Multiple one-electron oxidation and reduction of trinuclear bis(2,4-pentanedionato)ruthenium complexes with substituted diquinoxalino[2,3-a:2′,3′-c]phenazine ligands

Article abstract of DOI:10.1016/j.poly.2007.03.030

The complexes (μ₃-L¹/L²)[Ru(acac)₂]₃, acac^- = 2,4-pentanedionato, L¹ = 2,3,8,9,14,15-hexachlorodiquinoxalino[2,3-a:2′,3′-c]phenazine and L² = 2,3,8,9,14,15- hexamethyldiquinoxalino[2,3-a:2′,3′-c]phenazine, undergo stepwise one-electron oxidation involving a total of three electrons and stepwise one-electron reduction with three (L²) or four electrons (L¹). All reversibly accessible states were characterized by UV-Vis-NIR spectroelectrochemistry. Oxidation leads to mixed-valent intermediates {(μ₃-L)[Ru(acac)₂]₃}⁺ and {(μ₃-L)[Ru(acac)₂]₃}²⁺ of which the Ru^IIIRu^IIRu^II combinations exhibit higher comproportionation constants K_c than the Ru^IIIRu^IIIRu^II states - in contrast to a previous report for the unsubstituted parent systems {(μ₃-L³)[Ru(acac)₂]₃}^+/2+, L³ = diquinoxalino[2,3-a:2′,3′-c]phenazine. No conspicuous inter-valence charge transfer absorptions were observed for the mixed-valent intermediates in the visible to near-infrared regions. The monocations and monoanions were characterized by EPR spectroscopy, revealing rhombic ruthenium(III) type signals for the former. Electron addition produces ruthenium(II) complexes of the reduced forms of the ligands L, a high resolution EPR spectrum with ¹⁴N and ^35,37Cl hyperfine coupling and negligible g anisotropy was found for {(μ₃-L¹)[Ru(acac)₂]₃}^-. DFT calculations of (μ₃-L¹)[Ru(acac)₂]₃ confirm several ligand-centered low-lying unoccupied MOs for reduction and several metal-based high-lying occupied MOs for electron withdrawal, resulting in low-energy metal-to-ligand charge transfer (MLCT) transitions.

Full text of DOI:10.1016/j.poly.2007.03.030

Polyhedron 26 (2007) 3409–3418
www.elsevier.com/locate/poly
Multiple one-electron oxidation and reduction of trinuclear
bis(2,4-pentanedionato)ruthenium complexes with substituted
diquinoxalino[2,3-a:20,30-c]phenazine ligands
a
b
a
c
Sandeep Ghumaan , Biprajit Sarkar , Mahendra P. Patil , Jan Fiedler ,
a,
b,
a,
*
Raghavan B. Sunoj *, Wolfgang Kaim *, Goutam Kumar Lahiri
a
Department of Chemistry, Indian Institute of Technology – Bombay, Powai, Mumbai 400076, India
Institut fu¨r Anorganische Chemie, Universita¨t Stuttgart, Pfaffenwaldring 55, D – 70550 Stuttgart, Germany
J. Heyrovsky´ Institute of Physical Chemistry, v.v.i., Academy of Sciences of the Czech Republic, Dolejsˇkova 3, CZ – 18223 Prague, Czech Republic
b
c
Received 18 January 2007; accepted 15 March 2007
Available online 30 March 2007
Abstract
The complexes (l3-L1/L2)[Ru(acac)2]3, acacꢀ = 2,4-pentanedionato, L1 = 2,3,8,9,14,15-hexachlorodiquinoxalino[2,3-a:20,30-c]phena-
zine and L2 = 2,3,8,9,14,15- hexamethyldiquinoxalino[2,3-a:20,30-c]phenazine, undergo stepwise one-electron oxidation involving a total
of three electrons and stepwise one-electron reduction with three (L2) or four electrons (L1). All reversibly accessible states were char-
acterized by UV–Vis–NIR spectroelectrochemistry. Oxidation leads to mixed-valent intermediates {(l3-L)[Ru(acac)2]3}+ and {(l3-
L)[Ru(acac)2]3}2+ of which the RuIIIRuIIRuII combinations exhibit higher comproportionation constants Kc than the RuIIIRuIIIRuII
states – in contrast to a previous report for the unsubstituted parent systems {(l3-L3)[Ru(acac)2]3}+/2+, L3 = diquinoxalino[2,3-
a:20,30-c]phenazine. No conspicuous inter-valence charge transfer absorptions were observed for the mixed-valent intermediates in the
visible to near-infrared regions. The monocations and monoanions were characterized by EPR spectroscopy, revealing rhombic ruthe-
nium(III) type signals for the former. Electron addition produces ruthenium(II) complexes of the reduced forms of the ligands L, a high
resolution EPR spectrum with 14N and 35,37Cl hyperfine coupling and negligible g anisotropy was found for {(l3-L1)[Ru(acac)2]3}ꢀ. DFT
calculations of (l3-L1)[Ru(acac)2]3 confirm several ligand-centered low-lying unoccupied MOs for reduction and several metal-based
high-lying occupied MOs for electron withdrawal, resulting in low-energy metal-to-ligand charge transfer (MLCT) transitions.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: DFT calculations; EPR; Radical complexes; Ruthenium; Electrochemistry; Spectroelectrochemistry
1. Introduction
to two homovalent and to two different mixed-valent
states, the odd-electron form RuIIIRuIIRuII and the even-
electron species RuIIIRuIIIRuII. Electron transfer to or
from a trigonally symmetrical system can involve partial
occupation of degenerate orbitals with the resulting ten-
dency for distortion.
Following a previous report [8] on the synthesis and
spectroelectrochemical characterization of unsubstituted
(l3-L3)[Ru(acac)2]3 (3), a trinuclear complex of the
extended p ligand L3 = diquinoxalino[2,3-a:20,30-c]phena-
zine, we now describe the results for the redox systems
{(l3-L1)[Ru(acac)2]3}n, 1n, and {(l3-L2)[Ru(acac)2]3}n, 2n,
with the hexasubstituted ligands L1 and L2 (Scheme 1).
In contrast to the vast number of mixed-valent ligand-
bridged diruthenium complexes [1] there have been
relatively few reports on the electron transfer reactivity of
symmetrically trinuclear compounds [2–8]. If the common
RuIII/RuII redox pairs are involved, this arrangement leads
*
Corresponding authors. Tel.: +91 22 2576 7173; fax: +91 22 2572 3480
(R.B. Sunoj), tel.: +49 711 685 64170; fax: +49 711 685 64165 (W. Kaim),
tel.: +91 22 2576 7159; fax: +91 22 2572 3480 (G.K. Lahiri).
E-mail addresses: sunoj@chem.iitb.ac.in (R.B. Sunoj), kaim@iac.uni-
stuttgart.de (W. Kaim), lahiri@chem.iitb.ac.in (G.K. Lahiri).
0277-5387/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2007.03.030

3410
S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
R
metric measurements were carried out using a PAR model
R
273A electrochemistry system. Platinum wire working and
auxiliary electrodes and an aqueous saturated calomel ref-
erence electrode (SCE) were used in a three-electrode con-
figuration. The supporting electrolyte was 0.1 M Et4NClO4
and the solute concentration was ꢁ10ꢀ3 M. The half-wave
potential E2ꢂ98 was set equal to 0.5(Epa + Epc), where Epa
and Epc are the anodic and cathodic cyclic voltammetric
peak potentials, respectively. A platinum wire-gauze work-
ing electrode was used in the coulometric experiments. The
elemental analysis was carried out with a Perkin–Elmer
240C elemental analyzer. Electrospray mass spectra were
recorded on a Micromass Q-ToF mass spectrometer.
O
II
Ru
O
O
O
N
O
R
R
N
N
N
O
O
II
Ru
N
O
R
O
O
N
II
Ru
O
O
R
O
O
O
L1, R = Cl,
1
and
L2, R = Me, 2
O
L3, R = H,
3
2.3. Preparation of complexes 1 and 2
Scheme 1.
2.3.1. [{(acac)2RuII}3(l-L1)] (1)
DFT calculation results are reported for the hexachloro-
substituted 1, involving structure optimization and MO
description. Voltammetric methods were employed to
establish multistep reversible electron transfer while spec-
troelectrochemistry in the UV–Vis–NIR region as well as
EPR in the X band (9.5 GHz) were used to characterize
the intermediates.
The starting complex Ru(acac)2(CH3CN)2 (100 mg,
0.26 mmol) and the ligand L1 (52 mg, 0.087 mmol) were
added to 20 ml of ethanol, and the mixture was heated to
reflux for 6 h under a dinitrogen atmosphere. The initial
orange solution gradually changed to dark green. The sol-
vent was evaporated to dryness and the residue was puri-
fied by using a silica gel column. Initially, a red
compound corresponding to Ru(acac)3 was eluted by
CH2Cl2–CH3CN (25:1 v/v). With CH2Cl2–CH3CN (20:1
v/v), a green compound corresponding to 1 was then sepa-
2. Experimental
2.1. Materials
rated.
Yield:
45 mg,
35%.
Anal.
Calc.
for
C54H48Cl6N6O12Ru3: C, 43.55; H, 3.25; N, 5.65. Found:
C, 43.20; H, 3.10; N, 5.32%. 1H NMR (400 MHz) in
(CD3)2SO d (ppm), 8.29/8.27, 8.21/8.19 (s, (CH(L1)),
5.67/5.66/5.64 (s, CH(acac)), 2.34; 1.68/1.67 (s, CH3(acac)).
The positive ion electrospray mass spectrum of 1 shows the
molecular ion peak centered (m/z) at 1489.58, correspond-
ing to 1+ (calculated molecular mass, 1488.94) and a peak
centered (m/z) at 1388.53 corresponding to {1-acac}+ (cal-
culated molecular mass, 1389.83).
The precursor compounds Ru(acac)2(CH3CN)2 [9],
2,3,8,9,14,15-hexachlorodiquinoxalino[2,3-a:20,30-c]phena-
zine (L1) [10] and 2,3,8,9,14,15-hexamethyldiquinoxa-
lino[2,3-a:20,30-c]phenazine (L2) [11]1 were prepared
according to the reported procedures. Other chemicals
and solvents were reagent grade and used as received.
For spectroscopic and electrochemical studies HPLC grade
solvents were used.
2.2. Physical measurements
2.3.2. [{(acac)2RuII}3(l-L2)] (2)
The complex Ru(acac)2(CH3CN)2 (100 mg, 0.26 mmol)
and L2 (41 mg, 0.087 mmol) were added to 20 ml of ethanol
and the mixture was heated to reflux for 12 h under a dini-
trogen atmosphere. The initial orange solution gradually
changed to bluish green. The solvent was evaporated and
the residue purified by using a silica gel column. Initially,
a red compound corresponding to Ru(acac)3 was eluted
by CH2Cl2–CH3CN (25:1 v/v). With CH2Cl2–CH3CN
(5:1 v/v), a blue compound corresponding to 2 was sepa-
rated. Yield: 55 mg, 46%. Anal. Calc. for C60H66N6O12Ru3:
C, 52.62; H, 4.86; N, 6.14. Found: C, 52.83; H, 4.85; N,
UV–Vis–NIR spectroelectrochemical studies were per-
formed in CH3CN/0.1 M Bu4NPF6 at 298 K using an opti-
cally transparent thin layer electrode (OTTLE) cell [12] in
connection with a J&M TIDAS spectrophotometer. FT-
IR spectra were taken on a Nicolet spectrophotometer with
samples prepared as KBr pellets. Solution electrical con-
ductivity was checked using a Systronic 305 conductivity
1
bridge. H NMR spectra were obtained with a 400 MHz
Varian FT spectrometer. The EPR measurements were
made in a two-electrode capillary tube [13] with an X-band
Bruker system ESP300, equipped with a Bruker ER035M
gaussmeter and a HP 5350B microwave counter. Cyclic
voltammetric, differential pulse voltammetric and coulo-
1
6.22%. H NMR (400 MHz) in (CD3)2SO d (ppm), 8.34/
8.28; 8.15/7.94 (s, CH(L2)), 5.76/5.52 (s, CH(acac)), 2.41/
2.21/1.56 (multiplet, CH3). The positive ion electrospray
mass spectrum of 2 shows the molecular ion peak centered
(m/z) at 1365.95, corresponding to 2+ (calculated molecu-
lar mass, 1366.43).
1
The heterocyclic system was referred to as 5,6,11,12,17,18-hexaazatri-
naphthalene (Ref. [11]).

S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
3411
2.4. Computational details
Complete geometry optimization without symmetry
constraints was carried out for 1 in the D3/K3 configura-
tion, using the density functional theory method at the
B3LYP level. The D3/K3 configuration was chosen for con-
venience, isomerism was shown to affect the electronic
structure very little [1f,1i,1j,3]. All elements except ruthe-
nium were assigned a 6-31G* basis set. The Stuttgart–Dres-
den effective core potential (ECP), representing 19 core
electrons, along with valence basis sets (SDD) was used
for the ruthenium atoms [14a,14b]. Calculations were per-
formed with GAUSSIAN 98 [14c]. Vertical electronic excita-
tions based on B3LYP optimized geometries were
computed using the time-dependent density functional the-
ory (TD-DFT) formalism [14d,14e] with the B3LYP func-
tional using the above combination of basis sets. Extinction
coefficients were obtained from calculated oscillator
strengths using Gaussian line shape and a full width at half
maximum (fwhm) of 0.4 eV = 3228 cmꢀ1 [14f]. The molec-
ular orbital compositions have been determined using the
programme AOMIX using the B3LYP/SDD,6-31G* wave
function [14g]. Visual inspection of all key orbitals was
done with MOLDEN [14h] to assign the nature of various
electronic transitions.
3. Results and discussion
Regardless of the molar ratio of the reactants, the trinu-
clear complexes 1 and 2 were invariably obtained from
reactions between Ru(acac)2(CH3CN)2 and either of the
free ligands. The well established [9,15] p electron donor
capacity of the RuII(acac)2 complex fragment is held
responsible for enhancing the metal affinity of mono- and
dinuclear intermediates through extensive p conjugation
until full charge transfer supported coordinative saturation
[16] is achieved.
Fig. 1. Electrospray mass spectra of: (a) 1 and (b) 2 in CH3CN.
Positive ion electrospray mass spectroscopy of 1
(Fig. 1a) and 2 (Fig. 1b) support the identity of the new
hexasubstituted complexes. The 1H NMR spectra of 1
and 2 exhibit two sets of signals both for the aromatic
CH protons of L1 and L2 and for the CH/CH3 groups of
the acacꢀ ligands (Fig. 2, Section 2). The presence of two
sets of signals in each case indicates the existence of mix-
tures of two diastereoisomers (D3/K3 or D2K/K2D) in solu-
tion. The two closely spaced singlets for all protons
involved with a particular diastereoisomeric form result
from the pseudo C3-symmetry of the molecule. In related
cases such isomerism was shown to affect the electrochem-
ical and spectroscopic response only slightly [1f,1i,1j,3], we
therefore used the product distributions as obtained and as
characterized by NMR (Fig. 2).
Full geometry optimization of complex 1 in the D3/K3
configuration was carried out at the B3LYP level of theory
by using the SDD basis set for Ru and the 6-31G* basis for
all other atoms (Fig. 3). Important optimized structural
parameters of complex 1 are summarized in Tables S1
Fig. 2. 1H NMR spectrum of 1 in (CD3)2SO. Inserts show expanded parts
of the spectrum in the range of: (a) 5.5–8.5 ppm and (b) 1.5–2.5 ppm.

3412
S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
Fig. 3. Optimized structure of complex 1 at the B3LYP level of theory by using the SDD basis set for Ru and 6-31G* for all other atoms (no symmetry
constraints).
and S2. The Ru–N and Ru–O(acac) distances are found to
vary only slightly and range between 2.06 and 2.08 A. The
electronic nature of the substituents, Me or Cl. In case
of 1, four distinct reduction processes were observed
whereas only three waves were detected for 2 (Table 1,
Fig. 4). In addition, 1 and 2 display three successive oxi-
dation processes corresponding to the reversible transi-
˚
N(4)–C(5) and N(7)–C(6) p bonds are found to be elon-
˚
gated by about 0.023 A in 1 as compared to the free ligand
L1, and this N–C p bond elongation is nearly the same at
each ruthenium center which signifies the comparable
extent of p back-bonding from each RuII ion (vide infra).
The metal centers were found to be separated by about
tions
RuIIRuIIRuIII ꢀ RuIIRuIIRuII
(couple
I),
RuIIRuIIIRuIII ꢀ RuII RuIIRuIII (couple II), and RuIIIR-
uIIIRuIII ꢀ RuIIRuIIIRuIII (couple III) (Table 1, Fig. 4).
These potentials are lower for R = Me (complex 2) than
for R = Cl (complex 1) (Table 1), as expected from the
different electron donating and withdrawing abilities.
Remarkably, the separation between the redox potentials
for complex oxidation leads to larger comproportionation
˚
6.9 A (Table S1).
Electrochemically the complexes 1 and 2 exhibit multi-
ple reversible one-electron reduction waves which are
associated with the highly conjugated bridging ligands
(L1/L2); the potentials also vary systematically with the
Table 1
Electrochemical dataa
Compound
Eꢂ298=VðDEp; mVÞ
Couples
I
II
III
IV
V
VI
VII
1
0.16 (70)
0.01 (60)
0.34 (70)
0.56 (70)
0.41 (90)
0.68 (70)
0.79 (90)
0.62 (70)
1.16 (70)
ꢀ0.84 (70)
ꢀ1.09 (60)
ꢀ0.46 (80)
ꢀ0.98 (70)
ꢀ1.20 (80)
ꢀ1.07 (70)
ꢀ1.23 (80)
ꢀ1.58 (100)
ꢀ1.61 (80)
ꢀ1.71 (90)
2
3b
a
In CH3CN/0.1 M Et4NClO4 vs. SCE. DEp: peak potential difference.
b
Ref. [8].

S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
3413
is observed (Table 2) due to increasing Coulombic inter-
actions; for instance, Kc values for the corresponding par-
ent complex 3n+ with R = H have been reported at 105.7
(Kc1) < 108.0 (Kc2). In comparison to 3n Kc1 appears to be
slightly greater for 1n+ or 2n+, however, Kc2 is much lar-
ger for 3n+ than for 1n+ or 2n+
.
The electrolyte-dependent [17a,17b] comproportiona-
tion constants Kc reflect several different effects such as
the electrostatic interaction, an inductive factor, the reso-
nance exchange, the antiferromagnetic exchange stabiliza-
tion, and ion-pairing effects [17c,17d,17e]. Their
comparative interpretation is thus not always as straight-
forward as commonly believed. In the present case, the
smaller range for the dicationic, i.e. rather highly charged
RuIIRuIIIRuIII forms may reflect diminished Coulombic
effects due to enhanced charge delocalization by the six p
donating Cl (12+) or inductively electron donating Me sub-
stituents (22+). Not unexpectedly, the potentials for corre-
sponding diastereoisomers are rather similar [3,18,19].
Nevertheless, the profile of the cyclic voltammograms, in
particular for 2, suggests the existence of two diastereoiso-
Fig. 4. Cyclic voltammograms (—) and differential pulse voltammograms
(---) in CH3CN/0.1 M Et4NClO4 of: (a) 1 and (b) 2; potentials against
SCE.
1
mers in solution as has also been observed by H NMR
(Fig. 2, Section 2).
constants for the first intermediate (Kc1) than for the sec-
The EPR spectra taken of monocations and monoa-
nions are in agreement with expectations [8] for the fron-
tier orbital situation and with the DFT calculated orbital
ond form, RuIIRuIIIRuIII (Kc2), both for 1n+ and 2n+
,
respectively (RTlnKc = nF(DE) [1]). Usually, the opposite
Table 2
The variation of Kc and IVCT values in symmetrical triruthenium complexes
N
-S
S-
N(CH3)2
N
N
N
N
N
N
N
N
N
N
NH-
S-
L5
N
N
L4
N
L6
dmap
Formula
Kac1
Kac2
(IVCT)b1
(IVCT)b2
Ref.
k (nm) (e, Mꢀ1 cmꢀ1
)
k (nm) (e, Mꢀ1 cmꢀ1
)
c
c
c
c
c
c
[{(acac)2Ru}3(l3-L1)] (1)
[{(acac)2Ru}3(l3-L2)] (2)
[{(acac)2Ru}3(l3-L3)] (3)
D3/K3-[{(bpy)2Ru}3(l3-L4)]6+
D2K/K2D-[{(bpy)2Ru}3 (l3-L4)]6+
[{(phen)2Ru}3 (l3-L4)]6+
[{(bpy)2Ru}3 (l3-L5)]3+
106.8
106.8
105.7
103.9
103.7
104.5
103.7
103.2
1020
103.9
103.6
108.1
104.1
104.1
104.7
104.1
104.1
103.4
106.4
1027
1026
1014
1021
1013
1013
1011
this work
this work
[8]
[3b]
[3b]
[3d]
[4]
[4]
[5]
[6,7]
[2]
[2]
2150 (3315)
2155 (4180)
c
2130 (2990)
2236 (2150)
c
1900 (2300)
d
1900 (2400)
d
[{(phen)2Ru}3 (l3-L5)]3+
[{(pap)2Ru}3 (l3-L5)]3+
d
d
[{([9]-ane-S3)Ru(l-L6)}3]3+
[Ru3(O)(CH3CO2)6(CO)(dmap)2]
[Ru3(O)(CH3CO2)6(CO)(py)2]
[Ru3(O)(CH3CO2)6(CO)(CNXy)2]
[Ru3(O)(CH3CO2)6(CNXy)(py)2]
[Ru3(O)(CH3CO2)6(CNXy)2(py)]
[Ru3(O)(CH3CO2)6(CNXy)2(dmap)]
[Ru3(O)(CH3CO2)6(CNXy)3]
a
103.0
1016
950 (2950)
2085 (8400)
d
d
d
d
d
d
d
d
d
d
d
d
1014.6
1010.2
1013.5
1018
1022
108.8
[2]
[2]
[2]
[2]
1004 (12200)
1002 (6400)
[2]
Kc: comproportionation constant for Ru3III,II,II(Kc1) and Ru3III,III,II(Kc2).
b
c
II,III,III
IVCT: inter-valence charge-transfer transitions for the Ru3II,II,III(IVCT1) and Ru3
states (IVCT2).
Not observed.
Not reported.
d

3414
S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
Table 3
compositions (see Table 4). The cations exhibit rhombic
RuIII type signals with considerable g anisotropy in the
immobilized state (Fig. 5, Table 3); this and the average
g above 2 are indicative of low-spin d5 species in a dis-
torted octahedral situation [20].
EPR data of monocations and monoanions
Compound
g1
g2
g3
giso
Dg
Ref.
1Åꢀ
1+
2Åꢀ
2+
3Åꢀ
3+
a
2.0053
2.131b
1.9706
2.100b
1.969b
2.128b
<0.10
0.583
<0.10
0.589
0.133
0.455
this work
this work
this work
this work
[8]
a
a
a
2.398
a
2.140
a
1.815
a
In contrast, the monoanions exhibit EPR features with-
2.370
2.058
2.332
2.107
2.058
2.154
1.782
1.924
1.877
out detectable
g anisotropy at X band frequency
(9.5 GHz), the isotropic g values near 2.00 (Table 3) signify
vanishing participation of atoms with large spin–orbit cou-
pling constants at the spin distribution [21]. The complex
anion 1Åꢀ was even found to exhibit hyperfine structure
caused by coupling with 14N (I = 1) and 35,37Cl nuclei
(I = 3/2) at the order of about 1 G (Fig. 6). These values
are compatible with the composition of the lowest unoccu-
pied MO (see Fig. 8), small, unresolved hyperfine interac-
tion with the total of 18 methyl hydrogen atoms is likely
to obscure the hyperfine structure for 2Åꢀ. Evidence for
symmetry lowering on the EPR time scale has not been
observed, there is probably rapid equilibration of any
distortions.
[8]
Not resolved.
gav (calculated).
b
a
The signal detectability at room temperature, the small
line-width allowing for EPR resolution, the isotropic g fac-
tors close to the free electron value of 2.0023, and the
absence of metal (99,101Ru, I = 5/2) hyperfine splitting
and of g anisotropy at 9.5 GHz are all in agreement with
an almost exclusively ligand-centered spin [21]. In fact,
the results suggest much less metal participation than in
complexes of [Ru(bpy)2]2+ with anion radical ligands [13],
probably because of the donor characteristics of Ru(acac)2
3350
3400
3450
B/G
3500
3550
b
in comparison to [Ru(bpy)2]2+
.
Exp.
The energies and the composition of the frontier
Kohn–Sham orbitals (Figs. 7 and 8) have been examined
using the AOMIX programme [14g,22]. The filled levels
(between HOMO and HOMO ꢀ 6) are predominantly
metal-centered orbitals with total Ru contributions of
around 65% and additional mixing with the p orbitals
of acacꢀ (Table 4). The unfilled levels, on the other hand,
are primarily located over the L1 unit, having p*
Sim.
3410
3420
3430
3440
B/G
Fig. 6. EPR spectra of 1ꢀ at: (a) 110 K and (b) 298 K (top: experimental,
bottom: simulated) in CH3CN/0.1 M Bu4NPF6. Simulation parameters
1.8 G (6 · 14N) and 0.6/0.5 G (6 · 35Cl/37Cl in natural abundance).
characteristics2. Atomic orbital contributions from the
metal centers towards the LUMO are successfully adopted
in explaining the back-bonding within metal complexes in
more quantitative fashion [23]. In the present case, the con-
2
We have identified the p* orbital of L1 as the one in the xy-plane by
2000
2500
3000
3500
4000
4500
careful analysis of the MO coefficient of the LUMO, which is mainly
located on p* of L1 (66%) with minor contribution from dp on Ru.
Further, the d-orbital in the same plane as that of L1 is taken as the dxy
B/G
Fig. 5. EPR spectrum of 2+ in CH3CN/0.1 M Bu4NPF6 at 4 K (artefact
radical signal from cavity at 3500 G).
and the remaining orthogonal d-orbitals as dxz and dyz
.

S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
3415
HOMO
LUMO
HOMO-1
LUMO+1
HOMO-2
LUMO+2
Fig. 7. Representative orbital contour diagrams of the HOMOs of 1.
Fig. 8. Representative orbital contour diagrams of the LUMOs of 1.
Table 4
Orbital energies and atomic orbital contributions for the frontier molecular orbitals of 1 computed without symmetry constraints
MO
Orbital energy (eV)
%Ru1
%Ru2
%Ru3
%Rutotal
%L1
%acac
LUMO + 2
LUMO + 1
LUMO
ꢀ2.95 (23790)a
ꢀ2.96 (23870)
ꢀ3.16 (25490)
ꢀ4.86 (39200)
ꢀ4.91 (39600)
ꢀ4.98 (40170)
6.5
5.2
8.1
29.1
10.7
25.5
4.6
6.7
8.4
30.3
8.2
26.6
2.9
3.9
12.2
6.5
47.1
16.1
14.0
15.8
28.7
65.9
66.0
68.2
82.7
81.3
66.1
7.2
5.3
6.6
3.3
2.9
5.3
26.9
29.7
25.2
HOMO
HOMO ꢀ 1
HOMO ꢀ 2
a
Values in parentheses indicate orbital energy in cmꢀ1
.

3416
S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
tributions from the ruthenium atoms towards the LUMO
are found to be about the same on average, the effective tri-
gonal symmetrical situation being further confirmed by the
optimized geometrical parameters around the ruthenium
atoms. Nonetheless, the metals contribution especially to
the calculated LUMO seems to be overestimated, consider-
ing the EPR results with their suggestion of marginal ruthe-
nium influence.
UV–Vis–NIR absorption spectroelectrochemistry was
performed for the multistep series 13+–14ꢀ (Figs. 9 and
10) and 23+–23ꢀ (Table 5).
The results show similar trends for both series in agree-
ment with the donor effect of both kinds of substituents.
The free ligands have long-wavelength p–p* absorption
bands at about 412 nm (e ca. 6 · 104 Mꢀ1 cmꢀ1) whereas
the lowest energy absorption maxima of 1 and 2 lie around
660 nm (e ca. 1.8 · 104 Mꢀ1 cmꢀ1). These bands are
assumed to derive from allowed metal-to-ligand charge
transfer (MLCT) transitions, given by the experimentally
and computationally suggested presence of several high-
lying occupied d levels and low-lying unoccupied ligand
p* orbitals.
A selected list of computed vertical excitation energies
using the time-dependent density functional theory (TD-
DFT) is therefore given in Table S2. For 1, two transi-
tions with fairly high oscillator strengths and calculated
extinction coefficients have been obtained, one at
781 nm (e = 2914 Mꢀ1 cmꢀ1) and another at 771 nm
(e = 4993 Mꢀ1 cmꢀ1). The quantitative assessment of
these orbital overlap-favored MLCT processes compares
with experimental values for
1 of 890 nm(sh) and
Fig. 9. UV–Vis–NIR spectroelectrochemistry for the conversions: (a)
1 ! 1+, (b) 1+ ! 12+ and (c) 12+ ! 13+in CH3CN/0.1 M Bu4NPF6.
680 nm (e = 23900 Mꢀ1 cmꢀ1). As must be expected for
Fig. 10. UV–Vis–NIR spectroelectrochemistry for the conversions: (a) 1 ! 1ꢀ, (b) 1ꢀ ! 12ꢀ, (c) 12ꢀ ! 13ꢀ and (d) 13ꢀ ! 14ꢀ in CH3CN/0.1 M
Bu4NPF6.

S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
3417
Table 5
UV–Vis–NIR data for complexes 1 and 2 from spectroelectrochemistrya
Complex
kmax (nm) (e, Mꢀ1 cmꢀ1
)
13+
12+
1+
1
687 (18300), 440 (30000), 366 (38700), 290 (38600), 247 (52000)
865sh, 700sh, 602 (16800), 438sh, 362 (40900), 275sh, 248 (48400)
685sh, 620 (18800), 435sh, 350 (43100), 270 (46500), 252 (46900)
890sh, 680 (23900), 342 (47600), 271 (50500), 250sh
1ꢀ
12ꢀ
13ꢀ
14ꢀ
23+
22+
2+
2
1060sh, 783 (17200), 616 (12300), 340 (40600), 270 (54900)
1065sh, 785 (17400), 617 (12500), 511 (9500), 339 (39200), 272 (58600)
1018 (9800), 865 (14800), 589 (13500), 515 (12600), 410 (27200), 338 (37000), 271 (63000)
1145sh, 880 (12300), 573 (15700), 512 (17100), 380sh, 340 (34900), 270 (67100)
618 (7900), 424 (20400), 350 (28900), 289sh, 243 (29400)
584 (8800), 414 (19100), 347 (29100), 326sh, 276sh, 243 (28500)
614 (10700), 416sh, 334 (29100), 276 (25800), 243 (25300)
644 (12100), 331 (30400), 275 (27300), 242 (25700)
2ꢀ
22ꢀ
23ꢀ
a
723 (9400), 582 (8000), 337 (27400), 275 (28500), 247 (26500)
840sh, 747 (9500), 546 (7100), 415sh, 336 (25000), 274 (29800), 250sh
1017 (4700), 861 (8200), 787 (8900), 533 (8200), 411 (15800), 333 (23700), 274 (32600), 256sh
In CH3CN/0.1 M Bu4NPF6.
a situation with several filled d orbitals from three d6
metal centers and a number of ligand p* MOs (Fig. 4)
there are several other MLCT transitions of lower energy
but much smaller oscillator strengths (poor orbital over-
lap) which are not experimentally detectable (Table S2).
A similar reason may preclude [8] the detection of the
inter-valence charge transfer (IVCT) bands of the
mixed-valent forms 1+, 12+, 2+ and 22+ in the typical
near-infrared region [1,3]. In spite of careful spectral
search even in the mid-infrared region we could not
establish such bands (Table 5), although [Ru(bpy)2]2+
complexes of the related 1,4,5,8,9,12-hexaazatriphenylene
ligands exhibit such features [3b,3c]. This absence of
detectable long-wavelength absorption for otherwise
clearly identified RuII/RuIII mixed-valent intermediates
has been noted before not only for 3+ and 32+;8 dinucle-
ating chelate acceptor bridges connecting ruthenium com-
plex fragments, especially Ru(acac)2, have previously
been described to yield ‘‘IVCT silent’’ mixed-valent species
[24]. An alternative interpretation [3a] involving the
obscuring of the IVCT band by the MLCT absorptions
in the visible would imply an usually strong metal–metal
interaction which we consider less likely.
4. Conclusion
We have shown that EPR spectroscopy in conjunction
with electrochemistry can provide an alternative way to
identify mixed-valent intermediates, especially when the
expected inter-valence CT bands in the long-wavelength
region are not observable. The six fold substitution of
the large heterocyclic p system clearly affects the physical
properties, allowing not only for multiple oxidation at
the metals but also for several ligand based reduction
processes.
Acknowledgements
Financial support received from the Department of Sci-
ence and Technology (DST), Council of Scientific and
Industrial Research (CSIR)-New Delhi (India), the Deut-
sche Forschungsgemeinschaft (DFG), Fonds der Chemis-
chen Industrie (FCI), the Deutsche Akademische
Austauschdienst (DAAD) (Germany), is gratefully
acknowledged.
Appendix A. Supplementary material
Within the spectroelectrochemically monitored electron
transfer series (Figs. 9,10, and Table 5) it is remarkable
that the second oxidation and reduction involve very lit-
tle spectral changes. This may suggest spin pairing of the
doubly reduced or oxidized forms in agreement with the
absence of EPR triplet features. Although they can be
generally assigned as MLCT and, at higher energies, as
intra-ligand (IL) absorptions, the detailed identification
of the major bands of the spectroelectrochemically gener-
ated complex ions (Table 5) will not be attempted and
will have to await open-shell calculations. Given the
photophysical and intercalating potential of such large
p systems [25] further investigations of their electronic
structures are warranted.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.poly.2007.03.
030.
References
[1] (a) C. Creutz, Prog. Inorg. Chem. 30 (1983) 1;
(b) D.E. Richardson, H. Taube, Coord. Chem. Rev. 60 (1984) 107;
(c) K.D. Demadis, D.C. Hartshorn, T.J. Meyer, Chem. Rev. 101
(2001) 2655;
(d) J.T. Hupp, in: J.A. McCleverty, T.J. Meyer (Eds.), Comprehen-
sive Coordination Chemistry II, Elsevier Science, Amsterdam, 2004,
p. 709;
(e) M.D. Ward, Chem. Soc. Rev. 24 (1995) 121;

3418
S. Ghumaan et al. / Polyhedron 26 (2007) 3409–3418
(f) F.R. Keene, Chem. Soc. Rev. 27 (1998) 185;
(e) M. Casida, in: D.P. Chong (Ed.), Recent Advances in Density
Functional Methods, vol. 1, World Scientific Press, Singapore, 1995,
p. 155;
(f) J.E. Monat, J.H. Rodriguez, J.K. McCusker, J. Phys. Chem. 106
(2002) 7399;
(g) S.I. Gorelsky, AOMIX: Program for Molecular Orbital Analysis,
YorkUniversity:Toronto, Canada, 1997. <http://www.sg-chem.net/>;
(h) G. Schaftenaar, J.H. Noordik, J. Comput. – Aid. Mol. Des. 14
(2000) 123.
(g) G. Giuffrida, S. Campagna, Coord. Chem. Rev. 135 (1994) 517;
(h) K. Kalyanasundaram, M.K. Nazeeruddin, Inorg. Chim. Acta 226
(1994) 213;
(i) D.M. D’Alessandro, F.R. Keene, Chem. Soc. Rev. 35 (2006) 424;
(j) D.M. D’Alessandro, F.R. Keene, Chem. Rev. 106 (2006) 2270;
(k) R.J. Crutchley, in: J.A. McCleverty, T.J. Meyer (Eds.), Compre-
hensive Coordination Chemistry II, Vol. 2 (2004) 235;
(l) J.-P. Launay, Chem. Soc. Rev. 30 (2001) 386.
[2] K.-I. Ota, H. Sasaki, T. Matsui, T. Hamaguchi, T. Yamaguchi, T.
Ito, H. Kido, C.P. Kubiak, Inorg. Chem. 38 (1999) 4070.
[3] (a) B.D. Yeomans, L.S. Kelso, P.A. Tregloan, F.R. Keene, Eur. J.
Inorg. Chem. (2001) 239;
[15] S. Patra, B. Sarkar, S. Maji, J. Fiedler, F.A. Urbanos, R. Jimenez-
Aparicio, W. Kaim, G.K. Lahiri, Chem. Eur. J. 12 (2006) 489.
[16] W. Kaim, B. Olbrich-Deussner, R. Gross, S. Ernst, S. Kohlmann, C.
Bessenbacher, in: M. Chanon (Ed.), Importance of Paramagnetic
Organometallic Species in Activation, Selectivity and Catalysis,
Kluwer Academic Publishers, Dordrecht, 1989, p. 283.
(b) D.M. D’Alessandro, F.R. Keene, Chem. Eur. J. 11 (2005) 3679;
(c) D.M. D’Alessandro, M.S. Davies, F.R. Keene, Inorg. Chem. 45
(2006) 1656;
`
[17] (a) F. Barriere, W.E. Geiger, J. Am. Chem. Soc. 128 (2006) 3980;
(d) L. Jacquet, A. Kirsch-De Mesmaeker, J. Chem. Soc., Faraday
Trans. 88 (1992) 2471.
[4] S. Kar, T.A. Miller, S. Chakraborty, B. Sarkar, B. Pradhan, R.K.
Sinha, T. Kundu, M.D. Ward, G.K. Lahiri, Dalton Trans. (2003)
2591.
[5] S. Kar, B. Pradhan, R.K. Sinha, T. Kundu, P. Kodgire, K.K. Rao,
V.G. Puranik, G.K. Lahiri, Dalton Trans. (2004) 1752.
[6] N. Shan, S.J. Vickers, H. Adams, M.D. Ward, J.A. Thomas, Angew.
Chem. 116 (2004) 4028.
(b) D.M. Alessandro, F.R. Keene, Dalton Trans. (2004) 3950;
(c) R.J. Crutchley, Adv. Inorg. Chem. 41 (1994) 273;
(d) C.E.B. Evans, M.L. Naklicki, A.R. Rezvani, C.A. White, V.V.
Kondratiev, R.J. Crutchley, J. Am. Chem. Soc. 120 (1998) 13096;
(e) M. Al-Noaimi, G.P.A. Yap, R.J. Crutchley, Inorg. Chem. 43
(2004) 1770.
[18] B. Sarkar, S. Patra, J. Fiedler, R.B. Sunoj, D. Janardanan, S.M.
Mobin, M. Niemeyer, G.K. Lahiri, W. Kaim, Angew. Chem. 117
(2005) 5800.
[7] N. Shan, S.J. Vickers, H. Adams, M.D. Ward, J.A. Thomas, Angew.
Chem., Int. Ed. 43 (2004) 3938.
[8] S. Patra, B. Sarkar, S. Ghumaan, J. Fiedler, W. Kaim, G.K. Lahiri,
Dalton Trans. (2004) 754.
[9] T. Kobayashi, Y. Nishina, K. Shimizu, G.P. Satoˆ, Chem. Lett. (1988)
1137.
[10] G. Kestemont, V.D. Halleux, M. Lehmann, D.A. Ivanov, M.
Watson, Y.H. Geerts, Chem. Commun. (2001) 2074.
[11] V.J. Catalano, W.E. Larson, M.M. Olmstead, H.B. Gray, Inorg.
Chem. 33 (1994) 4502.
[19] B. Sarkar, S. Patra, J. Fiedler, R.B. Sunoj, D. Janardanan, S.M.
Mobin, M. Niemeyer, G.K. Lahiri, W. Kaim, Angew. Chem., Int. Ed.
44 (2005) 5655.
[20] R. Hariram, B.K. Santra, G.K. Lahiri, J. Organomet. Chem. 540
(1997) 155.
[21] W. Kaim, Coord. Chem. Rev. 76 (1987) 187.
[22] S.I. Gorelsky, A.B.P. Lever, J. Organomet. Chem. 635 (2001)
187.
[23] S.I. Gorelsky, A.B.P. Lever, M. Ebadi, Coord. Chem. Rev. 230 (2002)
97.
[12] M. Krejcik, M. Danek, F. Hartl, J. Electroanal. Chem. 317 (1991) 179.
[13] W. Kaim, S. Ernst, V. Kasack, J. Am. Chem. Soc. 112 (1990) 173.
[14] (a) M. Dolg, U. Wedig, H. Stoll, H. Preuss, J. Chem. Phys. 86 (1987)
866;
[24] (a) S. Chellamma, M. Lieberman, Inorg. Chem. 40 (2001) 3177;
(b) See also J. Poppe, M. Moscherosch, W. Kaim, Inorg. Chem. 32
(1993) 2640;
(c) S. Patra, B. Sarkar, S. Ghumaan, J. Fiedler, W. Kaim, G.K.
Lahiri, Inorg.Chem. 43 (2004) 6108.
[25] (a) A. Masschelein, A. Kirsch-De Mesmaeker, C. Verhoeven, R.N.
Hinkens, Inorg. Chim. Acta 129 (1987) L13;
(b) D. Andrae, U. Haussermann, M. Dolg, H. Stoll, H. Preuss, Theor.
Chim. Acta 77 (1990) 123;
(c) M.J. Frisch et al., GAUSSIAN 98, Revision A.6, Gaussian, Inc.,
Pittsburgh, PA, 1998;
(b) R. Blasius, C. Moucheron, A. Kirsch-De Mesmaeker, Eur. J.
Inorg. Chem. (2004) 3071.
(d) C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785;