Ruthenium(II) complexes of carboxylated terpyridines and dipyrazinylpyridines

Article abstract of DOI:10.1002/ejic.201000122

One-pot preparations of carboxylated 2,2′6′,2Prime;-terpyridine and 2,6-dipy:razin-2-ylpyridine ligands are reported, in free acid and ester forms, as well as their transformations to homoleptic Ru^II complexes in very good yields. Density functional theory calculations of their geometry-optimized structures enabled their comparison with crystal, structures; although the gross features of the crystal structures were reproduced, they showed much variability and distortion. The spectroscopic and electrochemical properties were analyzed from, the point of view of their geometry-optimized, structures, molecular orbital characteristics and electronic transitions, specifically to assess the effects of carboxylation and of inserting a phenylene spacer between the tridentate portion and the carboxyl group. The dipyrazinylpyridine complexes had HOMO levels stabilized by approximately 1 eV relative to the terpyridine analogues, and showed positive-shifted electrochemical potentials and redshifted electronic absorptions. Phenylene spacers were found to act as electron-donating groups, and the lowest-energy UV/Vis transitions showed intraligand character.

Full text of DOI:10.1002/ejic.201000122

FULL PAPER
DOI: 10.1002/ejic.201000122
Ruthenium(II) Complexes of Carboxylated Terpyridines and
Dipyrazinylpyridines
Arta Stublla[a] and Pierre G. Potvin*[a]
Keywords: Ruthenium / Ligand effects / Density functional calculations / Carboxylation / Electronic structure
One-pot preparations of carboxylated 2,2Ј;6Ј,2ЈЈ-terpyridine
and 2,6-dipyrazin-2-ylpyridine ligands are reported in free
acid and ester forms, as well as their transformations to
homoleptic RuII complexes in very good yields. Density func-
tional theory calculations of their geometry-optimized struc-
tures, molecular orbital characteristics and electronic transi-
tions, specifically to assess the effects of carboxylation and of
inserting a phenylene spacer between the tridentate portion
and the carboxyl group. The dipyrazinylpyridine complexes
had HOMO levels stabilized by approximately 1 eV relative
tures enabled their comparison with crystal structures; al- to the terpyridine analogues, and showed positive-shifted
though the gross features of the crystal structures were repro-
duced, they showed much variability and distortion. The
spectroscopic and electrochemical properties were analyzed
from the point of view of their geometry-optimized struc-
electrochemical potentials and redshifted electronic absorp-
tions. Phenylene spacers were found to act as electron-donat-
ing groups, and the lowest-energy UV/Vis transitions
showed intraligand character.
Unless a ligand is attached to the electrode surface, its
photoinduced dissociation is a concern with respect to the
Introduction
Carboxylated RuII complexes are of interest in several long-term stability of photoelectrodes and, in this regard,
contexts. We have reported how carboxyl groups can supra- tridentate ligands are preferred over bidentate ones. They
molecularly assist electron transfers from a photoexcited are, moreover, nonstereogenic. However, terpyridine (tpy)
complex to a cationic electron acceptor.[1] The Hanan group complexes have notoriously short excited-state lifetimes in
used carboxyl groups on such complexes to bridge RhI di- comparison with their bidentate analogues, as exemplified
mers in unique multinuclear assemblies.[2] Constable et al. by the parent complexes [Ru(tpy)2]2+ (sub-nanosecond
studied the solid-state hydrogen-bonding organization of timescale)[5] and [Ru(bpy)3]2+ (millisecond timescale; bpy =
two examples.[3] We also have an interest in using carboxyl- bipyridine).[6,7] Pyrazine-containing ligands have long been
ated RuII complexes and their derivatives in the self-as- known,[8] but the first dipyrazinylpyridine (dpp) was re-
sembly of linear oligomers.
ported in 2001,[9] and we found that its homoleptic RuII
However, carboxylated complexes of RuII are currently complex showed a longer excited-state lifetime than its ter-
of greatest interest as photovoltaic sensitizers, with so- pyridine analogue and faster rates of photoinduced electron
called N3 and “black dye” as the best known examples.[4] transfer in homogeneous solution.[10] This was attributed
In this context, the carboxyl groups serve to anchor the mainly to the presence of the additional nitrogen atoms.
complexes to the surfaces of the photoelectrodes, most The electron injections from photoexcited states into a con-
commonly consisting of semiconducting anatase (TiO2) duction band of a semiconductor have been found to be
coatings on indium/tin oxide conducting glass. The carb- extremely rapid (femtosecond timescale),[11] and this would
oxyl groups are also expected to affect the electronic struc- seem to diminish the importance of the excited-state life-
tures of the complexes and, by extension, their electrochem- time. Nevertheless, electron-withdrawing appendages have
ical and photophysical properties and, therefore, the driving been used to lengthen the lifetimes, albeit with a lower net
force for electron injection, as well as the degree of elec- exploitable excited-state energy.
tronic coupling between sensitizer and conduction band.
Parsing those effects is not a simple matter.
To address some of these questions, we have launched a
limited but systematic study of carboxylated RuII complexes
of terpyridines and dipyrazinylpyridines and their ester de-
rivatives, with and without a spacer separating the carboxyl
groups from the core ligand to mitigate the electronic im-
pact of carboxylation. This will entail studies of homoleptic
and heteroleptic sensitizers, in solution as well as adsorbed
on surfaces. We report here the preparation of carboxylated
terpyridine and dipyrazinylpyridine ligands and their
[a] Department of Chemistry, York University,
4700 Keele Street, Toronto, ON M3J 1P3, Canada
Fax: +1-416-736-5936
E-mail: pgpotvin@yorku.ca
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201000122.
3040
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 3040–3050

Complexes of Carboxylated Terpyridines and Dipyrazinylpyridines
homoleptic complexes, their spectroscopic and electrochem- dines.[9,10,16,17] This precipitates a salt form which, after re-
ical characterization in solution, and analyses based on rel- dissolving, was acidified to precipitate analytically pure H1
evant electronic structure computations using density func- in 80% yield. The salt form was also transformed to the
tional theory (DFT).
ethyl ester Et1 by standard Fischer esterification in 70%
yield. To prepare the corresponding complexes, RuCl3 was
pre-activated with AgBF4 in dmf, then treated with the li-
gand to afford, after precipitation by anion exchange, pure
[Ru(H1)2](PF6)2 (95% yield) or [Ru(Et1)2](PF6)2 (98%
yield).
Results and Discussion
Synthesis
Some of the terpyridine materials described here have
been reported previously or were reported during the course
of our own work, although the yields, the synthetic ease,
the ease of isolation and the levels of characterization that
have been reported were not always satisfactory. None have
previously been fully characterized in terms of spectral and
electrochemical properties. The dpp materials reported are
all new. There are two general approaches that have been
followed in the preparation of RuII complexes of carboxyl-
ated terpyridines, either by direct synthesis from the free
ligands or by the modification of precursor complexes, and
our work has used both approaches.
The terpyridinylbenzoic acid H1 has been prepared in
18% yield in boiling acetamide,[12] and was reported as the
dmso solvate of the HBr disalt. It was later prepared by a
two-step Kröhnke method, with the second step proceeding
in 65% yield (first step not described), and incorporated
into an IrI complex.[13] A third preparation used a process
similar to ours (49% yield) and converted H1 to neutral,
zwitterionic Ru(1)2 (66% yield),[3] the pentahydrate of
which was characterized by using crystallography. Finally,
the ligand has been prepared by base hydrolysis (93%
Scheme 1. Ligand synthesis: (i) OHCC6H4-4-COOH or
OHCCOOEt, 15% KOH, concd. NH4OH/CH3OH/H2O (1:1)/r.t./
3 d; then dilute HCl; (ii) EtOH, concd. H2SO4/reflux/2.5–3 d; (iii)
(a) RuCl3, AgBF4/dmf/reflux/3 h; (b) add L/reflux/3 d; (c) excess
amount of NH4PF6/H2O.
The free ligand lacking the spacer, terpyridine-4Ј-carbox-
yield)[14] of the methyl ester, itself obtained in two steps ylic acid H2, was first isolated as an unintended byproduct
(15% yield),[15] and incorporated into a heteroleptic RuII of a convergent preparation of its ethyl ester by Stille cou-
species. We had earlier prepared both homo- and heterolep- pling, in 4 steps and 28% overall yield from commercial
tic RuII complexes of H1 by KMnO4 oxidation of the corre- citrazinic acid.[18] Later, H2 was prepared by means of 4-
sponding homoleptic 4Ј-(p-tolyl) complex, albeit as a mix- and 5-step processes featuring two methods of oxidizing 4Ј-
ture requiring chromatography, in low yields of each, and methylterpyridine, in 18 and 16% yields, respectively, then
in an incompletely characterized state.[1] In accord with converted to [Ru(H2)2]2+ in 55% yield by heating with
Constable et al.,[3] part of the difficulty we encountered in [Ru(dmso)4Cl2], but none of the products or intermediates
1
manipulating these products was their variable protonation were characterized by anything more than H NMR spec-
state. The methyl ester of H1 (Me1) has also been prepared troscopy and sometimes MS.[14] Constable et al. prepared
in 25% yield after 8 d.[3] Using microwave irradiation in [Ru(H2)2]2+ in 48% yield by oxidative degradation of a 4Ј-
high-boiling solvent, Constable et al. converted this ester to furyl precursor, which itself was available in 86% yield.[3]
its homoleptic RuII complex, [Ru(Me1)2]2+, isolated as the This novel approach was based on earlier work by Beley et
–
PF6 disalt in 73% yield. Alternatively, autoclave reaction al.,[19] who prepared a heteroleptic RuII complex of H2 by
of the ester in MeOH at 150 °C provided the dichloride salt means of a similar oxidation. These authors had also pre-
(66% yield), characterized as a pentahydrate by using ele- pared the same complex directly from H2, but only in 6%
mental analysis but as a tetrahydrate by using crystallogra- yield, as well as by hydrolysis of the ester analogue. That
–
phy. The PF6 salt was hydrolyzed to a variable mixture of process was sluggish (18% yield) and was accompanied by
[Ru(H1)2]2+ and [Ru(1)(H1)]+ (61% yield) lacking confirm- ligand scrambling to also afford [Ru(H2)2]2+ as one of two
ation by elemental analysis.[3] Heteroleptic RuII complexes byproducts (3% yield), but none of these materials was
of H1 have also been described.[2,14]
completely characterized. The fully deprotonated Ru(2)2 re-
In this work, we used a one-pot synthesis of H1 at room sulted from an attempt to prepare a coordination polymer
temperature from 2-acetylpyridine and 4-carboxybenzalde- from [Ru(H2)2]2+ and Zn2+, and was only characterized by
hyde (the ethyl ester of which can be used just as well) in crystallography as the tetrahydrate.[3]
the presence of KOH and NH4OH in aqueous MeOH
In a similar fashion to that for H1, we prepared H2 in
(Scheme 1), a process that we have previously used to pre- one step from 2-acetylpyridine and ethyl glyoxylate and iso-
pare several 4Ј-substituted terpyridines and dipyrazinylpyri- lated it as the hydrochloride salt in 60% yield. Conversion
Eur. J. Inorg. Chem. 2010, 3040–3050
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3041

A. Stublla, P. G. Potvin
FULL PAPER
of the ester Et2[28] to [Ru(Et2)2](PF6)2 as before was uncom- has been amply described in the literature, and the relatively
plicated (82% yield) but not so with H2. The direct reaction simple spectra obtained with our high-symmetry complexes
to produce [Ru(H2)2](PF6)2 worked well and provided high- will not be detailed here, apart to report that they showed
purity material, the 1H NMR spectra of which matched the diagnostic upfield shift of the narrow 6/6ЈЈ-H doublets
those reported,[3,14] but drying the product to prepare an and 5/5ЈЈ-H doublets of doublets upon complexation, ow-
analytical sample resulted in a partial loss of HPF6, ing to the magnetic anisotropy of the perpendicular ligand.
1
whereby the material turned a darker shade of red and be- The H NMR spectroscopy of dipyrazinylpyridines differs
came insoluble in common solvents. A reliable measure of from that of terpyridines because of small coupling con-
the yield was therefore not possible. This precipitation was stants, such that the aromatic signals can appear as broad
reversed by stirring with aqueous HPF6 but no water-free singlets at low resolution, and at generally more downfield
sample could be prepared in this way. Instead, [Ru(H2)2]- positions. The 13C NMR spectra of RuII complexes are ra-
(PF6)2 was obtained analytically pure and water-free by rely reported but were obtained for all species. For both
base hydrolysis of [Ru(Et2)2](PF6)2 and precipitation with ligands and complexes, these were consistent with the struc-
acid (100% yield).
tures and, by use of distortionless enhancement by polariza-
By following the same one-pot procedure as that for H1 tion transfer (DEPT), heteronuclear single quantum coher-
and H2, the new ligands H3 and H4 were obtained from 2- ence (HSQC) and heteronuclear multiple bond correlation
acetylpyrazine in 80 and 55% yields, respectively. Conver- (HMBC) pulse sequences, all 13C NMR spectroscopic sig-
sion to Et3 (76% yield), [Ru(H3)2](PF6)2 (58% yield), nals were assigned. Because the spectra were not all ob-
[Ru(Et3)2](PF6)2 (30% yield) or [Ru(Et4)2](PF6)2 (38% tained in the same solvent, comparisons between them must
yield) was uncomplicated. As was the case with H2, a direct be made with caution. In comparing the spectra of the es-
preparation of [Ru(H4)2](PF6)2 provided material that had terified species, the 13C NMR spectroscopic signals gen-
1H NMR spectra consistent with the structure, but the ele- erally migrate downfield upon complexation (by up to
mental analyses were not. Instead, [Ru(H4)2](PF6)2 was ob- 3.8 ppm), including those for C-6/6ЈЈ and C-5/5ЈЈ. In other
tained analytically pure by base hydrolysis of [Ru(Et4)2]- words, the magnetic anisotropy felt by 6/6ЈЈ-H and 5/5ЈЈ-H
(PF6)2 and precipitation with acid (38% yield).
was not felt as strongly by the attached carbon nuclei. The
The yields of these dpp materials were generally lower only exceptions were upfield-migrating signals from those
than those of terpyridine analogues, which we ascribe to the carbon atoms lying along the symmetry axes of the ligands
generally lower reactivity of dipyrazinylpyridines in forming (through C-4Ј and on the side-chain), by 0.3–2.1 ppm, as
complexes, as well as difficulties in purifying and handling well as those for the methyl carbon atoms (by 0.7 ppm) but
the products. Although [Ru(Et3)2](PF6)2 was obtained es- not the CH2 carbon atoms. Analogous changes were found
sentially pure without chromatography, an analytically pure in comparison of the spectra of free carboxylic acid ligands
sample was obtained after preparative TLC. All of the car- and their complexes, although additional changes were at-
boxylic acid complexes tended to lose the elements of HPF6 tributable to the partially protonated states of the free li-
while in solution, as noted with other deprotonatable com- gands.
plexes,[1,20,21] turning darker and becoming insoluble, but
were redissolvable upon stirring with aqueous HPF6.
1
All products were characterized by a combination of H
Molecular Structure
and 13C NMR spectroscopy, IR spectroscopy, MS and ele-
mental analysis for new materials. The complete NMR
The complexes reported here were red, microcrystalline
spectroscopic signal assignments, supported by 2D experi- and stable as solids, but all attempts at crystal growth were
ments, and the structures assigned to the mass spectral ions unsuccessful. In particular, the prolonged standing of solu-
appear in the Supporting Information.
tions of the carboxylic forms resulted in the deposition of
The EI-MS mass spectra of the free ligands and their insoluble and powdery deprotonated forms, and the crystal
esters showed the molecular ions and the expected fragment structures of the deprotonated forms Ru(1)2 or Ru(2)2 have
ions. The monocation-selective LDI-TOF mass spectromet- already been reported.[3] Therefore, DFT-optimized struc-
ric technique was used for the complexes, which all exhib- tures were obtained for all the complexes reported here, and
ited the expected isotopic clustering at m/z values that were these provided insights into their electronic structures (see
in agreement with the assigned ion structures. The highest- below). Except for the obvious differences owing to the re-
mass clusters were variably of formula [M – 2 PF6]+, [M – placement of pyridine rings with pyrazine rings, the struc-
PF6]+ or [M]+. With [Ru(H1)2]2+ and [Ru(H2)2]2+, overlap- tures of the dpp complexes were virtually the same as those
ping clusters of ions differing by one mass unit were found. of the terpyridine analogues. The Supporting Information
In addition, the complexes yielded cascades of fragment reports some measurements. The crystal structures of six
ions from the predictable losses of side-chain components.
Because of their zwitterionic nature, the carboxylic acid nity to compare solid-phase and computed structures.
complexes were not soluble in ordinary solvents, so their In terms of the flatness of the tridentate moieties, there
relevant terpyridine complexes[2,3] also gave us the opportu-
NMR spectra were acquired in deuterated trifluoroacetic was an important difference. All six crystal structures
acid ([D]tfa) and thus reflect partially protonated forms. showed deviations in the coplanarity of the three pyridine
1
The H NMR spectroscopy of (terpyridine)RuII complexes rings in each terpyridine ligand (pyridine planes twisted by
3042
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 3040–3050

Complexes of Carboxylated Terpyridines and Dipyrazinylpyridines
up to 7°), which has been noted in earlier examples of ter- in an [Ru(1)(tpy)]-linked Rh dimer.[2] Both crystalline
pyridine complexes.[22] This was most pronounced with [Ru(2)2]·4H2O[3] and the computed structures of H2 and H4
Ru(2)2, and was coupled to a strong bowing of the meridio- showed analogous twists of the COO moieties relative to
nal planes by the Ru–pyridine–phenylene–carboxylate se- the central pyridine ring.
quences, no doubt because of packing pressures. In the
There were additional small distortions: Some examples
DFT-calculated structures of our complexes, any such devi- showed an in-plane bend of the pyridine-to-carboxyl or
ation from terpyridine coplanarity was very slight (1° or phenylene-to-carboxyl linkages off the central axis by 2–
less). There was no evident sign of a particular distortion 3.5°, perhaps to accommodate the OH/OEt bulk. Several
of the computed pyridine–pyridine linkages to accommo- showed a bend of these linkages off the attached pyridine or
date the greater coplanarity, neither in the interpyridine phenylene planes (by up to 0.141 Å at the carboxyl carbon
bond lengths, nor in the N–C–C–N, N–C–C–C or N–C– atoms). This was most severe with the benzoate esters. Sim-
C angles. However, the interpyridine C3–C3Ј and C5Ј–C3ЈЈ ilar bends by up to 0.142 Å were also found in the crystal
distances were more variable but noticeably shorter in the structures. As the side chains are not symmetrical, and in
crystal structures (3.070Ϯ0.016 Å) than in the computed any case twisted, bent in-plane and sometimes out-of-plane,
ones (3.107Ϯ0.004 Å). These shorter distances will trans- the tridentate cores themselves are also unsymmetrical, but
late into stronger steric repulsions between the attached hy- the bond-length differences between otherwise equivalent
drogen atoms and greater interplanar twisting, but this was bonds within any one tridentate moiety were small (at most
perhaps avoided in the computed structures by the cumu- 0.007 Å, but averaging Ͻ 0.0003 Å).
[3]
lated effects of several unremarkably small bond and angu-
Comparison between the crystalline Ru(1)2 and Ru(2)2
lar distortions. On the other hand, DFT systematically reveals that the insertion of a phenylene spacer caused each
overestimated the Ru–N bond lengths, but only by about of the central pyridine bonds to lengthen, on average by
0.04 Å.
(0.013Ϯ0.004) Å (C4Ј–C3Ј/5Ј), (0.007Ϯ0.008) Å (C2Ј/6Ј-
There were some differences in the amount of twisting C3Ј/5Ј) or (0.012Ϯ0.006) Å (C2Ј/6Ј-N). Only the first of
by the phenylene and carboxyl groups out of coplanarity these changes is statistically significant, and with only one
with the rings to which they are attached, some of which such comparison available between species under crystal-
can be ascribed to secondary interactions and crystal pack- packing pressures and showing high bond-length variabil-
ing. The four crystallographically unique ligands of ity, no definitive conclusion can be drawn. In contrast, the
[Ru(Me1)2]Cl2·4H2O[3] had half of the phenylene side DFT structures were more consistent. Introduction of a
chains rotated 36.2–39.6° from the planes of the attached spacer
lengthened
the
C-4Ј–C-3Ј/5Ј
bonds
by
pyridines, whereas those of the other half were essentially (0.011Ϯ0.001) Å, in accord with what was found in the
coplanar (interplanar angles 4.4–5.3°), apparently stabilized crystal structures, whereas all other central pyridine bonds
by π stacking with the pyridine rings on a neighbouring were shortened by an average (0.002Ϯ0.0005) Å. Compari-
complex of the unit cell, as well as by agostic H–π interac- son with the unsubstituted parent complexes, [Ru(tpy)2]2+
tions (2.75–3.00 Å) between a phenylene hydrogen atom and [Ru(dpp)2]2+ [17]
revealed that adding a COOH or CO-
,
and a pyridine ring belonging to the perpendicular ligand OEt appendage caused weaker changes to these bond
on that same neighbour. There were also hydrogen-bonding lengths, on the order of 0.003 Å or less. This is consistent
networks involving the C=O groups, the anions and water with the relatively strong effects of phenylene groups on or-
molecules that imposed some distortions, but the ester bital energies (see below).
COO moieties were essentially coplanar with the attached
Importantly, metal binding caused little change in the in-
phenylene groups (interplanar angles 2.0–8.8°) in all four terplanar twists: within a few degrees, the computed struc-
crystallographically unique ligands. The computed struc- tures of the free ester ligands had the same side-chain orien-
tures of the Et1 and Et3 analogues differed in that the phen- tations as in the complexes. Metal binding to the ligands
ylene and central pyridine rings were strongly twisted (30.0– caused bond-length changes that stretched throughout the
31.9°), and the carboxy and phenylene planes were also full length of the side chains (see the Supporting Infor-
twisted by similar amounts (32.8–33.2°), but in the opposite mation). There were bond elongations and compressions in
direction such that the carboxy moieties were virtually par- alternation within the central pyridine rings, which, in the
allel with the central pyridine rings. This appears to give Et1 case, were coupled to the phenylene spacer being drawn
rise to communication between these groups (see below). in, and furthermore, compressed along the central axis,
The crystal structure of Ru(1)2·5H2O[3] showed smaller suggestive of an overall shift of phenylene electron density
twists (phenylene–pyridine interplanar angles 18.7–29.1°, toward the metal atom. With or without a spacer, the ester
phenylene–COO interplanar angles 10.1–17.0°). In contrast, portions were at the same time drawn away from the metal
the COOH groups in the calculated H1 and H3 structures atom, suggestive of a decreased ester-to-pyridine coupling,
were virtually parallel with the spacers. Hydrogen-bonding balanced by what appears to be an increased overlap of the
networks may have been responsible for the lack of copla- ethoxy oxygen atom with the carbonyl group. There were
narity in the crystal. The crystal structures of three mixed- also changes in atomic charge (see the Supporting Infor-
ligand complexes of H1 were variable: The COOH group mation), namely, increases in positive charge (or reductions
was coplanar in [Ru(H1)(phtpy)]2+ (phtpy = 4Ј-phenyl- of negative charge) at all positions of the central pyridine
2,2Ј;6Ј,2ЈЈ-terpyridine) but twisted in [Ru(H1)(tpy)]2+ and ring and stretching out to the ester groups as well, similar
Eur. J. Inorg. Chem. 2010, 3040–3050
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3043

A. Stublla, P. G. Potvin
FULL PAPER
in both ester cases, with the notable exceptions of the
charges on the carboxyl carbon atoms and the phenylene
C-4 atom (linked to the pyridine C-4Ј atom), which became
more negative. Unfortunately, neither bond-length changes
nor atomic-charge shifts provided a ready, consistent expla-
nation for the complexation-induced upfield migrations of
the NMR spectroscopic signals from the carbon atoms ly-
ing on the C2 axis.
Electronic Structure
Figure 1 presents the calculated frontier orbital composi-
tions for [Ru(H1)2]2+ and [Ru(H3)2]2+, which are illustrative
of most of our complexes, and Figures 2 and 4 (below)
show a comparison of the energy levels of the frontier orbit-
als for all of the new complexes with those of the unsubsti-
tuted analogues [Ru(tpy)2]2+ and [Ru(dpp)2]2+
.
Figure 2. Frontier orbital energy levels (H stands for HOMO and
L for LUMO) and HOMO–LUMO gaps (in eV) in homoleptic
complexes of the indicated ligands. The dotted lines join the high-
est-lying ligand-centred π orbitals. The LUMO patterns are the
same for all species, and the upper three HOMO levels (mainly t2g)
appear in the same order with tpy, H2 and Et2.
butions from both terpyridine and carboxylate moieties
without significant mediation by the phenylene rings. The
orbital compositions for the H2 species (see the Supporting
Information) are similar: the three highest-energy HOMO
levels are less mixed than in the H1 case and are mainly
of Ru t2g character, but with only two high-lying orbitals
(HOMO–3 and HOMO–4) possessing much side-chain
character, as expected for the smaller substituent. In both
complexes, the two lowest LUMO levels are nearly degener-
ate and contain some metal and side-chain character, but
the lowest four are all principally of pyridine π* type. When
compared to the corresponding situation in [Ru(tpy)2]2+
,
Figure 2 reveals that the COOH group in the H2 complex
stabilizes to a greater extent the LUMO levels than the
HOMO levels, narrowing the HOMO–LUMO gap some-
what; it thus acts as a typical electron-withdrawing substitu-
Figure 1. Frontier orbitals of [Ru(H1)2]2+ (top) and [Ru(H3)2]2+
(bottom) decomposed into contributions from Ru s and p (black),
Ru 4d (dark grey), the central pyridine (medium grey), outer pyr-
idine/pyrazine ring (light grey) and side chain (hashed) orbitals.
For the H1 species, the three highest-energy occupied or-
bitals, two of which are nearly degenerate, are mixed but
with much Ru t2g character, and these are followed by a
suite of side-chain π-type orbitals, including two nearly de-
generate pairs, showing little mixing with pyridine π levels,
as expected given the significant twist angle (Figure 2). Yet,
Figure 3. Images of HOMO–10 in [Ru(H1)2]2+ (top) and HOMO–
11 in [Ru(Et1)2]2+ (bottom) showing electron density bypassing the
twisted phenylene spacers. Note the coplanarity of the COOH
group and the non-coplanarity of the COOEt group with the phen-
as Figure 3 shows, some occupied orbitals have contri- ylene spacer.
3044
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 3040–3050

Complexes of Carboxylated Terpyridines and Dipyrazinylpyridines
ent, even if not coplanar. In contrast, the carboxyphenyl cies (Figure 4) are not at all similar to the H4 case: As with
groups in [Ru(H1)2]2+ raised the metal-centred HOMO H1, the insertion of a phenylene spacer raised all of the
levels and, to a lesser extent, the pyridine-centred LUMO frontier orbital energies, and introduced many side-chain-
levels, such that the HOMO–LUMO gap was slightly nar- centred orbitals among the highest-energy occupied ones.
rowed; it thus acts as a typical electron-donating substituent The phenylene spacer therefore acts as an electron-donating
in spite of the attached, coplanar COOH.
group, but much more dramatically so than in the tpy ana-
One difference between the H1 and H2 cases is that the logue, with a much narrower HOMO–LUMO gap. Indeed,
lowest two LUMO levels have a greater total contribution six of the eleven highest-energy HOMO levels are almost
from the outer pyridine rings in the former, whereas they totally side chain in origin (HOMO, HOMO–1, HOMO–4,
are centred mostly over the central pyridine ring in the lat- HOMO–5, HOMO–9, HOMO–10). The Ru t2g character is
ter case. This difference is consistent with the different elec- more distributed, with HOMO–2 and HOMO–3 having
tronic effects of the two side chains.
only 21% and with more significant amounts lodged in
The frontier orbital compositions for the Et2 complex deeper orbitals (HOMO–6 to HOMO–8).
closely mimic those of its carboxylic acid parent, but the
levels are shifted up in energy, essentially cancelling out the
effect of the COOH group. This is in accord with the ester
having a less electron-withdrawing C=O group, even though
it was rotated into coplanarity. The Et1 case is more com-
plex. The phenylene–pyridine twists are identical in both
the H1 and Et1 cases, but the carboxyl group of the H1
complex is parallel to the phenylene plane, whereas it is
twisted out of plane in complexed Et1 to lie virtually paral-
lel to the central pyridine ring. Here, too, some occupied
ligand-centred orbitals reveal a terpyridine–carboxylate
overlap that bypasses the phenylene ring (Figure 3). Pre-
sumably, the poorer phenylene–C=O overlap caused the en-
ergy of the two highest-lying and nearly degenerate phenyl-
ene-centred MOs to lie above those of the H1 case and be-
come the nominal HOMO and HOMO–1 levels, with a sub-
stantially narrower calculated HOMO–LUMO gap re-
sulting. Thus, in comparison with [Ru(tpy)2]2+, the side
chain of Et1 also behaves as a typical electron-donating
group, but the LUMO and metal-centred HOMO levels Figure 4. Frontier orbital energy levels (H stands for HOMO and
L for LUMO) and HOMO–LUMO gaps in homoleptic complexes
were lower in energy than those with H1. There were no
of the indicated ligands. The dotted lines join the highest-lying li-
significant differences between the H1 and Et1 complexes
gand-centred π orbitals. The LUMO patterns are the same for all
in any of the bond lengths involving the metal atom, the
species, and the upper three HOMO levels (mainly t2g) appear in
terpyridine or the phenylene atoms, and the only significant
differences in atomic charges occurred at the phenylene
atoms.
the same order with dpp, H4 and Et4.
In both of the H3 and H4 cases, the six lowest LUMO
For the dpp analogues, the frontier orbitals are about levels have a greater total contribution from the pyrazine
1 eV lower in energy than in terpyridine analogues because rings than from the pyridine rings, in accord with the expec-
of the added nitrogen atoms. The situation with the H4 spe- tation that the extra nitrogen atom of a pyrazine ring would
cies (see the Supporting Information) was otherwise similar cause its π* levels to be lower in energy than those of a
to that of H2: the three highest-energy occupied orbitals pyridine ring and shift the principal electron sink from the
are close in energy (two are nearly degenerate) and are central pyridine ring (as it is in terpyridines) to the outer
mostly of Ru t2g character; these are followed by two nearly pyrazine rings. These orbitals have more pyridine character
degenerate side-chain π-type orbitals, showing little mixing in H4 than in H3, presumably because the electron-with-
with pyridine or pyrazine π levels owing to the significant drawing COOH group of H4 stabilizes the π* levels of the
twist angle, and then these are followed by four orbitals of pyridine ring, thus increasing its contribution to the lowest
mostly pyrazine π character. The LUMO and LUMO+1 unoccupied orbitals and resulting in a more distributed
levels are also nearly degenerate and show small contri- electron sink.
butions from metal and side-chain orbitals, but along with
As was the case with Et2 relative to H2, the frontier or-
LUMO+3 and LUMO+4, have mostly pyridine and pyr- bitals of the Et4 complex closely mimic those of the H4
azine π* character. Here, too, a comparison with the corre- complex, with the levels shifted higher in energy due to es-
sponding situation in [Ru(dpp)2]2+ shows that the COOH terification. Esterification of the H3 complex caused the
group acts as an electron-withdrawing substituent.
LUMO and metal-character HOMO levels to decrease in
The LUMO levels are very much the same in the three energy. Otherwise, the lowest LUMO levels were virtual re-
other dpp complexes, but the HOMO levels for the H3 spe- productions of those of its carboxylic acid analogue. Esteri-
Eur. J. Inorg. Chem. 2010, 3040–3050
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3045

A. Stublla, P. G. Potvin
FULL PAPER
fication to Et3 also caused a destabilization of the side-
chain π levels, such that the uppermost four HOMO levels
are almost purely side chain in character. As was the case
for the tpy analogues, the optimized structure of [Ru(H3)2]
2+
placed the COOH group coplanar with the spacer, but
the ester group of [Ru(Et3)2]2+ was rotated out of coplanar-
ity. This no doubt reduced the π-withdrawing ability of the
C=O group, giving the phenylene spacer higher-energy π
levels. At the same time, the lowering of the LUMO and
metal-character HOMO levels would appear to reflect an
increase in electron withdrawal by the side chain, or a de-
crease in electron donation, although the phenylene–pyr-
idine twist angle in the optimized structures actually de-
creased upon esterification.
Electronic Spectroscopy
UV/Vis absorption and electrochemical measurements
were performed in CH3CN to assess the effects of the sub-
stituents on the electronic properties of the complexes in
comparison with noncarboxylated analogues. Table 1 re- Figure 5. Experimental (solid lines) and TD-DFT calculated spec-
tra (dotted lines, arbitrary scale).
ports the visible-region absorption maxima (λMLCT) and the
measured half-wave potentials (E ). Figure 5 presents rep-
½
resentative UV/Vis spectra, with the corresponding time-
dependent (TD)-DFT-calculated spectra overlaid for com-
parison. These were typical of this class of complex, with
the λMLCT values matched best. The transition assignments
strong π–π* bands in the 250–350 nm range, and dominated
appear in the Supporting Information. In all cases, there
in the visible region by a metal-to-ligand charge transfer
are two transitions mainly responsible for the lowest-energy
(MLCT) band near 490 nm. There was a higher-energy
absorption envelopes. In the cases lacking a spacer, the least
shoulder that reveals the contribution of more than one
energetic transition is from degenerate HOMO/HOMO–1
transition. Although the λMLCT positions were relatively in-
(or HOMO–1/HOMO–2) of mainly t2g character to degen-
erate LUMO/LUMO+1, and the higher-energy shoulder
comes from the third, mainly t2g level to the LUMO+2, all
dependent of the substitution, all substituents caused a red-
shift relative to the λMLCT position reported for [Ru-
(tpy)2]2+, which is consistent with prior findings and with
tridentate-centred. In the H1 and H3 complexes, however,
our computation of narrower HOMO–LUMO gaps.
the HOMO and HOMO–1 levels, which are almost totally
side chain in character, contribute negligibly to the visible-
region transitions. The same is true of HOMO–4 and
HOMO–5. Instead, the lowest-energy transition originates
from the mixed t2g/side-chain orbitals (HOMO–2 and
HOMO–3) to the usual LUMO levels, and the two transi-
tions contributing to the higher-energy shoulder come from
the other orbitals with relatively high metal character
(HOMO–6 to HOMO–8) to these same LUMO levels. On
the contrary, the main low-energy transition for the Et1
complex involved the side-chain-centred HOMO/HOMO–1
levels and is thus very much intraligand in character (phen-
ylene π to mostly central pyridine π*) as well as having
MLCT character (Figure 6 illustrates this combined transi-
tion), whereas the higher-energy shoulder is a more classical
MLCT transition. In the Et3 case, in which six of the eight
highest-lying occupied orbitals are almost purely side chain
in character, the main low-energy transition also had a
Table 1. Visible-region absorption maxima and redox potentials in
CH3CN of [RuL2]2+ complexes (see Experimental Section for con-
ditions).
L
λMLCT [nm]
E
½
[V] vs. SCE
(ε [104 –1 cm–1])
Eox
Ered
H1
Et1
491 (2.76)
491 (2.50)
490 (3.10)
488 (2.97)
476[c]
+1.24
+1.27
–1.28,[a] –1.5[a]
–1.15, –1.37, –1.74
H2
+1.37 –0.8,[a] –1.03, –1.40, –1.65
Et2
+1.43
+1.27
–0.99, –1.21, –1.47
–1.27
(no clear waves)
–0.88, –1.07, –1.40
(no clear waves)
tpy[b]
H3
493 (0.625), 410 (sh.) +1.72
495 (2.00), 409 (sh.) +1.60
510 (2.00), 430 (sh.) +1.83
483 (1.95), 406 (sh.) +1.73 –0.76, –0.95, –1.36, –1.69
498 (1.66) +1.62 –0.83, –1.04, ca. –1.30
Et3
H4
Et4
tdpp[d]
[a] Irreversible process; peak cathodic potential (Epc) reported.
[b] From Morris et al.[23] [c] From Hecker et al.[24] [d] From
Liegghio et al.[9]
The gas-phase TD-DFT and experimental spectra were strong intraligand character, whereas the new low-energy
qualitatively the same in all cases, with the λMLCT values shoulder (calculated peak at 617 nm) is almost purely intra-
matching best for the complexes with phenylene spacers. In ligand in character (phenylene π to mostly pyrazine π*, also
general, the HOMO–LUMO gaps were systematically over- illustrated in Figure 6), originating from the four highest-
estimated except for the Et1 and Et3 complexes, for which lying, ligand-centred HOMOs.
3046
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 3040–3050

Complexes of Carboxylated Terpyridines and Dipyrazinylpyridines
In contrast to λMLCT, the E values were more sensitive
½
to the substituents: COOH or COOEt groups shifted the
RuIII/RuII couple in a more positive direction and the first
reductions with COOEt even more so, whereas the interses-
sion of a phenylene spacer softened these changes, such that
the RuIII/RuII couple of the Et1 species matched that of the
parent tpy complex. In spite of the fact that O-alkyl groups
reduce the π-withdrawing ability of the carboxyl C=O
group, the COOH groups were found to exert a weaker ef-
fect than the COOEt groups, part of which was attributable
to their acidities, as had been noted in NH-acidic
cases.[20,26] In general, the calculated ordering of the frontier
orbitals was more faithfully followed by the reduction po-
tentials than by the oxidation potentials, and by the dpp
cases than by the tpy analogues. Since the highest-occupied
MOs are side chain in origin in those species with phenylene
spacers, their oxidations may initially occur on the side
chain, but the oxidized forms perhaps relax to lower-energy
Figure 6. Kohn–Sham representation of the lowest-energy transi-
tions in the UV/Vis spectra of [Ru(Et1)2]2+, which illustrates the
combined metal-to-ligand and intraligand characters (top), and of
[Ru(Et3)2]2+, which has pure intraligand character (bottom). Dark
grey areas represent electron-density sources corresponding to
combinations of HOMO to HOMO–3 (top) and HOMO to
HOMO–1 (bottom). Light gray represents electron-density sinks
(LUMO and LUMO+1).
states that are measured by E values but are not reflected
½
by the MO levels of the unoxidized forms.
Electrochemistry
Conclusion
Figure 7 presents representative cyclic voltammetry (CV)
and differential pulse voltammetry (DPV) traces. CV
showed one reversible or quasireversible oxidation wave as-
The target ligands and complexes were prepared by short
routes in good to excellent yields, and were completely char-
acterized. The yields were generally lower in the dpp cases,
in part because of decreased stability, greater difficulties in
handling and increased levels of side products (and the need
for purification).
signed to the RuIII/RuII couple. As expected, the E values
½
of the dpp complexes were shifted positive from those of
the tpy analogues, no doubt owing to the additional nitro-
gen atoms in the tridentate moiety.[8,25] The free acid com-
plexes showed poorly discernible reduction waves, whether
in CH3CN or in dmf – no reversible CV waves and no DPV
reduction peaks were found in the H1 case – but the ester
complexes showed up to four reversible or quasireversible
reduction waves attributable to ligand-centred processes.
Bond lengths, bond angles and inter-ring torsions were
highly variable in the crystals containing the tpy ligands
described in the literature, owing to packing forces and sec-
ondary interactions, and comparison with the more consis-
tent DFT-calculated structures was difficult. DFT slightly
overestimated the bond lengths but correctly predicted the
side-chain conformations. In terms of electronic structures,
COOH or COOEt groups were found to act as electron-
withdrawing substituents, as expected, but phenylene spa-
cers acted as electron-donating substituents that moderated
the effects of COOH or COOEt groups. These structures
led to reasonably well-predicted UV/Vis spectra. A notable
finding is the intraligand character of the low-energy transi-
tions in the complexes with phenylene spacers, especially
when esterified. Although the experimental λMLCT positions
varied little, and were slightly redshifted in the dpp com-
plexes relative to their tpy analogues, the electrochemical
potentials were positively shifted in comparison with the
tpy analogues, in line with the DFT-predicted frontier or-
bitals.
It remains to be seen whether the TiO2-chemisorbed
complexes will more closely resemble the acid or the ester
forms studied here, insofar as these show differences in ex-
cited-state energies (as reflected by ground-state LUMO
levels) and in the quality of the electronic coupling between
the semiconductor conduction band and the ligand-centred
excited-state HOMO (as reflected by the degree of copla-
narity between the side chain and the central pyridine ring).
Figure 7. Cyclic voltammogram at 100 mVs–1 (top) and differential
pulse voltammogram (bottom) of [Ru(Et2)2]2+ in CH3CN contain-
ing 0.1  TBAH.
Eur. J. Inorg. Chem. 2010, 3040–3050
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3047

A. Stublla, P. G. Potvin
FULL PAPER
4Ј-(4-Carboxyphenyl)-2,2Ј:6Ј,2ЈЈ-terpyridine (H1): 2-Acetylpyridine
(290 mg, 2.4 mmol) and 4-carboxybenzaldehyde (180 mg,
1.2 mmol) were dissolved in CH3OH (8 mL) by stirring for 5 min,
followed by addition of 15% KOH (7.2 mL) and concentrated
NH4OH (0.8 mL). The mixture was allowed to stand at ambient
temperature for 3 d. The emulsion formed was filtered off and
washed with CHCl3 (4 mL) and cold CH3OH/H2O (1:1; 4 mL). The
crude product was suspended in CH3OH/H2O (80:20), and the mix-
ture was stirred and sonicated at 35 °C until a clear solution was
obtained. This was then acidified to pH = 2 by addition of 1 
HCl, resulting in the formation of a white precipitate that was col-
lected by vacuum filtration and rinsed with cold water. The solid
was dried by means of vacuum filtration to give the pure product
Phenylene spacers usefully shift the LUMO levels to higher
energies but likely diminish the degree of electronic cou-
pling. The localization of the LUMOs over the pyrazine
rings in the dpp cases may further hinder effective electronic
coupling. Preliminary results on the surface attachment,
surface spectral and electrochemical properties and photo-
voltaic performances of some of the tpy complexes have
appeared elsewhere.[27]
Experimental Section
1
as a white solid (340 mg, 80%). H NMR ([D]tfa/D2O capillary): δ
General: The preparation of ligands Et2 and Et4 is reported else-
where.[28] All reagents and solvents were reagent grade and used
without further purification, except dry CH3CN, which was ob-
tained by distillation with P2O5 prior to use. The 1H and 13C NMR
spectra of ligands and complexes were recorded at room tempera-
ture in [D]tfa with an external reference capillary containing D2O,
or in CDCl3, [D6]dmso or CD3CN, with 300 or 400 MHz Bruker
ARZ instruments. Chemical shifts (δ) are reported in parts per mil-
lion (ppm) downfield from tetramethylsilane (TMS). The resonance
assignments were supported by 2D NMR spectroscopic techniques
(1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC and DEPT-135) and
appear in the Supporting Information. Fourier-transform infrared
(FTIR) data were obtained by using a Genesis II spectrometer with
samples prepared as KBr pellets. UV/Vis spectra were obtained by
using an Ultraspec 4300 pro Biochrom spectrometer with samples
prepared in dry CH3CN at 10–5–10–6 . Electrochemical data were
obtained with an Autolab Eco-Chemie BV analyzer in a three-elec-
trode cell at room temperature in anhydrous CH3CN containing
0.1  nBu4NPF6 (TBAH) after purging with N2 for 15 min. Cyclic
voltammograms were obtained with a Pt disk working electrode, a
graphite counterelectrode and an Ag/AgCl wire as the pseudorefer-
ence electrode. The potentials were scanned from +1.6 to –2 V at
rates between 100 and 1000 mVs–1. E0 values were taken to be the
mean of anodic (Epa) and cathodic (Epc) peak potentials from the
tenth steady-state scan at 100 mVs–1, and verified by differential
pulse voltammetry. Waves were considered to be reversible if the
peak cathodic current (ic) equalled the peak anodic current (ia),
= 7.90 (d, 2 H), 8.13 (dd, 2 H), 8.30 (d, 2 H), 8.83 (d, 2 H), 8.65
(dd, 2 H), 8.74 (s, 2 H), 9.04 (d, 2 H) ppm. 13C NMR ([D]tfa/D2O
capillary): δ = 126.0, 127.2, 129.8, 130.7, 133.0, 133.9, 143.1, 144.9,
149.1, 149.9, 150.9, 156.9, 174.2 ppm. C22H15N3O2 (353.38): calcd.
C 74.78, H 4.28, N 11.89; found C 74.71, H 3.99, N 11.38. EI-MS:
m/z (%) = 353 (100), 336 (2), 325 (8), 308 (24). IR: ν = 1691 (C=O
˜
str.) cm–1.
4Ј-Carboxy-2,2Ј:6Ј,2ЈЈ-terpyridine (H2): 2-Acetylpyridine (290 mg,
2.4 mmol) and ethyl glyoxalate (0.238 mL, 1.2 mmol) were used in
the same manner as for H1. The crude product was dissolved and
acidified as before to give a suspension, which was collected by
filtration and carefully washed with a small amount of slightly
acidified water and a small amount of cold water. Drying under
vacuum provided H2 as the hydrochloride salt, which was a white-
tinged pink precipitate containing traces of free HCl (225 mg,
60%). 1H NMR ([D]tfa/D2O capillary): δ = 8.53 (dd, 2 H), 9.13 (s,
2 H), 9.21 (d, 2 H), 9.45 (d, 2 H) ppm. 13C NMR ([D]tfa/D2O
capillary): δ = 125.1, 125.4, 128.8, 142.7, 143.0, 146.0, 148.5, 148.7,
166.7 ppm. IR: ν = 1694 (C=O str.) cm–1.
˜
4-(4-Carboxyphenyl)-2,6-dipyrazin-2-ylpyridine (H3): By following
the same procedure used for H1, 2-acetylpyrazine (295 mg,
2.4 mmol) was used to produce a shiny white, flaky solid (347 mg,
80%). 1H NMR ([D]tfa/D2O capillary): δ = 8.36, (d, 2 H), 8.66 (d,
2 H), 9.30 (br. s, 2 H), 9.48 (br., 2 H), 9.77 (s, 2 H), 10.37 (br. s, 2
H) ppm. 13C NMR ([D]tfa/D2O capillary): δ = 123.5, 127.9, 131.2,
131.5, 136.2, 138.5, 140.3, 148.6, 150.4, 157.1, 171.8 ppm.
C20H13N5O2 (355.35): calcd. C 67.60, H 3.69, N 19.71; found C
½
and if DPV plots of ic vs. ν were linear, in which ν is the scan
rate. The potentials were referenced internally to the ferrocenium/
ferrocene couple [E(Fc+/Fc) = +0.64 V vs. normal hydrogen elec-
trode (NHE) in CH3CN].[29] High-resolution EI-MS was per-
formed with a Waters GCT Premier instrument. Matrix-free LDI-
MS was carried out with a Voyager-DE spectrometer (PerSeptive
Biosystems) equipped with a TOF detector in the positive ion
mode. Elemental analyses were performed by weighing samples un-
der N2 by Guelph Chemical Laboratories (Guelph, ON, Canada).
67.69, H 3.70, N 19.31. IR: ν = 1716 (C=O str.), 3422 (O–H str.)
˜
cm–1.
4-Carboxy-2,6-dipyrazin-2-ylpyridine (H4): As for H3, 2-acetylpyr-
azine (295 mg, 2.4 mmol) was used to produce a beige solid
(180 mg, 55%). 1H NMR ([D]tfa/D2O capillary): δ = 8.93 (d, 2 H),
9.37 (br. s, 2 H), 9.54 (d, 2 H), 10.06 (s, 2 H) ppm. 13C NMR
([D]tfa/D2O capillary): δ = 125.1, 133.1, 134.1, 141.3, 149.7, 152.2,
155.6, 168.3 ppm. C14H9N5O2 (279.25): calcd. C 60.21, H 3.25, N2
5.08; found C 60.52, H 3.10, N 25.34. EI-MS: m/z (%) = 279.15
Computational Details: Geometry-optimized structures were ob-
tained by using Gaussian 03 [G03W C.02 (v6.0)] employing DFT
calculations, using the hybrid B3LYP exchange-correlation func-
tional and the LANL2DZ basis set with spin-restricted wave func-
tions for closed-shell species and spin-unrestricted wave functions
for open-shell species.[30] A tight convergence (10–8 a.u.) was used
for all calculations. Vibrational frequency calculations were per-
formed on all optimized complexes to verify that an energy mini-
mum had been attained. The wave functions were also checked for
stability. All the ground states are spin singlets. The energies of the
predicted electronic transitions were obtained by using the TD-
DFT method.[31] The absorption profiles of the complexes were
calculated by using the SWIZARD program.[32] A natural popula-
tion analysis (G03W) was also carried out.
(100), 262.15 (2), 251.15 (5), 235.14 (12.1). IR: ν = 1710 (C=O str.)
˜
cm–1.
Esterification
4Ј-(4-Ethoxycarbonylphenyl)-2,2Ј:6Ј,2ЈЈ-terpyridine (Et1) and 4Ј-
+
Ethoxycarbonyl 2,2Ј:6Ј,2ЈЈ-terpyridine (Et3): The crude K+/NH4
salt form of acid H1 or H3 was dissolved in EtOH in the presence
of a catalytic amount of H2SO4 and heated to reflux for 2.5–3 d.
The pure product was isolated as white flakes after extraction with
CH2Cl2, affording ligands Et1 (yield 70%) or Et3 (yield 76%).
1
Ester Et1: H NMR (CDCl3): δ = 1.45 (t, 3 H), 4.44 (q, 2 H), 7.36
(dd, 2 H), 7.88 (dd, 2 H), 7.96 (d, 2 H), 8.18 (d, 2 H), 8.67 (d, 2
3048
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 3040–3050

Complexes of Carboxylated Terpyridines and Dipyrazinylpyridines
H), 8.74 (d, 2 H), 8.76 (s, 2 H) ppm. 13C NMR (CDCl3): δ = 14.3, 140.0, 146.0, 147.3, 148.0, 149.0, 153.6, 154.6, 167.4 ppm.
61.1, 118.9, 121.7, 124.0, 127.3, 130.1, 130.8, 136.9, 142.8, 149.1,
C40H26F12N10O4P2Ru (1101.70): calcd. C 43.61, H 2.38, N 12.71;
149.1, 155.9, 156.1, 166.2 ppm. C24H19N3O2 (381.43): calcd. C found C 43.82, H 2.13, N 12.25. LDI-MS: m/z (%) = 810.87 (100),
75.72, H 5.02, N 11.02; found C 75.72, H 5.18, N 10.90. EI-MS:
766.90 (100), 722.92 (38). IR: ν = 1707 (C=O), 3434 (O–H str.),
˜
m/z (%) = 381 (100), 352 (12), 336 (21), 308 (29). IR: ν = 1690
837 (PF6) cm–1.
˜
(C=O str.) cm–1.
[Bis(4-carboxy-2,6-dipyrazin-2-ylpyridine)]ruthenium(II) Bis(hexa-
fluorophosphate) [Ru(H4)2](PF2)6: This was produced in the same
manner as was used to obtain bis(4Ј-carboxy-2,2Ј:6Ј,2ЈЈ-terpyrid-
ine)ruthenium(II) bis(hexafluorophosphate),[5] by hydrolysis of
[Ru(Et4)2](PF6)2 (0.171 mg, 0.17 mmol) to produce a red powder
1
Ester Et3: H NMR (CDCl3): δ = 1.37 (t, 3 H), 4.41 (q, 2 H), 7.93
(d, 2 H), 8.21 (d, 2 H), 8.65 (br. s, 2 H), 8.69 (br. s, 2 H), 8.75 (s,
2 H), 9.88 (br. s, 2 H) ppm. 13C NMR (CDCl3): δ = 14.3, 61.3,
119.9, 127.2, 130.3, 131.3, 142.0, 143.6, 144.9, 149.7, 150.5, 154.6,
166.1 ppm. C22H17N5O2 (383.40): calcd. C 68.92, H 4.47, N 18.27;
found C 69.09, H 4.62, N 17.95. EI-MS: m/z (%) = 383.27 (27),
1
(293 mg, 38%). H NMR (CD3CN): δ = 7.47 (d, 4 H), 8.41 (d, 4
H), 9.43 (s, 4 H), 9.80 (br. s, 4 H) ppm. 13C NMR ([D6]dmso): δ =
124.7, 139.5, 145.8, 148.0, 148.7, 152.9, 154.7, 165.7 ppm.
C28H18F12N10O4P2Ru (949.51): calcd. C 35.42, H 1.91, N 14.75;
found C 35.22, H 2.22, N 14.42. LDI-MS: m/z (%) = 803.14 (50),
338.27 (5), 284.21 (16.9). IR: ν = 1726 (C=O str.) cm–1.
˜
Ruthenium Complexes
Bis[4Ј-(4-Carboxyphenyl)-2,2Ј:6Ј,2ЈЈ-terpyridine]ruthenium(II) Bis-
(hexafluorophosphate) [Ru(H1)2](PF6)2: RuCl3·3H2O (147 mg,
0.56 mmol) and AgBF4 (330 mg, 1.69 mmol) were dissolved in dmf
(30 mL), and the mixture was kept at reflux for 3 h until a sand-
coloured suspension appeared and the solution had turned bright
red. Upon addition of ligand H1 (400 mg, 1.13 mmol), the mixture
turned black, at which point more dmf (5 mL) was added, and
reflux was maintained for 3 d, yielding a red solution. This was
filtered through Celite, and the filtrate was treated with 0.5  aque-
ous NH4PF6 (18 mL) and allowed to stand at ambient temperature
overnight. The red precipitate was isolated by filtration through
Celite, rinsed several times with water and twice with diethyl ether,
and finally collected with hot CH3CN containing a few drops of
aqueous HPF6 (60% tech. grade solution). The solution was con-
centrated, and a red microcrystalline product was obtained after
the addition of diethyl ether (590 mg, 95%). 1H NMR ([D6]dmso):
δ = 7.28 (dd, 4 H), 7.56 (d, 4 H), 7.98 (dd, 4 H), 8.29 (d, 4 H), 8.55
(d, 4 H), 9.14 (d, 4 H), 9.54 (s, 4 H) ppm. 13C NMR ([D6]dmso):
δ = 121.9, 125.3, 128.2, 128.3, 130.6, 132.5, 138.5, 140.6, 146.1,
152.7, 155.6, 158.3, 167.3 ppm. C44H30F12N6O4P2Ru (1097.75):
calcd. C 48.14, H 2.75, N 7.66; found C 47.53, H 2.70, N 7.45.
LDI-MS: m/z (%) = 1097.34 (13), 808.17 (23), 763.17 (32), 718.18
587.06 (40), 570.07 (95). IR: ν = 1712 (C=O str.), 838 (P–F str.)
˜
cm–1.
Bis[4Ј-(4-ethoxycarbonylphenyl)-2,2Ј:6Ј,2ЈЈ-terpyridine]ruthenium(II)
Bis(hexafluorophosphate) [Ru(Et1)2](PF6)2: By following the same
procedure used for [Ru(H1)2](PF6)2, but without the use of HPF6,
RuCl3·3H2O (172 mg, 0.655 mmol), AgBF4 (383 mg, 1.97 mmol)
and ligand Et1 (500 mg, 1.13 mmol) were converted to complex
1
[Ru(Et1)2](PF6)2, a red microcrystalline product (0.74 g, 98%). H
NMR (CD3CN): δ = 1.48 (t, 6 H), 4.48 (q, 4 H), 7.22 (dd, 4 H),
7.46 (d, 4 H), 7.99 (t, 4 H), 8.33 (d, 4 H), 8.40 (d, 4 H), 8.68 (d, 4
H), 9.07 (s, 4 H) ppm. 13C NMR (CD3CN): δ = 13.6, 61.3, 121.9,
124.6, 127.5, 128.0, 130.3, 132.1, 138.1, 141.0, 147.0, 152.5, 155.5,
158.0, 165.7 ppm. C48H38F12N6O4P2Ru (1153.86): calcd. C 49.96,
H 3.32, N 7.28; found C 49.54, H 3.30, N 6.99. LDI-MS: m/z (%)
= 1153.28 (13), 863.22 (22), 835.18 (12), 807.16 (80). IR: ν = 1712
˜
(C=O str.), 838 (P–F str.) cm–1.
Bis(4Ј-ethoxycarbonyl-2,2Ј:6Ј,2ЈЈ-terpyridine)ruthenium(II) Bis-
(hexafluorophosphate) [Ru(Et2)2](PF6)2: As for [Ru(Et1)2](PF6)2,
RuCl3·3H2O (85 mg, 0.325 mmol), AgBF4 (191 mg, 0.98 mmol)
and ligand Et2 (200 mg, 0.655 mmol) were converted to [Ru(Et2)2]-
(PF6)2, also a red microcrystalline product (269 mg, 82%). 1H
NMR (CD3CN): δ = 1.59 (t, 3 H), 4.68 (q, 2 H), 7.21 (dd, 4 H),
7.37 (d, 4 H), 7.98 (dd, 4 H), 8.69 (d, 4 H), 9.24 (s, 4 H) ppm. 13C
NMR (CD3CN): δ = 13.6, 62.9, 122.8, 125.0, 127.8, 138.4, 138.4,
152.5, 155.7, 157.3, 164.8 ppm. C36H30F12N6O4P2Ru (1001.67):
calcd. C 43.17, H 3.02, N 8.39; found C 43.20, H 3.08, N 8.42.
LDI-MS: m/z (%) = 782.83 (100), 711.83 (100), 684.86 (43), 639.85
(61). IR: ν = 1713 (C=O str.), 3431 (O–H str.), 842 (P–F str.) cm–1.
˜
Bis(4Ј-carboxy-2,2Ј:6Ј,2ЈЈ-terpyridine)ruthenium(II) Bis(hexafluoro-
phosphate) [Ru(H2)2](PF6)2: Complex [Ru(Et2)2](PF6)2 (0.17 mg,
0.17 mmol) was suspended in water (20 mL) to which 0.1  NaOH
(6 mL) was added. The suspension was heated to reflux for 1.5 d.
The resulting red solution was concentrated to remove EtOH, then
water (4 mL) was added, and the solution was acidified to pH = 2
by addition of 1  perchloric acid or HCl, resulting in the immedi-
ate formation of a precipitate. This was collected by vacuum fil-
tration, then washed with cold water and copiously with diethyl
ether to give the pure product as a red powder in 100% yield
(100), 611.95 (98), 566.88 (86). IR: ν = 1727 (C=O str.), 843 (P–F
˜
str.) cm–1.
{Bis[4-(4-ethoxycarbonylphenyl)-2,6-dipyrazin-2-ylpyridine]}-
ruthenium(II) Bis(hexafluorophosphate) [Ru(Et3)2](PF2)6:
RuCl3·3H2O (137 mg, 0.52 mmol), AgBF4 (305 mg, 1.56 mmol)
and Et3 (400 mg, 1.04 mmol) were used by following the same pro-
cedure as with ligand H3. The crude, dark red-brown hexafluoro-
phosphate salt was purified by preparatory TLC on 2 mm-thick
Macherey–Nagel Polygram Sil G/UV silica gel plates, developed
with CH3CN/saturated KNO3/H2O (65:2:1). The most mobile red
band was scraped off and extracted with CH3CN and CH3OH. The
solvents were evaporated in vacuo. The resulting dark red solid was
dissolved in a minimum amount of CH3CN and poured into a
saturated aqueous solution of NH4PF6. The mixture was allowed
to stand at ambient temperature overnight to give a red precipitate,
1
(160 mg). H NMR ([D6]dmso): δ = 7.19 (dd, 4 H), 7.36 (d, 4 H),
7.96 (dd, 4 H), 8.67 (d, 4 H), 9.25 (s, 4 H) ppm. 13C NMR ([D6]-
dmso): δ = 123.7, 125.6, 128.4, 138.2, 138.8, 152.7, 155.5, 157.6,
165.9 ppm. C32H22F12N6O4P2Ru (945.56): calcd. C 40.65, H 2.35,
N 8.89; found C 40.09, H 2.38, N 9.17. LDI-MS: m/z (%) = 800.18
(20), 798.17 (80), 583.08 (70), 566.09 (80). IR: ν = 1694 (C=O str.),
˜
3447 (O–H str.) cm–1.
{Bis[4-(4-carboxyphenyl)-2,6-dipyrazin-2-ylpyridine]}ruthenium(II)
Bis(hexafluorophosphate) [Ru(H3)2](PF2)6: Analogously to 4Ј-(4-
carboxyphenyl)-2,2Ј:6Ј,2ЈЈ-terpyridine,[5] RuCl3·3H2O (73 mg,
0.29 mmol), AgBF4 (163 mg, 0.837 mmol) and ligand H3 (200 mg, which was isolated by vacuum filtration through a Celite layer and
0.563 mmol) were used to obtain a dark red microcrystalline prod-
washed with cold water and copiously with diethyl ether (180 mg,
uct (307 mg, 58%). 1H NMR ([D6]dmso): δ = 7.74 (br. s, 4 H), 8.34 30%). 1H NMR (CD3CN): δ = 1.48 (t, 6 H), 4.49 (q, 4 H), 7.56 (s,
(d, 4 H), 8.49 (br. s, 4 H), 8.57 (d, 4 H), 9.72 (s, 4 H), 10.2 (br. s, 4 H), 8.37 (d, 4 H), 8.42 (br. s, 4 H), 8.44 (d, 4 H), 9.25 (s, 4 H),
4 H) ppm. 13C NMR ([D6]dmso): δ = 122.5, 128.4, 130.7, 133.0,
9.80 (br. s, 4 H) ppm. 13C NMR (CD3CN): δ = 13.6, 61.4, 123.0,
Eur. J. Inorg. Chem. 2010, 3040–3050
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3049

A. Stublla, P. G. Potvin
FULL PAPER
[11] N. A. Anderson, T. Lian, Coord. Chem. Rev. 2004, 248, 1231–
1246.
128.1, 130.5, 133.0, 140.0, 145.0, 147.6, 148.3, 149.0, 153.1, 154.3,
165.8 ppm. C44H34F12N10O4P2Ru (1157.81): calcd. C 45.64, H
2.96, N 12.10; found C 45.54, H 3.28, N 11.94. LDI-MS: m/z (%)
[12] G. D. Storrier, S. B. Colbran, Inorg. Chim. Acta 1999, 284, 76–
84.
= 939.7 (100), 866.75 (100), 839.74 (30). IR: ν = 1708 (C=O str.),
˜
[13] F. Neve, M. La Deda, F. Puntoriero, S. Campagna, Inorg.
Chim. Acta 2006, 359, 1666–1672.
846 (P–F str.) cm–1.
[14] H. Wolpher, S. Sinha, J. Pan, A. Johansson, M. J. Lundqvist,
P. Persson, R. Lomoth, J. Bergquist, L. Sun, V. Sundström, B.
Åkermark, T. Polívka, Inorg. Chem. 2007, 46, 638–651.
[15] E. Figgemeier, V. Aranyos, E. C. Constable, R. W. Handel,
C. E. Housecroft, C. Risinger, A. Hagfeldt, E. Mukhtar, Inorg.
Chem. Commun. 2004, 7, 117–121.
[Bis(4-ethoxycarbonyl-2,6-dipyrazin-2-ylpyridine)]ruthenium(II) Bis-
(hexafluorophosphate) [Ru(Et4)2](PF2)6: This was prepared in the
same way as [Ru(H3)2](PF2)6 by using RuCl3·3H2O (213 mg,
0.814 mmol), AgBF4 (476 mg, 2.44 mmol) and ligand Et4 (500 mg,
1.63 mmol) to give a red microcrystalline product (350 mg, 38%).
1H NMR (CD3CN): δ = 1.56 (t, 6 H), 4.72 (q, 4 H), 7.44 (dd, 4
H), 8.39 (d, 4 H), 9.41 (s, 4 H), 9.75 (d, 4 H) ppm. 13C NMR
(CD3CN): δ = 13.6, 63.2, 124.1, 139.1, 145.4, 147.7, 148.6, 152.4,
154.6, 163.3 ppm. C32H26F12N10O4P2Ru (1005.62): calcd. C 38.22,
H 2.61, N 13.93; found C 38.35, H 2.73, N 13.59. LDI-MS: m/z
[16] S. Vaduvescu, P. G. Potvin, Eur. J. Inorg. Chem. 2004, 1763–
1769.
[17] M. Abboud, D. Kalinina, P. G. Potvin, Inorg. Chim. Acta 2009,
362, 4953–4959.
[18] R.-A. Fallahpour, Synthesis 2000, 1138–1142.
[19] J. Husson, M. Beley, G. Kirsch, Tetrahedron Lett. 2003, 44,
1767–1770.
[20] J. Zadykowicz, P. G. Potvin, Inorg. Chem. 1999, 38, 2434.
[21] P. G. Potvin, P. U. Luyen, F. Al-Mutlaq, New J. Chem. 2001,
25, 839–846.
(%) = 714.94 (100), 687.97 (43). IR: ν = 1724 (C=O str.), 834 (P–
˜
F str.) cm–1.
Supporting Information (see footnote on the first page of this arti-
cle): 1H and 13C NMR spectroscopic assignments and coupling
constants; 1D and 2D NMR spectra; mass spectral assignments;
frontier molecular orbital compositions; experimental and pre-
dicted optical spectra; transition assignments for near-UV and vis-
ible absorptions; cyclic and differential pulse voltammograms; se-
lected orbital images and orbital difference diagrams; tables of
atomic charges from natural population analysis; and MOL2 coor-
dinates for all geometry-optimized complexes.
[22] J. Zadykowicz, P. G. Potvin, J. Coord. Chem. 1999, 47, 395–
407.
[23] D. E. Morris, K. W. Hanck, M. K. DeArmond, J. Electroanal.
Chem. 1983, 149, 115–130.
[24] C. R. Hecker, A. K. I. Gushurst, D. R. McMillin, Inorg. Chem.
1991, 30, 538–541.
[25] R. J. Crutchley, A. B. P. Lever, Inorg. Chem. 1982, 21, 2276–
2282.
[26] Y. Luo, P. G. Potvin, Y.-H. Tse, A. B. P. Lever, Inorg. Chem.
1996, 35, 5445–5452.
[27] A. Sepehrifard, A. Stublla, S. Haftchenary, S. Chen, P. G. Pot-
vin, S. Morin, J. New Mater. Electrochem. Syst. 2008, 11, 281–
285.
Acknowledgments
[28] G. C. Vougioukalakis, T. Stergiopoulos, G. Kantonis, A. G.
Kontos, K. Papadopoulos, A. Stublla, P. G. Potvin, F. Falaras,
manuscript in preparation.
[29] G. N. Connelly, E. W. Geiger, Chem. Rev. 1996, 96, 877–910;
J. R. Aranzaes, M. C. Daniel, D. Astruc, Can. J. Chem. 2006,
84, 288–299.
We thank the Natural Sciences and Engineering Research Council
(Canada) for funding, Professor A. B. P. Lever for helpful dis-
cussions, Professor S. Morin for help with the electrochemical
work, and Wylie O, Dr. G. Koyanagi and D. Kalinina for the DFT
calculations.
[30] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T.
Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, 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, P. Y. Ayala, K. Moro-
kuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G.
Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O.
Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cios-
lowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Ko-
maromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill,
B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, J. A. Pople,
Gaussian 03 (G03W revision C.02), Gaussian, Inc., Wall-
ingford, CT, 2004.
[1] P. G. Potvin, P. U. Luyen, J. Bräckow, J. Am. Chem. Soc. 2003,
125, 4894–4906.
[2] M. W. Cooke, G. S. Hanan, F. Loiseau, S. Campagna, M. Wat-
anabe, Y. Tanaka, J. Am. Chem. Soc. 2007, 129, 10479–10488.
[3] E. C. Constable, E. L. Dunphy, C. E. Housecroft, M. Neu-
burger, S. Schaffner, F. Schapera, S. R. Batten, Dalton Trans.
2007, 4323–4332.
[4] M. Graetzel, J. Photochem. Photobiol. C: Photochem. Rev.
2003, 4, 145–153.
[5] F. Barigelletti, L. Flamigni, V. Balzani, J.-P. Collin, J.-P. Sauv-
age, A. Sour, E. C. Constable, A. M. W. Cargill Thompson, J.
Chem. Soc., Chem. Commun. 1993, 942–944.
[6] M. Graetzel, Energy Resources through Photochemistry and Ca-
talysis, Academic Press, New York, NY, USA, 1983.
[7] L. J. Nurkkala, R. O. Steen, H. K. J. Friberg, J. A. Haeggs-
troem, P. V. Bernhardt, M. J. Riley, S. J. Dunne, Eur. J. Inorg.
Chem. 2008, 4101–4110.
[8] R. J. Crutchley, A. B. P. Lever, J. Am. Chem. Soc. 1980, 102,
7128–7129.
[31] S. Grimme, Rev. Comput. Chem. 2004, 20, 153–218.
[32] S. I. Gorelsky, SWizard program, CCRI, University of Ottawa,
Ottawa, Canada, 2008, http://www.sg-chem.net/.
Received: February 3, 2010
[9] R. Liegghio, P. G. Potvin, A. B. P. Lever, Inorg. Chem. 2001,
40, 5485–5486.
[10] F. A. Mutlaq, P. G. Potvin, A. I. Philippopoulos, P. Falaras,
Eur. J. Inorg. Chem. 2007, 2121–2128.
Published Online: May 12, 2010
3050
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 3040–3050