Structure-photoluminescence quenching relationships of iridium(III)- tris(phenylpyridine) complexes

Article abstract of DOI:10.1002/ejic.201101315

The synthesis, structural, photophysical, theoretical, and electrochemical characterization of four tris(2-phenylpyridine)-based Ir^III complexes are reported. The complexes were functionalized on the pyridine or on the phenyl rings with amide moieties substituted with a tris(ethyl)amine or ethyl groups, thereby yielding a family of compounds with hemicaged or open (without a capping unit but with similar functional groups on the ligand) structure. Within the context of the parent tris(2-phenylpyridine) and the full-cage iridium(III) complexes, structure-photoluminescence quenching relationships (SPQR) of the four complexes have been investigated. Luminescence quenching by oxygen has been studied with Stern-Volmer plots and through evaluation of the thermodynamic parameters involved in the quenching process. Density functional theory (DFT) and time-dependent DFT (TD-DFT) calculations have been performed on the complexes to gain insights into structural and electronicfeatures and the nature of the excited states involved in the electronic absorption processes. Interestingly, shielding by the capping unit of moieties in which the LUMO orbital is mostly localized (on the pyridines) results in a dramatic 40 % decrease in oxygen quenching. Conversely, shielding ofmoieties in which the HOMO orbital is partially localized (on the phenyl rings) does not induce any change in the oxygen quenching degree. In both sets of compounds, the thermodynamic feasibility of oxygen quenching is the same for the hemicaged and open compounds, thus giving evidence of the structural origin of such quenching decrease. The SPQRopens up new routes to the design of tailored, more or less sensitive to oxygen, luminescent iridium complexes (e.g., for use as biolabels). A family of four tris(2-phenylpyridine)-based Ir^III complexes with hemicaged or open (without capping unit but with similar functional groups on the ligand) structure are reported. Within the context of the parent tris(2-phenylpyridine) and the full-cage iridium(III) complexes, structure-photoluminescence quenching relationships (SPQR) of the four complexes have been investigated. Copyright

Full text of DOI:10.1002/ejic.201101315

FULL PAPER
DOI: 10.1002/ejic.201101315
Structure–Photoluminescence Quenching Relationships of
Iridium(III)–Tris(phenylpyridine) Complexes
Albert Ruggi,[a] Matteo Mauro,[b] Federico Polo,[b] David N. Reinhoudt,[a]
Luisa De Cola,*[a,b] and Aldrik H. Velders*[a,c]
Keywords: Iridium / Luminescence / Photoluminescence / Electrochemistry / Oxygen sensitivity / Density functional
calculations
The synthesis, structural, photophysical, theoretical, and
electrochemical characterization of four tris(2-phenylpyr-
idine)-based IrIII complexes are reported. The complexes
were functionalized on the pyridine or on the phenyl rings
with amide moieties substituted with a tris(ethyl)amine or
ethyl groups, thereby yielding a family of compounds with
hemicaged or open (without a capping unit but with similar
functional groups on the ligand) structure. Within the context
of the parent tris(2-phenylpyridine) and the full-cage iridi-
um(III) complexes, structure–photoluminescence quenching
relationships (SPQR) of the four complexes have been inves-
tigated. Luminescence quenching by oxygen has been
studied with Stern–Volmer plots and through evaluation of
the thermodynamic parameters involved in the quenching
complexes to gain insights into structural and electronic
features and the nature of the excited states involved in the
electronic absorption processes. Interestingly, shielding by
the capping unit of moieties in which the LUMO orbital is
mostly localized (on the pyridines) results in a dramatic 40%
decrease in oxygen quenching. Conversely, shielding of
moieties in which the HOMO orbital is partially localized (on
the phenyl rings) does not induce any change in the oxygen
quenching degree. In both sets of compounds, the thermo-
dynamic feasibility of oxygen quenching is the same for the
hemicaged and open compounds, thus giving evidence of the
structural origin of such quenching decrease. The SPQR
opens up new routes to the design of tailored, more or less
sensitive to oxygen, luminescent iridium complexes (e.g., for
process. Density functional theory (DFT) and time-depend- use as biolabels).
ent DFT (TD-DFT) calculations have been performed on the
and, more recently, also for several biomedical applica-
tions.[5–8] Moreover, the interaction of the excited states of
Introduction
Here we report the evidence for a structure-related transition-metal complexes with dioxygen (which results in
shielding effect of the oxygen quenching of the lumines- the quenching of luminescence and the generation of highly
cence of iridium(III)–tris(phenylpyridine) derivatives with reactive species like singlet oxygen and superoxide radical
hemicaged (i.e., with a capping unit) or open (i.e., without anion) has received increasing interest due to the possible
a capping unit) ligand structures. Iridium(III)-based lumi- realization of diagnostic and therapeutic agents.[9–14] De-
nophores have been the object of extensive studies in the spite the studies directed towards understanding the theo-
last decades, especially due to their wide spectrum of pos- retical relationship between structure and optical proper-
sible applications.[1] The possibility of tuning the emission ties,[15,16] little is known about the structure–photolumines-
wavelength and the generally good emission properties cence quenching relationship (SPQR) of these complexes.
make iridium(III) complexes ideal luminophores for the re- Transition-metal complexes with a cagelike ligand structure
alization of organic light-emitting diodes (OLEDs),[2–4] show a remarkable decrease in oxygen quenching. For in-
stance, an RuII caged complex shows an 80% decrease in
oxygen quenching with respect to [Ru(bpy)3]2+ (bpy = 2,2Ј-
bipyridine),[17] and an IrIII caged complex (ii, Scheme 1)
[a] Laboratory of Supramolecular Chemistry and Technology,
MESA+ Institute for Nanotechnology, University of Twente,
P. O. Box 217, Enschede, The Netherlands
that shows a similar decrease in oxygen quenching com-
pared to the archetypical [Ir(ppy)3] (ppy = 2-phenylpyr-
idine) (i, Scheme 1); this has been recently described by our
group.[18] However, the origin of this quenching decrease is
not yet clear. Moreover, many parameters (e.g., triplet en-
ergy, oxidation potential) are different from the caged to
the respective reference complexes {i.e., [Ru(bpy)3]2+ and
[Ir(ppy)3]}, thus an univocal interpretation of the observed
quenching decrease is not trivial.[13]
E-mail: decola@uni-muenster.de
a.h.velders@utwente.nl
[b] Westfälische Wilhelms-Universität Münster, Physikalisches
Institut and Center for Nanotechnology (CeNTech),
Heisenbergstrasse 11, 48149 Münster, Germany
[c] BioMedical Chemistry, MIRA Institute of Biomedical
Technology and Technical Medicine, University of Twente
P. O. Box 217, Enschede, The Netherlands
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201101315.
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1025

L. De Cola, A. H. Velders et al.
FULL PAPER
2) or on the phenyl (3, 4) rings, respectively. These com-
plexes are synthetically more accessible than the related
fully caged system (ii),[18] and they show interesting struc-
tural shielding behavior. The introduction of a capping unit
on the pyridine side (1) of the [Ir(ppy)3] core, on which the
LUMO orbital is mostly localized, induces a different
shielding towards oxygen quenching than the open (un-
shielded) complex 2. We in fact find that for 1 the quench-
ing from dioxygen is lower than for complex 2. On the con-
trary, the presence of a capping unit on the phenyl side (3),
on which the HOMO is mostly localized, does not influence
the degree of oxygen quenching and, therefore, the behavior
of the hemicaged (3) and open (4) complexes towards oxy-
gen quenching is the same. It is interesting to note that the
molecules of each pair are in principle electronically and
energetically equivalent in terms of excited state, and there-
Scheme 1. Structure of the archetypical [Ir(ppy)3] (i) and Ir caged
complex (ii).
The remarkably low oxygen quenching of the lumines- fore the differences in their oxygen quenching can derive
cence observed for caged RuII[17] and IrIII[18] complexes sug- only from a structural effect. The oxygen quenching effi-
gests a possible structural contribution to the shielding ciency has been evaluated by Stern–Volmer analysis and
mechanism of the excited state (i.e., of the atoms on which compared with the thermodynamic parameters involved in
the LUMO is mostly localized). A selective modification the possible quenching pathways (i.e., energy transfer and
of the HOMO or of the LUMO properties in homoleptic electron transfer) to obtain clear evidence of the structure-
[Ru(bpy)3]2+-like complexes is challenging, since the related shielding against oxygen quenching. From this sys-
HOMO is mostly localized on the metal center and the tematic study it was found that the introduction of a cap-
LUMO is localized on the (symmetric) bipyridine ligands. ping unit on the pyridine side of IrIII–tris(phenylpyridine)
On the other hand, for iridium complexes, and in particular derivatives induces a remarkable decrease in oxygen
for the [Ir(ppy)3] derivatives, the 2-phenylpyridines are not quenching with respect to the open (uncapped) structure,
symmetric and therefore are ideal candidates for a system- whereas only a minor decrease in oxygen quenching is ob-
atic study of the SPQR. In homoleptic [Ir(ppy)3]-like com- served upon introduction of the same capping unit on the
plexes, the HOMO is usually localized on the metal ion and phenyl side of IrIII–tris(phenylpyridine) derivatives with re-
on the phenyl rings, whereas the LUMO (which can be con- spect to the corresponding open structure.
sidered a first approximation of the electronic distribution
of the molecule in the excited state) is mostly localized on
the pyridine rings.[1,15] The synthesis of caged complexes re-
mains a great challenge and the versatility in tuning the
ligands (and therefore properties) for these systems is com-
plicated. Hemicaged systems are, on the other hand, much
more accessible, and they might show similar shielding be-
havior but with a more accessible synthesis. In fact, upon
introduction of suitable groups, either on the pyridine or
phenyl side of the [Ir(ppy)] core, it is possible to selectively
modify the electronic and/or shielding properties of the
LUMO or HOMO, respectively, with a consequent change
of the redox and optical properties like the emission ener-
gies.[1,16] In this paper we describe a series of IrIII–tris(phen-
ylpyridine) derivatives with a hemicaged or open structure
in which we have selectively modified the HOMO and the
LUMO orbitals.[11,19] Such a strategy allows us to under-
stand the SPQR and to rationalize the behavior towards
dioxygen quenching upon the shielding of the HOMO or
of the LUMO.
Scheme 2. Structures of the hemicaged (1, 3) and open (2, 4) IrIII
complexes.
More precisely, we present here the synthesis, characteri-
zation, and photophysical properties of a series of phenyl-
pyridine-based IrIII complexes functionalized with a tris(2-
amidoethyl)amine capping unit or with an ethylamide moi-
Results and Discussion
ety (i.e., with a hemicaged and open structure) (Scheme 2).
The complexes possess a hemicaged (1, 3) or an open (2, 4)
Iridium(III) complexes 1–4 were synthesized according
structure in which the substituents are on the pyridine (1, to the synthetic pathway shown in Scheme 3. The phenyl-
1026
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 1025–1037

Iridium(III)–Tris(phenylpyridine) Complexes
pyridine-based building blocks were obtained upon Suzuki
All the intermediates and the target complexes were
coupling[20] of bromonicotinic acid with phenylboronic acid characterized by IR and NMR (1H and 13C) spectroscopy
(to give the phenylpyridine derivative 5) or upon Suzuki and mass spectrometry, and the characteristic data are re-
coupling of 4-carboxymethylboronic acid with 2-bromopyr- ported in the Experimental Section and in the Supporting
idine (to give the phenylpyridine derivative 8) with subse- Information. Complexes 1–4 show only one set of NMR
quent amide formation through amide coupling with, spectroscopic signals for the three phenylpyridine ligands,
respectively, tris(2-aminoethyl)amine (Tren) (6, 9) or ethyl- which proves that the threefold symmetry expected for fac
1
amine (7, 10). The so-obtained ligands were treated with complexes is maintained. The H NMR spectra of the he-
IrIII to obtain the desired target complexes. The hemicaged micaged complexes 1 and 3 (see the Supporting Infor-
complexes were synthesized by direct reaction of the suit- mation) show, in addition, that upon formation of the iridi-
able tripodal ligand (6, 9) with IrCl3 in the presence of um(III) complexes, the hemicage ligands become quite ri-
CF3CO2Ag in ethylene glycol heated at reflux to give the gid: due to the hampered rotation around the C–C bond of
hemicaged complexes 1 and 3.[18] This strategy was not suc- the ethyl linkers, the four protons on each linker unit be-
cessful in the case of the open complexes; they were synthe- come magnetically nonequivalent, thereby resulting in four
sized in two steps.[21] First the dichlorido-bridged dimer was different multiplet signals in the 1H NMR spectrum. A sim-
obtained by reaction of the desired ligand (7, 10) with IrCl3 ilar behavior has been observed for the hemicaged and
in 2-ethoxyethanol, and the so-obtained intermediate was caged IrIII complexes previously described by our group.[18]
1
treated with an excess amount of ligand in the presence of Conversely, the H NMR spectra of the open complexes 2
CF3SO3Ag in toluene heated at reflux (see Scheme 3) to and 4 do not show any nonequivalent CH2 or CH3 protons,
give the open complexes 2 and 4.
as expected for ethyl moieties with rotational freedom. The
Scheme 3. Synthetic pathways for the synthesis of the hemicaged (1, 3) and open (2, 4) IrIII complexes. The phenylpyridine derivatives 5
and 8 (obtained by Suzuki coupling) are coupled with Tren or ethylamine, thus giving the tripodal (6, 9) or simple ligands (7, 10), which
are consecutively bound to IrIII, thus giving the target molecules 1–4.
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1027

L. De Cola, A. H. Velders et al.
FULL PAPER
Figure 1. Aromatic region of the 1H NMR spectra and peak assignment of hemicaged (1, 3) and open (2, 4) iridium(III) complexes;
[D7]DMF (* = residual solvent).
introduction of the capping unit induces a remarkable similar coupling of the amide proton with three aliphatic
change in the signal of the amide proton (NH). Compared protons on the ethyl linker was also observed for the he-
to the open complex 2, the NH proton of the hemicaged micaged complex 3 (Figure 2), thus giving proof of the he-
complex 1 shows an upfield shift of 1 ppm (Figure 1). A micaged structure of the two complexes 1 and 3. In the
similar upfield shift is also shown for the amide proton aromatic region of the ROESY spectrum of the hemicaged
when comparing 4 to 3. The observed change in the chemi- complexes (Figure 2), the cross-peak between the protons
cal shift of the amide proton (NH) is probably due to the Hc and Hd, which are located on the pyridine and phenyl
change of the local polarity induced by the presence of the ring, respectively, allows for a straightforward assignment
capping unit (vide infra). Interestingly, the proton Ha shows of the remaining peaks from COSY data and the heteronu-
an upfield shift of 0.40 ppm from 2 to 1, whereas only a clear HMBC and HMQC experiments.
modest (δ = 0.13 ppm) downfield shift is observed for Ha
Careful analysis of the aromatic region of the ROESY
from 4 to 3. The upfield shift of the Ha proton observed spectrum of the hemicaged complex 1 (Figure 2) reveals the
from 2 to 1 is possibly attributable to a magnetic shielding remarkable orientation of the amide moiety: the cross-
exerted by the aromatic rings located on the other branches peaks between the amide proton (NH) with both the in-
of the molecule as a result of a higher degree of twisting of ternal (Ha) and the external (Hb) protons, ortho to the
the helical structure of the [Ir(ppy)3] core in the hemicage amide group, indicate that the NH group is not coplanar
1: because of this higher twisting, the proton Ha in the he- with the pyridine ring. To investigate the possibility of a
micaged complex 1 faces the pyridine ring, thereby resulting certain degree of fluxionality in the capping unit of the he-
in a downfield shift. On the other hand, the modest down- micage system 1 to cause the characteristic two cross-peaks
field shift observed for the proton Ha between 3 and 4 indi- of the amide protons, variable-temperature (VT) NMR
cates that the structure of the [Ir(ppy)3] core in the he- spectroscopy was performed. Neither the 1D VT measure-
micaged complex 3 is only slightly distorted with respect to ments nor the ROESY performed at 250 K (see the Sup-
the open complex 4.
porting Information) indicate the presence of more con-
To further investigate the geometry of the hemicaged formers of hemicage 1 to be present in solution on the
complexes (1, 3), a series of 2D NMR spectroscopic experi- NMR spectroscopic timescale.
ments (H,H-COSY, H,H-ROESY, H,C-HMBC and H,C-
A different behavior was observed for hemicage 3 (Fig-
HMQC) was performed. The aromatic–aliphatic region of ure 2), the amide proton of which gives a coupling only
the ROESY spectrum of the hemicaged complex 1 (Fig- with the internal proton (Ha) on the phenyl ring; therefore,
ure 3) shows the coupling between the amide proton (NH) the NH group in 3 is likely coplanar with the phenyl ring
and three nonequivalent aliphatic protons on the ethyl and oriented towards the inner side of the molecule.
linker (Hh, HhЈ, Hi), analogously to what was found for
Two possible structures can be drawn for the hemicage 1
the hemicaged complex previously described by our on the basis of the observed NOE cross-peaks. Either the
group.[18] The fourth proton (HiЈ) is too far from the amide entire amide moiety is not coplanar with the pyridine ring,
proton and does not give any through-space coupling. A or only the NH moiety of the amide group is bent out of
1028
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 1025–1037

Iridium(III)–Tris(phenylpyridine) Complexes
Figure 2. Downfield region of the H,H-ROESY spectra of the hemicaged complexes 1 (left) and 3 (right), recorded in [D7]DMF. The
residual solvent peaks are indicated with an asterisk.
the pyridine plane, with the carbonyl still coplanar with the the corresponding open derivatives (2 and 4), the computed
pyridine. The first structure, with a lower conjugation of bond displacements being Յ0.005 Å. As already reported
the pyridine ring and the carbonyl moiety, should influence by Nozaki et al.[23] and by Hay et al.,[15] BeckeЈs three-pa-
the LUMO level less than the second structure in which the rameter together with the Lee–Yang–Parr exchange corre-
planarization of the system allows a stronger perturbation, lation hybrid functional (namely, B3LYP) tends to slightly
and in particular a stabilization of the LUMO. Therefore, overestimate the Ir–N bond lengths. Also, a similar agree-
in the case of a lower conjugation of the pyridine ring and ment between theoretical and experimental geometrical val-
the carbonyl moiety, the emission of the hemicaged com- ues can be envisaged for the complexes 2–4. This finding
plex 1 is expected to be blueshifted compared to the open nicely highlights the suitability of the employed theoretical
complex 2 (analogously to what has been observed in the model for describing the geometrical parameters of the
case of the RuII caged complex).[17] Since complex 1 shows complexes investigated here. Also, the B3LYP functional
a small but clear bathochromic shift with respect to com- has been already widely proven to properly describe the
plex 2 (vide infra), it can be concluded that the carbonyl electronic and optical properties of phosphorescent cyclo-
moiety of the amide group is conjugated with the pyridine metalated iridium complexes.[25,26] In Figure 3, the sche-
ring. Therefore, most likely, the NH group is bent out of matic representation of the energy levels of the orbitals
the pyridine plane.
Molecular structures of 1–4 were optimized at their elec-
tronic ground state (S0) by means of DFT at the B3LYP/
(6-31G(d,p) + SDD) level. The most meaningful computed
geometrical parameters are reported in the Supporting In-
formation (Table S7). The calculated S0 structures are in
good agreement when compared with experimental values
of the closely related archetypical complex fac-[Ir(ppy)3][22]
within the known limitation of the density functional calcu-
lations used.[23] All the computed geometries investigated
here possess a distorted octahedral arrangement around the
heavy metal center, with C3 point-group symmetry, and ex-
cellently reproduce the Ir–C bond length, with 2.036 Å be-
ing the value for 1 and 2.032–2.060 Å the theoretical and
experimental values, respectively, whereas the Ir–N bond
lengths are 2.192 (computed) and 2.071–2.095 Å (experi-
mental).[24] It is also worth noting that the presence of the
capping unit in the hemicaged complexes (1 and 3), seems
Figure 3. Partial molecular orbital diagram for complexes 1–4 and
fac-[Ir(ppy)3]. The arrows are intended to highlight the HOMO–
not to affect the metal–ligand bond lengths with respect to LUMO energy gaps. All the DFT energy values are given in eV.
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1029

L. De Cola, A. H. Velders et al.
FULL PAPER
closer to the frontier region for the complexes 1–4 is re-
The three HOMOs mainly correspond to the 5d orbitals
ported, together with fac-[Ir(ppy)3], which is used as refer- of the d6 IrIII center in a (distorted) octahedral environ-
ence. To provide deeper insight into the electronic structure ment, and are denoted d1a, d1b, and d2, for which d1a and
of the complexes 1–4, the isodensity surface plots of the
d1b are the twofold degenerated orbitals with “e” symmetry,
most relevant Kohn–Sham molecular orbitals are shown in and d2 is the highest energy orbital with “a” symmetry.
Figure 4 (for a more extended view, see Figures S5 and S6 Nonetheless, a sizeable admixing between the three filled d
in the Supporting Information). A list of the energies orbitals of the iridium and the π-orbitals of the ligand can
HOMO – 4 to LUMO + 4, along with the corresponding be clearly envisaged, as is typical of cyclometalated iridi-
HOMO–LUMO energy gap, is reported in Table 1 (an ex- um(III) complexes. In particular, the HOMO (namely, d2)
tended version is available as Table S8 in the Supporting is an antibonding combination of Ir(t2g) and π orbitals of
Information).
two of the cyclometalating phenyl rings of A symmetry. The
corresponding energy was calculated to be –5.268, –5.261,
–5.352, and –5.384 eV for complex 1, 2, 3, and 4, respec-
tively. It is worth noting that in both couples of complexes
the presence of the capping unit causes a negligible differ-
ence in the HOMO energy level. On the other hand, due to
the nature of the HOMO, the position of the amido substit-
uent (phenyl versus pyridyl ring) has a larger influence on
the stabilization of the orbital. Complex 4 is stabilized by
as much as 0.123 eV with respect to complex 2. At slightly
more negative energies, two degenerate d1a and d1b, namely,
HOMO – 1 and HOMO – 2, are encountered. Their nature
can be mainly described as the two antibonding combina-
tions of the Ir(t2g) and the π orbitals of two of the cyclomet-
alating phenyl rings of E symmetry. These orbitals are stabi-
lized with respect to the corresponding HOMO level by
0.046, 0.128, 0.029, and 0.01 eV, and lie at –5.314, –5.389,
–5.382, and –5.394 eV, for 1, 2, 3, and 4, respectively. Going
from the occupied to the virtual molecular orbitals, the
three lowest-lying unoccupied molecular orbitals can be de-
noted, in order of increasing energy, as π*1, π*2a, and π*2b.
The LUMO, namely, π*1, can be described as the combina-
tion π* orbitals of the ligands of “a” symmetry, mainly lo-
calized on the pyridyl rings, and extends over the carbonyl
group of the amido substituents in complexes 1 and 2. Such
an orbital is calculated to lie at –1.844, –1.842, –1.728, and
–1.763 eV for 1, 2, 3, and 4, respectively. As far as com-
plexes 1 and 2 are concerned, the presence of the electron-
withdrawing amido groups, which are strongly involved in
the LUMO, induces a stabilization of the corresponding en-
ergy level with respect to the fac-[Ir(ppy)3] as high as 0.623
and 0.621 eV, for 1 and 2, respectively. On the other hand,
such a stabilization effect is lower for complexes 3 and 4,
with values of 0.507 and 0.542 eV, respectively. Finally, a
set of twofold degenerate orbitals of “e” symmetry lies at
–1.758 (1), –1.719 (2), –1.620 (3), and –1.630 eV (4), which
can be described as the combination of π* of the pyridyl
moieties, and extends over the carbonyl group of the amido
substituents in complexes 1 and 2.
Figure 4. Isodensity surface plots of some selected frontier molecu-
lar orbitals for complexes 1–4 at their optimized S0 geometry in
the gas phase. Isodensity value: 0.035 eBohr–3. A larger version is
given in Figures S4–S6 of the Supporting Information).
Table 1. List of selected molecular orbital energies [eV] for com-
plexes 1–4 and fac-[Ir(ppy)3] at their S0-optimized geometry in
vacuo, and HOMO–LUMO energy gap.
Orbital
fac-[Ir(ppy)3]
1
2
3
4
LUMO + 4 (e)
LUMO + 3 (a)
LUMO + 2 (e)
LUMO + 1 (e)
LUMO (a)
–0.653
–0.894
–1.131
–1.131
–1.221
–4.881
–5.018
–5.018
–5.862
–5.862
3.661
–1.192 –1.169 –1.054
–1.501 –1.337 –1.277
–1.758 –1.719 –1.620
–1.758 –1.719 –1.620
–1.844 –1.842 –1.728
–5.268 –5.261 –5.352
–5.314 –5.389 –5.382
–5.314 –5.389 –5.382
–1.082
–1.306
–1.630
–1.630
–1.763
–5.384
–5.394
–5.394
The UV/Vis absorption spectra of both pairs of hemicage
and open complexes are reported in Figure 5, and the typi-
cal absorption maxima and shoulder are reported in Table 2
together with their molar extinction coefficients. Also, to
gain deeper insight into the nature of the electronic transi-
tion involved in the absorption spectra, complexes 1–4 were
investigated by means of time-dependent density functional
theory (TD-DFT) calculations, and the computed vertical
transitions described in terms of molecular orbitals of the
HOMO (a)
HOMO – 1 (e)
HOMO – 2 (e)
HOMO – 3 (e)
HOMO – 4 (e)
HOMO–LUMO gap
–6.165 –6.215 –6.207 –6.213[a]
–6.165 –6.215 –6.207
3.424 3.419 3.624
–6.221
3.620
[a] “a” symmetry.
1030
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 1025–1037

Iridium(III)–Tris(phenylpyridine) Complexes
corresponding ground-state geometry. The most relevant
lowest-lying singlet-to-singlet (S0 Ǟ Sn, n = 1–30) and the
three lowest-lying singlet-to-triplet (S0 Ǟ Tn, n = 1–3) cal-
culated transitions are respectively listed in the Supporting
Information together with the nature of the involved orbit-
als and their expansion coefficients (see Tables S7 and S8).
In the experimental spectra, all the complexes show a
strong absorption band (εՆ104 m– 1 cm–1) between 280 and
320 nm and weaker absorption bands (εՅ103 m– 1 cm–1) be-
tween 350 and 450 nm. By comparison with the typical ab-
sorption shown by fac-[Ir(ppy)3] derivatives,[27,28] the ab-
sorption bands centered around 300 nm can be assigned to
ligand-centered π Ǟ π* transitions, whereas the weaker
bands centered around 400 nm can be assigned to the con-
volution of the lowest-lying spin-allowed singlet-to-singlet
metal-to-ligand charge transfers (1MLCTs). It is worth not-
ing that the hemicaged complex 1 showed a slight bathoch-
romic shift of its MLCT absorption bands compared to the
open complex 2, whereas no similar shift was observed for
derivatives 3 and 4. These findings nicely agree with the
TD-DFT calculations. In particular, the lowest-energy tran-
sitions are the S0 Ǟ S1 and mainly involve the iridium d
orbitals (partially mixed with the π orbitals of the cyclomet-
alating phenyl rings) and π* orbitals of the pyridyl moieties
Figure 5. UV/Vis absorption of 1 (black), 2 (red) (top), and 3
(black), and 4 (red) (bottom) in DMF at room temp., and corre-
sponding lower-lying computed excitation energies and oscillator
strengths (vertical lines).
1
(HOMO Ǟ LUMO, MLCT1), and are computed to lie at
450, 453, 427, and 429 nm for complex 1, 2, 3, and 4,
respectively. At slightly higher energies, two doubly gener-
ated MLCT transitions with moderate oscillator strengths
Table 2. UV/Vis absorption data of 1–4 and their molar extinction
coefficient in DMF at room temp.
1
(f = 0.018–0.047) were computed, which can be ascribed to
linear combinations of HOMO – 1/HOMO – 2 Ǟ LUMO
(1MLCT2, with E symmetry) and HOMO – 1/HOMO – 2
Ǟ LUMO + 1/LUMO + 2 (1MLCT3, with A symmetry).
Such transitions are calculated to occur at 435 and 423 nm
(1); 425 and 410 nm (2); 414 and 399 nm (3); and 418 and
398 nm (4), respectively. At higher energies, such transitions
are followed by A-symmetry excitation with intensities
similar to the above-described processes and involve
Absorption[a]
λ [nm], (ε [103 m–1 cm–1])
1
292 (45), 302 (sh. 15), 361 (sh. 5.3),420 (6.5), 487 (sh. 2.1)
297 (32), 400 (6.9), 487 (sh. 2.1)
287 (52), 387 (9.4), 474 (sh. 2.1)
2
3
4*
288 (45), 387 (6.3), 474 (sh. 2.6)
[a] sh. = shoulder; * = estimated value because of the poor solubil-
ity of 4 in DMF.
HOMO Ǟ LUMO + 3 (computed at 399 and 379 nm, the emission of the complexes 3 and 4 compared to
for 1 and 2, respectively) and linear combination of [Ir(ppy)3].[1] This different effect between 1 and 2 on the one
HOMO – 1/HOMO – 2 Ǟ LUMO + 1/LUMO + 2 (com- hand and 3 and 4 on the other is due to the low atomic
puted at 389 and 388 nm, for 3 and 4, respectively). More- contribution given to the HOMO orbital by the atom on
over, the weak shoulder at lower energies (centered around which the amide moiety is located.[29] The hemicaged com-
470 nm) can be assigned to spin-forbidden singlet-to-triplet plex 1 showed an evident redshift of the emission compared
3MLCT. The latter transitions are usually observed in com- to the open complex 2. Conversely, the hemicaged complex
plexes that contain heavy atoms (like iridium), which show 3 showed a slight ipsochromic shift of the emission com-
a remarkable spin–orbit coupling that makes the singlet-to- pared to the open complex 4. As already mentioned, the
triplet transitions partially allowed. Such transitions are in rigidity and lack of flexibility induced by the capping unit
very good agreement with the theoretical (S0 Ǟ T1) exci- blocks the rotation of the amido group, which is almost
tation energies, which are computed to be 502, 498, 485, coplanar with the pyridine ring (see Table S7 in the Sup-
and 486 nm, for 1, 2, 3, and 4, respectively.
porting Information).[17] Furthermore, the locally different
All the reported complexes 1–4 showed intense lumines- polarity induced by the presence of the N atom of the cap-
cence at room temperature (Figure 6). As expected, the ping unit can influence the degree of shift (e.g., a kind of
presence of electron-withdrawing groups on the pyridine local solvatochromism).[30] The situation is different when
ring induced a bathochromic shift in the emission of com- the phenyl is substituted with the amido group, since a de-
plexes 1 and 2 compared to the archetypical fac-[Ir(ppy)3] stabilization of the HOMO orbital is expected. The slight
as a consequence of the stabilization of the LUMO orbital. blueshift observed for the hemicage 3 with respect to 4 can
Conversely, the same groups on the phenyl ring minimally again be explained by the structural rigidity induced by the
stabilize the HOMO orbital to result in a slight redshift of capping and by the locally different polarity induced by the
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1031

L. De Cola, A. H. Velders et al.
FULL PAPER
presence of the N atom of the capping unit.[30] The increase quenching was obtained by a Stern–Volmer analysis of the
in energy for the HOMO–LUMO gap in the phenyl-substi- luminescence quenching (vide infra).
tuted complexes 3 and 4 implies that the lowest excited
3
state, 3MLCT, is more mixed with the LC energetically ac-
Table 3. Photophysical properties of hemicaged and open com-
plexes in DMF at 25 °C (unless stated).[a]
cessible, as can be seen by the structured emission and
λem
[nm]
φ0
τ0
[ns]
τ
[ns]
λem
τ
[μs]
longer excited-state lifetimes. On the other hand, the struc-
tureless emission bands shown by complexes 1 and 2 indi-
cate that the excited state of those complexes has mainly an
3MLCT character.[27]
[nm]
(77 K)[a] (77 K)[a]
1
2
3
4
580
556
532
537
0.54
0.62
0.81
0.63
0.048 1013 100
540
532
521
519
5.56
5.06
5.03
4.68
0.040 1110
0.035 1775
0.034 1620
78
78
77
[a] Measured in CH2Cl2/MeOH (1:1) glass. Values of φ0 and φ are
quantum yields in degassed and aerated solutions, respectively; τ0
and τ are lifetimes in degassed and aerated solutions, respectively.
The emission spectra obtained in rigid matrices at 77 K
(Figure 7) show that all the compounds 1–4 undergo a blue-
shift of the emission maximum with respect to the room-
temperature emission. This observation confirms that there
is a 3MLCT character of the emitting state.[31] The long
lifetime recorded at 77 K for all the complexes and the
structure of the emission again indicate the mixing with the
3LC states and the triplet character of the emitting state.
Figure 6. Normalized emission profile of 1 (black), 2 (red), 3 (blue),
and 4 (green) in DMF at room temperature and upon excitation at
λ = 400 nm.
The photophysical properties of 1–4 are summarized in
Table 3; all complexes were highly luminescent in oxygen-
free solvents, with quantum yields Φ0 of 0.54 up to 0.81.
Long excited-state lifetimes (τ0 = 1.0–1.7 μs) were observed
for all the complexes, thereby confirming the triplet nature
of the emission. In air-equilibrated solutions, a general de-
crease of quantum yields and lifetimes was observed as a
consequence of the oxygen quenching. Under these condi-
tions, the hemicaged 1 showed a higher quantum yield (Φ
= 0.048) than the open form 2 (Φ = 0.040), whereas com-
plexes 3 and 4 show basically the same quantum yield (Φ
= 0.035 and 0.034, respectively). The same trend was ob-
served for the emission lifetimes in deaerated solution. The
hemicage 1 showed a longer lifetime (τ = 100 ns) than the
open complex 2 (τ = 78 ns), whereas the two complexes
Figure 7. Normalized emission profile of 1 (black), 2 (red), 3 (blue),
and 4 (green) in CH2Cl2/MeOH (1:1) glass at 77 K upon excitation
at λ = 400 nm.
All the relevant electrochemical data are summarized in
substituted on the phenyl ring showed practically the same Table 4. For complex 1, the main aspect that emerges from
lifetime (τ = 78 and 77 ns for 3 and 4, respectively). The the electrochemical analysis is the effect played by the he-
different trend observed for the quantum yields and excited- micage formation on lowering the LUMO level of 1 by
state lifetimes of the hemicaged 1 and the open form 2 in about 0.1 eV compared to 2. The presence of the hemicage
the absence and presence of oxygen provides a first indica- seems to induce a more effective conjugation (coplanarity)
tion of the different behavior towards oxygen quenching. In of the amide bond in the meta position in the pyridyl
contrast, a quite similar decrease is observed for the two moiety of 1 compared to that in 2, thus stabilizing the
complexes 3 and 4, which suggests a comparable behavior LUMO. In fact, 1 is more easily reduced than its related
of these two compounds towards oxygen quenching. There- open-form 2 (Eored,1 Ͼ Eroed,2). This finding could also be
fore, comparison of the two pairs of complexes shielded on ascribed to a sizeable difference in solvation between 1 and
the pyridine (1, 2) or on the phenyl (3, 4) ring provides 2, but such stabilization is not evidenced by DFT calcula-
an explanation for the different oxygen sensitivity, which is tions even if the solvation effect is taken into account by
induced by the shielding of the atoms on which the LUMO means of a conductor-like polarizable continuum model
(1, 2) or the HOMO (3, 4) orbitals are localized. According (PCM) as single-point calculation at the S0-optimized ge-
to these qualitative data, the shielding of the LUMO exerts ometry (see Table S11 in the Supporting Information).
an important effect on the oxygen sensitivity, whereas the However, while comparing theoretical and electrochemical
shielding of the HOMO does not have any relevant effect. HOMO and LUMO energy levels, one should keep in mind
A more quantitative evaluation of the difference in oxygen that calculated values are approximations of the ionization
1032
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 1025–1037

Iridium(III)–Tris(phenylpyridine) Complexes
potential (HOMO) and electron affinity (LUMO) (i.e., ver- quencher, and [O2] the concentration of oxygen in solu-
tical oxidation and reduction potentials), whereas the elec- tion.[32] The Stern–Volmer plots for the four complexes are
trochemical values are the result of adiabatic (equilibrium) reported in Figure 8.
processes, which lead to the formation of radical cation and
anion, respectively. The HOMO level is not affected by the
hemicage structure (Eoox,1 ≈ Eoox,2), which confirms that it is
mainly centered on the Ir metal core, as also supported by
DFT calculations (vide supra).
Table 4. Electrochemical potential values (versus SCE) for samples
1–4 in DMF/0.1 m tetrabutylammonium hydroxide (TBAH).[a][b][c]
Eored
[LUMO]
Eoox
[HOMO]
ΔEH–L
1
2
3
–1.74 V, r
[–2.60 eV]
–1.87 V, r
[–2.46 eV]
–1.78 V, r
[–2.56 eV]
+0.92 V, r
[–5.26 eV]
+0.91 V, r
[–5.25 eV]
+1.06 V
2.66 eV
2.78 eV
2.84 eV
[–5.40 eV]
4
n/a[d]
n/a[d]
n/a[d]
[a] The sample solutions under investigation were 1 mm in DMF/
0.1 m TBAH. A glassy carbon was employed as the working elec-
trode, a platinum ring as the counter, and a silver wire as the refer-
ence. The scan rate was varied in the range of 0.2–5 Vs–1. E = Eo
calculated as the mean value between the cathodic and the anodic
peak averaged in the scan rate range 0.2–0.5 Vs–1. [b] The peak-to-
peak separation was ranging between 70 and 80 mV at 0.2 Vs–1,
which is larger than expected for an ideal Nernstian behavior
(59 mV). However, the behavior of the redox couple ferrocene/fer-
ricenium (Fc+|Fc0), used as internal standard, showed the same
trend. Therefore, we can attribute the observed effect to the ohmic
drop of the system, as previously reported for aprotic media by
Bard and co-workers.[47] The peak currents instead were found to
be linearly dependent on the square root of the scan rate as ex-
pected for a diffusion-controlled redox process. [c] Reversible peak.
[d] The solubility of 4 was poor under these experimental condi-
tions, and it was not possible to estimate the potential values.
Figure 8. Stern–Volmer plot of complexes 1–4 in DMF at room
temperature. Comparison of the oxygen quenching of the two cou-
ples of complexes shows that introduction of a capping unit on the
pyridine side (hemicaged 1) induces a decrease in oxygen quenching
with respect to the open form 2. Conversely, only a minor differ-
ence is observed between 3 and 4, functionalized on the phenyl
side.
Furthermore, by comparison of 1 and 3, the reduction
potentials (Eored) of these two hemicaged systems differ by
about 40 mV. The presence of the amide bond in the meta
position in the pyridyl moiety is more effective in stabilizing
the LUMO than in the phenyl moiety of 3, thus corroborat-
ing the evidence of the redshift observed in the emission
spectra (vide supra). Compound 4 is poorly soluble under
the conditions used for the electrochemical measurements,
and it was not possible to estimate the HOMO and LUMO
values from electrochemical measurements. Nevertheless,
because of the similar photophysical characteristics of 3
and 4, it is possible to make an educated guess also for 4
(vide infra).
The quenching constant kq was calculated according to
Equation (1) from the value of the slope obtained by fitting
the Stern–Volmer plot. By consecutive comparison of the
quenching constants (Table 5), it was possible to determine
the efficiency of oxygen quenching. The pyridine-substi-
tuted complexes showed a remarkable decrease in oxygen
quenching (40%) going from the open (2) to the hemicaged
(1) complex, whereas no noticeable changes were observed
in the case of the phenyl-substituted complexes (3, 4). To
exclude any difference in terms of energy or electron trans-
Table 5. Oxygen quenching constant (kq) and thermodynamic pa-
rameters involved in the oxygen quenching mechanisms.
The luminescence oxygen quenching processes of com-
plexes 1–4 were studied by monitoring the luminescence in-
tensity of solutions with different concentrations of oxygen
and by plotting the obtained results according to the Stern–
Volmer equation [Equation (1)].
ΔGet [kJmol–1]
E00 [eV]
ΔGel [kJmol–1] kq [m–1 s–1])
[a]
1
2
3
4
2.42
2.43
2.47
2.47
–
–139
–140
–144
–144
–
–69
–71
–60
–
1.0ϫ1010
1.7ϫ1010
1.5ϫ1010
1.6ϫ1010
0.5ϫ1010
2.4ϫ1010[b]
IrCage[b]
[Ir(ppy)3][c]
–
–95
(1)
2.52
–149
for which I0 and I are the emission intensities in the ab-
sence or presence of quencher, respectively, kq is the
[a] Calculated assuming EO *(1Δg) = 0.98 eV. [b] From the litera-
2
ture.[18] [c] Calculated on the basis of the photophysical data from
quenching constant, τ0 is the lifetime in the absence of the literature[21] unless otherwise stated.
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1033

L. De Cola, A. H. Velders et al.
FULL PAPER
fer, the free energies of energy transfer (ΔGet) and of elec- in the hemicaged complexes 1 and 2 the LUMO orbital is
tron transfer (ΔGel) were calculated according to Equa- partly localized on the carbonyl moiety (vide supra). There-
tions (2) and (3).
fore, the orientation of the pyridine carbonyl towards the
inner side of the cage would likely result in a higher degree
of LUMO shielding and a consequent decrease in the oxy-
gen quenching degree, since the lowest excited state is an
MLCT in which the ligand involved is indeed the pyridine
unit.
(2)
(3)
E00 is the energy of the 0–0 transition and EO * the energy
of the excited state of singlet oxygen: EO *(1Σg2) = 1.63 eV
2
and EO *(1Δg) = 0.98 eV depending on which excited state
Conclusion
2
of oxygen is initially produced during quenching, F is Fara-
We have reported the synthesis, characterization, and
photochemical properties of two pairs of iridium(III)–tris-
(phenylpyridine) derivatives with a hemicaged (1, 3) or open
(2, 4) structure. The results obtained from this systematic
study of the oxygen quenching of hemicages and open com-
plexes shielded on the pyridine (1, 2) or on the phenyl (3,
4) ring show the hemicaged complex that bears a capping
unit on the pyridine ring (1) to have a much lower oxygen
day’s constant, EoFx is the oxidation potential of the fluoro-
phore, ErOed is the reduction potential of oxygen (–0.78 V),
2
and C is a Coulomb term (usually neglected in polar sol-
vents).[33–35] The calculated values of ΔGet and ΔGel are re-
ported in Table 5. The evaluation of the ΔGel of the open
complex 4 was not possible because of its low solubility
in DMF, which precludes the possibility of recording the
potentiometric data. However, since both kq and ΔGet ap-
peared to be identical for 3 and 4 within the experimental
error, it is reasonable to assume that also ΔGel will have the
same value for both compounds. From the analysis of these
data, it is possible to conclude that both the hemicaged and
open molecule of each pair show only a modest difference
in terms of thermodynamic feasibility of energy or electron
transfer. The basically equal thermodynamic feasibility of
energy and electron transfer between the pair of complexes
1, 2 and 3, 4 suggests that the remarkable difference ob-
served in the quenching constant kq between the hemicaged
(1) and the open (2) complex functionalized on the pyridine
side of the [Ir(ppy)3] is due to a structural effect.
quenching constant (kq = 1.0ϫ1010
open complex 2 with ethylamide on the pyridine rings (kq
= 1.7ϫ1010 –1 s–1). Moreover, the hemicaged complex 3,
m
–1 s–1) than its parent
m
in which the capping unit is connected to the phenyl rings,
shows basically the same degree of oxygen quenching as
the open complex 4 functionalized with ethylamides on the
phenyl rings (kq = 1.5ϫ1010 and 1.6ϫ1010 –1 s–1, respec-
m
tively). The ΔGet and ΔGel parameters show that in terms
of thermodynamic feasibility only a modest difference exists
between the members of each pair of compounds (1, 2 and
3, 4) and, therefore, the lower quenching shown by 1 with
respect to 2 can be ascribed to a structural effect. The struc-
tural (shielding) effect induced by the presence of the cap-
ping unit is due to the shielding of the LUMO orbital ex-
erted by the capping unit, which is possibly enforced by the
high degree of twisting of the hemicaged complex 1. It is
reasonable to hypothesize that both these parameters give
a contribution that results in a lower degree of oxygen
quenching of 1 with respect to 2, but on the basis of the
structural and photophysical data available it was not pos-
sible to quantify the respective contributions of the two ef-
fects. The introduction of a capping unit on the phenyl ring
induced only a minor change in the oxygen sensitivity of
the hemicaged complex 3 with respect to the open complex
4, which is explained by the fact that the LUMO of the
hemicage 3 is mostly localized on the pyridine ring, and the
luminescent state is an MLCT that involves the pyridine.
The structure–photoluminescence quenching relationship
obtained from this study together with thermodynamic con-
siderations for the substituents on the phenylpyridine ring
nicely explain the observed shielding mechanism in the
hitherto only reported iridium(III) caged complex.
The caged complex (ii, Scheme 1) previously reported by
our group shows a further 50% decrease in the oxygen
quenching (kq = 0.5ϫ1010 –1 s–1) with respect to 1.[18] On
m
the basis of the emission maximum (λem = 570 nm), it is
reasonable to hypothesize that the energy of the triplet state
(E00) of the cage has likely an intermediate value between
the E00 of 1 (λem = 580 nm) and 2 (λem = 556 nm), therefore
the value of the free energy of the energy transfer ΔGet can
be estimated to be similar to the ΔGet of 1 and 2. On the
other hand, because of the presence of two amide moieties
on the phenylpyridine ligands, the oxidation potential of
the cage is probably higher than the oxidation potential of
1.[1] Consequently, the value of the free energy of the elec-
tron transfer ΔGel is probably higher than the value found
for 1, thereby resulting in a less favorable thermodynamic
feasibility of electron transfer. In conclusion, on the basis
of the estimated thermodynamic parameters, it is possible
to hypothesize that the low degree of oxygen quenching ob-
served for the caged complex ii probably derives from a de-
creased thermodynamic feasibility of the electron transfer
with respect to 1. We believe, however, that structural fac-
tors also play an important role in the low degree of oxygen
quenching of the caged IrIII complex ii: it has been observed
that in the caged complex the amide moiety located on the
pyridine side shows the carbonyl oriented towards the inner
Experimental Section
General: The NMR spectroscopic experiments were performed
with
a Bruker Avance II NMR spectrometer operating at
side of the molecule. According to computational models, 600.35 MHz for 1H and 150.09 MHz for 13C. Chemical shifts are
1034
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 1025–1037

Iridium(III)–Tris(phenylpyridine) Complexes
given in ppm using the residual solvent signal as reference. The
multiplicity of the peaks is reported by using the following abbrevi-
ations: s = singlet, d = doublet, t = triplet, quint = quintuplet, m
= multiplet. The chemical shifts and assignments of 13C signals are
reported in the Supporting Information. Mass spectra were mea-
sured with a Micromass LCT (ESI-HRMS) spectrometer. IR spec-
tra were measured with a Thermo Scientific Nicolet 6700 FTIR
spectrometer equipped with a Smart Orbit diamond ATR access-
ory. Main bands are reported and assigned to functional groups by
using the following abbreviations: br. = broad band, str. = stretch-
ing band, def. = deformation band. UV/Vis spectra were measured
with a Perkin–Elmer Lambda 850 UV/Vis spectrophotometer by
using a quartz cuvette with 1 cm path length. Steady-state lumines-
cence spectra were measured with an Edinburgh FS900 fluorospec-
trometer. A 450 W xenon arc lamp was used as excitation source.
mixtures prepared with a Brooks 5850S Mass Flow control and by
purging the fluorophore solutions for 40 min.
Computational Details: Geometries were optimized by means of
DFT and employing the exchange correlation hybrid functional
B3 LY P. [15,37,38] The standard valence double-ζ polarized basis set
6-31G(d,p)[39] was used for C, H, N, and O. For Ir, the Stuttgart–
Dresden (SDD) effective core potentials were employed along with
the corresponding valence triple-ζ basis set. All the calculations
were done assuming C3 symmetry. The nature of all the stationary
points was checked by computing vibrational frequencies, and all
the species were found to be true potential energy minima, as no
imaginary frequency was obtained (NImag = 0). Single-point calcu-
lations at the S0-optimized geometry were done also in the solvent
used for the electrochemical characterization (N,N-dimethylform-
amide), described by means of a conductor-like polarizable contin-
uum model (PCM).[40–42]
Luminescence quantum yields at room temperature (Φ and Φair
)
were evaluated by comparing wavelength-integrated intensities (I)
of isoabsorptive optically diluted solutions (abs. Ͻ 0.1) with refer-
ence to [Ru(bpy)3]Cl2 (ΦR = 0.028 in air-equilibrated water) or flu-
orescein (ΦR = 0.95 in 0.1 m NaOH) standard for the complexes 1,
2 and 3, 4, respectively, and by using the Equation (4), in which n
and nR are the refractive index of the sample and reference solvent,
respectively.[36]
To simulate the absorption electronic spectrum down to 300 nm,
for each complex the lowest 30 singlet (S0 Ǟ Sn) as well as the four
lowest triplet excitation energies (S0 Ǟ Tn) were computed on the
optimized geometry at the S0 by means of TD-DFT calcula-
tions.[43,44] Oscillator strengths were deduced from the dipole tran-
sition-matrix elements (for single states only). Isodensity surfaces
plots of selected orbitals for the investigated structures in the gas
phase at their optimized ground-state geometry were plotted at iso-
density value of 0.035 eBohr–3. All the calculations were performed
with Gaussian 09 program package.[45]
(4)
Luminescence lifetimes of the compounds were determined by re-
cording the decay curves of the luminescence intensity at the emis-
sion maximum using the TCSPC option on a HORIBA Jobin Yvon
Fluoromax 4 instrument and a pulsed solid-state LED as excitation
source at 462 nm wavelength. The recorded data were analyzed
using the DAS6 software package of HORIBA Jobin Yvon. De-
gassed solutions were prepared by four freeze–pump–thaw cycles.
Electrochemical measurements were done in N,N-dimethylform-
amide (Acros, extra dry over molecular sieves, 99.8%), which was
used as received without any further purification. Tetrabutylammo-
nium hexafluorophosphate (electrochemical grade, Ն99%, Fluka)
was used as supporting electrolyte, which was recrystallized from a
1:1 ethanol/water solution and dried at 60 °C under vacuum. For
Materials: Oxygen-sensitive reactions were carried out by using
standard Schlenk techniques. Commercial-grade reagents were pur-
chased from Sigma–Aldrich and used without further purification.
4-(Pyridin-2-yl)benzoic acid (8) was prepared according to the lit-
erature procedure.[46]
6-Phenylnicotinic Acid (5): Phenylboronic acid (3 g, 24.6 mmol), 6-
bromonicotinic acid (3.66 g, 18.2 mmol), and tetrakis(triphenyl-
phosphane)palladium(0) (900 mg, 0.78 mmol) were dissolved in a
mixture (180 mL) of Na2CO3 (0.2 m) and acetonitrile (1:1). After
several argon/vacuum cycles, the reaction was heated at reflux for
48 h. The hot reaction mixture was filtered through Celite, the ace-
tonitrile removed, and the resulting aqueous solution was extracted
several times with CH2Cl2. The aqueous solution was then acidified
with acetic acid and the white precipitate was filtered and dried
with P2O5; yield 3.1 g (15.6 mmol; 85%) of pure compound. 1H
NMR ([D6]DMSO): δ = 9.15 (s, 1 H), 8.33 (d, J = 12 Hz, 1 H),
8.15 (d, J = 6 Hz, 2 H), 8.10 (d, J = 12 Hz, 1 H), 7.54–7.49 (m, 3
H) ppm. 13C NMR ([D6]DMSO): see the Supporting Information.
the electrochemical experiments,
a CHI750C Electrochemical
Workstation (CH Instruments, Inc., Astin, TX, USA) was used.
The electrochemical experiments were performed in a glass cell un-
der an Ar atmosphere. To minimize the ohmic drop between the
working and the reference electrodes, feedback correction was em-
ployed. The electrochemical experiments were performed by using
a 3 mm diameter glassy carbon disk electrode (homemade from
Tokai glassy carbon rod). Before the experiments, the working elec-
trode was polished with a 0.05 μm diamond suspension (Metadi
Supreme Diamond Suspension, Buehler) and ultrasonically rinsed
with ethanol for 5 min. The electrode was electrochemically acti-
vated in the background solution by means of several voltammetric
cycles at 0.5 Vs–1 between the anodic and cathodic solvent/electro-
lyte discharges until the same quality features were obtained. The
reference electrode was a silver quasi-reference electrode (Ag-
QRE), which was separated from the catholyte by a glass frit
(Vycor). The reference electrode was calibrated at the end of each
experiment against the ferrocene/ferricenium couple, the formal po-
tential of which in N,N-dimethylformamide is 0.464 V against the
standard calomel electrode (SCE) with satd. aqueous KCl; all po-
tential values are reported against SCE. A platinum ring or coil
served as the counter electrode.
IR (neat): ν = 1673 (C=O str.), 1417 (OH def. acid), 935 (OH def.
˜
acid) cm–1. ESI-HRMS: calcd. for C12H9NO2 199.063 [M+]; found
199.065.
General Synthesis of Hemicaged Ligands: Acid (600 mg, 3 mmol),
N,NЈ-dicyclohexylcarbodiimide (620 mg, 3 mmol), and N-hydroxy-
benzotriazole (400 mg, 3 mmol) were dissolved in dry N,NЈ-dimeth-
ylacetamide (50 mL). Tris(2-aminoethyl)amine (120 μL, 0.8 mmol)
was added dropwise and the resulting mixture was stirred overnight
at 60 °C. The reaction mixture was then filtered and added to ethyl
ether (ca. 400 mL) under vigorous stirring. The white precipitate
was then collected by filtration and washed with hot acetonitrile.
In a typical experiment, 212 mg (0.3 mmol, 40%) of pure tripodal
ligand was obtained.
1
Hemicaged Ligand (6): H NMR ([D4]MeOH): δ = 8.90 (s, 1 H),
7.93 (dd, J = 12 Hz, 1 H), 7.73 (d, J = 6 Hz, 2 H), 7.43 (d, J =
6 Hz, 1 H), 7.37 (t, J = 6 Hz, 1 H), 7.29 (t, J = 6 Hz, 2 H), 3.60 (t,
J = 6 Hz, 2 H), 2.82 (t, J = 6 Hz, 2 H) ppm. 13C NMR ([D4]-
Solutions with different oxygen concentration, suitable for the
Stern–Volmer quenching studies, were prepared by using N2/O2
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1035

L. De Cola, A. H. Velders et al.
FULL PAPER
MeOH): see the Supporting Information. IR (neat): ν = 3396 (NH
6.25 (d, J = 6 Hz, 1 H), 3.90 (t, J = 6 Hz, 1 H), 3.05 (t, J = 12 Hz,
˜
str.), 2948 (CH2 str.), 2815 (N–CH2/CH2 str.), 1630 (C=O str.), 1 H), 2.79 (t, J = 12 Hz, 1 H), 2.37 (d, J = 12 Hz, 1 H) ppm. 13C
1587 (NH bend.), 1465 (CH2–N str.), 743 (N–C–O str.) cm–1. ESI-
HRMS: calcd. 690.319 [M + H+]; found 690.320.
NMR ([D ]CH Cl ): see the Supporting Information. IR (neat): ν
˜
2 2 2
= 3290 (br., NH str.), 3066 (CH2 str.), 2921 (CH2 str.), 2850 (N–
CH2/CH2 str.), 1643 (C=O str.), 1540 (NH bend.), 1470 (CH2–N
str.), 756 (N–C–O str.) cm–1. ESI-HRMS: calcd. 880.259 [M + H+];
found 880.262.
1
Hemicaged Ligand (9): H NMR ([D4]MeOH): δ = 8.46 (s, 1 H),
7.69 (d, J = 6 Hz, 2 H), 7.66 (d, J = 6 Hz, 2 H), 7.59 (t, J = 12 Hz,
1 H), 7.49 (d, J = 6 Hz, 1 H), 7.24 (t, J = 6 Hz, 1 H), 3.59 (t, J =
6 Hz, 2 H), 2.82 (t, J = 6 Hz, 2 H) ppm. 13C NMR ([D4]MeOH):
General Synthesis of Open IrIII Complexes: Ligand (100 mg,
0.42 mmol) and IrCl3 (66 mg, 0.22 mmol) were dissolved in a mix-
ture of 2-ethoxyethanol and water (3:1). After several cycles of ar-
gon/vacuum, the mixture was heated to reflux overnight. The reac-
tion mixture was cooled to room temperature, half of the solvent
was evaporated, and the resulting solution was poured on ice. The
resulting precipitate was filtered, washed with water and diethyl
ether, and dried under vacuum. The solid (≈70 mg) was then sus-
pended in toluene, ligand (36 mg, 0.15 mmol) and CF3SO3Ag
(80 mg, 0.30 mmol) were added, several cycles argon/vacuum were
performed to remove oxygen, and the mixture was heated to reflux
overnight under inert atmosphere. The crude was then washed with
methanol and then dissolved in CH2Cl2 (10% MeOH), filtered
through Celite, and eventually purified by preparative TLC
(CH2Cl2/MeOH, 95:5).
see the Supporting Information. IR (neat): ν = 3293 (NH str.), 3055
˜
(CH2 str.), 2925 (CH2 str.), 2790 (N–CH2/CH2 str.), 1631 (C=O
str.), 1537 (NH bend.), 1463 (CH2–N str.), 752 (N–C–O str.) cm–1.
HRMS: calcd. 690.319 [M + H+]; found 690.317.
General Synthesis of Ethylamide Ligands: Acid (1 g, 5 mmol)
and
1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide
(1.9 g,
12.5 mmol) were dissolved in a mixture (60 mL) of CH2Cl2/pyr-
idine (7:3). Ethylamine (3 mL, 2.0 m in THF) was added dropwise,
and the mixture was stirred dropwise at room temperature. The
solvent was removed under vacuum, and the crude was then sus-
pended in water and extracted with ethyl acetate. The organic phase
was washed with a satd. solution of CuSO4, then with satd. NH4Cl,
and eventually with brine. The organic phase was then dried with
Na2SO4 and the solvent removed by rotavapor. Typically, 1 g
(4.25 mmol; 85%) of pure compound was obtained.
1
Ir-73 (2): H NMR ([D7]DMF): δ = 8.60 (t, J = 6 Hz, 1 H), 8.35
(d, J = 6 Hz, 1 H), 8.31–8.28 (m, 2 H), 7.90 (d, J = 12 Hz, 1 H),
6.87 (t, J = 6 Hz, 1 H), 6.79–6.74 (m, 2 H), 3.27 (quint., J = 6 Hz,
2 H), 1.08 (t, J = 6 Hz, 3 H) ppm. 13C NMR ([D7]DMF): see the
N-Ethyl-6-phenylnicotinamide (7): 1H NMR ([D6]DMSO): δ = 9.21
(s, 1 H), 8.38 (d, J = 6 Hz, 1 H), 8.21 (d, J = 6 Hz, 2 H), 8.13 (d,
J = 6 Hz, 1 H), 7.57–7.50 (m, 3 H), 3.43 (q, J = 6 Hz, 2 H), 1.21
(t, J = 6 Hz, 3 H) ppm. 13C NMR ([D4]MeOH): see the Supporting
Supporting Information. IR (neat): ν = 3290 (N–H str.), 3035 (CH
˜
2
str.), 2970 (CH3 str.), 2930 (CH2 str.), 2873 (CH3 str.), 1635 (C=O
str.), 1540 (N–H bend.), 1473 (CH2–N str.), 1257 (CH3 str.), 748
(N–C–O str.) cm–1. ESI-HRMS: calcd. 891.261 [M + Na+]; found
891.262.
Ir-103 (4): 1H NMR ([D7]DMF): δ = 8.26 (d, J = 12 Hz, 1 H), 7.92
(t, J = 6 Hz, 1 H), 7.86–7.88 (m, 2 H), 7.63 (d, J = 6 Hz, 1 H),
7.37 (s, 1 H), 7.31 (d, J = 12 Hz, 1 H), 7.20 (t, J = 6 Hz, 1 H), 3.22
(quint., J = 6 Hz, 2 H), 1.05 (t, J = 6 Hz, 3 H) ppm. 13C NMR
Information. IR (neat): ν = 3322 (N–H str.), 2975 (CH str.), 2931
˜
3
(CH2 str.), 2877 (CH3 str.), 1627 (C=O str.), 1522 (N–H str.), 1469
(CH2–N str.), 1267 (CH3 str.), 746 (N–C–O str.) cm–1. ESI-HRMS:
calcd. 226.111 [M+]; found 226.113.
N-Ethyl-4-(pyridin-2-yl)benzamide (10): 1H NMR ([D6]DMSO): δ
= 8.71 (d, J = 6 Hz, 1 H), 8.57 (t, J = 6 Hz, 1 H), 8.18 (d, J =
6 Hz, 2 H), 8.05 (d, J = 6 Hz, 1 H), 7.96 (d, J = 6 Hz, 2 H), 7.94
(t, J = 12 Hz, 1 H), 7.41 (t, J = 6 Hz, 1 H), 3.31 (m), 1.15 (t, J
= 6 Hz, 3 H) ppm. 13C NMR ([D4]MeOH): see the Supporting
([D ]DMF): see the Supporting Information. IR (neat): ν = 3297
˜
7
(N–H str.), 3060 (CH2 str.), 2966 (CH3 str.), 2927 (CH2 str.), 2852
(CH3 str.), 1633 (C=O str.), 1540 (N–H bend.), 1467 (CH2–N str.),
1257 (CH3 str.), 754 (N–C–O str.) cm–1. ESI-HRMS: calcd. 869.279
[M + H+]; found 869.278.
Information. IR (neat): ν = 3297 (N–H str.), 2973 (CH str.), 2931
˜
3
(CH2 str.), 2870 (CH3 str.), 1627 (C=O str.), 1548 (N–H bend.),
1463 (CH2–N str.), 1288 (CH3 str.), 746 (N–C–O str.) cm–1. ESI-
HRMS: calcd. 226.111 [M+]; found 226.115.
Supporting Information (see footnote on the first page of this arti-
cle): HH-COSY, HH-ROESY and 13C NMR data, isodensity plots
and energies of selected molecular orbitals, excitation energies and
oscillator strength of the lowest transitions.
General Synthesis of IrIII Hemicages: Tripodal ligand (100 mg,
0.14 mmol), IrCl3 (41 mg, 0.14 mmol), and CF3CO2Ag (89 mg,
0.4 mmol) were stirred in ethylene glycol (10 mL) previously purged
with nitrogen. After several cycles of argon/vacuum, the resulting
mixture was heated to reflux overnight. The reaction mixture was
cooled to room temperature, diluted with water, and extracted sev-
eral times with ethyl acetate. The organic phases were collected,
washed with brine, and dried with Na2SO4. The products were then
purified by column chromatography (SiO2, CH2Cl2/MeOH, 95:5).
Acknowledgments
M. M. thanks the Alexander Humboldt Fundation for financial
support.
1
Ir-6 (1): H NMR ([D2]CH2Cl2): δ = 7.99 (d, J = 12 Hz, 1 H), 7.96
[1] L. Flamigni, A. Barbieri, C. Sabatini, B. Ventura, F. Barigel-
letti, Photochemistry and Photophysics of Coordination Com-
pounds II, vol. 281, 2007, pp. 143–203.
[2] A. Tsuboyama, H. Iwawaki, M. Furugori, T. Mukaide, J. Kam-
atani, S. Igawa, T. Moriyama, S. Miura, T. Takiguchi, S. Ok-
ada, M. Hoshino, K. Ueno, J. Am. Chem. Soc. 2003, 125,
12971–12979.
(d, J = 6 Hz, 1 H), 7.76 (s, 1 H), 7.71 (m, 1 H), 7.02 (d, J = 6 Hz,
1 H), 6.95–6.96 (m, 2 H), 6.34 (t, J = 6 Hz, 1 H), 3.62 (d, J =
18 Hz, 1 H), 3.41 (t, J = 12 Hz, 1 H), 2.91 (t, J = 6 Hz, 1 H), 2.30
(d, J = 12 Hz, 1 H) ppm. 13C NMR ([D2]CH2Cl2): see the Support-
ing Information. ν = IR (neat): 3274 (br., NH str.), 3041 (CH str.),
˜
2
2925 (CH2 str.), 2815 (N–CH2/CH2 str.), 1633 (C=O str.), 1539
(NH bend.), 1471 (CH2–N str.), 748 (N–C–O str.) cm–1. ESI-
HRMS: calcd. 880.259 [M + H+]; found 880.260.
[3] J. I. Kim, I. S. Shin, H. Kim, J. K. Lee, J. Am. Chem. Soc. 2005,
127, 1614–1615.
[4] M. A. Baldo, D. F. O’Brien, Y. You, A. Shoustikov, S. Sibley,
M. E. Thompson, S. R. Forrest, Nature 1998, 395, 151–154.
[5] V. Fernandez-Moreira, F. L. Thorp-Greenwood, M. P. Coogan,
Chem. Commun. 2010, 46, 186–202.
1
Ir-9 (3): H NMR ([D2]CH2Cl2): δ = 7.98 (d, J = 6 Hz, 1 H), 7.75
(d, J = 12 Hz, 1 H), 7.72 (d, J = 12 Hz, 1 H), 7.59 (d, J = 6 Hz, 1
H), 7.40 (d, J = 6 Hz, 1 H), 7.33 (s, 1 H), 7.09 (t, J = 6 Hz, 1 H),
1036
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 1025–1037

Iridium(III)–Tris(phenylpyridine) Complexes
[6]
[29] I. Avilov, P. Minoofar, J. Cornil, L. De Cola, J. Am. Chem. Soc.
2007, 129, 8247–8258.
[30] K. D. Oyler, F. J. Coughlin, S. Bernhard, J. Am. Chem. Soc.
2007, 129, 210–217.
[31] A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, A.
Vonzelewsky, Coord. Chem. Rev. 1988, 84, 85–277.
[32] J. R. Lakowicz, Principles of Fluorescence Spectroscopy,
Springer, Singapore, 2006.
[33] Q. G. Mulazzani, H. Sun, M. Z. Hoffman, W. E. Ford, M. A. J.
Rodgers, J. Phys. Chem. 1994, 98, 1145–1150.
[34] A. A. Abdel-Shafi, P. D. Beer, R. J. Mortimer, F. Wilkinson,
Helv. Chim. Acta 2001, 84, 2784–2795.
[35] M. Bodesheim, M. Schutz, R. Schmidt, Chem. Phys. Lett.
1994, 221, 7–14.
A. Ruggi, D. N. Reinhoudt, A. H. Velders, in: Bioinorganic Me-
dicinal Chemistry (Ed.: E. Alessio), Wiley-VCH, Weinheim,
Germany, 2011.
D. Crespy, K. Landfester, U. S. Schubert, A. Schiller, Chem.
Commun. 2010, 46, 6651–6662.
J. Kuil, P. Steunenberg, P. T. K. Chin, J. Oldenburg, K. Jalink,
A. H. Velders, F. W. B. van Leeuwen, ChemBioChem 2011, 12,
1897–1903.
J. N. Demas, E. W. Harris, R. P. McBride, J. Am. Chem. Soc.
1977, 99, 3547–3551.
[7]
[8]
[9]
[10]
M. P. Coogan, J. B. Court, V. L. Gray, A. J. Hayes, S. H. Lloyd,
C. O. Millet, S. J. A. Pope, D. Lloyd, Photochem. Photobiol. Sci.
2010, 9, 103–109.
[36] M. Montalti, A. Credi, L. Prodi, M. T. Gandolfi, Handbook of
Photochemistry, CRC press, Boca Raton, 2006.
[11]
[12]
[13]
[14]
J. N. Demas, D. Diemente, E. W. Harris, J. Am. Chem. Soc.
1973, 95, 6864–6865.
S. J. Zhang, M. Hosaka, T. Yoshihara, K. Negishi, Y. Iida, S.
Tobita, T. Takeuchi, Cancer Res. 2010, 70, 4490–4498.
A. Ruggi, F. W. B. van Leeuwen, A. H. Velders, Coord. Chem.
Rev. 2011, 255, 2542–2554.
C. Li, M. Yu, Y. Sun, Y. Wu, C. Huang, F. Li, J. Am. Chem.
Soc. 2011, 133, 11231–11239.
P. J. Hay, J. Phys. Chem. A 2002, 106, 1634–1641.
[37] C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–
789.
[38] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[39] M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M. S.
Gordon, D. J. Defrees, J. A. Pople, J. Chem. Phys. 1982, 77,
3654–3665.
[40] M. Cossi, N. Rega, G. Scalmani, V. Barone, J. Comput. Chem.
2003, 24, 669–681.
[41] M. Cossi, V. Barone, J. Chem. Phys. 2001, 115, 4708–4717.
[42] V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102, 1995–2001.
[43] R. E. Stratmann, G. E. Scuseria, M. J. Frisch, J. Chem. Phys.
1998, 109, 8218–8224.
[44] M. E. Casida, C. Jamorski, K. C. Casida, D. R. Salahub, J.
Chem. Phys. 1998, 108, 4439–4449.
[15]
[16]
M. S. Lowry, S. Bernhard, Chem. Eur. J. 2006, 12, 7970–7977.
[17] F. Barigelletti, L. De Cola, V. Balzani, P. Belser, A. Von Zelew-
sky, F. Voegtle, F. Ebmeyer, S. Grammenudi, J. Am. Chem. Soc.
1989, 111, 4662–4668.
[18] A. Ruggi, M. B. Alonso, D. N. Reinhoudt, A. H. Velders,
Chem. Commun. 2010, 46, 6726–6728.
[45] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M.
Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Starov-
erov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell,
J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M.
Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Ad-
amo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Mar-
tin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09, re-
vision B.1, Gaussian, Inc., Wallingford CT, 2009.
[46] Y. Gong, H. W. Pauls, Synlett 2000, 829–831.
[47] A. J. Bard, L. R. Faulkner, Electrochemical Methods, Funda-
mentals and Applications, Wiley, New York, 2001.
Received: November 24, 2011
[19] R. F. Beeston, S. L. Larson, M. C. Fitzgerald, Inorg. Chem.
1989, 28, 4187–4189.
[20] N. Miyaura, A. Suzuki, Chem. Rev. 2002, 95, 2457–2483.
[21] A. B. Tamayo, B. D. Alleyne, P. I. Djurovich, S. Lamansky, I.
Tsyba, N. N. Ho, R. Bau, M. E. Thompson, J. Am. Chem. Soc.
2003, 125, 7377–7387.
[22] J. Breu, P. Stossel, S. Schrader, A. Starukhin, W. J. Finkenzeller,
H. Yersin, Chem. Mater. 2005, 17, 1745–1752.
[23] K. Nozaki, K. Takamori, Y. Nakatsugawa, T. Ohno, Inorg.
Chem. 2006, 45, 6161–6178.
[24] Experimental values are referred to fac-[Ir(ppy)3], see: J. Breu,
P. Stossel, S. Schrader, A. Starukhin, W. J. Finkenzeller, H. Yer-
sin, Chem. Mater. 2005, 17, 1745–1752.
[25] S. Fantacci, F. De Angelis, Coord. Chem. Rev., DOI: 10.1016/
j.ccr.2011.1003.1008.
[26] A. Vlcek, S. Zalis, Coord. Chem. Rev. 2007, 251, 258–287.
[27] S. Lamansky, P. Djurovich, D. Murphy, F. Abdel-Razzaq, H. E.
Lee, C. Adachi, P. E. Burrows, S. R. Forrest, M. E. Thompson,
J. Am. Chem. Soc. 2001, 123, 4304–4312.
[28] W. Holzer, A. Penzkofer, T. Tsuboi, Chem. Phys. 2005, 308, 93–
102.
Published Online: January 11, 2012
Eur. J. Inorg. Chem. 2012, 1025–1037
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1037