Influence of halogen atoms on a homologous series of bis-cyclometalated iridium(III) complexes

Article abstract of DOI:10.1021/ic2011474

A series of homologous bis-cyclometalated iridium(III) complexes Ir(2,4-di-X-phenyl-pyridine)₂(picolinate) (X = H, F, Cl, Br) HIrPic, FIrPic, ClIrPic, and BrIrPic has been synthesized and characterized by NMR, X-ray crystallography, UV-vis abso

Full text of DOI:10.1021/ic2011474

Article
pubs.acs.org/IC
Influence of Halogen Atoms on a Homologous Series of
Bis-Cyclometalated Iridium(III) Complexes
Etienne Baranoff,*,† Basile F. E. Curchod,‡ Filippo Monti,§ Frederic Steimer,‡ Gianluca Accorsi,*,§
́
́
Ivano Tavernelli,‡ Ursula Rothlisberger,‡ Rosario Scopelliti,† Michael Gratzel,†
̈
and Md. Khaja Nazeeruddin†
‡
†Laboratory of Photonics and Interfaces and Laboratory of Computational Chemistry and Biochemistry, Institute of Chemical
́
Sciences and Engineering, School of Basic Sciences, Ecole Polytechnique Fed
́
er
́
ale de Lausanne, CH-1015 Lausanne, Switzerland
§Molecular Photoscience Group, Istituto per la Sintesi Organica e la Fotoreattivita,
̀
Consiglio Nazionale delle Ricerche,
Via P. Gobetti 101, 40129, Bologna, Italy
S
* Supporting Information
ABSTRACT: A series of homologous bis-cyclometalated
iridium(III) complexes Ir(2,4-di-X-phenyl-pyridine)2-
(picolinate) (X = H, F, Cl, Br) HIrPic, FIrPic, ClIrPic, and
BrIrPic has been synthesized and characterized by NMR, X-ray
crystallography, UV−vis absorption and emission spectros-
copy, and electrochemical methods. The addition of halogen
substituents results in the emission being localized on the main
cyclometalated ligand. In addition, halogen substitution induces
a blue shift of the emission maxima, especially in the case of the fluoro-based analogue but less pronounced for chlorine and
bromine substituents. Supported by ground and excited state theoretical calculations, we rationalized this effect in a simple manner
by taking into account the σp and σm Hammett constants on both the highest occupied molecular orbital (HOMO) and the
lowest unoccupied molecular orbital (LUMO) energy levels. Furthermore, in comparison with FIrPic and ClIrPic, the impact of
the large bromine atom remarkably decreases the photoluminescence quantum yield of BrIrPic and switches the corresponding
lifetime from mono to biexponential decay. We performed theoretical calculations based on linear-response time-dependent
density functional theory (LR-TDDFT) including spin−orbit coupling (SOC), and unrestricted DFT (U-DFT) to obtain
information about the absorption and emission processes and to gain insight into the reasons behind this remarkable change in
photophysical properties along the homologous series of complexes. According to theoretical geometries for the lowest triplet
state, the large halogen substituents contribute to sizable distortions of specific phenylpyridine ligands for ClIrPic and BrIrPic,
which are likely to play a role in the emissive and nonradiative properties when coupled with the heavy-atom effect.
properties. This is because the properties of molecules are
expected to vary in a systematic way with their chemical struc-
ture. Hammett parameters have been highly successful as a
structural parameter which can be related to chemical and
physical properties of the molecules.7 While the Hammett
equation8 has been initially used for gaining insight into chem-
ical reactions and their mechanism, Hammett parameters are
very powerful to correlate chemical structures to electronic
properties such as redox potentials9 and absorption and emis-
sion maxima.10 Besides inductive and resonance effects, another
influence of substituents on photophysical properties of organic
molecules is through the heavy atom effect, which enhances the
rate of spin-forbidden processes. In this case, halogen-based
homologous series have been classically used.11 On the other
hand, the effect of halogen substituents on organometallic com-
plexes has not been investigated as much as in the case of
organic chromophores. As being part of the same family of
INTRODUCTION
■
Photochemistry and photophysics of transition metal complexes
continue to attract interest in particular because of their relevant
practical applications for solving energy issues.1 Alteration of the
excited-state properties of the transition metal complexes
through systematic chemical modification of the ligands resulted
over decades in enormous number of photoactive molecules.2,3
General design rules have been derived from these results, and it
is nowadays relatively straightforward to predict and therefore to
adjust at will any given property of transition-metal complexes
such as redox potentials, emission or absorption maximum
wavelength, or quantum yield of photoinduced processes. How-
ever, because of the close interplay of all these properties,
the difficulty is to control all these aspects at once and obtain
fully optimized materials with desired properties. In this respect,
the accurate understanding of structure-properties relation-
ships4,5 is highly sought after.6 Since very few parameters are
modified, preferably one only, the study of homologous series
are of fundamental interest to unequivocably understand the
effect of a specific chemical or structural change on the sample
Received: May 29, 2011
Published: December 22, 2011
© 2011 American Chemical Society
799
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
harmonic vibrational frequencies at the same level of theory. At each
ground state (singlet) geometry, LR-TDDFT calculations were
performed using the same basis sets and xc-functional. To cover the
experimental range, the first 140 vertical excitations (singlet and triplet)
were considered. Condensed-phase effects were taken into account
using a self-consistent reaction-field (SCRF) model in which the
solvent is implicitly represented by a dielectric continuum characterized
by its relative static dielectric permittivity ε. The solute, which is placed
in a cavity created in the continuum after spending some cavitation
energy, polarizes the continuum, which in turn creates an electric field
inside the cavity. Within the different approaches that can be followed
to calculate the electrostatic potential created by the polarized
continuum in the cavity, we have employed the integral equation
formalism of the polarizable continuum model (IEFPCM).35 As the
solvent molecules have no time to geometrically rearrange within the
time of a vertical excitation, nonequilibrium solvation36 has been used
for the LR-TDDFT calculations of absorption spectra. A relative
permittivity of 8.93 (35.69) was employed to simulate dichloromethane
(acetonitrile),34 the solvents used in the experimental work. SOC LR-
TDDFT was performed with the ADF200937,38 package using the
ZORA methodology.39 All electron basis sets have been used for all
the atoms (iridium: TZP, all other atoms: DZP). To gain insights into
the phosphorescence behavior of the different iridium compounds, we
optimized the geometry of the first triplet state using unrestricted DFT
(U-DFT) with the same basis set as described before. As suggested by a
recent work,40 we used the xc-functional M05-2X41 for this task,
because of its excellent performance for the emission spectra for a series
of iridium-based compounds. At the minimum energy structure, we
computed the difference in energy between the triplet (T1) and singlet
(S0) with the inclusion of implicit solvent and obtained an estimation
of the first phosphorescence band. See Supporting Information for
additional details on the calculations.
substituents, we believe that homologous series based on
halogen substituents are of particular interest to improve the
understanding of photophysical processes in transition metal
complexes.
Because of their high photoluminescence quantum yields
(PLQY), wide color tunability, relatively short excited-state
lifetime, and general thermal and electrochemical stability,
cyclometalated iridium(III)12−15 complexes are attracting wide-
spread interest for a large range of luminescence-based
applications.16 Chlorine and bromine atoms have been used as
substituents for charged complexes.17 In these complexes, the
triplet levels are known to be localized on the N∧N ancillary
ligand when bipyridine or phenanthroline units are also involved
in the metal core complexation.18 Therefore the substituents on
the main ligands affect primarily the highest occupied molecular
orbital (HOMO) of the complex without significant impact on
the lowest unoccupied molecular orbital (LUMO). Limited
studies exist on neutral cyclometalated iridium complexes with
bromine as substituents19,20 and none using chlorine.
Here, we report the results on a homologous neutral series of
biscyclometalated iridium(III) complexes Ir(2,4-di-X-phenyl-
pyridine)2(picolinate) (X = H, F, Cl, Br). Data for the
nonsubstituted complex HIrPic is scarce in the literature21−25
despite its simple structure. By contrast, fluorine that is arguably
the most utilized halogen substituent has been used because of
its peculiarity to shift emission maxima of the cyclometalated
iridium complexes to higher energy. In particular, FIrPic26,27 is
one of the commercially available and widely studied neutral
cyclometalated iridium complexes and therefore can be taken as
a benchmark for comparing properties of other similar iridium
complexes. Thus, we prepared in addition the homologous
chloro and bromo complexes, ClIrPic and BrIrPic, that have
not yet been reported in the literature. In all cases, the structural
modifications along the series may be considered minor since
chemically similar units have been introduced in the same
position of the main ligand of the complexes. However, we
observed significant changes in the photophysical properties along
this series, in particular for BrIrPic. We used linear-response
time-dependent density functional theory (LR-TDDFT) includ-
ing spin−orbit coupling (SOC), and unrestricted DFT (U-DFT)
to obtain information about this effect. Overall, in addition of the
expected heavy atom effect, the changes in emission properties
are attributed to a deformation of the excited state geometry
caused by the large size of the halogen.
Spectroscopic Measurements. Absorption spectra were re-
corded with a Perkin-Elmer λ950 spectrophotometer. For lumines-
cence experiments, the samples were placed in fluorimetric 1-cm path
cuvettes and, when necessary, purged from oxygen by bubbling with
argon. Uncorrected emission spectra were obtained with an Edinburgh
FLS920 spectrometer equipped with a peltier-cooled Hamamatsu
R928 photomultiplayer tube (185−850 nm). An Edinburgh Xe900
450 W xenon arc lamp was used as exciting light source. Corrected
spectra were obtained via a calibration curve supplied with the
instrument. Photoluminescence quantum yields (PLQY, Φem) in
solution obtained from spectra on a wavelength scale (nm) were
measured according to the approach described by Demas and Crosby42
using air-equilibrated [Ru(bpy)3]Cl2 in aqueous solution [Φem
=
0.028]43 as standard. Emission lifetimes in the ns-μs range were
determined with the single photon counting technique by means of the
same Edinburgh FLS920 spectrometer using a laser diode as excitation
source (1 MHz, λexc = 407 nm, 200 ps time resolution after
deconvolution) and the above-mentioned PMT as detector. Or, with
an IBH single photon counting spectrometer equipped with a thyratron
gated nitrogen lamp working in the range 2−40 kHz (λexc = 337 nm,
0.5 ns time resolution) or by using pulsed NanoLED excitation sources
at 278 nm, 331 nm, 465 nm, and 560 nm (pulse width ≤0.3 ns); the
detector was a red-sensitive (185−850 nm) Hamamatsu R-3237-01
photomultiplier tube. Analysis of the luminescence decay profiles versus
time was accomplished with the DAS6 Decay Analysis Software
provided by the manufacturer. To record the 77 K luminescence spectra,
the samples were put in glass tubes (2 mm diameter) and inserted in a
special quartz dewar, filled up with liquid nitrogen.
EXPERIMENTAL PROCEDURES
■
Materials and General Considerations. The solvents (puriss.
grade) and commercially available starting materials were used as
received. NMR spectra were measured with AV-400 spectrometers,
and the reported chemical shifts were referenced to tetramethylsilane,
Si(CH3)4. Mass spectra and elemental analysis have been performed
by the services of analysis at EPFL. 2-(2,4-difluorophenyl)pyridine,28
2,4-dibromoiodobenzene29 and 2-(allyldimethylsilyl)pyridine30 have
been prepared following the literature. We reported the synthesis of
2-(2,4-chlorophenyl)pyridine and HIrPic previously.25
Cyclic Voltammetric Measurements. A PC controlled AutoLab
PSTAT10 electrochemical workstation has been employed. Cyclic
voltammograms (CV) were obtained at a scan rate of 100 mV/s using
0.1 M TBAPF6 as supporting electrolyte in acetonitrile. Glassy carbon,
sputtered platinum, and platinum wire were employed as working,
counter, and reference electrodes, respectively. At the end of each
measurement, the ferrocinium/ferrocene (Fc+/Fc) potential was
measured and used as an internal reference.
Theoretical Calculations. Full geometry optimizations of the
iridium compounds in their singlet ground state were performed with
DFT using the M06 functional,31 with the relativistic effective core
potential and basis set LANL2DZ32 for the iridium and the 6-311G*33
basis set for the remaining atoms. The use of more flexible basis sets
including diffuse functions showed no significant changes in the main
geometrical features of the optimized compounds. No symmetry
constraints were applied during the geometry optimizations, which
were carried out with the Gaussian 09 package.34 The nature of the
stationary points located was further checked by computations of
X-ray Crystal Structure Determination. The data collections
for the three crystal structures were measured at low temperature using
800
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
Mo Kα radiation. An Oxford Diffraction Sapphire/KM4 CCD was
employed for BrIrPic while the remaining samples were measured on
a Bruker APEX II CCD. Both diffractometers have a kappa geometry
goniometer. Data reduction were carried out by Crysalis PRO44
(BrIrPic), and EvalCCD45 (FIrPic, ClIrPic) and then corrected for
absorption.46 The solutions and refinements were performed by
SHELX.47 The structures were refined using full-matrix least-squares
based on F2 with all non hydrogen atoms anisotropically defined.
Hydrogen atoms were placed in calculated positions by means of the
“riding” model. Disorder problems dealing with the solvent (CH2Cl2)
were found during the refinement of BrIrPic. In this case some
restraints were applied (SHELX cards: ISOR and DFIX) to get
reasonable parameters.
to room temperature, the solution was filtered on cellite, and
evaporated to dryness to give a yellow viscous oil which precipitated
upon addition of methanol (30 mL). The suspension was kept in the
fridge for 2 h, filtered off, and washed with cold methanol (30 mL) and
hexane (100 mL). The solid was adsorbed on silica, deposited on
the top of a silica gel chromatography column, and eluted with
dichloromethane. Finally, the main yellow fraction was dissolved in a
minimum amount of a methanol/dichloromethane mixture (10/90,
v:v) and slowly precipitated with hexane. The suspension was filtered
off, washed with hexane and dried to afford FIrPic as a bright yellow
solid (473 mg, yield 82%). 1H NMR (CDCl3, 400 MHz): δ 8.75 (1H,
d, J = 4.8 Hz); 8.34 (1H, d, J = 7.6 Hz); 8.30 (1H, d, J = 4.4 Hz); 8.24
(1H, d, J = 8.4 Hz); 7.95 (1H, dt, J = 7.6, 1.2 Hz); 7.78−7.74 (3H, m);
7.45−7.41 (2H, m); 7.19 (1H, t, J = 6.0 Hz); 6.97 (1H, t, J = 6.0 Hz);
6.50 (1H, t, J = 9.6 Hz); 6.40 (1H, t, J = 9.6 Hz); 5.82 (1H, dd, J = 7.6,
1.6 Hz); 5.57 (1H, dd, J = 8.2, 2.0 Hz). TOF MS ES: MH+ m/z: calc.
696.0881 found: 696.0895. Anal. Calcd. for C28H16F4IrN3O2: C, 48.41;
H, 2.32; N, 6.05. Found: C, 48.38; H, 2.48; N, 5.92.
ClIrPic. As FIrPic using 6 (130 mg, 0.10 mmol), picolinic acid
(36 mg, 0.30 mmol) and tetrabutyl ammonium hydroxide (388 mg,
0.50 mmol). ClIrPic was obtained as a yellow solid (127 mg, 87%). 1H
NMR (CDCl3, 400 MHz): δ 9.11 (1H, d, J = 6.8 Hz); 9.02 (1H, d, J =
6.8 Hz); 8.81 (1H, d, J = 5.6 Hz); 8.33 (1H, d, J = 7.6 Hz); 7.95 (1H,
dt, J = 7.6, 1.2 Hz); 7.85−7.79 (2H, m); 7.72 (1H, d, J = 4.8 Hz); 7.50
(1H, d, J = 5.2 Hz); 7.43 (1H, td, J = 5.2, 1.2 Hz); 7.26 (1H, td, J =
5.2, 1.2 Hz); 7.08 (1H, d, J = 2.0 Hz); 7.02 (1H, td, J = 6.0, 1.2 Hz);
7.02 (1H, d, J = 2.0 Hz); 6.16 (1H, d, J = 2.0 Hz); 5.84 (1H, d, J =
1.6 Hz). 13C NMR (CDCl3, 100 MHz): δ 175.99, 171.93, 165.62,
164.19, 152.35, 151.57, 151.06, 148.47, 147.78, 147.49, 138.41, 138.26,
137.87, 137.24, 134.55, 134.06, 131.07, 130.88, 130.12, 129.92, 128.07,
127.91, 124.69, 124.13, 124.02, 123.56, 122.96, 122.72. TOF MS ES:
MH+ m/z: calc. 759.9683 found: 759.9684 Anal. Calcd. for
C28H16Cl4IrN3O2·CH3OH: C, 43.95; H, 2.54; N, 5.30. Found: C,
44.18; H, 2.48; N, 4.98.
2-(2,4-Bromophenyl)pyridine 3. A mixture of 2-(allyldimethyl-
silyl)pyridine (1.77 g, 10.0 mmol), 2,4-dibromoiodobenzene (4.70 g,
13.0 mmol), Ag2O (3.47 g, 15.0 mmol), and Pd(PPh3)4 (635 mg,
0.55 mmol) in dry tetrahydrofuran (THF, 50.0 mL) was stirred at
60 °C for 10 h under Ar. After cooling the reaction mixture to room
temperature, the mixture was filtered on a short silica gel pad. The
crude mixture was chromatographed on silica gel (hexane/EtOAc =
50/50 as eluent) to afford 2-(2,4-bromophenyl)pyridine (2.08 g, 66%)
1
as white solid. H NMR (CDCl3, 400 MHz): δ 8.72 (1H, dt, J = 7.0,
2.0 Hz); 7.86 (1H, d, J = 5.5 Hz); 7.77 (1H, td, J = 7.6, 2.0 Hz); 7.60
(1H, dt, J = 7.2, 1.5 Hz); 7.55 (1H, dd, J = 7.0, 2.2 Hz); 7.43 (1H, d,
J = 7.2 Hz); 7.32 (1H, ddd, J = 7.0, 3.5, 1.2 Hz). TOF MS ES: MH+
m/z: calc. 313.9003 found: 313.8951.
[Ir(2-(2,4-difluorophenyl)pyridine)2Cl]2 5. A solution of
IrCl3·xH2O (1.07 g, 3.03 mmol) in a mixture 2-ethoxyethanol/water
(60 mL/20 mL) was degassed by bubbling argon for 15 min while
heating at 80 °C. 1 (1.28 g, 6.70 mmol) was added as a solid, and the
mixture heated at 135 °C for 20 h. After cooling down to room
temperature, water (150 mL) was added, and the mixture kept in the
fridge for 2 h. The precipitate was filtered on fritted (G4) glass,
washed with water (6 × 50 mL), MeOH (50 mL), and hexane (3 ×
50 mL) and dried under vacuum to afford 5 as a bright yellow solid
BrIrPic. As FIrPic using 7 (130 mg, 0.076 mmol), picolinic acid
(38 mg, 0.31 mmol) and tetrabutyl ammonium hydroxide (470 mg,
0.59 mmol). BrIrPic was obtained as a yellow solid (113 mg, 79%). 1H
NMR (CDCl3, 400 MHz): δ 9.30 (1H, d, J = 6.8 Hz); 9.22 (1H, d, J =
6.8 Hz); 8.80 (1H, d, J = 5.6 Hz); 8.33 (1H, d, J = 7.6 Hz); 7.95 (1H,
dt, J = 7.6, 1.2 Hz); 7.87−7.80 (2H, m); 7.71 (1H, d, J = 4.8 Hz); 7.50
(1H, d, J = 5.2 Hz); 7.49 (1H, d, J = 2.0 Hz); 7.43 (1H, td, J = 5.2, 1.2
Hz); 7.41 (1H, d, J = 2.0 Hz); 7.27 (1H, td, J = 5.2, 1.2 Hz); 7.06 (1H,
td, J = 6.0, 1.2 Hz); 6.33 (1H, d, J = 2.0 Hz); 5.99 (1H, d, J = 1.6 Hz).
13C NMR (CDCl3, 100 MHz): not measured because of low solubility.
TOF MS ES: MH+ m/z: calc. 939.7654 found: 939.7634 Anal. Calcd.
for C28H16Br4IrN3O2: C, 35.84; H, 1.72; N, 4.48. Found: C, 35.77; H,
2.08; N, 4.08.
1
(1.31 g, yield 71%). H NMR (CDCl3, 400 MHz): δ 9.12 (1H, d, J =
6.0 Hz); 8.31 (1H, d, J = 7.6 Hz); 7.83 (1H, t, J = 7.2 Hz); 6.83 (1H, t,
J = 6.4 Hz); 6.34 (1H, t, J = 10.4 Hz); 5.28 (1H, dd, J = 7.4, 1.2 Hz).
TOF MS ES: [Ir(2-(2,4-difluorophenyl)pyridine)2]+ m/z: calc.
573.0566 found: 573.0672; [Ir(2-(2,4-difluorophenyl)-
pyridine)2(MeCN)]+ m/z: calc. 614.0831 found: 614.0510; [Ir(2-
(2,4-difluorophenyl)pyridine)2(MeCN)2]+ m/z: calc. 655.1097 found:
655.1145; {[Ir(2-(2,4-difluorophenyl)pyridine)2](μ-Cl)[Ir(2-(2,4-
difluorophenyl)pyridine)2(MeCN)]}+ m/z: calc. 1222.1086 found:
1222.0740.
[Ir(2-(2,4-dichlorophenyl)pyridine)2Cl]2 6. As 5, using
IrCl3·xH2O (210 mg, 0.60 mmol), and 2 (290 mg, 1.30 mmol). 6
was obtained as a yellow solid (361 mg, yield 90%). 1H NMR (CDCl3,
400 MHz): δ 9.16 (1H, d, J = 5.6 Hz); 9.08 (1H, d, J = 8.0 Hz); 7.87
(1H, t, J = 7.6 Hz); 6.93 (1H, d, J = 2.0 Hz); 6.89 (1H, td, J = 2.8, 1.2
Hz); 5.55 (1H, d, J = 2.0 Hz). TOF MS ES: [Ir(2-(2,4-dichloro-
phenyl)pyridine)2]+ m/z: calc. 636.9383 found: 636.9857; [Ir(2-(2,4-
dichlorophenyl)pyridine)2(MeCN)]+ m/z: calc. 677.9649 found:
677.9795; [Ir(2-(2,4-dichlorophenyl)pyridine)2(MeCN)2]+ m/z: calc.
718.9915 found: 719.0273; {[Ir(2-(2,4-dichlorophenyl)pyridine)2]-
(μ-Cl)[Ir(2-(2,4-dichlorophenyl)pyridine)2(MeCN)]}+ m/z: calc.
1349.8722 found: 1349.8344.
[Ir(2-(2,4-dibromophenyl)pyridine)2Cl]2 7. As 5, using
IrCl3·xH2O (155 mg, 0.44 mmol), and 3 (303 mg, 0.97 mmol). 7
was obtained as a yellow solid (318 mg, yield 85%). 1H NMR (CDCl3,
400 MHz): δ 9.27 (1H, d, J = 8.4 Hz); 9.13 (1H, dd, J = 6.0, 1.2 Hz);
7.89 (1H, t, J = 7.6 Hz); 7.31 (1H, d, J = 2.0 Hz); 6.91 (1H, td, J = 2.8,
1.2 Hz); 5.70 (1H, d, J = 2.0 Hz). TOF MS ES: [Ir(2-(2,4-dibro-
mophenyl)pyridine)2(MeCN)]+ m/z: calc. 853.7629 found: 853.5371;
[Ir(2-(2,4-dibromophenyl)pyridine)2(MeCN)2]+ m/z: calc. 894.7894
found: 894.5612.
FIrPic. Picolinic acid (154 mg, 1.25 mmol, ∼3 equiv) and tetrabutyl
ammonium hydroxide (1.66 g, 2.1 mmol, ∼5 equiv) were added to a
solution of 5 (505 mg, 0.42 mmol) in dichloromethane (150 mL). The
yellow solution was refluxed under argon overnight. After cooling down
RESULTS AND DISCUSSION
■
Synthesis and X-ray Crystal Structures. The ligands 1
and 2 were prepared using Suzuki coupling. The ligand 3
requires a longer synthesis using an allylsilane pyridine inte-
rmediate for avoiding mixtures of multisubstituted compounds
(Scheme 1). Complexes were prepared in a usual two-step
synthesis according to Scheme 2. Reaction of IrCl3·xH2O with
2.2 equiv of ligand in a mixture 2-ethoxy-ethanol/water at reflux
overnight gives the corresponding chloro-bridged iridium(III)
dimer in good yield (70−90%) after precipitation, and washing
with water, MeOH, and hexane. Subsequent reaction in
dichloromethane with 3 equiv of picolinic acid in the presence
of 5 equiv of tetrabutyl ammonium hydroxide at reflux overnight
gives after workup and purification, the expected XIrPic (X = F,
Cl, Br) compounds in good yields (80−90%).
Single crystals of FIrPic, ClIrPic, and BrIrPic have been
grown by slow diffusion of hexane into a dichloromethane
solution of the complexes. Structures are shown in Figure 1 and
selected crystallographic data are provided in Table 1 and Table 2.
801
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
Scheme 1. Synthesis of 2,4-Dibromo-phenyl-pyridine 3 (i) NaNO2, H2SO4, KI; (ii) n-BuLi, allylchlorodimethylsilane; (iii)
Pd(PPh3)4, Ag2O
Scheme 2. Chemical Structures of Molecules; (i) 2-ethoxy-ethanol/H2O, reflux; (ii) CH2Cl2, TBAOH, picolinic acid, reflux
Figure 1. X-ray crystal structure of FIrPic, ClIrPic, and BrIrPic.
a
Table 1. Comparison between Experimental and Theoretical (DFT/M06) Ground State Geometries
c
FIrPic
ClIrPic
BrIrPic
X-ray
M06
X-ray
M06
X-ray
M06
Bond Distance
Ir−N(1)
Ir−N(2)
Ir−N(3)
Ir−C(7)
Ir−C(12)
Ir−O(1)
C(11)-X(1)
2.049(3)
2.056(3)
2.135(3)
2.002(3)
2.010(3)
2.158(2)
1.369(4)
2.05
2.06
2.19
2.00
2.01
2.18
1.34
2.038(6)
2.05
2.06
2.19
2.00
2.01
2.18
1.76
2.039(7)
2.041(6)
2.137(7)
1.989(9)
1.996(8)
2.139(6)
1.897(10)
2.05
2.06
2.19
2.00
2.01
2.18
1.91
2.038(6)
2.154(6)
1.999(7)
2.003(7)
2.162(5)
1.772(8)
b
Bond Angle
C(7)−Ir−C(12)
C(7)−Ir−N(1)
C(12)−Ir−N(1)
C(7)−Ir−N(3)
C(12)−Ir−O(1)
N(1)−Ir−N(2)
88.76(12)
80.91(11)
94.63(12)
96.57(12)
97.81(11)
175.17(10)
88.6
80.4
95.1
98.2
97.4
174.7
88.9(3)
80.1(3)
95.9(3)
97.7(3)
96.8(2)
174.5(2)
89.1
79.9
98.7
96.6
98.7
174.8
89.7(3)
79.3(3)
95.7(3)
98.8(3)
94.6(3)
173.7(3)
90.4
79.9
95.4
96.5
97.3
174.2
a
b
c
Selected bond distances (Å) and angles (deg). X = F, Cl, or Br. Asymmetric unit contains two independent molecules, the selected values belong
to molecule A.
The three complexes have the expected slightly deformed
octahedral coordination geometry around the iridium center
with the cis-C,C trans-N,N chelate configuration. Overall, there
are no large geometric differences among the complexes, the size
802
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
Table 2. Crystallographic Data for FIrPic, ClIrPic, and BrIrPic
FIrPic·MeOH
ClIrPic
C28H16Cl4IrN3O2
BrIrPic·1/2CH2Cl2
empirical formula
formula weight
temperature, K
wavelength (Å)
crystal system
space group
unit cell dimensions
a (Å)
C29H20F4IrN3O3
726.68
C28.5H17Br4ClIrN3O2
980.74
760.44
100(2)
100(2)
140(2)
0.71073
0.71073
orthorhombic
Pbca
0.71073
monoclinic
C2/c
monoclinic
I2/a
30.723(6)
15.0279(14)
30.6777(6)
b (Å)
10.570(2)
15.440(2)
11.2485(2)
c (Å)
15.488(3)
28.284(3)
35.0954(7)
α (deg)
90
90
90
β (deg)
92.40(3)
90
90.412(2)
γ (deg)
volume (Å3)
90
90
90
5025.3(17)
6562.6(13)
12110.3(4)
Z
8
8
16
density, calcd (g/cm3)
absorption coefficient (mm−1
F(000)
1.921
1.539
2.152
)
5.382
4.421
9.808
2816
2928
7344
crystal size (mm3)
0.87 × 0.41 × 0.22
3.31 to 25.02
47584
0.69 × 0.63 × 0.32
3.30−25.03
98253
0.35 × 0.31 × 0.22
2.66−25.03
44681
θ range for data collection (deg)
reflections collected
independent reflections
absorption correction
4445 [R(int) = 0.0452]
semiempirical from equivalents
full-matrix least-squares on F2
4445/0/361
1.158
5710 [R(int) = 0.0577]
semiempirical from equivalents
full-matrix least-squares on F2
5710/0/344
10690 [R(int) = 0.0446]
semiempirical from equivalents
full-matrix least-squares on F2
10690/39/739
refinement method
data/restraints/parameters
goodness-of-fit on F2
1.128
1.077
final R indices [I > 2σ(I)]
R1 = 0.0185
wR2 = 0.0377
R1 = 0.0259
wR2 = 0.0412
R1 = 0.0479
R1 = 0.0379
wR2 = 0.0896
R1 = 0.0635
wR2 = 0.0883
R indices (all data)
R1 = 0.0648
wR2 = 0.0958
wR2 = 0.1067
of the halogen atoms having apparently no significant impact.
BrIrPic shows a more compact octahedral geometry around the
central iridium atom than FIrPic as bonds involving the iridium
are generally shorter with the bromine substituents. ClIrPic
stands out with a more symmetrical environment around the
iridium center since the bonds Ir−N(1) and Ir−N(2) are
identical, which is not the case in the other complexes. In
addition, the ancillary ligand is less tightly bound as the distances
Ir−N(3) and Ir−O(1) are longer than in the other complexes.
The only notable consequence of the halogen substituents lies in
the carbon−halogen bond lengths and in the distortion of the
phenyl-pyridine ligand (see below). The crystal structure of FIrPic
has been recently reported.48 The octahedral coordination geo-
metry around the iridium center is slightly more compact than
our report, an effect that might be attributed to the absence of
cocrystallized solvent. We have reported the X-ray crystal
structure of HIrPic previously.25
Figure 2. Electronic absorption spectra of HIrPic (black), FIrPic
(blue), ClIrPic (green), and BrIrPic (red) in dichloromethane
solutions at room temperature. The lowest energy T1 transitions are
zoomed in at the top-right inset.
Photophysical Properties. The UV−visible electronic
absorption spectra of HIrPic, FIrPic, ClIrPic, and BrIrPic
were recorded in CH2Cl2 solution (Figure 2). Though the
spectral profiles of the four complexes are quite similar, the
onset of absorption of the fluorine-substituted complex shows a
significant blue shift compared to the others. The UV region is
dominated by intense absorption bands (ε ≈ 3.8 × 104 M−1
5000 M−1 cm−1) are of charge-transfer (CT) nature and are
related to electronic transitions occurring from the metal center
to the cyclometalated ligands (MLCT) (see theoretical
calculations below).50
As reported for similar complexes by Thompson et al., this
class of compounds exhibits a weak bathochromic shift of
MLCT absorption features upon decrease of the solvent polarity
(see Supporting Information, Figure S1 to S3), suggesting that
the CT state is less polar than the ground state.51 Moreover,
49
1
cm−1) assigned to ligand centered (LC) (π−π*) transitions.
The peaks at 256 (FIrPic), 262 (HIrPic), 264 (ClIrPic), and
268 (BrIrPic) nm are due to the absorption of cyclometalating
ligands F2ppy, H2ppy, Cl2ppy, and Br2ppy, respectively. The
weaker and broader bands at lower energies (350−440 nm, ε ≈
803
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
the virtually solvent-independent and extremely weak (ε ≈
450 M−1 cm−1) absorption bands detected at 455 (FIrPic),
474 (ClIrPic), 475 (BrIrPic), and 480 (HIrPic) nm can be
attributed to the direct population of the emitting triplet state
(T1). Our theoretical calculations reproduce correctly these low
intensity absorption tails by calculating the direct singlet−triplet
transitions taking into account the SOC effects (see below). The
small spectral shift between the T1 absorption and emission profiles
(<0.1 eV, for all the complexes) at room temperature suggests that
the emitting state has some *π−π character (see below).
(λmax = 505 nm) is the most red-shifted one. Our theoretical
calculations (see below) and previous studies indicate that the
picolinate triplet levels lie at rather high energy and are not
involved in the electronic transition causing luminescence.49 For
this reason, the picolinate moiety is essentially an ancillary ligand
3
and the luminescence properties are mainly of (MLCT−LC)
nature, involving iridium d orbitals and π−π* orbitals of the
halogen-substituted phenylpyridine (ppy) cyclometalated units.
On the other hand, the LUMO of HIrPic has been found to be
mainly located on the picolinate ligand.24 Therefore the use of
halogen substituents on the phenyl ring of the main ligand
induces a change of the localization of the LUMO of the
complexes by stabilizing the LUMO energy levels located on the
main cyclometalated ligand (Figure 4, 5, and 6).
The luminescence spectra of the complexes in CH2Cl2 are
depicted in Figure 3. As for the absorption, the emission profiles
The luminescence properties of FIrPic, ClIrPic, BrIrPic, and
HIrPic are unaffected by changes in the excitation wavelength,
and the spectral position of the emission band is not depending
on the solvent polarity (Supporting Information, Figure S1 to
S3). This last feature and the well-resolved vibronic progressions
displayed by all the emission profiles indicate that, in all the
cases, the emitting state (T1) has a predominant 3LC character.
In addition, at 77 K (Figure 3), all the samples show very
intense and highly resolved bands that exhibit small rigid-
ochromism; a hypsochromic shift of less than 5 nm is observed
for FIrPic and BrIrPic while in the case of ClIrPic no shift is
detected. The strong similarity between room temperature (RT)
and 77 K emission bands, as well as the comparable excited state
lifetimes, confirm the weak MLCT character of T1 at room
temperature.54,55 Finally, the triplet nature of the emitting states
Figure 3. Normalized emission spectra (λexc= 407 nm) of HIrPic,
FIrPic, ClIrPic, and BrIrPic in CH2Cl2, at room temperature (solid
line) and 77 K rigid matrix (dashed line).
of ClIrPic (λmax = 491 nm) and BrIrPic (λmax = 494 nm) are
red-shifted with respect to FIrPic (λmax = 468 nm) while HIrPic
Figure 4. Selected Kohn−Sham molecular orbitals for the characterization of LR-TDDFT excited states of FIrPic. Isovalue is set to 0.03 au.
804
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
Figure 5. Selected Kohn−Sham molecular orbitals for the characterization of LR-TDDFT excited states of ClIrPic. Isovalue is set to 0.03 au.
Figure 6. Selected Kohn−Sham molecular orbitals for the characterization of LR-TDDFT excited states of BrIrPic. Isovalue is set to 0.03 au.
Table 3. Photophysical Data of the Investigated Complexes at Room and Low Temperature
a
cd
,
c
298 K emission
deactivation rate constants
77 K emission
λmax [nm]
b
sample
λmax [nm]
Φem [%]
τ [ns]
kr [105 s−1
]
knr [105 s−1
]
τ [μs]
HIrPic
FIrPic
505
14.7 (2.2)
62.0 (3.9)
65.9 (5.1)
9.1 (4.0)
514
1722
2408
e
(74)
2.86
3.60
2.73
biexp.
19.17
2.20
496, 519
463, 495
491, 522
490, 524
2.98
2.24
3.10
2.92
468, 495
491, 517
494, 519
(103)
(185)
(145)
ClIrPic
BrIrPic
1.42
biexp.
a
b
Oxygen-free CH2Cl2 solutions, from spectra corrected for the detector response; in brackets, air-equilibrated solutions. Although we are aware that
the reference value used for luminescence quantum yield of [Ru(bpy)3]2+ in aerated water has been recently re-evaluated and found to be 0.04 (see
refs 52 and 53), we prefer to use the value reported for the air equilibrated [Ru(bpy)3]Cl2 in aqueous solution [Φem = 0.028] (see ref 43) as the
c
d
standard, which was also reproduced in house. λexc = 407 nm. Radiative constant: kr = Φem·τ−1; nonradiative constant: knr = τ−1 − kr (assuming
e
unitary intersystem crossing efficiency). t1 = 1100 (7%), t2 = 330 (93%).
(highly oxygen sensitive) leads to a strong decrease of PLQY
and lifetimes in air-equilibrated solution, as reported in Table 3.
To further support the LC character of the emission, we
measured the photophysical properties of the ligands 2-(2,4-di-
X-phenyl)pyridine (X = H, F, Cl, Br), H2ppy, F2ppy, Cl2ppy,
and Br2ppy (Supporting Information, Figure S4, S5 and S6).
As reported in the literature,56,57 2-phenylpyridine is not emit-
ting in common organic solvents and only the protonated
805
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
Table 4. Electrochemical Properties of Complexes HIrPic, FIrPic, ClIrPic, and BrIrPic
a
b
c
d
e
f
Eox /V
EHOMO /eV
Eopt‑LUMO /eV
σm
σp
computed −IE/eV
computed EA/eV
computed Eopt‑LUMO /eV
HIrPic
FIrPic
0.58
0.92
0.94
0.95
−5.38
−5.72
−5.74
−5.75
−2.92
−3.07
−3.22
−3.24
0
0
−5.46
−5.81
−5.85
−5.82
−2.01
−2.10
−2.34
−2.54
−2.68
−2.85
0.34
0.37
0.39
0.06
0.23
0.23
ClIrPic
BrIrPic
−2.35
−2.84
a
b
c
d
e
0.1 M TBAPF6 in acetonitrile, potential vs ferrocene. EHOMO = −(Eox + 4.8). Eopt‑LUMO = EHOMO + E0−0. Ionization energy DFT/M06. Electron
affinity DFT/M06. Computed Eopt‑LUMO = −IE + ΔELR‑TDDFT(S1−S0).
f
2-phenylpyridinium ion shows significant emission, both in
fluorescence and phosphorescence. This is due to a shift in the
shows that the HOMOs are stabilized as expected because of
the replacement of H with the stronger acceptor fluoro, chloro,
and bromo substituents (Table 4). While the reduction of
FIrPic is quasi-reversible (−2.29 vs Fc+/0), it is not the case for
ClIrPic and BrIrPic, where the reductions are irreversible. This
makes the direct comparison of the LUMOs energies based on
electrochemical measurements poorly reliable. Therefore we
calculated the optical LUMOs energy, that is, the energy of the
LUMOs based on the optical gap as obtained experimentally
from the emission spectra of the complexes. In our study
focusing on the photophysical properties of the complexes the
optical LUMO energy is anyway more relevant than the redox
LUMO. Generally the optical LUMO is lower than the redox
LUMO as the optical gap is smaller than the redox gap mainly
because of the exciton binding energy. Several other factors
should be taken into account to compare precisely the optical
and redox gaps,58 hence to compare the optical LUMO energy
with the redox LUMO energy. Nevertheless, as a first approxi-
mation supported by calculations, it can be assumed that the
difference between the optical and redox gaps is constant for
closely related molecules.58 Using this simplification, the trend
observed in the calculated optical LUMO energies is assumed to
be reasonably valid for the redox LUMO energies. Computed
vertical ionization energies (energy of the oxidized molecule
minus the energy of the neutral molecule) and vertical electron
affinities (energy of the reduced molecule minus the energy
of the neutral molecule) can be taken as an estimator of the
experimentally determined −EHOMO and ELUMO, respectively.
Both quantities follow closely the experimental trend along the
halogen series, with almost constant −IE values for FIrPic,
ClIrPic, and BIrPic. Even though the absolute EA values are
shifted with respect to Eopt‑LUMO, the trend between the different
compounds is well reproduced with a first energy stabilization
step between HIrPic and FIrPic, and a second one between
FIrPic and ClIrPic/BrIrPic. A way to approximate Eopt‑LUMO
consists in calculating the difference between the computed
transition energy of the first singlet state and the ionization
energy (see Table 4). This procedure is in fact more closely
related to how the experimental Eopt‑LUMO is obtained, as we add
to the computed Eox (= −IE) the optical gap obtained from LR-
TDDFT. These theoretical Eopt‑LUMO provide a further good
agreement with the observed experimental trend. As the trends
for experimental energies of redox LUMO and optical LUMO
are expected to be similar and as the trends for computed EA
and EoptLUMO are similar, in the discussion about trends hereafter
we will use LUMO without distinction.
1
1
nature of the transition from n−π* to π−π* upon protona-
tion. Therefore the measurements have been performed in
0.1 N H2SO4. In fluorescence, all the protonated ligands display
an unstructured emission band around 370 nm. While H2ppyH+
and F2ppyH+ are characterized by high fluorescent quantum
yields (50.0% and 36.0% respectively), it drops considerably
for Cl2ppyH+ (5.2%) and Br2ppyH+ (0.3%). Although a
quantitative evaluation of the phosphorescence quantum yield in
frozen matrixes is not straightforward, qualitatively the intensity
of the phosphorescence emission significantly increases along
the series H < F < Cl < Br concomitantly with the drop in
fluorescence. This points to an enhanced intersystem crossing
because of the increasing heavy atom effect along the halogen
series. With heavier halogen substituents, the deactivation of
the first singlet-excited state through the triplet is faster. The
phosphorescence emission of F2ppyH+ (λmax = 445 nm) is
14 nm blue-shifted compared to H2ppyH+ (λmax = 459 nm),
while the triplet-excited states of Cl2ppyH+ and Br2ppyH+ are
very close in energy (λmax = 453 and 455 nm, respectively). The
energy trend follows the one observed for the emission of the
XIrPic series. In addition the spectral profiles of the ligand
triplet emission reflect the structured emission of the cor-
responding complexes. This evidence confirms that the T1
emitting state of the XIrPic complexes arises from a triplet state
3
with a strong LC character, involving mainly the X2ppy ligands.
ClIrPic and FIrPic show remarkable photophysical proper-
ties compared to HIrPic, that is, high emission intensities
(PLQY > 0.6) and long lifetimes (τ > 1.5 μs). On the other
hand, the PL performance of BrIrPic is poorer (PLQY < 0.1).
Assuming unitary intersystem crossing quantum yield, it is
possible to calculate radiative (kr) and nonradiative (knr) decay
rates from phosphorescence quantum yields and lifetimes.
This leads to a radiative lifetime τrad (respectively kr) of 3.49 μs
(2.86 × 105 s−1) for HIrPic, 2.77 μs (3.6 × 105 s−1) for FIrPic,
3.65 μs (2.73 × 105 s−1) for ClIrPic, and 3.62 μs (2.76 ×
105 s−1) for BrIrPic (using the major component of the
biexponential). The radiative constants are similar for all com-
plexes and cannot explain the poor performance of BrIrPic.
Therefore nonradiative processes are significantly more active
in BrIrPic than in the other complexes. In addition of the heavy
atom effect, they appear to be favored by the significant
geometrical distortion of the excited state induced by the large
bromine atom (see below).
Electrochemistry. The oxidation potentials of the proto-
type complex HIrPic and of the three halogenated compounds
have been measured by cyclic voltammetry in acetonitrile solu-
tion containing 0.1 M tetrabutyl ammonium hexafluorophos-
phate (TBAPF6). The complexes show reversible processes
at 0.58, 0.92, 0.94, and 0.95 V vs ferrocene for the non-
halogenated, the fluoro, the chloro, and the bromo complexes,
respectively, because of the oxidation of the iridium center. This
On the basis of these experimental HOMOs and LUMOs
energy levels, we discuss next the observed red shift of the
emission maximum for ClIrPic and BrIrPic compared to
FIrPic. Besides varying the skeleton of the cyclometalated
ligand, strategies for emission color tuning by modifying the
HOMO−LUMO gap with substituents are based on simple
rules. Indeed the theory (see theoretical part below) tells that
806
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
a
Table 5. (Top) DFT/M06 Optimized Geometries of FIrPic, ClIrPic, and BrIrPic, and (Bottom) Most Distorted ppy Ligand
Observed in the X-ray Structures in Comparison with the Geometries for the Ground State (S0) and the First Excited Triplet
State (T1)
FIrPic
ClIrPic
BrIrPic
HIrPic
dihedral angle
X-ray
S0 M06 T1 M05-2X X-ray S0 M06 T1 M05-2X
X-ray
S0 M06 T1 M05-2X X-ray S0 M06 T1 M05-2X
δ1(X1−C2−C3−C4)
δ2(C2−C3−C4−C5)
−0.1
4.0
1.2
3.2
0.2
1.7
2.4
9.1
4.3
21.9
12.1
6.2
9.4
7.2
25.7
13.5
1.6
3.8
0.4
4.8
−1.5
0.6
11.0
14.5
a
Distortions on a specific X2ppy ligand are highlighted by the use of a Ball and Stick representation.
HOMO−LUMO is important in this context for absorption
and emission. First, a substituent with donor (acceptor)
character will destabilize (stabilize) the energy of the molecular
orbital (HOMO and LUMO). Second, the HOMO being
mostly localized on the anionic phenyl ring and the LUMO
being mostly localized on the neutral part of the cyclometalated
ligand (see theoretical part below),59 that is the pyridine in the
present case, a substituent on the anionic part of the ligand will
influence the HOMO more than the LUMO; and a substituent
on the neutral part of the ligand will influence the LUMO more
than the HOMO. In this respect, fluorine substituents on the
phenyl ring of a cyclometalated ligand are used to induce a large
blue shift of the emission maximum by stabilizing the HOMO
much more than the LUMO, overall leading to an increased
HOMO−LUMO gap. It should be noted that, experimentally,
the position of the substituent is generally critical for the tuning
of the HOMO and LUMO energy levels.
As observed from electrochemical data, the slightly stronger
acceptor effect of chlorine and bromine compared to fluorine
leads to a further stabilization of the HOMO, which translates
into a higher oxidation potential. Therefore, the observed red
shifts for ClIrPic (23 nm) and for BrIrPic (26 nm) with respect
to FIrPic (λem = 468 nm) indicate that the LUMO levels of the
chloride and the bromine derivatives are significantly lower than
the LUMO of the fluoro compound. This conclusion is further
supported by the theoretical calculations (see below). This effect
can be simply rationalized by using the Hammett parameters σm
and σp (≈ σo), denoting the electronic character of the
substituent toward the meta and para (≈ ortho) position
respectively.7 The Hammett parameters have been limited to
correlation with oxidation potentials of cyclometalated iridium
complexes.12,60,61 Along the halogens series studied here, σm
correlates well with the energy of the HOMO as obtained from
electrochemical measurements, and therefore cannot explain the
observed red-shift of the emission. On the other hand, the σp
values show that the chlorine and the bromine have a much
stronger acceptor character than fluorine toward the para
(≈ ortho) position. The pyridine, where the LUMO of the
complex is mostly localized (see theoretical part below), is in
this particular position relative to the halogen substituents. The
impact of the chlorine and bromine substituents on the energy
of the LUMO will be much stronger than in the case of the
fluorine which has a σp value about 1/4 of that of the chlorine
and the bromine. This stronger stabilization of the LUMO by
chlorine and bromine compared to fluorine, while having similar
HOMO energy level, qualitatively explains the observed red
shift of emission maxima of ClIrPic and BrIrPic compared
to FIrPic. Furthermore correlations of Hammett parameters
with computed −IE and computed EoptLUMO (Supporting
Information, Figure S7) allow a first step toward a quantitative
structure−property relationship based solely on Hammett
parameters for predicting the emission maxima of cyclo-
metalated iridium complexes with good accuracy (Supporting
Information, Figure S8).
DFT Calculations. To gain insight into the electronic
structures and the optical properties of the halogen-based
complexes we performed DFT/TDDFT calculations on both
the ground and the excited states. Structures of HIrPic, FIrPic,
ClIrPic, and BrIrPic have been optimized in their electronic
(singlet) ground state using the exchange and correlation (xc)
functional M06 (see computational details). The geometrical
predictions are in good agreement with the experimental X-ray
structures (Table 1) as the root-mean-square deviation (rmsd)
between the theoretical and the X-ray coordinates for iridium
and the 6 metal-coordinated atoms indicates rather small
deviations: 0.04 Å for FIrPic and ClIrPic, and 0.05 Å for
BrIrPic. The theoretical geometries present distortions of the
ppy ligands, which are due to the steric hindrance induced by
the presence of different halogen substituents. The distortion
along the halogen series follows the trend observed in the X-ray
structures (the most distorted ppy ligands are reported in
Table 5) and can significantly impact the optical properties of
the compounds, as observed in a previous work (see below for a
more extended discussion).62
The analysis of excitation energies of the different iridium
complexes requires a characterization of the molecular orbitals
involved in the different transitions. For all three complexes, the
HOMO Kohn−Sham (KS) molecular orbital has an important
5d contribution of 56% in FIrPic and of 41% in ClIrPic and
BrIrPic (Table 6). The remaining contributions come from π
807
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
Table 6. Main Vertical Excitations of FIrPic, ClIrPic,
BrIrPic, Computed at the Singlet Electronic Ground State
a
Minimum Energy Structure (LR-TDDFT/M06)
singlet state E (nm) oscillator strength dominant character
FIrPic
1
S1
396.5
338.7
294.0
280.8
264.3
234.1
0.0491
0.0597
0.1177
0.1045
0.1745
0.1317
H→L (85.8%)
2
S5
H-1→L+1 (67.2%)
H-1→L+3 (32.0%)
H-1→L+4 (64.2%)
H-2→L+5 (43.8%)
H-8→L+2 (23.6%)
3
S15
S19
S30
S49
4
5
6
ClIrPic
1
S1
413.6
348.5
298.3
280.7
235.3
0.0595
0.1005
0.1302
0.1993
0.0486
H→L (93.6%)
2
S5
H-1→L+1 (77.0%)
H-1→L+3 (57.2%)
H-1→L+4 (32.4%)
H-8→L+2 (25.6%)
3
S15
S22
S53
Figure 7. Absorption spectrum of FIrPic. Experimental spectrum
(black lines) is superimposed on LR-TDDFT/M06 convoluted
absorption spectrum (dashed blue line). Singlet (triplet) vertical
excitations energies are represented by blue (red) bars. Inset: zoom on
the low energy tail of the absorption spectrum, on which ZORA/LR-
TDDDT first vertical transitions are superimposed.
4
5
BrIrPic
1
2
3
4
5
6
S1
416.2
353.2
309.8
304.8
269.7
234.3
0.0668
0.1128
0.1783
0.1774
0.3646
0.0661
H→L (93.8%)
S5
H-1→L+1 (69.4%)
H-2→L+2 (59.6%)
H→L+5 (40.0%)
H→L+8 (72.6%)
H-8→L+2 (15.6%)
S12
S14
S29
S62
a
The excited singlet state is labeled “Sn”, where “n” is the electronic
state number according to LR-TDDFT/M06. Only the dominant
orbital character is reported.
orbitals located on ppy ligands. In a similar way, the HOMO-1
and HOMO-2 KS orbitals of all studied complexes are
combinations of 5d(Ir) and π(ppy) orbitals. All LUMOs can
be described as π*-ligand centered orbitals. It is interesting to
note that the LUMO of FIrPic is more delocalized over the two
ppy moieties than in ClIrPic and BrIrPic, and also presents a
contribution from the pic ligand (Figure 4, 5, and 6). LUMO+1
is symmetrically equivalent to the LUMO with respect to the
ppy ligand in both ClIrPic and BrIrPic. Within the
energetically lower lying unoccupied KS orbitals we also find
contributions from MOs at the pic ligand. This is the case in all
studied complexes and is particularly strong in FIrPic. Finally,
contributions from the σ*(C-X) ppy-based orbitals of ClIrPic
and BrIrPic appear already at LUMO+8 (see for example
Figure 6).
Figure 8. Absorption spectrum of ClIrPic. Experimental spectrum
(black lines) is superimposed on LR-TDDFT/M06 convoluted
absorption spectrum (dashed blue line). Singlet (triplet) vertical
excitations energies are represented by blue (red) bars. Inset: zoom on
the low energy tail of the absorption spectrum, on which ZORA/LR-
TDDDT first vertical transitions are superimposed.
LR-TDDFT63,64 has been used to study the absorption
spectra of FIrPic, ClIrPic, and BrIrPic. The first vertical singlet
and triplet excitations have been computed to reproduce the
full experimental spectra and validate the methodology. The
main contributions of the different theoretical peaks with
significant oscillator strength are listed in Table 6. The LR-
TDDFT excitations are analyzed in terms of electronic transi-
tions between occupied and unoccupied KS orbitals, and only
the dominant character of each excited state is listed in the
same table. However, in the most general case, several KS
molecular orbital transitions characterize a simple LR-TDDFT
excitation. In Figures 7, 8, and 9, LR-TDDFT excitation energies
are compared with the experimental spectra. To facilitate the
comparison, a broadening of the transitions with Gaussians of
width 0.37 eV is applied to all spectra.
Figure 9. Absorption spectrum of BrIrPic. Experimental spectrum
(black lines) is superimposed on LR-TDDFT/M06 convoluted
absorption spectrum (dashed blue line). Singlet (triplet) vertical
excitations energies are represented by blue (red) bars. Inset: zoom on
the low energy tail of the absorption spectrum, on which ZORA/LR-
TDDDT first vertical transitions are superimposed.
The agreement between theory and experiment is generally
good. Even if the calculated oscillator strengths do not always
match perfectly the experimental absorption, the positions of
the different bands and their shoulders are rather well described
808
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
by the M06 functional. The first experimental bands (350 to
440 nm) were assigned to MLCT transitions, which is in good
agreement with the theoretical predictions (Table 6). The three
compounds present indeed a common first transition (toward
S1) of HOMO→LUMO type, where the HOMO is mainly a
5d(Ir) orbital and the LUMO is mainly a π* orbital localized on
one ppy ligand (FIrPic: 396.5 nm, ClIrPic: 413.6 nm, BrIrPic:
416.2 nm). The second computed band is attributed to a
transition to the fifth singlet excited state (S5), which involves
another 5d(Ir)-based orbital (HOMO-1) and a π* orbital
mainly localized on the second ppy ligand (LUMO+1). We
may assign this transition to the experimentally observed
shoulder at 340 nm (FIrPic), 352 nm (ClIrPic), and 354 nm
(BrIrPic). Note that the measured red shift along the halogen
series is also reproduced by the calculations. Other transitions
with MLCT character are obtained at higher energies in all three
compounds. These states have often an important contribu-
tion from the pic ligand, and in the case of ClIrPic and BrIrPic
they are responsible for the shoulder at around 300 nm. As
observed experimentally, 1(π−π*) transitions become more
important in the UV region of the spectra. Of interest is, for
example, the π−π* contribution of the pic ligand to the band
around 235 nm. The first transitions to singlet states of metal-
centered (MC) character are observed at rather high energies:
above 275 nm for FIrPic, 281 nm for ClIrPic, and 292 nm for
BrIrPic.
band at 2.66 eV (exp: 2.65 eV), ClIrPic at 2.44 eV (exp: 2.52
eV), and BrIrPic at 2.40 eV (exp: 2.51 eV). The experimentally
observed red shift along the series of compounds is also
captured by the calculations while the average discrepancy in
the peak positions with respect to the experimental values lies
within the accuracy of the method. The same singlet-to-triplet
energy gap can also be computed with a LR-TDDFT calculation
with spin-flip. In this case, we obtain excitation energies of
2.36 eV for FIrPic, 2.18 eV for ClIrPic, and 2.15 eV for
BrIrPic. Even though the LR-TDDFT values are clearly down-
shifted, the energy differences between the FIrPic transition and
the ones of the two other compounds are in good agree-
ment with the Δ-SCF calculation. For this transition, LR-
TDDFT shows for all three compounds a dominant (>70%)
HOMO→LUMO character with a tiny (15−20%) HOMO→
LUMO+1 contribution. In agreement with the Δ-SCF DFT
calculations, LR-TDDFT/M05-2X predicts therefore a tran-
sition to an orbital mainly localized on a single ppy ligand. This
result is also confirmed by the analysis of the spin density
obtained with the Δ-SCF approach and, equivalently, by the
transition density difference computed with LR-TDDFT
(Figure 10). These findings are in good agreement with the
Concerning singlet-to-triplet transitions, the first transitions
occur at 442 and 435 nm (FIrPic), 463 and 454 nm (ClIrPic),
468 and 460 nm (BrIrPic). These are mainly composed by
HOMO→LUMO and HOMO→LUMO+1 transitions, respec-
tively, and can thus be considered to be of MLCT type. Previous
studies65,66 have proposed that because of the strong SOC of
iridium complexes these singlet-to-triplet transitions could
contribute to the low energy tail of the absorption spectra. To
validate this proposition in the case of our three compounds, we
have performed a series of additional calculations including the
effect of SOC using the ZORA approach (see computational
details). The results are presented in the inset of Figures 7, 8,
and 9, which shows the first six transitions with their cor-
responding oscillator strengths. Because of the size of SOC,
these transitions become slightly allowed and contribute indeed
to the experimentally observed absorption tail at low energy,
even though an overall blue shift of these weak absorptions is
observed in the calculation. As in the case of the corresponding
singlet-to-triplet transitions, the character of these spin−orbit
allowed excitations is mainly of MLCT type.
Figure 10. Spin density (upper part) resulting from the unrestricted
DFT calculations of the first triplet state. Density differences (lower
part) between the first LR-TDDFT triplet state electronic density and
the ground state electronic density (positive part in red, negative one
in blue). Isovalue is set to 0.002 (0.001) for the spin density (density
difference).
To gain some information about the emission processes in
FIrPic, ClIrPic, and BrIrPic, we performed, starting from the
ground state singlet geometry, unrestricted DFT (U-DFT)
geometry optimizations of the first triplet state (T1) using the
xc-functional M05-2X. This functional has indeed provided
singlet-to-triplet DFT energies in very good agreement with
experimental emission spectra for a series of iridium com-
plexes.40 At the optimized T1 geometries, we computed both
lowest-energy triplet and singlet state energies using an implicit
solvent model. The difference of these energies gives an
estimate of the position of the first phosphorescence peak. It is
important to note here that these Δ-SCF calculations do not
take the overlap between the excited and ground state
vibrational states into account, which would be needed for
the calculation of the Franck−Condon factor. Nonetheless the
theoretical values obtained are in rather good agreement with
the experimental emission spectra. FIrPic shows a theoretical
experimental observations, where the presence of vibronic
structures on the emission spectra suggests an important LC
character associated with this transition.
As discussed above, the ground state geometries of FIrPic,
ClIrPic, and BrIrPic exhibit already some distortion compared
to the optimized nonhalogenated HIrPic ground state structure.
This distortion is very small in FIrPic and becomes more
important as the size of halogen atom increases (Table 5).
At the optimized T1 geometry, there is a strong increase of
the distortion angle δ1 in the case of ClIrPic and BrIrPic. In
addition, the deformation is now specifically located on the ppy
ligand carrying the populated π* molecular orbital characterizing
the T1 state (see Figure 10). The analysis of the potential energy
surface (PES) near the optimized T1 geometries shows that the
triplet excited state PES is much shallower than the ground state
809
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
(4) Katritzky, A. R.; Kuanar, M.; Slavov, S.; Hall, C. D.; Karelson, M.;
Khan, I.; Dobchev, D. A. Chem. Rev. 2010, 110, 5714.
(5) Verma, R. P.; Hansch, C. Chem. Rev. 2011, 111, 2865.
(6) Makedonas, C.; Mitsopoulou, C. A. Eur. J. Inorg. Chem. 2007,
4176.
(7) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165.
(8) Hammett, L. P. Chem. Rev. 1935, 17, 125.
(9) Alexiou, C.; Lever, A. B. P. Coord. Chem. Rev. 2001, 216−217, 45.
(10) Chen, L. Q.; Yang, C. L.; Li, S. Y.; Qin, J. H. Spectrochim. Acta,
Part A 2007, 68, 317.
(11) McClure, D. S. J. Chem. Phys. 1949, 17, 905.
(12) Flamigni, L.; Barbieri, A.; Sabatini, C.; Ventura, B.; Barigelletti,
F. Top. Curr. Chem. 2007, 281, 143.
(13) Baranoff, E.; Yum, J. H.; Graetzel, M.; Nazeeruddin, M. K.
J. Organomet. Chem. 2009, 694, 2661.
PES, and therefore a distortion of the ppy ligands along the δi
angles can have an important contribution for the tuning of
the emission bands (in addition to the chemical substitution
discussed previously). Furthermore, this increased ligand
flexibility coupled with a larger heavy-atom effect of bromine
could possibly open novel nonradiative deactivation channels for
BrIrPic, as observed experimentally.
CONCLUSION
■
In summary, a homologous series of bis-cyclometalated
iridium(III) complexes Ir(2,4-di-X-phenyl-pyridine)2-
(picolinate) (X = H, F, Cl, Br) HIrPic, FIrPic, ClIrPic, and
BrIrPic has been synthesized and fully characterized. In the
present case, a simple variation of the substituent along the
halogen group leads to dramatic changes of photophysical
properties. Theoretical calculations accurately reproduce the
absorption trend for the different halogenated species, provide
assignments of the main bands, and confirm the role of SOC in
the low energy tail of the absorption spectra. The combined
experimental and theoretical results show that Hammett
parameters should be separated in σm and σp and their in-
fluence on both HOMO and LUMO energy levels should be
taken into account for a more accurate and quasi-quantitative
assessment of the influence of a substituent on the HOMO−
LUMO gap. Furthermore, theoretical results highlight sizable
distortions of the populated ppy ligand of the lowest triplet state,
which, in combination with the expected heavy atom effect, are
likely to play a role in the nonradiative relaxation processes of
BrIrPic. Overall our results show that the substituents impact
the optical properties of the XIrPic complexes with both
electronic and geometric effects. To be able to understand in
detail the impact of the halogen atoms, it is now necessary to
separate those two effects, which is the focus of future work.
(14) Chi, Y.; Chou, P. T. Chem. Soc. Rev. 2010, 39, 638.
(15) You, Y.; Park, S. Y. Dalton Trans. 2009, 1267.
(16) Ulbricht, C.; Beyer, B.; Friebe, C.; Winter, A.; Schubert, U. S.
Adv. Mater. 2009, 21, 4418.
(17) Lowry, M. S.; Hudson, W. R.; Pascal, R. A.; Bernhard, S. J. Am.
Chem. Soc. 2004, 126, 14129.
(18) De Angelis, F.; Fantacci, S.; Evans, N.; Klein, C.; Zakeeruddin, S.
M.; Moser, J.-E.; Kalyanasundaram, K.; Bolink, H. J.; Gratzel, M.;
̈
Nazeeruddin, M. K. Inorg. Chem. 2007, 46, 5989.
(19) Bronstein, H. A.; Finlayson, C. E.; Kirov, K. R.; Friend, R. H.;
Williams, C. K. Organometallics 2008, 27, 2980.
(20) Coughlin, F. J.; Westrol, M. S.; Oyler, K. D.; Byrne, N.; Kraml,
C.; Zysman-Colman, E.; Lowry, M. S.; Bernhard, S. Inorg. Chem. 2008,
47, 2039.
(21) Gupta, D.; Katiyar, M.; Deepak.; Hazra, T.; Verma, A.;
Marloharan, S. S.; Biswas, A. Opt. Mater. 2006, 28, 1355.
(22) Xu, M. L.; Zhou, R.; Wang, G. Y.; Xiao, Q.; Du, W. S.; Che,
G. B. Inorg. Chim. Acta 2008, 361, 2407.
(23) Zhang, M.; Li, Z. S.; Li, Y.; Liu, J.; Sun, J. Z. Int. J. Quantum
Chem. 2009, 109, 1167.
(24) Minaev, B.; Minaeva, V.; Agren, H. J. Phys. Chem. A 2009, 113,
726.
(25) Baranoff, E.; Jung, I.; Scopelliti, R.; Solari, E.; Gratzel, M.;
Nazeeruddin, M. K. Dalton Trans. 2011, 40, 6860.
(26) Holmes, R. J.; Forrest, S. R.; Tung, Y. J.; Kwong, R. C.; Brown,
J. J.; Garon, S.; Thompson, M. E. Appl. Phys. Lett. 2003, 82, 2422.
(27) Sudhakar, M.; Djurovich, P. I.; Hogen-Esch, T. E.; Thompson,
M. E. J. Am. Chem. Soc. 2003, 125, 7796.
̈
ASSOCIATED CONTENT
■
S
* Supporting Information
Absorption and emission spectra in various solvents, additional
figures and computational details. This material is available free
of charge via the Internet at http://pubs.acs.org.
(28) Thesen, M. W.; Krueger, H.; Janietz, S.; Wedel, A.; Graf, M.
J. Polym. Sci., Part A: Polym. Chem. 2010, 48, 389.
(29) Jensen, T.; Pedersen, H.; Bang-Andersen, B.; Madsen, R.;
Jørgensen, M. Angew. Chem., Int. Ed. 2008, 47, 888.
(30) Kamei, T.; Fujita, K.; Itami, K.; Yoshida, J.-I. Org. Lett. 2005, 7,
4725.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: etienne.baranoff@epfl.ch (E.B.), gianluca.accorsi@isof.
cnr.it (G.A.).
(31) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215.
(32) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(33) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639.
(34) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J., J. A.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
ACKNOWLEDGMENTS
■
We acknowledge financial support for this work by Solvay S.A.,
the European Union (CELLO, STRP 248043), and the Swiss
National Science Fundation (NCCR-MUST interdisciplinary
research program), and thank Dr. Davide Di Censo for his kind
assistance with electrochemical measurements.
REFERENCES
■
(1) Armaroli, N.; Balzani, V. Energy for a Sustainable World: From the
Oil Age to a Sun-Powered Future; Wiley-VCH: Weinheim, Germany,
2010; p 368.
(2) Balzani, V.; Campagna, S. Photochemistry and Photophysics of
Coordination Compounds I; Springer-Verlag: Heidelberg, Germany,
2007; Vol. 280.
(3) Balzani, V.; Campagna, S. Photochemistry and Photophysics of
Coordination Compounds II; Springer-Verlag: Heidelberg, Germany,
2007; Vol. 281.
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision A.02; Gaussian, Inc.: Wallingford, CT, 2009.
810
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811

Inorganic Chemistry
Article
(35) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999.
(36) Scalmani, G.; Frisch, M. J.; Mennucci, B.; Tomasi, J.; Cammi, R.;
Barone, V. J. Chem. Phys. 2006, 124, 094107.
(37) te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.; Fonseca
Guerra, C.; Baerends, E. J.; Snijders, J. G.; Ziegler, T. J. Comput. Chem.
2001, 22, 931.
(38) Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J.
Theor. Chem. Acc. 1998, 99, 391.
(39) Wang, F.; Ziegler, T.; van Lenthe, E.; van Gisbergen, S. J. A.;
Baerends, E. J. J. Chem. Phys. 2005, 122, 204103.
(40) Swiderek, K.; Paneth, P. J. Phys. Org. Chem. 2009, 22, 845.
(41) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. J. Chem. Theory Comput.
2006, 2, 364.
(42) Demas, J. N.; Crosby, G. A. J. Phys. Chem. 1971, 75, 991.
(43) Nakamaru, K. Bull. Chem. Soc. Jpn. 1982, 55, 2697.
(44) CrysAlis PRO, 171.34.49; Agilent Technologies: Santa Clara,
CA, 2010.
(45) Duisenberg, A. J. M.; Kroon-Batenburg, L. M. J.; Schreurs, A. M. M.
J. Appl. Crystallogr. 2003, 36, 220.
(46) Blessing, R. H. Acta Crystallogr., Sect. A 1995, 51, 33.
(47) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112.
(48) Xu, M.-L.; Che, G.-B.; Li, X.-Y.; Xiao, Q. Acta Crystallogr. 2009,
E65, m28.
(49) Lamansky, S.; Djurovich, P.; Murphy, D.; Abdel-Razzaq, F.;
Kwong, R.; Tsyba, I.; Bortz, M.; Mui, B.; Bau, R.; Thompson, M. E.
Inorg. Chem. 2001, 40, 1704.
(50) Schmid, B.; Garces, F. O.; Watts, R. J. Inorg. Chem. 1994, 33, 9.
(51) Li, J.; Djurovich, P. I.; Alleyne, B. D.; Yousufuddin, M.; Ho,
N. N.; Thomas, J. C.; Peters, J. C.; Bau, R.; Thompson, M. E. Inorg.
Chem. 2005, 44, 1713.
(52) Suzuki, K.; Kobayashi, A.; Kaneko, S.; Takehira, K.; Yoshihara,
T.; Ishida, H.; Shiina, Y.; Oishic, S.; Tobita, S. Phys. Chem. Chem. Phys.
2009, 11, 9850.
(53) Ishida, H.; Tobita, S.; Hasegawa, Y.; Katoh, R.; Nozaki, K.
Coord. Chem. Rev. 2010, 254, 2449.
(54) Tsuboyama, A.; Iwawaki, H.; Furugori, M.; Mukaide, T.;
Kamatani, J.; Igawa, S.; Moriyama, T.; Miura, S.; Takiguchi, T.; Okada,
S.; Hoshino, M.; Ueno, K. J. Am. Chem. Soc. 2003, 125, 12971.
(55) Orselli, E.; Kottas, G. S.; Konradsson, A. E.; Coppo, P.; Frohlich,
̈
R.; De Cola, L.; van Dijken, A.; Buchel, M.; Borner, H. Inorg. Chem.
2007, 46, 11082.
̈
(56) Kubin, J.; Testa, A. C. J. Photochem. Photobiol., A 1994, 83, 91.
(57) Sarkar, A.; Chakravorti, S. J. Lumin. 1995, 65, 163.
(58) Vlcek, A. A.; Dodsworth, E. S.; Pietro, W. J.; Lever, A. B. P.
Inorg. Chem. 1995, 34, 1906.
(59) Hay, P. J. J. Phys. Chem. A 2002, 106, 1634.
(60) Lin, C. F.; Huang, W. S.; Chou, H. H.; Lin, J. T. J. Organomet.
Chem. 2009, 694, 2757.
(61) Laskar, I. R.; Chen, T.-M. Chem. Mater. 2004, 16, 111.
(62) Baranoff, E.; Fantacci, S.; De Angelis, F.; Zhang, X.; Scopelliti,
R.; Gratzel, M.; Nazeeruddin, M. K. Inorg. Chem. 2011, 50, 451.
̈
(63) Casida, M. E. In Recent Advances in Density-Functional Methods;
World Scientific: Singapore, 1995, 155.
(64) Petersilka, M.; Gossmann, U. J.; Gross, E. K. U. Phys. Rev. Lett.
1996, 76, 1212.
(65) De Angelis, F.; Belpassi, L.; Fantacci, S. J. Mol. Struct. Theochem.
2009, 914, 74.
(66) Di Censo, D.; Fantacci, S.; De Angelis, F.; Klein, C.; Evans, N.;
Kalyanasundaram, K.; Bolink, H. J.; Gratzel, M.; Nazeeruddin, M. K.
Inorg. Chem. 2008, 47, 980.
811
dx.doi.org/10.1021/ic2011474 | Inorg. Chem. 2012, 51, 799−811