davies, Organometal-
lics 2002, 21, 2107–2118.
[18] W. Wenseleers, W. Gerbrandij, E. Goovaerts, M. H. Garcia,
M. P. Robalo, M. J. Mendes, C. J. Rodrigues, R. A. Dias, J.
Mater. Chem. 1998, 8, 925–930.
[19] B. J. Coe, Chem. Eur. J. 1999, 5, 2464–2471.
[20] I. Asselberghs, K. Clays, A. Persoons, M. D. Ward, J. McClev-
erty, J. Mater. Chem. 2004, 14, 2831–2839.
[21] B. J. Coe, Acc. Chem. Res. 2006, 39, 383–393.
[22] K. A. Green, M. P. Cifuentes, M. Samoc, M. G. Humphrey,
Coord. Chem. Rev. 2011, 255, 2530–2541.
[23] F. Mançois, J.-L. Pozzo, J. Pan, F. Adamietz, V. Rodriguez, L.
Ducasse, F. Castet, A. Plaquet, B. Champagne, Chem. Eur. J.
2009, 15, 2560–2571.
[24] M. Moreno Oliva, J. Casado, J. T. López Navarrete, G.
Hennrich, M. C. Ruiz Delgado, J. Orduna, J. Phys. Chem. C
2007, 111, 18778–18784.
[25] Y. Liu, G.-C. Yang, C.-G. Liu, S.-L. Sun, Y.-Q. Qiu, Int. J.
Quantum Chem. 2012, 112, 779–788.
ing Information) are in accord with the Taylor series expansion
(T convention). To compare with experimental HRS values, which
follows the power series expansion of the dipole moment as a func-
tion of the external fields (B convention), the calculated βHRS pre-
sented in Table 5 were previously converted into the B convention.
The relation between the hyperpolarizabilities in these conventions
is 1/2βT = βB.[43] TD-DFT[63,64] was used to compute the electronic
spectra of the studied molecules applying the same level of theory
and basis sets as used for the calculation of the hyperpolarizabili-
ties. The first 24 lowest excitation energies were computed and the
simulated absorption bands were obtained by convolution of
Gaussian functions cantred at the calculated excitation energies by
using the GaussSum[65] (version 2.2.4) software. The Chemcraft[66]
program (version 1.6) was used for visualization of the computed
results, including the representation of the geometries and the or-
bitals.
Supporting Information (see footnote on the first page of this arti-
cle): The raw NMR spectroscopic data, optimized structures, an
example of a simulated absorption spectrum, selected calculated
structural data, β tensor components and orientational averages of
the tensor components resulting from DFT calculations.
[26] B. Champagne, A. Plaquet, J.-L. Pozzo, V. Rodriguez, F.
Castet, J. Am. Chem. Soc. 2012, 134, 8101–8103.
[27] C.-H. Wang, N. N. Ma, X.-X. Sun, S.-L. Sun, Y.-Q. Qiu, P.-J.
Liu, J. Phys. Chem. A 2012, 116, 10496–10506.
Acknowledgments
[28] S. Di Bella, I. P. Oliveri, A. Colombo, C. Dragonetti, S. Righ-
etto, D. Roberto, Dalton Trans. 2012, 41, 7013–7016.
[29] C.-G. Liu, Z.-M. Su, X.-H. Guan, S. Muhammad, J. Phys.
Chem. C 2011, 115, 23946–23954.
The authors thank the Fundação para a Ciência e Tecnologia
(FCT) for funding (project number FCOMP-01-0124-FEDER-
007433). T. J. L. S. is also grateful to the FCT for his Ph.D. grant.
[30] G. R. Hutchison, M. A. Ratner, T. J. Marks, J. Phys. Chem. B
2005, 109, 3126–3138.
[31] Y. Qin, J. Y. Kim, C. D. Frisbie, M. A. Hillmyer, Macromole-
cules 2008, 41, 5563–5570.
[32] A. Drury, S. Burbridge, P. A. Davey, J. W. Blau, J. Mater.
Chem. 1998, 8, 2353–2355.
[33] R. D. A. Hudson, I. Asselberghs, K. Clays, L. P. Cuffe, J. F.
Gallagher, A. R. Manning, A. Persoons, K. Wostyn, J. Or-
ganomet. Chem. 2001, 637–639, 435–444.
[34] P. J. Mendes, T. J. L. Silva, M. H. Garcia, J. P. P. Ramalho,
A. J. P. Carvalho, J. Chem. Inf. Comput. Sci. 2012, 52, 1970–
1983.
[35] R. H. L. Kiebooms, P. J. A. Adriaensens, D. J. M. Vanderz-
ande, J. M. J. V. Gelan, J. Org. Chem. 1997, 62, 1473–1480.
[36] P. Amaladass, J. A. Clement, A. K. Mohanakrishnan, Eur. J.
Org. Chem. 2008, 3798–3810.
[1] D. J. Williams, Angew. Chem. 1984, 96, 637; Angew. Chem. Int.
Ed. Engl. 1984, 23, 690–703.
[2] H. S. Nalwa, Appl. Organomet. Chem. 1991, 5, 349–377.
[3] N. J. Long, Angew. Chem. 1995, 107, 37; Angew. Chem. Int. Ed.
Engl. 1995, 34, 21–38.
[4] T. Verbiest, S. Houbrechts, M. Kauranen, K. Clays, A. Per-
soons, J. Mater. Chem. 1997, 7, 2175–2189.
[5] I. R. Whittall, A. M. McDonagh, M. G. Humphrey, M. Sa-
moc, Organometallic Complexes, in: Nonlinear Optics, I: Sec-
ond-Order Nonlinearities, in: Advances in Organometallic Chem-
istry (Eds.: F. G. A. Stone, W. Robert), Elsevier Inc., 1998, vol.
42, p. 291–362.
[6] S. Di Bella, Chem. Soc. Rev. 2001, 30, 355–366.
[7] E. Goovaerts, W. E. Wenseleers, M. H. Garcia, G. H. Cross,
Design and characterization of organic and organometallic mole-
cules for second order non-linear optics, in: Handbook of Ad-
vanced Electronic and Photonic Materials and Devices (Ed.:
H. S. Nalwa), Academic Press, Burlington, 2001, p. 127–191
[8] C. E. Powell, M. G. Humphrey, Coord. Chem. Rev. 2004, 248,
725–756.
[37]
A. K. Mohanakrishnan, P. Amaladass, Tetrahedron Lett. 2005,
46, 4225–4229.
[38]
L. l. Bonniard, S. Kahlal, A. K. Diallo, C. t. Ornelas, T. Rois-
nel, G. Manca, J. o. Rodrigues, J. Ruiz, D. Astruc, J.-Y. Saillard,
Eur. J. Inorg. Chem. 2010, 50, 114–124.
[39]
C. Ornelas, C. Gandum, J. Mesquita, J. Rodrigues, M. H. Gar-
cia, N. Lopes, M. P. Robalo, K. Nättinen, K. Rissanen, Inorg.
Chim. Acta 2005, 358, 2482–2488.
[9] E. Cariati, M. Pizzotti, D. Roberto, F. Tessore, R. Ugo, Coord.
Chem. Rev. 2006, 250, 1210–1233.
[10] H. Bozec, V. Guerchais, Molecular Organometallic Materials
for Optics. Springer, Berlin, Heidelberg, 2010, p. 230.
[11] J. Heck, M. Dede, Ferrocene-Based Electro-Optical Materials,
in: Ferrocenes, Wiley, New York, 2008, p. 319–392.
[12] M. H. Garcia, P. Florindo, M. F. M. Piedade, M. T. Duarte,
M. P. Robalo, E. Goovaerts, W. Wenseleers, J. Organomet.
Chem. 2009, 694, 433–445.
[13] M. H. Garcia, P. J. Mendes, M. P. Robalo, M. T. Duarte, N.
Lopes, J. Organomet. Chem. 2009, 694, 2888–2897.
[14] M. H. Garcia, P. J. Mendes, M. P. Robalo, A. R. Dias, J.
Campo, W. Wenseleers, E. Goovaerts, J. Organomet. Chem.
2007, 692, 3027–3041.
[40]
[41]
G. Zotti, S. Zecchin, A. Berlin, S. Grimoldi, M. C. Pasini,
M. M. M. Raposo, Chem. Mater. 2005, 17, 6492–6502.
P. Hapiot, F. Demanze, A. Yassar, F. Garnier, J. Phys. Chem.
1996, 100, 8397–8401.
[42]
[43]
J. L. Oudar, D. S. Chemla, J. Chem. Phys. 1977, 66, 2664–2668.
A. Willetts, J. E. Rice, D. M. Burland, D. P. Shelton, J. Chem.
Phys. 1992, 97, 7590–7599.
[44]
T. Verbiest, K. Clays, V. Rodriguez, Second-Order Nonlinear
Optical Characterization Techniques – An Introduction, CRC
Press, New York, 2009.
[45]
[46]
L. Jensen, P. T. van Duijnen, J. Chem. Phys. 2005, 123, 074307–
7.
[15] M. H. Garcia, M. P. Robalo, A. R. Dias, M. F. M. Piedade, A.
Galvão, W. Wenseleers, E. Goovaerts, J. Organomet. Chem.
2001, 619, 252–264.
W. L. F. Armarego, C. L. L. Chai, Purification of Laboratory
Chemicals, 5th ed., Butterworth-Heinemann, 2003.
Eur. J. Inorg. Chem. 0000, 0–0
12
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I30048
/KAP1
Date: 28-05-13 17:19:49
Pages: 13
www.eurjic.org
FULL PAPER
[47]
[48]
lam, 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, Ö. Farkas, J. B. Fore-
sman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09, rev.
A.01, Gaussian, Ltd., Wallingford, CT, 2009.
G. S. Ashby, M. I. Bruce, I. B. Tomkins, R. C. Wallis, Aust. J.
Chem. 1979, 32, 1003–1016.
K. Willinger, K. Fischer, R. Kisselev, M. Thelakkat, J. Mater.
Chem. 2009, 19, 5364–5376.
N. G. Connelly, W. E. Geiger, Chem. Rev. 1996, 96, 877–910.
S. Stadler, R. Dietrich, G. Bourhill, C. Bräuchle, A. Pawlik, W.
[49]
[50]
Grahn, Chem. Phys. Lett. 1995, 247, 271–276.
J. Ipaktschi, J. Mohseni-Ala, A. Dülmer, S. Steffens, C. Witten-
[57] T. Yanai, D. P. Tew, N. C. Handy, Chem. Phys. Lett. 2004, 393,
51–57.
[51]
burg, J. Heck, Organometallics 2004, 23, 4902–4909.
[52] C. Lambert, G. Nöll, E. Schmälzlin, K. Meerholz, C. Bräuchle,
Chem. Eur. J. 1998, 4, 2129–2135.
[53] R. Bersohn, Y.-H. Pao, H. L. Frisch, J. Chem. Phys. 1966, 45,
3184–3198.
[54] D. A. Kleinman, Phys. Rev. 1962, 126, 1977–1979.
[55] S. J. Cyvin, J. E. Rauch, J. C. Decius, J. Chem. Phys. 1965, 43,
4083–4095.
[58] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283.
[59] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[60] C. E. Check, T. O. Faust, J. M. Bailey, B. J. Wright, T. M. Gil-
bert, L. S. Sunderlin, J. Phys. Chem. A 2001, 105, 8111–8116.
[61] D. Feller, J. Comput. Chem. 1996, 17, 1571–1586.
[62] K. L. Schuchardt, B. T. Didier, T. Elsethagen, L. Sun, V. Guru-
moorthi, J. Chase, J. Li, T. L. Windus, J. Chem. Inf. Comput.
Sci. 2007, 47, 1045–1052.
[56] M. J. Frisch, G. W. 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. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bear-
park, 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. Mil-
[63] R. Nalewajski, Density Functional Theory, in: Perspectives in
Electronic Structure Theory, Springer, Berlin, Heidelberg, 2011,
p. 255–368.
[64] M. E. Casida, C. Jamorski, K. C. Casida, D. R. Salahub, J.
Chem. Phys. 1998, 108, 4439–4449.
[65] N. M. O’Boyle, A. L. Tenderholt, K. M. Langner, J. Comput.
Chem. 2008, 29, 839–845.
[66] http://www.chemcraftprog.com.
Received: January 16, 2013
Published Online: ᭿
Eur. J. Inorg. Chem. 0000, 0–0
13
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim