Hydroxyphenyl- and octoxyphenyl-substituted dipyrazinylpyridine complexes of ruthenium(II), iron(II) and nickel(II)

Article abstract of DOI:10.1016/j.ica.2011.04.018

A water-soluble, hydroxyphenyl-substituted dipyrazinylpyridine (HOL) and its liposoluble, n-octylated derivative (C₈OL) were prepared in short sequences and good yields. The monocarbonyldichlororuthenium(II) complex (C₈OL)Ru(CO)Cl₂ were obtained in cis and trans isomeric forms, which could be distinguished by their absorption, ¹³C NMR and ¹⁵N NMR spectra. The octahedral homoleptic complexes [(C ₈OL)₂M]²⁺ (M = Ru^II, Fe ^II) and five-coordinate (HOL)NiCl₂ are also described. Density functional theory is used to derive optical geometries, optical spectra, a molecular orbital description and the electronic structure. The time-dependent calculations reveal that the strong visible-region absorption responsible for the color of the homoleptic Fe^II and Ru^II species arises mainly from an internal ligand charge-transfer process and only to a small degree from the expected metal-to-ligand CT process. The crystal structures of free C₈OL, trans-(C₈OL)Ru(CO)Cl₂, [(C₈OL)₂Fe](FeCl₄)₂ and (HOL)NiCl₂ were also obtained. Electrochemical data are also presented and discussed.

Full text of DOI:10.1016/j.ica.2011.04.018

Inorganica Chimica Acta 374 (2011) 606–619
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Hydroxyphenyl- and octoxyphenyl-substituted dipyrazinylpyridine complexes
of ruthenium(II), iron(II) and nickel(II)
⇑
⇑
Christopher Dares, Tharsini Manivannan, Pierre G. Potvin , A.B.P. Lever
Department of Chemistry, York University, 4700 Keele Street, Toronto, ON, Canada M3J 1P3
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 24 April 2011
A water-soluble, hydroxyphenyl-substituted dipyrazinylpyridine (HOL) and its liposoluble, n-octylated
derivative (C8OL) were prepared in short sequences and good yields. The monocarbonyldichlororutheni-
um(II) complex (C8OL)Ru(CO)Cl2 were obtained in cis and trans isomeric forms, which could be distin-
guished by their absorption, 13C NMR and 15N NMR spectra. The octahedral homoleptic complexes
[(C8OL)2M]2+ (M = RuII, FeII) and five-coordinate (HOL)NiCl2 are also described. Density functional theory
is used to derive optical geometries, optical spectra, a molecular orbital description and the electronic
structure. The time-dependent calculations reveal that the strong visible-region absorption responsible
for the color of the homoleptic FeII and RuII species arises mainly from an internal ligand charge-transfer
process and only to a small degree from the expected metal-to-ligand CT process. The crystal structures of
free C8OL, trans-(C8OL)Ru(CO)Cl2, [(C8OL)2Fe](FeCl4)2 and (HOL)NiCl2 were also obtained. Electrochemical
data are also presented and discussed.
Dedicated to Professor W. Kaim
Keywords:
DFT
Ruthenium
Dipyrazylpyridine
Optical spectra
X-ray
Ó 2011 Elsevier B.V. All rights reserved.
1. Introduction
Fe](FeCl4)2 and (HOL)NiCl2. We also report electrochemical studies
of both isomers of (C8OL)Ru(CO)Cl2 and of both Fe and Ru homo-
We have reported one-step preparations of novel dipyrazinyl-
pyridine (dpp) ligands [1–4]. As with 2,20;60,200-terpyridine (tpy)
and other tridentate ligands, dpp ligands are non-stereogenic, but
they feature an electron deficiency, relative to tpy, owing to the
peripheral nitrogen atoms, which promotes very positive electro-
chemical oxidation and less negative reduction events. Carbonylat-
ed RuII complexes were of interest as potential photosensitizers
and electrocatalysts of CO2 reduction once attached to appropriate
electrode surfaces, but the Ru(CO)Cl2 complexes of p-tolyl- and
p-4-tert-butylphenyl-substituted dpp ligands proved to be very
insoluble and impossible to purify and fully characterize. We
therefore opted for a solubilizing long-chain substituent in prepar-
ing the n-octyloxy analog C8OL (Scheme 1). We report the prepara-
tion of this ligand via hydroxy precursor HOL, and the potentially
water-soluble potassium salt (KOL) and synthesis of the carbony-
lated RuII complexes in two isomeric forms, cis- and trans-(C8OL)-
Ru(CO)Cl2. The homoleptic [(C8OL)2M]2+ species (M = RuII and
M = FeII) are also reported along with the five-coordinate (HOL)-
NiCl2. The spectroscopic determination of the structures of the
(C8OL)Ru(CO)Cl2 isomers was supported by DFT calculations and
confirmed by crystallographic determination of the trans isomer.
Crystallographic data were also obtained for free C8OL, [(C8OL)2-
leptic species.
2. Results and discussion
2.1. Synthesis and identity
By the same one-pot route used before [1–4], 2-acetylpyrazine
and p-hydroxybenzaldehyde were condensed in the presence of
NH3 and KOH to form the yellow-orange K+ salt of 4-p-hydroxy-
phenyl-2,6-dipyrazinylpyridine (KOL) in 61% isolated yield
(Scheme 1). This salt could be quantitatively transformed with
mild acid to the light tan colored neutral form (HOL), which was
probably zwitterionic as its solubility in organic solvents was
limited. Either KOL or HOL could be alkylated to give the n-octyl
derivative (C8OL) in 90% yield, for which
determination was made.
a crystallographic
The homoleptic RuII complex was formed in high yield by stan-
ꢀ
dard means from RuCl3 and was isolated as the PF6 salt. The cor-
responding iron species was similarly obtained using ferric
chloride and isolated both as a PF6 salt and, as a crystal for XRD
containing the tetrachloroferrate(III) anion.
ꢀ
The ruthenium carbonyl species were prepared using
a
modified literature procedure [5], by heating RuCl3 at reflux in
DMF to generate a red-coloured carbonylated intermediate, then
treated with C8OL with further heating at reflux. Addition into
⇑
Corresponding authors.
E-mail address: blever@yorku.ca (A.B.P. Lever).
0020-1693/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2011.04.018

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
607
Table 1
Some spectroscopic properties distinguishing the cis- and trans-(C8OL)Ru(CO)Cl2
isomers.
k (nm)a
dCO (ppm)b
dN1, dN1 (ppm)b
0
Isomer
trans
cis
486 626
458
201
189
272 272
247 253
a
In dichloromethane.
Signals from carbonyl C, pyridine (N1) and coordinated pyrazine (N1 ) nitrogen
b
0
atoms in CDCl3.
in a p
-richer trans ligand such as Clꢀ causes an upfield shift, while a
p-poorer one such as CO causes a downfield shift [10] (Table 1).
This assignment was supported by UV–Vis spectroscopy in that
the number, relative intensities and positions of the visible-region
absorptions (Table 1) were consistent with those reported for the
analogous tpy isomers. Indeed, the spectrum of the trans species
exhibits two absorptions in the visible region, with the lower-
energy band being approximately half as strong as the other.
Trans-(tpy)Ru(CO)Cl2 had initially been reported to absorb at
460–461 nm in dichloromethane [6,8] but more recently has been
reported to show two visible-region absorption bands in CH3CN at
450 and 570 nm, with the higher-energy band being twice as in-
tense [9]. Cis-(C8OL)Ru(CO)Cl2 also displays two visible-region
transitions, in dichloromethane (DCM), but the lower-energy one
is very weak so that a cursory examination would detect only
one band at 458 nm. Indeed in acetonitrile this lower energy band
is completely obscured. Cis-(tpy)Ru(CO)Cl2 has been reported to
only show a single visible-region band at 417–419 nm in dichloro-
methane [6,8] or 408 nm in acetonitrile [9]. The band assignments
for both isomers are discussed below. Finally, the isomeric identi-
ties were confirmed by a determination of the crystal structure of
the trans isomer.
Scheme 1. Conditions: (i) CH3OH/15% KOH/conc. NH4OH/ reflux 8 h; (ii) 5% HOAc;
(iii) NaH/DMSO/1-BrC8H17/16 h.
H2O precipitated a crude material consisting of both cis and trans
isomers of the carbonyldichloro complex, the proportions of which
depended on the heating time but reached as much as 50% of each.
Purification by chromatography on silica gel with very polar
solvent (CH3CN/H2O/KNO3) led to isolation of pure trans-(C8OL)-
Ru(CO)Cl2, a black solid, in high yield (80%), with apparent simul-
taneous isomerization of the cis species. Alternatively, use of
CH3CN alone as eluent allowed the isolation of the pure cis isomer
in 19% yield, a yellow solid that very slowly isomerized upon
standing in CH3CN (the isomerization conditions were not
optimized).
Finally, a mixture of HOL and NiCl2 in MeOH produced a red–
orange solid, from which a crystal suitable for XRD was grown.
Owing to instability in various solvents, especially H2O, it was
not further investigated.
Except for the nickel species which was identified by crystallog-
raphy (see below), the identities of all new compounds were estab-
lished by appropriate combinations of mass spectrometry, IR
spectroscopy, 1H and 13C NMR spectroscopy, including 2D experi-
ments to assist signal assignments, and their purities were addi-
tionally verified by elemental analyses.
3. Crystallography
The free ligand C8OL (Fig. 1), in the solid state, is essentially flat,
with a planar dpp core (pyridine–pyrazine ring dihedral angles of
2.5° and 5.6°), with the pyrazine rings oriented so as to minimize
lone pair-lone pair repulsion with the central pyridine ring. The
phenyl ring, however, is not coplanar with the dpp system (dihe-
dral angle of 14.8°), likely due to steric interactions between the
hydrogen atoms ortho to the phenyl–pyridyl bond.
2.2. Structures of the (C8OL)Ru(CO)Cl2 isomers
The crystal packing (Fig. 2) is stabilized by 1-dimensional weak
IR spectra could not readily be used to identify the (C8OL)
Ru(CO)Cl2 isomers. Both showed a single CO stretch in the FTIR
spectra at virtually the same frequency (1977 and 1979 cmꢀ1 for
cis and trans, respectively). Two examples of this type of complex
had been previously reported, with tpy [6] and with its 4,40,400-
inter-molecular hydrogen bonds and
p–p stacking. The hydrogen
tri-tert-butyl analog [7] (tBu3tpy) as ligands. The
mCO data reported
in the literature for tpy analogs are too variable to allow a clear dif-
ferentiation of the trans isomer (1943 cmꢀ1 [8], 1950 cmꢀ1 [6,9] or
1955 cmꢀ1 [6]) from the cis isomer (1943 cmꢀ1 [6], 1948 cmꢀ1 [6,9]
or 1957 cmꢀ1 [8]). With tBu3tpy, the IR data (mCO 1946 and
1917 cmꢀ1, respectively) [7] are inconsistent with the expected ef-
t
fects of the Bu groups.
13C NMR spectroscopic data are not routinely acquired for RuII
complexes, and 15N NMR data have been reported for very few,
but both proved to be very useful in establishing the isomeric iden-
tities. The positions of the 13CO resonances (dCO) were consistent
with the expected electronic effects of the ligands lying trans to
the CO ligands, specifically
p-poor pyridine N (in C8OL) in the trans
isomer vs.
p
-richer Clꢀ in the cis isomer. As well, the 15N NMR reso-
nances (dN) of the coordinated nitrogen atoms in the trans isomer
were downfield of those from the cis isomer, entirely consistent
with the reported variation of such chemical shifts with the elec-
tronic nature of the ligands lying trans to the nitrogen atoms, where-
Fig. 1. ORTEP rendering of C8OL in the crystal. Hydrogen atoms have been omitted for
clarity. Displacement ellipsoids are drawn at the 50% probability level.

608
C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
Fig. 2. Crystal packing of C8OL illustrating (upper) the one-dimensional intermolecular hydrogen-bonding contacts.
The crystal structure of trans-(C8OL)Ru(CO)Cl2 is similar to that
reported earlier by Ziessel et al. for trans-(tpy)Ru(CO)Cl2 [9] Both
ꢀ
compounds crystallize in P1 and include one acetonitrile solvate
molecule. The Ru–CCO bond lengths are identical (1.88 Å), while
the remaining bonds to the Ru centre are longer in the tpy analog
by 0.02–0.03 Å. The C–O bond length is longer in C8OL by 0.01 Å,
which is consistent with the lower-frequency mCO. The crystal pack-
ing (Fig. 4) shows sheets of parallel molecules whose C8OL moieties
lie essentially flat in the sheet plane, and antiparallel to the mole-
cules in adjacent layers, but there is no evidence of
p–p stacking.
The acetonitrile molecules are inserted within every other layer, ly-
ing anti-parallel to their nearest neighbours within the same layer,
and contribute to rigidify the structure through a network of
CH3ꢁꢁꢁO, CH3ꢁꢁꢁN and C„NꢁꢁꢁHC interactions, alongside three CHꢁꢁꢁCl
contacts.
Fig. 3. A view of trans-(C8OL)Ru(CO)Cl2ꢁCH3CN showing the molecular structure.
The solvate molecule and all hydrogen atoms have been omitted for clarity.
Displacement ellipsoids are shown at the 50% probability level.
The crystal structure of [(C8OL)2Fe]2+ as a tetrachloroferrate(III)
salt has been obtained (Fig. 5). The central metal is in the 2+ oxida-
tion state, and is coordinated in an octahedral fashion. The complex
possesses near-C2 symmetry, with a local symmetry about the iron
centre which is approximately D2d. The dpp moieties are essen-
tially planar, forming an angle between each other that is slightly
off perpendicular (83.6°). This is likely due to the FeII centre being
slightly too large for a perfect octahedral coordination. The Fe–pz
bonds (1.97 Å) are 0.09 Å longer than the Fe–py bonds (1.88 Å),
but are typical for FeII. The py–ph dihedral angle is 34.3° for one li-
gand, and 38.2° for the other, while the C–C bond connecting the
two rings is 0.01 Å shorter than in the free ligand.
bonds form zipper-like connections between molecules, which
construct planar layers, with ring cores and tails facing anti-
parallel to each other. The
p–p stacking between electron-rich
phenyl rings and electron-poor pyrazine rings forms sheets similar
to graphite, separated by 3.29 Å.
Trans-(C8OL)Ru(CO)Cl2 crystallized as an acetonitrile solvate
(Fig. 3), confirming the earlier assignment of the geometry by spec-
troscopic means. This structure shows several differences from the
free C8OL structure. Metal coordination caused the phenyl-pyridyl
interplanar angle to increase from 14.8° to 20.8°, which decreases
the delocalization between the two ring systems. This is likely due
to a shortening of the interannular C–C bond upon complexation,
which results in a stronger steric interaction between the hydrogen
atoms flanking the interannular bond. The bite angle in the free
ligand is 114.8° but this decreases to 103.0° in the complex. The
octyl chain was straight in the sheeted free ligand crystal but
shows a kink in the complex and a high degree of thermal motion,
likely due to packing requirements.
The crystal packing of [(C8OL)2Fe](FeCl4)2 is a grid-like pattern
of parallel and essentially perpendicular layers, with one such layer
roughly along the b axis (Fig. 6). Unlike in the crystals of C8OL or
trans-(C8OL)Ru(CO)Cl2, the aliphatic tails do not exactly match
ꢀ
up, resulting in a tighter packing arrangement. The FeCl4 anions
are intercalated into the layers at the end of each tail, as well as
in between the layers, aligned in a linear fashion, in close proximity
to the dpp moiety. The layers along the b axis are separated by
6.67 Å, while the perpendicular layers have [Fe(C8OL)2]2+ units that
Fig. 4. Crystal packing of trans-(C8OL)Ru(CO)Cl2ꢁCH3CN showing pertinent intermolecular contacts, both (left) within a single sheet and (right) between sheets. The hydrogen
atoms that are not involved in these contacts have been omitted. Displacement ellipsoids are drawn at the (upper) 50% or (lower) 30% probability levels.

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
609
ꢀ
Fig. 5. A view of [(C8OL)2Fe](FeCl4)2 showing the molecular structure. The FeCl4 anions and all hydrogen atoms have been omitted for clarity. Displacement ellipsoids are
shown at the 50% probability level.
Fig. 6. A view of [(C8OL)2Fe](FeCl4)2 showing the packing of a layer approximately along the crystal’s b axis. All hydrogen atoms have been omitted for clarity. Displacement
ellipsoids are shown at the 50% probability level.
than the equatorial Ni–Cl bond, which is consistent with a square
pyramidal d8 complex. It forms planar layers with anti-parallel sets
of the complex (Fig. 8). Parallel layers show
p–p stacking, with a
separation of 3.46 Å, and the Ni atoms line up in a nearly linear
chain with a Ni–Ni separation of 3.97 Å. The HOL ligand is essen-
tially planar, with a phenyl-pyridyl twist angle of only 2.4°. This
is curious, considering that it showed the same CPh–Cpy bond
length as in free C8OL (1.48 Å), which showed an angle of 14.8°.
This coplanarity may be explained by electron demand from the
metal and, in the absence of the alkyl chain, a stronger crystal
packing owing to the p-p and Ni–Ni interactions.
Table 2 gathers some important measurements. All of the crys-
ꢀ
talline compounds crystallize within the P1 space group. They also
all exhibit packing in layers, which incorporate C8OL units nearly
completely in the layer. The angles between the phenyl ring and
central pyridine vary from nearly 0° up to 38°. Evidently, there is
free rotation in solution, and the conformation which is frozen
out in the crystal structures is the result of a combination of steric
and crystal packing effects. The M–Npz and M–Npy bond lengths in
the Fe complex are 0.09 Å and 0.12 Å shorter than those found in
trans-(C8OL)Ru(CO)Cl2. This difference is expected due to the
difference in size of the metal centre, which also accounts for the
slight difference in bite angle. There appear to be no significant
differences among the homoleptic complex and cis and trans
carbonyl species in the Ru–Npz and Ru–Npy bond distances. This
is despite, as will be seen below, a significant difference in the
Fig. 7. A view of (HOL)NiCl2ꢁ1.5CH3OH showing the molecular structure. The
solvate molecules and all hydrogen atoms have been omitted for clarity. Displace-
ment ellipsoids are shown at the 50% probability level.
are offset from each other, resulting in a very small 2.40 Å inter-
layer separation. The packing is further strengthened by non-
classical hydrogen bonds between the cation and anions through
weak ClꢁꢁꢁH–Cpz interactions, where the angle is 141.5° and the
Cl–H distance is 2.86 Å.
The five-coordinate (HOL)NiCl2 has a slightly distorted square
pyramidal geometry (Fig. 7). The axial Ni–Cl bond is 0.28 Å longer

610
C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
Fig. 8. Crystal packing diagram of (HOL)NiCl2ꢁ1.5CH3OH showing a single planar crystal layer. All hydrogen atoms have been omitted for clarity. Thermal displacement
ellipsoids are shown at the 50% probability level.
Table 2
Select bond lengths and angles in the crystal structures and in DFT-optimized structures.
Species
M–C (Å)
M–Npz (Å)
M–Npy (Å)
M–Cl (Å)
C@O (Å)
Bitea (°)
py–Ph (°)
Cpy–Cph (Å)
C8OL
–
–
–
–
–
–
–
–
–
–
114.8
120.1
102.0b
102.3
104.3
104.1
14.8
30.31
36.3b
26.8
27.0
2.4
1.481
1.486
1.471b
1.479
1.480
1.480
C8OLc
[(C8OL)2Fe][FeCl4]2
ꢀ
1.970b
2.004
2.100
2.054b
1.883b
1.911
2.012
1.948
ꢀ
ꢀ
[(C8OL)2Fe]2+
–
–
–
c
[(C8OL)2Ru]2+
–
ꢀ
–
–
ꢀ
c
(HOL)NiCl2
2.224 (eq)
2.503 (ax)
2.392b
2.507
2.531 (eq)
2.536 (ax)
2.4136b
trans-(C8OL)Ru(CO)Cl2
trans-(C8OL)Ru(CO)Cl2
cis-(C8OL)Ru(CO)Cl2
1.881
1.888
1.863
2.057b
2.090
2.095
2.007
2.032
1.992
1.140
1.184
1.185
103.0
103.4
103.8
20.8
27.9
27.3
1.475
1.482
1.482
c
c
d
trans-(tpy)Ru(CO)Cl2
1.882
2.086b
2.024
1.125
103.6
ꢀ
ꢀ
a
b
c
Npy–Npz–Npy angle.
Averaged value.
DFT-optimized structure (PCM, dichloromethane).
From Ref. [9].
d
natural population charges on ruthenium between the homoleptic
species and the cis/trans isomers.
the RCH2O of the octoxyphenyl residue is co-planar with the
4-phenyl group. Fig. 9 shows how the frontier molecular orbitals
are shared between the pyridine, pyrazine and octoxyphenyl frag-
ments of the ligand. Most dramatically, the HOMO is dominantly
located on the octoxyphenyl fragment, while the first five virtual
p⁄ MOs are localized on the central dpp unit but all five are polar-
ized more towards the pyrazine units, rather than the pyridine
ring, due to the greater electron acceptor strength of pyrazine
vis-a-vis pyridine. This organization of MOs dominates the proper-
ties of the resulting complexes.
4. Electronic structures
4.1. Frontier molecular orbitals of the free ligand C8OL
DFT computations of the molecular and electronic structures
were carried out for several of the new molecules reported. The
geometry of the DFT-optimized free ligand is slightly different
from its crystal structure since the pyrazine units twist out of plane
(37°–38°), presumably to minimize steric hindrance from the ortho
hydrogen on the adjacent pyridine ring. In the crystal, packing
forces apparently constrained this twist. In the complexes, the pyr-
azine units were co-planar with the pyridine units (Figs. 3, 5 and
7). In the optimized and coordinated ligand conformations, the
4-phenyl group is canted 29o out of the plane of the dpp core while
4.2. Frontier molecular orbitals of the homoleptic complexes
Fig. 10 displays the frontier orbital compositions of [(C8OL)2M]2+
.
These species have effective D2d symmetry at the metal binding site
and the HOMO and HOMOꢀ1 (MOs 240 and 241) transformas ‘‘e’’, as
do the LUMO and LUMO+1 (MOs 242 and 243). Following on from
the free ligand discussion above, MOs 240 and 241 are dominantly

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
611
accepting pyrazine groups. The Ru d6 levels (t2g in Oh) are fairly
clearly identified, being mostly localized in three MOs. There is little
contribution from the central dpp core in the upper filled orbitals, at
least not in the first seven shown in Fig. 10.
6
4.3. Frontier molecular orbitals of the cis/trans (C8OL)Ru(CO)Cl2
complexes
Fig. 11 displays the frontier orbital compositions of the carbonyl
complexes. The lower-energy virtual orbitals are localized on the
central dpp core, as was the case with the bis-ligand complexes.
These complexes have quite low symmetry. The HOMO is now
dominantly Ru-localized, with
a
significant Clꢀ contribution.
HOMOꢀ1 and HOMOꢀ2 contain moderate contributions from oct-
oxyphenyl, which no longer dominates, and the three filled Ru d
orbitals (t2g in Oh) are distributed over at least 5 MOs. Again, the
first seven filled MOs (except for MO 134 in the cis species) have
little dpp component. There is very little contribution from the car-
bonyl group in the frontier orbitals.
Fig. 9. Frontier molecular orbitals of the free ligand, coded according to contribu-
tions from the pyridine (py) (yellow), pyrazine (pz) (blue), and octoxyphenyl
(magenta) fragments (DFT, B3LYP/LANL2DZ).
The bonding and anti-bonding interaction between the various
components, Cl, CO and dpp with the frontier d orbitals is most
easily displayed by using the Angular Overlap Model (AOM)
[11–14]. The AOM easily reveals (Table 3) that the three d orbitals
localized on the octoxyphenyl group, especially in the Fe case, and
the next set of orbitals are virtual p⁄ in character and all localized
on the central dpp unit, with emphasis on the more electron-
(MOs 136, 138 and 139) interact repulsively with
p-chloride and
Fig. 10. Dominant contributions to the frontier orbitals of [(C8OL)2Fe]2+ (left) and [(C8OL)2Ru]2+ (right), colour-coded as to origin: Ru (black), pyridine (yellow), pyrazine
(blue), octoxyphenyl (magenta) (DFT, B3LYP/LANL2DZ, PCM (dichloromethane)).
Fig. 11. Dominant contributions to the frontier orbitals of cis-(C8OL)Ru(CO)Cl2 (left) and its trans isomer (right). The colour-coding is as in Fig. 10 with Clꢀ (red) and CO
(green) (DFT, B3LYP/LANL2DZ, PCM (dichloromethane)).

612
C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
Table 3
footnotes to Table 4. The values obtained are very closely similar
to those for the C8OL species.
d Orbital assignment and AOM energy contributions.a
(C8OL)Ru(CO)Cl2
MO
Designationb
AOM results
cis Isomer
134
136
138
139
136
138
139
dxy
dxy
dyz
dxz
dxz
dyz
dxy
ꢀ
pCl
pCl
pCl
pCl
ꢀ
pCO
pCO
pCO
5. Electrochemistry
ꢀ
2
ꢀ
+
+
pdpp
Within the solvent window, both homoleptic species display
one oxidation wave, rather broad with Fe but more well defined
with Ru, and four reversible or quasi-reversible reduction waves
(Fig. 12, Table 5). Differential pulse voltammetry (not shown) re-
veals that these are all one-electron processes. The spacing of the
two pairs of reduction processes probably relates to the first two
electrons going into the LUMO/LUMO+1 (‘e’ in D2d) representation
pair of MOs followed likely by the third electron going into
LUMO+2. LUMO+3 is close in energy to LUMO+2, especially in
the iron case, and so the fourth electron probably goes to LUMO+3
or pairs one of the ‘e’ electrons. Gas-phase DFT calculations of the
trans Isomer
ꢀ
pCO
2
2
pCl
pCl
ꢀ
pCO
pdpp
a
Angular Overlap Model (see Ref. [6]). Positive sign is anti-bonding, negative sign
is bonding (all AOM parameters here are assumed to be positive).
b
The dpp plane is designated xz with the z axis bisecting the pyridine ring.
p
-dpp, and attractively towards p-CO. In the cis isomer, MO 134 is
a bonding interaction between a Clꢀ and the dxy orbital. This infor-
mation will become important in considering the optical spectra
(below). Although the DFT energies of these orbitals are known,
we do not derive values of these AOM parameters because of
extensive mixing with other MOs. Others, however, have shown
relationships linking the AOM and DFT approach [12,14].
Table 4 shows some bond orders (B) [16,17] and natural popu-
lation charges (q) [15] in these species. The M–N bond orders are
rather greater with Fe than with Ru in the homoleptic species,
due to better matching of the frontier orbitals, better orbital over-
lap and somewhat greater
r-donation. The p back-donation is
small, though somewhat greater from Ru than from Fe (note the
very small metal contributions to the lower LUMO levels). In agree-
ment, the net charge on Fe is slightly less positive than on Ru. In
the carbonyl complexes, the net charge on Ru is much more nega-
tive, due to the presence of two Clꢀ ligands and despite the pres-
ence of the CO ligand. (This charge may well be overestimated by
DFT because the charges on the Clꢀ ligands is probably more dis-
persed into the solvent than is computed.) The Ru–Npy bond order
in the cis carbonyl complex is substantially larger than in the trans
isomer and the Ru–Npy bond is apparently shorter (Table 2) though
the difference is small. There is a clear trans effect operational here
where the chloride ion opposite the Ru–Npy bond facilitates ruthe-
nium-nitrogen bonding in the cis isomer, while the carbonyl group
which is opposite the Ru–Npy bond in the trans isomer, withdraws
electron density and weakens the Ru–Npy bond. Interestingly, there
is apparently no effect on the Ru–Npz bond order, which is the
same in both isomers (Table 4). Similar bond length variations
are observed [6] in the cis/trans isomers of (tpy)Ru(CO)Cl2.
Fig. 12. Cyclic voltammograms of [(C8OL)2M](PF6)2 (M = Ru, in black, or M = Fe, in
green) in CH3CN and trans-(C8OL)Ru(CO)Cl2 (red) in acetonitrile. The inset shows
the cis isomer in dichloromethane. Bu4NPF6 as supporting electrolyte. Scan rate
100 mV sꢀ1
.
Table 5
Electrochemical potentials (in V vs. SCE) in CH3CN at 100 mV sꢀ1 with Bu4NPF6 as
supporting electrolyte.
The final column of Table 4 shows net bond orders between
molecular fragments. For [(C8OL)2M]2+, this is between the metal
centre (f1) and the rest (f2). For the carbonyl species, the fragments
are Ru(CO)Cl2 (f1) and C8OL (f2). In all cases, B(f1–f2) > [4B(M–
Npz) + 2B(M–Npy)] in reflection of interactions between the
metal-centred MOs with those of other ligand atoms. NPA charge
and M–N bond orders for [M(2,20-bpy)3]2+ calculated by DFT at
the same level as for the C8OL complexes are shown in the
Oxidation
Reduction
[(C8OL)2Fe]2+
1.43
1.56
1.28
1.24
ꢀ0.87, ꢀ0.99, ꢀ1.54, ꢀ1.77
ꢀ0.84, ꢀ1.02, ꢀ1.35, ꢀ1.69
ꢀ1.06a
[(C8OL)2Ru]2+
cis-(C8OL)Ru(CO)Cl2
trans-(C8OL)Ru(CO)Cl2
ꢀ0.98a
b
a
Irreversible.
In DMF.
b
Table 4
Summary of DFT calculated (PCM, dichloromethane) natural chargesa q and Mayer bond ordersb B (M is Fe or Ru; py is pyridine, and pz is pyrazine).c
Species
q(M)
0.25
0.31
ꢀ0.27
ꢀ0.30
B(M–Npy
)
q(Npy
)
B(M–Npz
)
q(Npz
)
B(f1–f2)d
[(C8OL)2Fe]2+
0.54
0.375
0.53
ꢀ0.44
ꢀ0.42
ꢀ0.38
ꢀ0.40
0.55
0.41
0.41
0.40
ꢀ0.40
ꢀ0.40
ꢀ0.38
ꢀ0.37
3.61
3.08
1.82
1.64
[(C8OL)2Ru]2+
cis-(C8OL)Ru(CO)Cl2
trans-(C8OL)Ru(CO)Cl2
0.345
a
b
c
Natural population analysis charges, see Ref. [15].
See Refs. [16,17] and text.
q(M) and B(M–Npy) for [Fe(2,20-bipyridine)3]2+ are 0.31 and 0.53, respectively, and 0.28 and 0.385, respectively for the RuII analog.
d
Net fragment bond orders; see text.

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
613
Fig. 13. The S = 1/2 Mulliken spin density (B3LYP/LANL2DDZ/PCM,dichloromethane) of [(C8OL)2Ru]3+. The spin density lies approximately equally on Ru and on the two
octoxyphenyl residues (see text). In this view, one C8OL, is seen side-on.
triply reduced species do show that the spin quartet (no pairing in
‘e’) is more stable than the spin doublet (pairing in ‘e’), but the dif-
ference is small (bigger for Fe than for Ru) and the calculation can
only be regarded as an indicator and not a proof of the spin state of
these species.
The oxidation process is more interesting. One might suppose
that we observe MIII/II oxidation but as we note above the
HOMO/HOMOꢀ1 ‘e’ pair is mostly localized on the octoxyphenyl
group of the ligand, less so on the metal. Thus, a geometry-
optimized DFT calculation was carried out on the one-electron-
oxidized species. The formal RuIII product shows spin density of
0.41e (Mulliken) on the metal (NBA shows 0.39e) and the remain-
ing spin dominantly on the octoxyphenyl residues of the C8OL
ligands (Fig. 13). NPA 0.28e on each octoxyfragment and 0.04e
mostly on each central pyridine unit. Thus, according to DFT, oxi-
6.1. Electronic spectra of the free C8OL ligand
The calculated optical spectra of the optimized free ligand
geometry and that of the ligand in its metal-binding geometry,
are shown in Fig. 14. They are basically very similar, with a rela-
tively low energy, intense absorption which is shifted significantly
to the red in the geometry adopted by the homoleptic complexes
versus the optimized free ligand geometry. Time-dependent DFT
reveals that the assignment for this band, (transition #2) is HOMO
to LUMO (MO 117 ? MO 118), i.e. an internal ligand charge trans-
fer band (iLCT). The terminus of this transition lies primarily on the
two pyrazine units. Thus in comparing the non-planar free ligand
and the planar bound ligand, a red shift is entirely reasonable. This
transition is not so obvious in the experimental spectrum of the
free ligand (Fig. 14, in red) and presumably lies under the rising
shoulder of the asymmetric main peak (centered at 34 000 cmꢀ1).
Transition #7 at 33 500 cmꢀ1 is of the same nature (MO
117 ? MO 120). The cluster of higher-energy transitions, such as
dation yields rather
a species roughly described by [(C8OL)
(C8OL+0.5)Ru+2.5 3+ instead of [(C8OL)2RuIII]3+. In silico oxidation of
]
the iron homoleptic species yields a similar product. Unfortunately
the oxidation products are not sufficiently stable on the spectro-
electrochemical time frame to derive their physical properties
experimentally.
transition #15, are
p–
p⁄ transitions within the central dpp core.
The overall agreement between predicted and experimental spec-
tra is quite good, especially at the higher energies. A similar (iLCT)
band was reported recently for a pyrrole substituted dipyrazylpyri-
dine ruthenium species [3].
The cyclic voltammograms of cis- and trans-(C8OL)Ru(CO)Cl2
displayed a reversible or quasi-reversible oxidation wave and an
irreversible reduction wave (Fig. 12, Table 5). In this case, we have
noted that the HOMOs of these species do have substantial Ru con-
tent and so the oxidation wave is predominantly RuIII/II. The net
NPA charge on ruthenium in these species is much more negative
than in the homoleptic species (Table 4), consistent with the oxida-
tion process occurring at a substantially less positive potential. The
irreversibility of the reduction waves may be due to the lability of
the Clꢀ ligands [18], though spiking the solution with LiCl was not
effective in forcing reversibility. On the other hand, in both sys-
tems, there is deposition of an insoluble surface layer on the elec-
trode, following the first scan beyond the irreversible reduction
wave. This is likely the primary cause of irreversibility. A similar
insoluble product was noted in the corresponding terpyridine ser-
ies [19–22]. Both oxidation and reduction processes are shifted
approximately 300 mV to more positive potentials, relative to the
terpyridine analogs [9], due to the greater acceptor property of
the C8OL ligand.
6.2. Electronic spectra of the homoleptic species
The two homoleptic species have very similar spectra (Fig. 15),
with the visible-region band shifted further to the red for the Fe
species. Polypyridine FeII and RuII exhibit visible absorptions gen-
erally assigned to metal-to-ligand charge transfer (MLCT) transi-
tions (Ru/Fe
dp ) [26]. However the first strong
? ligand p⁄
visible region band in these systems (transition #3 for Ru and #4
for Fe) is not MLCT but is best described again as an internal ligand
charge transfer (iLCT) transition from octoxyphenyl to dpp core
(Table 6). The resulting, rather dramatic, change in charge distribu-
tion as a consequence of this transition is shown in Fig. 16 for the
ruthenium species. Clearly, the metal binding stabilizes the dpp
core to accept an electron from the more distant octyloxyphenyl
group. This transition is extremely sensitive to the acceptor charac-
ter of the dpp core, experimentally shifting from ca 34 000 cmꢀ1 in
the free ligand to 19 800 cmꢀ1 in the Ru homoleptic species to
17 500 cmꢀ1 in the Fe analog. The HOMO–LUMO gaps are: opti-
mized (non-binding) ligand (4.04 eV), ligand with the same geom-
etry as the homoleptic Ru complex (3.81 eV), homoleptic Ru
complex (2.97 eV), homoleptic iron complex (3.05 eV).
6. Optical spectroscopy
Time dependent density functional theory (TD-DFT, PCM
(dichloromethane)) [23–25] was used to derive the optical transi-
tions occurring in each species. The excited states are numbered
sequentially and a listing of energies, oscillator strengths and
assignments is given in Supplementary information. Table 6 lists
selected transition data mostly the more obvious strong transi-
tions. The transitions are identified in Table 6 and in figures by
their sequence number.
Since the HOMOs of the two species contain a small contribu-
tion from metal dp, there is a small MLCT contribution to the
strong visible region absorption. The first, primarily MLCT transi-
tion is transition #11 in RuL2 and #17 in FeL2 comprising excitation
from the localized Ru/Fe dp
(MOs 237 and 238) to dpp p⁄ orbitals,
but it is submerged by the much more intense iLCT transition. As in
the free ligand spectrum, transitions #23 (RuL2) and #28 (FeL2) are

614
C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
Table 6
Selected experimental and TD-DFT-predicted transitions and assignments.a
Transitionb
Energy (log
Expt.c
e
) (cmꢀ1
)
Orbital compositione
Assignmentf
TD-DFTd
trans-(C8OL)Ru(CO)Cl2
1
2
3
16100 (3.42)
17420 (sh)
20920 (3.69)
25640 (sh)
26600 (sh)
29070 (sh)
29600 (4.55)
16530 (0.035)
18520 (0.036)
21100 (0.056)
22815 (0.210)
26030 (0.044)
27600 (0.081)
28750 (0.035)
H-0 ? L+0
H-1 ? L+0
H-0 ? L+1
H-2 ? L+0
H-0 ? L+3
H-1 ? L+3
H-6,7 ? L+0
Cl2Ru ? p⁄ dpp MLCT
iLCT, Cl2Ru ? p⁄ dpp MLCT
Cl2Ru ? p⁄ dpp MLCT
iLCT
6
10
15
17
Cl2Ru ? p⁄ dpp MLCT
iLCT, Cl2Ru ? p⁄ dpp MLCT
ILCT,
p ?
p⁄ dpp
cis-(C8OL)Ru(CO)Cl2
1
2
7
16800 (2.57)
22120 (3.74)
16850 (0.001)
19520 (0.10)
23415 (0.130)
27800 (0.125)
H-0 ? L+0
H-1 ? L+0
H-2 ? L+0
H-1 ? L+3
Cl2Ru ? p⁄ dpp MLCT
Cl2Ru ? p⁄ dpp MLCT, iLCT
Cl2Ru ? p⁄ dpp MLCT, iLCT
Cl2Ru ? p⁄ dpp MLCT, iLCT
16
27200 (sh)
[(C8OL)2Ru]2+
3
19890 (4.21)
24350 (sh)
28870 (4.60)
32460 (4.66)
19170 (0.469)
23890 (0.146)
27400 (0.509)
32150 (0.435)
32190 (0.413)
32260 (0.217)
33500 (0.624)
35550 (0.547)
H-0,1 ? L+0,1
H-3,4 ? L+0,1
H-0,1 ? L+5,6
H-8 ? L+0,1
iLCT
11
23
50
Ru d
iLCT
p
?
p⁄ dpp MLCT
p–
p⁄ dpp
51
58
74
H-3,4 ? L+0,1
Very mixed
H-7,8 ? L+0,1
Ru d
iLCT,
p⁄ dpp
p ?
p
p⁄ dpp MLCT
?
p⁄ op
34730 (4.71)
p ?
[(C8OL)2Fe]2+
4
17510 (4.05)
18890 (sh)
21900 (sh)
28420 (sh)
29920 (4.49)
34470 (4.65)
20330 (0.563)
H-0,1 ? L+0,1
iLCT
17
28
24510 (0.106)
28940 (0.473)
H-3,4 ? L+0,1
H-0,1 ? L+6,7
Ru d
iLCT
p ?
p⁄ dpp MLCT
67
34390 (1.14)
H-0,1 ? L+8,9
iLCT, p ?
p⁄ op
a
b
c
d
e
f
A more complete listing of predicted TD-DFT data can be found in the Supplementary Information.
Column 1 contains the numbered transitions identified in Figs. 15 and 16.
Experimental data recorded in dichloromethane (sh = shoulder).
TD-DFT calculations include solvent using the polarized continuum model (PCM) and assuming dichloromethane.
Only the dominant contribution is shown; H = HOMO, L = LUMO.
Only the dominant assignment is provided; ‘dpp’ implies dipyrazylpyridine core and generally excludes the octoxyphenyl residue, iLCT is an intramolecular charge
transfer from the octoxyphenyl residue to the dpp core,
p
?
p⁄ op is an internal
p
?
p⁄ transition within the octoxyphenyl residue, MLCT is metal-to-ligand charge transfer.
higher-energy iLCT transitions. Bands in the vicinity of
35 000 cmꢀ1 comprise both further MLCT transitions and internal
dpp
predicted and experimental spectra is very good on the basis of en-
ergy, but the DFT-predicted spectrum is substantially more
intense.
p–
p⁄ transitions. The overall agreement between TD-DFT
6.3. Electronic spectra of the cis/trans-(C8OL)Ru(CO)Cl2 species
The lowest-energy TD-DFT-predicted visible-region transition
#1, is the HOMO to LUMO transition in both isomers (Fig. 17)
but, unlike the homoleptic species, the frontier filled orbitals in
these complexes contain appreciable metal content (Fig. 11). Thus,
transition #1 is an MLCT from Cl2Ru to dpp p⁄. This band has much
greater intensity in the trans isomer than in the cis isomer. As
noted above, the HOMO d orbital differs in these two isomers
(dxz for cis, dxy for trans, see Table 3) and, crucially, the HOMO d
orbital overlaps the dpp p⁄ orbitals in the trans isomer, but not
in the cis isomer. Hence, the HOMO to LUMO transition is very
much weaker in the cis isomer than in the trans isomer. As a corol-
lary, the dxz orbital in the trans isomer contributes to MO 136 and,
hence, MLCT transitions from this orbital to dpp p⁄ are weaker in
the trans isomer. Transition #2 in the trans isomer (#7 in the cis
isomer), is a second, similar MLCT transition but it does have an
appreciable octoxyphenyl to dpp iLCT character, more so in the
trans species (see Fig. 11). Transition #3 in the trans isomer (HOMO
to LUMO+1) is MLCT while #7 in the cis isomer is of mixed MLCT
and iLCT character (Table 6).
Fig. 14. Optical spectra of the free ligand (C8OL) in dichloromethane using B3LYP/
LANL2DZ/PCM. Experimental spectrum (red), DFT-calculated spectrum at opti-
mized geometry (blue), and DFT-calculated spectrum of the ligand in the same
geometry as in [(C8OL)2Ru]2+ (green). In this and subsequent optical spectra Figures,
the blue vertical lines show the locations of the TD-DFT predicted transitions scaled
by their oscillator strengths. The more important bands are numbered by their
location in the sequence of transitions. Some lines will appear to be missing because
their oscillator strength is too small to be observable on this scale or because they
occur at the same energy as another transition.
The visible-region bands of the (C8OL)Ru(CO)Cl2 species are sol-
vatochromic, varying in wavelength in the case of the trans isomer
from as high as 454 nm and 580 nm in a 95:5 acetone–water mix-
ture to as low as 490 nm and 648 nm in ethyl acetate (Table 7).

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
615
Fig. 15. Experimental and TD-DFT-predicted electronic spectra of [(C8OL)2M]2+ (M is Ru (left) or Fe (right)). The experimental spectra (red) were acquired in dichloromethane,
while the spectrum of the DFT-optimized structure (blue) was modeled in a dielectric medium matching that of dichloromethane. The DFT-calculated spectrum (blue) is
substantially more intense than the experimental spectrum and utilizes the right hand intensity axis. The vertical bars are the locations of the electronic transitions with their
correct relative oscillator strengths, numbered according to their relative energy.
Fig. 16. TD-DFT predicted excited state charge redistribution for transition #3 of [(C8OL)2Ru]2+ in dichloromethane. H-atoms are omitted for clarity. Red means a deficiency
and green an excess of electron density relative to the ground state (B3LYP/LANL2DZ/PCM (dichloromethane)).
Fig. 17. TD-DFT-computed (blue) and observed UV–Visi spectra (red) of trans- (left) and cis-(C8OL)Ru(CO)Cl2 (right) in CH2Cl2. The vertical bars are the locations of the
electronic transitions with their correct relative oscillator strengths. Both spectra use the left hand intensity axis.
The solvatochromism provides useful information about the
nature of the transition. MLCT transitions in polypyridine type
Table 7
Solvent-dependent transition energies for trans-(C8OL)Ru(CO)Cl2.
Ru(II) species are not generally very solvatochromic because of sig-
Solvent
E1 (cmꢀ1
a
)
E2 (cmꢀ1
a
)
nificant mixing between the ground and excited states, i.e. often
the same orbitals are involved in both states. This is especially true
of complexes with non-innocent ligands such as benzoquinonedii-
mine [27]. Indeed, significant solvatochromism is associated with a
large difference in dipole moment between the ground and excited
states [28–30]. Both the octoxyphenyl-to-dpp iLCT transition and
the Cl2Ru ? p⁄ (dpp) MLCT transition would be expected to show
significant solvatochromism, in the latter case because the solvent
will disperse the negative charge on the Clꢀ ligands to a greater or
lesser degree, depending on the polarity of the solvent. The polar
95:5 Acetone–water
Acetonitrile
DMSO
DMF
Acetone
Pyridine
Dichloromethane
Ethyl acetate
17240
16390
16290
16080
15970
15870
15970
15430
22030
21370
21370
21370
20830
20830
20580
20410
a
Energies of the first two experimental absorption bands in the visible region in
Fig. 17 (left).

616
C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
Fig. 18. The excited state charge re-distribution, for transitions #29 (MLCT) and #6 (iLCT) of trans-(C8OL)Ru(CO)Cl2. Red means a deficiency and green an excess of electron
density relative to the ground state (B3LYP/LANL2DZ/PCM (dichloromethane)).
solvents will reduce the effective charge on Cl, thereby increasing
the net positive charge on Ru and shifting the MLCT transition to
the blue, as observed. Fig. 18 shows the excited state charge re-dis-
tribution, for transitions #1 and #6 of the trans isomer, clearly
indicative of a change in dipole moment. Indeed the calculated
changes in dipole moment are 1.48 and 3.32 D, respectively. Thus,
the solvatochromic behavior is expected although one cannot eas-
ily conclude whether it is primarily due to the octoxyphenyl iLCT
contribution or to the solvation of chloride, likely both mecha-
nisms are in play.
8. Experimental
8.1. General
Microanalysis was carried out by Guelph Chemical Laboratories,
Guelph, ON, Canada. NMR spectra were acquired on Bruker AV 300,
AV 400 and DRX 600 instruments at 300, 400 or 600 MHz, respec-
tively. The 15N data were acquired on the 600 MHz spectrometer
using a z-gradient 1H–15N HMBC sequence with low-pass filter ad-
justed to observe 4 Hz coupling. There were 1024 and 256 data
points with spectroscopic widths of 6000 and 24,331 Hz in the
1H and 15N domains, respectively. The data were processed using
an unshifted sine-bell window function with the 15N domain
zero-filled twice. The indirect domain was externally referenced
In the trans isomer, we tentatively associate both transitions #1
and #2 to the first main visible-region feature, at ca 16 000 cmꢀ1
,
while transitions #3, and #6 agree in energy with the second
experimental feature at ca 21 000 cmꢀ1 but the predicted intensity
of transition #6 is much too high. In the 25 000–30 000 cmꢀ1 re-
gion, the overall experimental and predicted band shapes agree
but now it is the experimental intensities which are high relative
to predicted. In the cis isomer, transition #1 is energetically well
predicted and is of very low intensity. Transition #2 then corre-
sponds with the first intense absorption at ca 22 000 cmꢀ1 which
may also encompass transition #7. Transition #16 may then corre-
late with the shoulder at 27 200 cmꢀ1. Overall agreement with
higher-energy bands, in both isomers, is disappointingly not as
good as in the cases with the free ligand and homoleptic species.
This may be possibly be ascribed to a poor description of the
charge on the chloride groups.
to
a
sealed capillary of neat nitromethane assigned to
380.23 ppm. FT-IR spectra were obtained from KBr discs with a
Genesis II instrument. A Hewlett–Packard HP8452A diode array
was used to obtain UV–Vis spectra. Cyclic voltammetry used either
a Cypress Systems potentiostat (software version 5.5) in a Faraday
cage, or an Obbligato Objectives Faraday MP potentiostat, or a Pine
RDE3 bipotentiostat employing an EDAQ e-corder and software.
The working electrode was a Pt wire; a graphite rod was used as
counter-electrode and a pseudo-reference Ag/AgCl electrode was
employed. Ferrocene served as an internal reference. The correc-
tions employed were those provided by Geiger [34].
Density functional theory (DFT) calculations[i] were carried out
using the B3LYP functional [35,36] and LanL2DZ basis set [37–40]
within the GAUSSIAN G09 (revision B.01) software suite [41]. The spe-
cies were geometry-optimized while modeled in both gas and
solution phases (using a polarized continuum model, PCM), and a
frequency calculation was carried out to ensure that a true ground
state had been achieved. Time-dependent DFT (TD-DFT) [24,42]
was used to derive the energies of the excited states, from which
the calculated electronic spectra were derived using the SWIZARD
program [43]. The AOMIX suite of programs [44] was used to extract
numerical data from the Gaussian output files. Charge redistribu-
tion diagrams utilised the SWARLOCK program [45].
7. Summary and conclusions
The dipyrazinylpyridine core is a useful alternative to terpyri-
dine as a framework to build functional complexes. It is syntheti-
cally easy to prepare in a one pot synthesis and can readily be
modified by various substituents. We have demonstrated here
the ready synthesis of complexes of both first and second row tran-
sition elements, containing one or two dpp ligands. Complexes of
ligands containing the pyrazine fragment offer several advantages.
In this system the frontier virtual p⁄ orbitals are more heavily
localized on the pyrazine fragment, rather than the central pyri-
dine because of the greater acceptor character of pyrazine versus
pyridine. Thus excited state electron density can be directed along
the pyrazine coordinates when designing electron transfer supra-
molecular species. Previously we have shown that other metal con-
taining moieties can be attached to the far nitrogen of pyrazine
containing ligands [31,32]. This will be further developed in the fu-
ture with other electron-donating substituents [33].
8.2. Crystallography
X-ray diffraction data (Table 8) were collected by Dr. Alan Lough
(University of Toronto) on a Nonius Kappa CCD diffractometer at
150 K, using a fine-focus sealed tube Mo K
a
(k = 0.71073 Å) radia-
scan mode.
tion source with a graphite monochromator with a
x
The remarkable shift in the internal ligand charge transfer tran-
sition from the octoxyphenyl to the central dpp core illustrates an
unusual feature of these species, the ability to tune an internal li-
gand transition over a wide spectroscopic range. This will be fur-
ther developed in the future with other electron donating
substituents placed in the 4-position.
An absorption correction (Denzo-SMN) was applied to all unique
reflections prior to use in calculations [46]. The initial structure
solutions were obtained by direct methods using either the SHELXTL
V5.1 SHELX V6.12 or SHELXS97 packages, while full-matrix least-
squares refinement of the initial structure solution was performed
on F2. All non-hydrogen atoms were refined anisotropically. The

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
617
Table 8
Summary of crystal data and refinement results.
C8OL
[(C8OL)2Fe](FeCl4)2
trans-(C8OL)Ru(CO)Cl2ꢁCH3CN
(HOL)NiCl2ꢁ1.5CH3OH
Empirical formula
C
27H29N5O
C54H58Cl8N10O2Fe3
1330.25
C30H32Cl2N6O2Ru
680.59
C20.5H19Cl2N5NiO2.5
505.02
Molecular weight (g molꢀ1
Crystal dimensions (mm)
Crystal shape, color
Crystal system
Space group
)
439.55
0.24 ꢂ 0.16 ꢂ 0.12
block, colorless
triclinic
0.35 ꢂ 0.30 ꢂ 0.30
block, purple
triclinic
0.25 ꢂ 0.05 ꢂ 0.02
needle, dark green
triclinic
0.10 ꢂ 0.08 ꢂ 0.04
plate, pale green
triclinic
ꢀ
ꢀ
ꢀ
ꢀ
P1
P1
P1
P1
a (Å)
b (Å)
c (Å)
a
8.7734(4)
9.1450(7)
15.3210(7)
103.160(2)
94.510(2)
102.610(I)
1144.73(9)
2
11.0848(2)
15.5053(3)
18.7202(4)
98.466(1)
98.315(1)
109.520(1)
2934.0(1)
2
11.4711(4)
11.7394(3)
12.5506(5)
82.640(2)
76.6333(14)
66.408(2)
1505.69(9)
2
7.8610(3)
11.8140(5)
11.8790(5)
104.730(2)
106.410(2)
92.654(2)
1015.15(7)
2
(°)
b (°)
c
(°)
V (Å3)
Z
Dcalc (g cmꢀ3
)
1.275
0.080
1.506
1.146
1.501
0.74
1.652
1.252
l
(mmꢀ1
)
F(0 0 0)
h Range (°)
Limiting indices
468
1364
696
518
2.6–25.0
ꢀ10 < h < 10
ꢀ10 < k < 10
ꢀ18 < l < 17
9466/4038/2935
99.2
2.8–25.0
ꢀ13 < h < 13
ꢀ18 < k < 17
ꢀ22 < l < 21
32355/10178/8855
98.9
2.6–27.5
ꢀ14 < h < 14
ꢀ15 < k < 15
ꢀ15 < l < 16
15252/6739/4624
97.4
3.2–25.1
ꢀ9 < h < 9
ꢀ14 < k < 14
ꢀ14 < l < 14
12274/3584/2710
98.7
Reflections collected/unique reflections/for I > 2
Completeness to h (%)
r(I)
Maximum and minimum transmission
Data/parameters/restraints
0.981, 0.990
4039/300/0
1.06
0.676, 0.710
10178/696/0
1.06
0.839, 0.988
6739/372/0
1.02
0.643, 0.980
3584/285/7
1.08
Goodness-of-fit (GOF) on F2
R1 for F2 > 2
r
(F2)
(F2)a
0.0474
0.0404
0.0518
0.0634
wR2 for F2 > 2
R1 for all data
Largest residual difference in peak and hole (e Åꢀ3
r
0.125b
0.101c
0.126d
0.152e
0.0728
0.20, ꢀ0.22
0.0485
1.31, ꢀ1.12
0.0912
1.14, ꢀ0.89
0.0874
0.906, ꢀ0.746
)
(a) w = 1/[r
2(F2o ) +(pP)2 + qP] where P = (F2o + 2F2c )/3 and where (b) p = 0.0481, q = 0.167, (c) p = 0.0490, q = 3.3547, (d) p = 0.0376, q = 3.5152, or (e) p = 0.0869, q = 1.1106.
hydrogen atoms were assigned idealized positions according to a
riding model and refined isotropically.
of the solvent under vacuum, the residue was partitioned between
alkalized H2O and CH2Cl2. The organic phase was dried (MgSO4)
and rid of solvent to provide 1.12 g (90% yield) of cream-coloured
4-(4-(1-octoxy)phenyl)-2,6-di(pyrazin-2-yl)pyridine. The neutral
HOL could also be used instead of the KOL salt in this reaction.
1H NMR (CDCl3): d 9.86 (s, 2H), 8.69 (s, 2H), 8.65 (2s, 4H), 7.82
(d, 2H, J = 8.9 Hz), 7.03 (d, 2H, J = 8.9 Hz), 4.03 (t, 2H, J = 6.4 Hz),
1.83 (quintet, 2H, J = 6.9 Hz), 1.49 (m, 2H), 1.31–1.35 (m, 8H),
0.90 (t, 3H, J = 5.9 Hz) ppm. 13C NMR (CDCl3): d 160.4, 154.2,
150.9, 150.2, 144.7, 143.6, 143.5, 129.7, 128.4, 119.1, 115.0, 68.2,
31.8, 29.4, 29.24, 29.22, 26.0, 22.7, 14.11 ppm. 15N NMR (CDCl3):
d 333 (N-10), 322 (N-40) ppm (N-1 was not detected). Anal. Calc.
for C27H29N5O: C, 73.78; H, 6.65; N, 15.93. Found: C, 73.68; H,
6.82; N, 15.88%.
8.3. Synthetic procedures
8.3.1. 4-(4-Hydroxyphenyl)-2,6-di(2-pyrazinyl)pyridine, potassium
salt (KOL)
2-Acetylpyrazine (155.2 mg, 1.28 mmol), 4-hydroxybenzalde-
hyde (75.8 mg, 0.62 mmol), concentrated NH4OH (0.4 mL) and
15% aqueous KOH (3.6 mL) were heated to reflux for 8 h. The pre-
cipitate was collected, rinsed with CHCl3 and dried to yield 162 mg
(61% yield) of HOL as its orange hydrated K+ salt. 1H NMR
(CD3SO2CD3): d 9.79 (s, 2H), 8.78 (s, 2H), 8.73 s, 2H), 8.46 (s, 2H),
7.46 (d, 2H, J = 7.9 Hz), 6.18 (d, 2H, J = 7.9 Hz) ppm. EI-MS m/z
327; C19H13N5O requires 327.4. Anal. Calc. for C19H12N5OKꢁ3.5H2O:
C, 53.26; H, 4.47; N, 16.34. Found: C, 53.45; H, 4.29; N, 15.95%.
8.3.4. [(C8OL)2Ru](PF6)2
C8OL (100 mg, 0.23 mmol) and RuCl3ꢁ3H2O (29.7 mg,
0.11 mmol) were heated in ethylene glycol (15 mL) at reflux for
20 h under Ar. The dark orange mixture was cooled to room
temperature, then treated with excess aq. NH4PF6. The precipitate
was collected and washed with H2O, then dried under vacuum,
yielding 133 mg of red solid (95%). 1H NMR (CD3CN): d 9.76 (s,
2H), 9.15 (s, 2H), 8.39 (s, 2H), 8.23 (d, 2H, J = 7.9 Hz), 7.53 (s, 2H),
7.33 (d, 2H, J = 7.9 Hz), 4.21 (t, 2H, J = 6.4 Hz), 1.89 (quintet,
2H, J = 6.9 Hz), 1.56 (m, 2H), 1.37–1.42 (m, 8H), 0.95 (t, 3H, J =
6.9 Hz) ppm. ESI-MS: m/z 490.3 (cluster max., C54H58N10O2Ru2+ re-
quires 490.2). Anal. Calc. for C54H58N10O2P2F12Ru: C, 51.07; H, 4.60;
N, 11.03. Found: C, 52.06; H, 4.69: N, 10.58%.
8.3.2. 4-(4-Hydroxyphenyl)-2,6-di(2-pyrazinyl)pyridine (HOL)
The KOL salt (1.03 g, 2.4 mmol) was finely divided and sus-
pended in a warm (ca. 35 °C) 200 mL solution of 5% HOAc, and stir-
red for 10 min. The suspended product was then suction-filtered,
and this treatment was repeated an additional four times with
fresh acid, rinsing the final solid with copious amounts of cold,
doubly distilled H2O. The light tan product was dried under vac-
uum at room temperature, and a flame test of the dried material
revealed no potassium. 1H NMR (DMSO-d6): d 9.97 (s, 1H), 9.86
(s, 2H), 8.80 (m, 4H), 8.63 (s, 2H), 7.81 (d, 2H, J = 8.4 Hz), 6.97 (d,
2H, J = 8.3 Hz) ppm. Anal. Calc. for C19H13N5O: C, 69.71; H, 4.00;
N, 21.39. Found: C, 69.96; H, 4.39; N, 20.75.
8.3.5. Trans- and cis-(C8OL)Ru(CO)Cl2
8.3.3. 4-(4-(1-Octoxy)phenyl)-2,6-di(2-pyrazinyl)pyridine (C8OL)
The KOL salt (1.21 g, 2.8 mmol) was dissolved in freshly distilled
DMSO (10 mL), treated with NaH (0.18 g, 7.5 mmol) and 1-bromo-
octane (0.78 g, 4.0 mmol), and left stirring overnight. After removal
RuCl3ꢁ3H2O (65.8 mg, 0.32 mmol) was heated to reflux in 15 mL
DMF until the dark black-brown color changed to a transparent red
color (0.5–3 h), then C8OL (137.6 mg, 0.31 mmol) was added. An
immediate green color was observed. Heating was continued for

618
C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
8 h. The addition of H2O produced a black precipitate (89% crude
yield), which was purified by column chromatography on silica
gel, using a 167:1:2 mixture of CH3CN, saturated aq. KNO3 and
H2O as eluent, to afford a black solid in 80% yield. 1H NMR (CDCl3,
high-dilution limit): d 9.34 (bs, 2H), 9.01 (s, 2H), 8.63 (bs, 2H), 8.39
(s, 2H), 7.75 (d, 2H, J = 9 Hz), 7.17 (d, 2H, J = 9 Hz), 4.11 (bt, 2H),
1.86 (m, 2H), 1.62–1.38 (m, 10H), 0.96 (bt, 3H) ppm. 13C NMR
(CDCl3): d 200.4, 161.9, 153.1, 152.3, 150.6, 147.3, 143.5, 129.0,
124.8, 118.1, 114.8, 68.4, 31.9, 29.4, 29.4, 29.0, 25.9, 22.8,
Acknowledgements
We thank the Natural Sciences and Engineering Research Coun-
cil for financial support, Johnson Matthey Inc. for the generous loan
of ruthenium trichloride, Dr. Alan Lough for acquisition of the
X-ray diffraction data, and structure solutions for C8OL, [(C8OL)2Fe]
(FeCl4)2 and (HOL)NiCl2 1.5CH3OH, Dr. Howard Hunter for the 15N
NMR data, Patrick Lewis for the C8OL and [(C8OL)2Fe](FeCl4)2 crys-
tals, and Leonid Chagal for the work with (HOL)NiCl2. Computing
was carried out by courtesy of the Shared Hierarchical Academic
Research Computing Network [SHARCNET, Ontario, Canada].
15
14.2 ppm. N NMR (CDCl3): d 272 (N-1), 272 (N-10), 335 (N-40)
ppm. Anal. Calc. for C28H29Cl2N5O2Ruꢁ0.5H2O: C, 51.86; H, 4.66;
N, 10.80. Found: C, 51.94; H, 4.66; N, 10.72%. mCO (KBr)
1979 cmꢀ1. FAB-MS: m/z (%) 639 (39, M), 611(94, M–CO), 604
(64, M–Cl), 576 (100, M–CO–Cl). Small, dark green needles of
trans-(C8OL)Ru(CO)Cl2ꢁCH3CN suitable for XRD were formed by
allowing a hot CH3CN solution to cool to room temperature and
stand for 24–48 h.
When column chromatography was performed on neutral alu-
mina, packed with dichloromethane and using a mixture of CH3CN
and CH2Cl2 as eluent, the dark green band of trans isomer (48%
yield) was followed by a yellow band of the cis isomer (19% yield).
1H NMR (CDCl3): d 9.35 (s, 2H), 9.17 (dd, 2H, J = 0.9, 2.9 Hz), 8.77 (d,
2H, J = 2.9 Hz), 8.36 (s, 2H), 7.79 (d, 2H, J = 8.8 Hz), 6.98 (d,
2H, J = 8.8 Hz), 4.05 (t, 2H, J = 6.7 Hz), 1.87 (m, 2H), 1.33 (m, 10H),
0.91 (m, 3H) ppm. 13C NMR (CDCl3): d 188.9, 162.1, 155.9, 153.5,
149.6, 148.6, 148.1, 143.6, 129.1, 126.5, 119.9, 115.7, 68.6, 31.9,
29.4, 29.3, 29.1, 26.1, 22.8, 14.1 ppm. 15N NMR (CDCl3): d 247 (N-
1), 253 (N-10), 337 (N-40) ppm. Anal. Calc. for C28H29N5O2R-
uCl2ꢁ0.5CH3CN: C, 52.77; H, 4.66; N, 11.67. Found: C, 52.69; H,
4.73; N, 11.63%. mCO (KBr) 1977 cmꢀ1. LDI-MS: m/z 611.1 (M–CO),
576.1 (M–CO–Cl).
Appendix A. Supplementary material
Tables of crystal data and refinement results, CCDC Codes:
[Fe(C8OL)2][FeCl4] – 2822236, trans-Ru(C8OL)(CO)Cl2 – 822237,
Ni(HOL)Cl2-1.5MeOH – 822238, C8OL – 822239. More complete
Tables of TD-DFT predicted electronic spectra. Supplementary data
associated with this article can be found, in the online version, at
doi:10.1016/j.ica.2011.04.018.
References
[1] R. Liegghio, P.G. Potvin, A.B.P. Lever, Inorg. Chem. 40 (2001) 5485.
[2] S. Vaduvescu, P.G. Potvin, Eur. J. Inorg. Chem. (2004) 1763.
[3] M. Abbouda, D. Kalinina, P.G. Potvin, Inorg. Chim. Acta 362 (2009) 4953.
[4] F. Al-mutlaq, P.G. Potvin, Eur. J. Inorg. Chem. (2007) 2121.
[5] D.D. Choudhury, R.F. Jones, G. Smith, D.J. Cole-Hamilton, J. Chem. Soc., Dalton
Trans. (1982) 1143.
[6] G.B. Deacon, J.M. Patrick, B.W. Skelton, N.C. Thomas, A.H. White, Aust. J. Chem.
37 (1984) 929.
[7] T.T. Ben-Hadda, C. Mountassir, H. Le Bozec, Polyhedron 14 (1995) 953.
[8] B.P. Sullivan, J.M. Calvert, T.J. Meyer, Inorg. Chem. 19 (1980) 1404.
[9] R. Ziessel, V. Grosshenny, M. Hissler, C. Stroh, Inorg. Chem. 43 (2004)
4262.
8.3.6. [(C8OL)2Fe](FeCl4)2 and [(C8OL)2Fe](PF6)2
Ferric chloride hydrate (0.20 g, 0.74 mmol) dissolved in CH3OH
(25 mL) was added dropwise over the course of 20 min with con-
[10] G. Otting, B.A. Messerle, L.P. Soler, J. Am. Chem. Soc. 119 (1997) 5425.
[11] C.E. Schäffer, Struct. Bonding 5 (1968) 67.
[12] C.E. Schäffer, C. Anthon, J. Bendix, Aust. J. Chem. 62 (2009).
[13] K.R. Kittilstved, A. Hauser, Coord. Chem. Rev. 254 (2010) 2663.
[14] C.E. Schäffer, C. Anthon, J. Bendix, Coord. Chem. Rev. 253 (2009) 575.
[15] A.E. Reed, L.A. Curtiss, F. Weinhold, Chem. Rev. 88 (1988) 899.
[16] I. Mayer, Chem. Phys. Lett. 97 (1983) 270.
[17] A.J. Bridgeman, G. Cavigliasso, L.R. Ireland, J. Rothery, J. Chem. Soc., Dalton
Trans. (2001) 2095.
[18] M.-N. Collomb-Dun-Sauthier, A. Deronzier, R. Ziessel, Inorg. Chem. 33 (1994)
2961.
[19] S. Chardon-Noblat, P.D.A. Da Costa, S. Maniguet, R. Ziessel, J. Electronal. Chem.
529 (2002) 135.
[20] H. Nagao, T. Mizukawa, K. Tanaka, Inorg. Chem. 33 (1994) 3415.
[21] S. Chardon-Noblat, A. Deronzier, R. Ziessel, D. Zsoldos, Inorg. Chem. 36 (1997)
5384.
[22] H. Ishida, K. Tanaka, T. Tanaka, Organometallics 6 (1987) 181.
[23] A. Vlcek, S. Zalis, Coord. Chem. Rev. 251 (2007) 258.
[24] F.M. Bickelhaupt, E.J. Baerends, in: K.B. Lipkowitz, D.R.E. Boyd (Eds.), Reviews
in Computational Chemistry, vol. 15, Wiley, New York, 2000, pp.
1–86.
stant stirring to
a CH2Cl2 solution (25 mL) containing C8OL
(0.21 g, 0.48 mmol). The solution changed colour from a cloudy
cream colour, to orange, to red then to black, as the ferric chloride
solution was added. The black solution was stirred at room tem-
perature for a further 48 h, after which, the solvent was removed,
and the collected black solid was re-dissolved in hot DMF. Upon
cooling, an excess of ethyl ether was added to the purple solution,
which was subsequently filtered after standing for 24 h, to remove
a white precipitate. The solvents were removed, and [(C8OL)2Fe]
(FeCl4)2 was collected as a purple/black solid. Crystals suitable
for X-ray diffraction were obtained by dissolving the precipitate
in DMF, diluting in EtOH and layering with a small amount of
CH2Cl2.
Conversion to the hexafluorophosphate salt was achieved by
dissolving the purple solid in CH3CN and added dropwise to an
aqueous solution of NH4PF6. The purple precipitate was filtered,
re-dissolved in CH3CN, and treated twice more with NH4PF6. A pur-
ple powder of analytical purity was finally collected (0.16 g, 55%
yield). 1H NMR (CD3CN): d 9.66 (s, 2H), 9.25 (s, 2H), 8.30 (d, 2H,
J = 8.8 Hz), 8.26 (d, 2H, J = 3.0 Hz), 7.34 (d, 2H, J = 8.8 Hz), 7.3 (d,
2H, J = 2.5 Hz), 4.21 (t, 2H, J = 6.5 Hz), 1.55 (m, 2H), 1.40 (m, 8H),
1.27 (m, 2H), 0.93 (m, 3H) ppm. LDI-MS: m/z 934.4 (FeL2+),
440.25 (L+). Anal. Calc. for C54H58N10O2FeP2F12: C, 52.9; H, 4.8; N,
11.5. Found: C, 52.81; H, 5.12; N, 10.89%.
[25] J. Li, L. Noodleman, D.A. Case, in: E.I. Solomon, A.B.P. Lever (Eds.), Inorg.
Electron. Struct. Spectrosc., vol. 1, John Wiley Sons, New York, 1999.
[26] A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, A.V. Zelewsky, Coord.
Chem. Rev. 84 (1988) 85.
[27] S.I. Gorelsky, A.B.P. Lever, Can. J. Anal. Sci. Spectrosc. 48 (2003) 93.
[28] A.B.P. Lever, Inorganic Electronic Spectroscopy, Elsevier Science, Amsterdam,
1984.
[29] E.S. Dodsworth, M. Hasegawa, M. Bridge, W. Linert, in: J.A. McCleverty, T.J.
Meyer (Eds.), Comprehensive Coordination Chemistry II, vol. 1, Elsevier, 2003.
pp. 351–365.
[30] E.S. Dodsworth, A.B.P. Lever, Coord. Chem. Rev. 97 (1990) 271.
[31] H.E. Toma, A.B.P. Lever, Inorg. Chem. 25 (1986) 176.
[32] H.E. Toma, P.S. Santos, A.B.P. Lever, Inorg. Chem. 27 (1988) 3850.
[33] A.B.P. Lever, P.G. Potvin, Unpublished Observations, 2011.
[34] N.G. Connelly, W.E. Geiger, Chem. Rev. 96 (1996) 877.
[35] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
8.3.7. (HOL)NiCl2
[36] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[37] T.H. Dunning Jr., P.J. Hay, in: H.F. Schaefer (Ed.), Modern Theoretical Chemistry,
vol. 3, Plenum, New York, 1976, p. 1.
[38] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 270.
[39] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 284.
[40] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299.
A solution of HOL (50 mg) in CH3OH was added to a CH3OH
solution of NiCl2ꢁ6H2O (42 mg). An orange-red precipitate was col-
lected by filtration, rinsed and dried (50% yield). Pale green plates
of (HOL)NiCl2ꢁ1.5CH3OH were grown from a methanol solution.

C. Dares et al. / Inorganica Chimica Acta 374 (2011) 606–619
619
[41] 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. Sonnenberg, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro,
M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M.
Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth,
P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman,
J.V. Ortiz, J. Cioslowski, D.J. Fox, GAUSSIAN 09, Revision B.01, Gaussian, Inc.,
Wallingford, CT, 2009.
[42] S.J.A. Van Gisbergen, E.J. Baerends, in: J.A. McCleverty, T.J. Meyer (Eds.),
Comprehensive Coordination Chemistry II, Elsevier, Oxford, 2003, pp. 511–517.
[43] S.I. Gorelsky, SWIZARD
<http://www.sg-chem.net/>.
[44] S.I. Gorelsky, AOMIX CDA Program, 2005, <http://www.sg-Chem.Net/>.
, CCRI, University of Ottawa, Ottawa, Canada, 2008,
-
[45] C. Dares, SWARLOCK, 2010, <http://www.ballparkchemistry.com>.
[46] R.H. Blessing, Acta Crystallogr., Sect. A 51 (1995) 33.