Design rules for charge-transport efficient host materials for phosphorescent organic light-emitting diodes

Article abstract of DOI:10.1021/ja305310r

The use of blue phosphorescent emitters in organic light-emitting diodes (OLEDs) imposes demanding requirements on a host material. Among these are large triplet energies, the alignment of levels with respect to the emitter, the ability to form and sustain amorphous order, material processability, and an adequate charge carrier mobility. A possible design strategy is to choose a π-conjugated core with a high triplet level and to fulfill the other requirements by using suitable substituents. Bulky substituents, however, induce large spatial separations between conjugated cores, can substantially reduce intermolecular electronic couplings, and decrease the charge mobility of the host. In this work we analyze charge transport in amorphous 2,8- bis(triphenylsilyl)dibenzofuran, an electron-transporting material synthesized to serve as a host in deep-blue OLEDs. We show that mesomeric effects delocalize the frontier orbitals over the substituents recovering strong electronic couplings and lowering reorganization energies, especially for electrons, while keeping energetic disorder small. Admittance spectroscopy measurements reveal that the material has indeed a high electron mobility and a small Poole-Frenkel slope, supporting our conclusions. By linking electronic structure, molecular packing, and mobility, we provide a pathway to the rational design of hosts with high charge mobilities.

Full text of DOI:10.1021/ja305310r

Article
pubs.acs.org/JACS
Design Rules for Charge-Transport Efficient Host Materials for
Phosphorescent Organic Light-Emitting Diodes
Falk May,†,‡ Mustapha Al-Helwi,‡,§,⊥ Bjorn Baumeier,† Wolfgang Kowalsky,⊥,∥ Evelyn Fuchs,§
̈
Christian Lennartz,*,‡,§ and Denis Andrienko*,†
†Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany
‡Innovation Lab Heidelberg, Speyerer Strasse 4, 69115 Heidelberg, Germany
§BASF SE, B009, 67056 Ludwigshafen, Germany
⊥Kirchhoff-Institut fur Physik der Universitat Heidelberg, Im Neuenheimer Feld 227, 69120 Heidelberg, Germany
̈
̈
∥Institut fur Hochfrequenztechnik, Technische Universitat Braunschweig, Schleinitzstrasse 22, 38106 Braunschweig, Germany
̈
̈
S
* Supporting Information
ABSTRACT: The use of blue phosphorescent emitters in
organic light-emitting diodes (OLEDs) imposes demanding
requirements on a host material. Among these are large triplet
energies, the alignment of levels with respect to the emitter,
the ability to form and sustain amorphous order, material
processability, and an adequate charge carrier mobility. A
possible design strategy is to choose a π-conjugated core with a
high triplet level and to fulfill the other requirements by using suitable substituents. Bulky substituents, however, induce large
spatial separations between conjugated cores, can substantially reduce intermolecular electronic couplings, and decrease the
charge mobility of the host. In this work we analyze charge transport in amorphous 2,8-bis(triphenylsilyl)dibenzofuran, an
electron-transporting material synthesized to serve as a host in deep-blue OLEDs. We show that mesomeric effects delocalize the
frontier orbitals over the substituents recovering strong electronic couplings and lowering reorganization energies, especially for
electrons, while keeping energetic disorder small. Admittance spectroscopy measurements reveal that the material has indeed a
high electron mobility and a small Poole−Frenkel slope, supporting our conclusions. By linking electronic structure, molecular
packing, and mobility, we provide a pathway to the rational design of hosts with high charge mobilities.
and the resulting high triplet energy, the number of promising
compatible host materials is limited, since an even higher triplet
energy is necessary for the host to ensure trapping of the
exciton on the emitter.10,11 Among possible charge-transporting
units fulfilling this criterion are dibenzofurans and N-phenyl-
carbazoles.12 Substituents must be attached to these materials
to prevent crystallization,13 suppress emitter aggregation, and
optimize their molecular weight for vacuum deposition. The
main role of the substituents, however, is to adjust the relative
positions of electron- and hole-transporting levels as well as
singlet and triplet excited states of the host to those of the
emitter. The relative level alignment, together with the
processes occurring in the EML, is shown in Figure 1a,b for
a hole-conducting emitter. Synthetically, level adjustment can
be achieved by inductive effects14 when strongly electronegative
substituents (e.g., trifluoromethyl) are used. An alternative
approach is to exploit the mesomeric effect14 when the frontier
orbital densities delocalize (by using, e.g., triphenylsilyl
substituents).
1. INTRODUCTION
Organic light-emitting diodes (OLEDs) have recently entered
the market of flat panel displays and lighting applications.1,2 In
spite of this successful commercialization, the field still has a
number of open issues, such as insufficient stability3 of OLEDs
based on deep-blue (λ < 460 nm) emitters.4,5
In a prototypical phosphorescent OLED, holes and electrons
are injected from electrodes on opposite sides into transport
and blocking layers that provide a balanced charge transport
into an emission layer (EML). The EML itself consists of a
charge-transporting organic semiconductor and an organo-
metallic emitter that allows for triplet harvesting. To avoid
triplet quenching and triplet−triplet annihilation, the charge-
transporting material as the majority component (host) is
doped with the emitter. An exciton can be formed on a host
molecule with a subsequent energy transfer to the dopant.
Alternatively, one of the charge carriers can be trapped on the
emitter and form a neutral exciton on-site by attracting a charge
of the opposite sign. This direct charge transfer to the emitter is
argued to lead to more efficient OLEDs than excitation by
energy transfer.6−9 The balance between the two routes is
determined by the respective energy level alignment of the
emitter and the host, and can therefore be rationally designed.
However, due to the large band gap of the deep-blue emitter
Apart from a suitable level alignment, processability,
amorphousness, and stability, an adequate charge carrier
Received: June 6, 2012
Published: July 30, 2012
© 2012 American Chemical Society
13818
dx.doi.org/10.1021/ja305310r | J. Am. Chem. Soc. 2012, 134, 13818−13822

Journal of the American Chemical Society
Article
Figure 1. (a) Hole- (HT) and electron-transport (ET) levels of emitter and host in the EML. (b) Energy levels of singlets (S) and triplets (T) of the
emitter emission spectrum and the host excitation spectrum. The following processes are shown assuming a hole (electron)-transporting emitter
(host): (1) Hole transfer from emitter to host is avoided by a large barrier, Δh. (2) Electron transfer is prevented by the barrier Δe. In order to ensure
exciton formation on the emitter, Δh has to be larger than Δe and the Coulomb attraction between the hole on the emitter and the electron on the
neighboring host should be capable of overcoming Δe, attracting the electron to the emitter cation. In practice Δh ≈ 1 eV and Δe ≈ 0.3 eV. (3) Back-
transfer of the exciton from the emitter to the host is avoided by the barrier Δt for triplet excitons. This also prevents re-absorption of an emitted
photon by the host. (4) Carrier recombination and emission of blue light. (c) Chemical structure of BTDF with eight “soft” dihedrals (red). (d)
Isosurfaces of the LUMO. An isovalue of 0.007 au allows to visualize small fractions of orbital density on the substituents.
el
Figure 2. (a) Distributions of differences in electrostatic energies including polarization, ΔE
(from a neighbor list), conformational energies,
e(h)
cf
ΔE , and reorganization energies, λe(h), for electrons (holes). Mean and variance σ given in eV. Inset shows the distribution of the dihedral angle δ
e(h)
introduced in Figure 1c. (b) Distributions of the logarithm of the transfer integral J for electron and hole transport constructed from diabatic states
based on dibenzofuran (DBF) core or using the whole molecule (BTDF). The same neighbor list and morphology of 4096 BTDF molecules was
used for all distributions. Inset: Radial distribution functions g(r) of the centers of mass of the DBF core (DBF) and of phenyl rings of the
triphenylsilyl groups (PHE).
mobility of the host is required in order to prevent ohmic
losses. Due to its complexity, the effect of the attached
substituents on charge carrier mobility has rarely been
addressed. It is, however, obvious that bulky substituents can
lead to large spatial separations of π-conjugated systems of
neighboring molecules. Since electronic couplings decrease
exponentially with intermolecular separations, one might expect
very poor charge carrier mobility of the host. The aim of this
study is to show that the use of the mesomeric effect can
remedy the situation by delocalizing the frontier orbitals over
the substituents. To do this, we perform a combined
experimental (admittance spectroscopy) and computer simu-
lation study of charge transport in 2,8-bis(triphenylsilyl)-
dibenzofuran (BTDF), a typical electron-conducting host
used in combination with hole-conducting deep-blue emitters.5
The maximal external quantum efficiencies of such OLEDs are
above 17% (see the Supporting Information for the device
charateristics).
13819
dx.doi.org/10.1021/ja305310r | J. Am. Chem. Soc. 2012, 134, 13818−13822

Journal of the American Chemical Society
Article
DBF
The chemical structure of BTDF is shown in Figure 1c. The
two triphenylsilyl groups are attached to a dibenzofuran core,
which lowers its electron-transport level below that of the
emitter. The details of the synthesis are described in the
Supporting Information.
energy (the dibenzofuran core has λ
increases electron-hopping rates.
= 0.27 eV) and hence
e(h)
The distributions of conformational energy differences,
ΔEcef(h), are of Gaussian type with a moderate variance of σ
cf
e(h)
= 0.04 (0.05) eV and are uncorrelated in space, which allows us
to draw them from such a distribution in the larger box of 4096
molecules. Simulations are then additionally averaged over two
realizations of this disorder.
2. RESULTS
2.1. Computer Simulations. To relate charge carrier
mobility to the chemical structure, atomistic molecular
dynamics (MD) is used to simulate material morphologies.
Then the high-temperature limit of Marcus theory15,16 is
employed to evaluate charge-transfer rates between molecules i
and j according to
Electrostatic contributions to site energy differences are
calculated using partial charges for charged and neutral
molecules in the ground state obtained from DFT. Polarization
contributions are taken into account self-consistently using the
Thole model33,34 with a cutoff of 3.5 nm between molecular
centers of mass. This results in polarized electrostatic site
J2
(ΔEij − λij)2
4λijkBT
⎡
⎤
⎥
el
energy differences, ΔE , that are Gaussian distributed with a
ij
π
e(h)
⎢
ω =
exp −
el
e(h)
ij
variance of σ
= 0.12 (0.11) eV when computed from the
ℏ
λijkBT
⎢
⎥
⎦
⎣
(1)
neighbor list. Rather small energetic disorder and weak spatial
correlations are due to small variations of atomic partial charges
(local dipole moments) as well as the total dipole moment of
BTDF of less than 1 D. Note that the attachment of the
substituents does not affect the molecular dipole moment. All
distributions are shown in Figure 2a.
where T is the temperature, Jij is the electronic coupling
element, or transfer integral, and ΔEij is the site energy
difference which has contributions due to an applied electric
field, electrostatics including polarization, ΔEel, and internal
energy differences due to molecular conformations, ΔEcf.
Finally, λij is the reorganization energy which is dominated by
intramolecular contributions due to a small Pekar factor.17
More information about the transport parameters can be found
in the Supporting Information.
The rates and molecular centers of mass are used in kinetic
Monte Carlo simulations to solve the master equation for a
charge drift-diffusing in a box with periodic boundary
conditions in an applied electric field F. The charge carrier
mobility is then determined as μ = ⟨v⟩/F, where ⟨v⟩ is the
averaged projection of the carrier velocity on the direction of
the field.
Simulated mobilities are averaged over two MD snapshots,
ten injection points, and six different spatial directions of the
field. More details are given in the Supporting Information.
Simulations are performed using the VOTCA package.17,18
This approach has been used to calculate mobility in columnar
discotic mesophases,19−24 amorphous systems,25−28 self-
assembled monoloayers,29 and conjugated polymers.30,31
An amorphous morphology of 4096 BTDF molecules is
obtained by first annealing the system at 700 K, well above the
glass transition temperature, Tg = 380 K, followed by fast
quenching to room temperature. The final length of the cubic
box is L = 16 nm. To determine intermolecular charge-hopping
rates in this morphology, a neighbor list based on the closest
approach of centers of mass between phenyl rings or
dibenzofuran cores is constructed using a cutoff of 0.7 nm.
The parameters entering the rate expression eq 1 are then
calculated for each molecular pair from the neighbor list.
Since BTDF has soft degrees of freedom, such as dihedrals δ
in Figure 1c, molecules in the amorphous phase have different
conformations. The distribution of this dihedral angle is shown
in the inset of Figure 2a. These conformations are frozen on the
time scale of charge transport (see the Supporting Information
The remaining ingredient entering the rate expression, eq 1,
is the transfer integral J, which relies on the definition of
diabatic states of a pair of molecules. The latter are usually
constructed from representative orbitals of the π-conjugated
parts, since the effect of attached substituents on the diabatic
states is rather small (e.g., in case of alkyl or glycol side
chains20,24,26). Following this approach, the diabatic states are
evaluated by substituting triphenylsilyl by a hydrogen (without
modifying the rest of the morphology). Reorganization energies
of the dibenzofuran core are used, and transfer integrals are
then calculated on DFT level with the PBE functional and a
TZVP basis set using the dimer projection method.35,36 The
distribution of the logarithm of transfer integrals J (see Figure
2b) for pairs of the neighbor list is very broad. This can be
rationalized in terms of morphology, as the transfer integral
depends exponentially on the intermolecular separation. The
distance between the dibenzofuran cores is large due to the
attached bulky substituents, as illustrated by the radial
distribution function for centers of mass of dibenzofuran
cores, shown in the inset of Figure 2b. The onset of this
function is at ca. 0.5 nm and has a peak g(r) > 1 at a separation
larger than 1 nm, which eventually leads to the broad
distribution of J. The small number of high transfer integrals
due to a few close-lying cores is apparently not sufficient for
charge percolation. As a consequence, simulations predict low
mobilities at experimentally relevant electric fields, μe(h) < 4 ×
10−7 (3 × 10−8) cm2/V·s, which would lead to ohmic losses and
poor device performance.
The above assumptions on the nature of the diabatic states
seem logical but are ultimately invalid. Indeed, if the diabatic
states are constructed using the frontier orbitals of the entire
BTDF molecule, the distributions of transfer integrals become
significantly less broad and peak at much larger values, as
shown in Figure 2b. As a result, predicted mobilities are much
higher, μe(h) ≈ 5 × 10−4 (10−5) cm2/V·s, which is in agreement
with experiments performed by admittance spectroscopy (AS).
2.2. Admittance Spectroscopy. AS allows to extract
mobilities in an organic film sandwiched between two
electrodes by applying dc and ac voltages and finding the
transit time of the carriers in the film from a maximum in the
negative differential susceptance −ΔB = ω(C − C0), where C
for details). Reorganization energies λij and internal energy
cf
ij
differences ΔE are therefore computed from potential energy
surfaces of 512 molecules in neutral and charged states making
use of density functional theory (DFT). We find a small
variance in reorganization energies which does not affect the
mobility. Hence, the mean values of λe(h) = 0.19 (0.27) eV for
electrons (holes) are used. Due to delocalization effects,32
attaching the triphenylsilyl groups decreases the reorganization
13820
dx.doi.org/10.1021/ja305310r | J. Am. Chem. Soc. 2012, 134, 13818−13822

Journal of the American Chemical Society
Article
and C0 are the ω-dependent capacitance of the organic layer
and its zero-frequency analogue37,38 (see the Supporting
Information for details of device preparation and mobility
extraction). Both experimental and simulated mobilities for
electrons and holes are shown in Figure 3. The agreement is
beneficial, since it practically does not change the charge
distribution of the conjugated core. One might argue that the
mesomeric effect leads to an additional, conformational
disorder of site and reorganization energies due to soft
molecular degrees of freedom frozen in an amorphous
morphology. This disorder is, however, small compared to
the electrostatic disorder, is uncorrelated in space, and therefore
has a minor effect on charge transport.17
To summarize, we suggest using the mesomeric effect to
adjust the energy levels via side group attachments to
conjugated cores. It substantially improves electronic couplings
in the host by delocalizing the frontier orbitals, reduces
reorganization energy, and does not lead to significant
additional energetic disorder.
ASSOCIATED CONTENT
* Supporting Information
■
S
Details of synthesis, MD simulations, force-field parameters,
charge-transport simulations, and admittance spectroscopy.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
christian.lennartz@basf.com; denis.andrienko@mpip-mainz.
mpg.de
■
Notes
Figure 3. Simulated electron (black) and hole (red) mobility μ as
function of applied field F, with diabatic states based on the whole
BTDF molecule (solid) or the DBF core (dashed), respectively. For
BTDF, error bars computed from a bootstrap analysis are smaller than
the size of the data points. Experimental data obtained from
admittance spectroscopy at room temperature on films of thickness
d are shown for comparison.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was partially supported by the DFG program IRTG
1404, DFG grant SPP 1355, and BMBF grant MESOMERIE.
We are grateful to Mara Jochum, Kostas Daoulas, Carl
Poelking, Pascal Kordt, and Nicolle Langer for critical reading
of the manuscript.
excellent for the relative electron/hole mobilities, and
experimentally measured values are closer to the scenario
where the substituents are incorporated in the diabatic states.
Both simulations and experiments have small Poole−Frenkel
slopes, indicating small energetic disorder. Simulated electron
mobilities are higher due to stronger delocalization leading to λe
REFERENCES
(1) Forrest, S. R. Nature 2004, 428, 911−918.
■
(2) Reineke, S.; Lindner, F.; Schwartz, G.; Seidler, N.; Walzer, K.;
Lussem, B.; Leo, K. Nature 2009, 459, 234−238.
(3) Giebink, N. C.; D’Andrade, B. W.; Weaver, M. S.; Brown, J. J.;
Forrest, S. R. J. Appl. Phys. 2009, 105, 124514.
2
< λh, and Je2 ≫ Jh , while σe ≈ σh.
(4) Thompson, M. E.; Burrows, P. E.; Forrest, S. R. Curr. Opin. Solid
State Mater. Sci. 1999, 4, 369−372.
3. DISCUSSION AND CONCLUSIONS
(5) Erk, P.; Bold, M.; Egen, M.; Fuchs, E.; Gessner, T.; Kahle, K.;
Lennartz, C.; Molt, O.; Nord, S.; Reichelt, H.; Schildknecht, C.;
Johannes, H.-H.; Kowalsky, W. SID Symposium Digest of Technical
Papers 2006, 37, 131−134.
We now discuss the discrepancy between the two approaches,
that is, including/excluding the substituents in the definition of
diabatic states. The reason for the much higher mobility in the
first case is the mesomeric effect, which delocalizes the frontier
orbitals over the silicon atom to the substituents, as shown in
Figure 1d. A Mulliken population analysis39 indicates that a
fraction of only 10 (4)% of LUMO (HOMO) populates the
substituents (see the Supporting Information). Although such
small delocalization can be easily overlooked on a single-
molecule level, the effect on electronic couplings is much more
pronounced since the substituents are in a closer contact than
the cores. This is illustrated in the inset of Figure 2b, where the
radial distribution function for centers of mass of phenyl rings is
shown. Smaller separations boost electronic couplings
exponentially and dramatically increase charge mobility.
(6) 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.
(7) Yersin, H. Top. Curr. Chem. 2004, 241, 1−26.
(8) Adachi, C.; Kwong, R. C.; Djurovich, P.; Adamovich, V.; Baldo,
M. A.; Thompson, M. E.; Forrest, S. R. Appl. Phys. Lett. 2001, 79,
2082.
(9) Holmes, R. J.; D’Andrade, B. W.; Forrest, S. R.; Ren, X.; Li, J.;
Thompson, M. E. Appl. Phys. Lett. 2003, 83, 3818.
(10) Sasabe, H.; Takamatsu, J.-i.; Motoyama, T.; Watanabe, S.;
Wagenblast, G.; Langer, N.; Molt, O.; Fuchs, E.; Lennartz, C.; Kido, J.
Adv. Mater. 2010, 22, 5003−5007.
(11) Hsieh, C.; Wu, F.; Fan, C.; Huang, M.; Lu, K.; Chou, P.; Yang,
Y. O.; Wu, S.; Chen, I.; Chou, S.; Wong, K.; Cheng, C. Chem.Eur. J.
2011, 17, 9180−9187.
(12) Turro, N. J.; Ramamurthy, V.; Scaiano, J. Modern Molecular
Photochemistry of Organic Molecules; Palgrave Macmillan: 2010.
(13) Fuhrmann, T.; Salbeck, J. Adv. Photochem. 2002, 27, 83−166.
The charge-transfer rate also depends on site energy
differences. It is therefore essential to minimize variations of
local dipole moments which lead to spatially correlated
energetic disorder.40,41 Here the mesomeric effect is also
13821
dx.doi.org/10.1021/ja305310r | J. Am. Chem. Soc. 2012, 134, 13818−13822

Journal of the American Chemical Society
Article
̌ ́ ̌
(14) Nic, M., Jirat, J., Kosata, B., Jenkins, A., McNaught, A., Eds.
IUPAC Compendium of Chemical Terminology, 2.1.0; IUPAC: Research
Triagle Park, NC, ed., 2009.
(15) Marcus, R. A. Rev. Mod. Phys. 1993, 65, 599.
(16) Hutchison, G. R.; Ratner, M. A.; Marks, T. J. J. Am. Chem. Soc.
2005, 127, 2339−2350.
(17) Ruhle, V.; Lukyanov, A.; May, F.; Schrader, M.; Vehoff, T.;
̈
Kirkpatrick, J.; Baumeier, B.; Andrienko, D. J. Chem. Theory Comput.
2011, 7, 3335−3345.
(18) Ruhle, V.; Junghans, C.; Lukyanov, A.; Kremer, K.; Andrienko,
̈
D. J. Chem. Theory. Comput. 2009, 5, 3211−3223.
(19) Kirkpatrick, J.; Marcon, V.; Nelson, J.; Kremer, K.; Andrienko,
D. Phys. Rev. Lett. 2007, 98, 227402.
(20) Feng, X.; Marcon, V.; Pisula, W.; Hansen, M. R.; Kirkpatrick, J.;
Grozema, F.; Andrienko, D.; Kremer, K.; Mullen, K. Nat. Mater. 2009,
̈
8, 421−426.
(21) Marcon, V.; Breiby, D. W.; Pisula, W.; Dahl, J.; Kirkpatrick, J.;
Patwardhan, S.; Grozema, F.; Andrienko, D. J. Am. Chem. Soc. 2009,
131, 11426−11432.
(22) Olivier, Y.; Muccioli, L.; Lemaur, V.; Geerts, Y. H.; Zannoni, C.;
Cornil, J. J. Phys. Chem. B 2009, 113, 14102−14111.
(23) Di Donato, E.; Fornari, R. P.; Di Motta, S.; Li, Y.; Wang, Z.;
Negri, F. J. Phys. Chem. B 2010, 114, 5327−5334.
(24) May, F.; Marcon, V.; Hansen, M. R.; Grozema, F.; Andrienko,
D. J. Mater. Chem. 2011, 21, 9538.
́
(25) Kwiatkowski, J. J.; Nelson, J.; Li, H.; Bredas, J. L.; Wenzel, W.;
Lennartz, C. Phys. Chem. Chem. Phys. 2008, 10, 1852.
(26) Nelson, J.; Kwiatkowski, J. J.; Kirkpatrick, J.; Frost, J. M. Acc.
Chem. Res. 2009, 42, 1768−1778.
(27) Nagata, Y.; Lennartz, C. J. Chem. Phys. 2008, 129, 034709.
(28) Lukyanov, A.; Andrienko, D. Phys. Rev. B 2010, 82, 193202.
(29) Vehoff, T.; Baumeier, B.; Troisi, A.; Andrienko, D. J. Am. Chem.
Soc. 2010, 132, 11702−11708.
(30) Ruhle, V.; Kirkpatrick, J.; Andrienko, D. J. Chem. Phys. 2010,
̈
132, 134103.
(31) Vukmirovic,
́
N.; Wang, L.-W. Nano Lett. 2009, 9, 3996−4000.
(32) Bredas, J.-L.; Beljonne, D.; Coropceanu, V.; Cornil, J. Chem. Rev.
́
2004, 104, 4971−5004.
(33) Thole, B. Chem. Phys. 1981, 59, 341−350.
(34) Ponder, J. W.; Wu, C.; Ren, P.; Pande, V. S.; Chodera, J. D.;
Schnieders, M. J.; Haque, I.; Mobley, D. L.; Lambrecht, D. S.; DiStasio,
R. A.; Head-Gordon, M.; Clark, G. N. I.; Johnson, M. E.; Head-
Gordon, T. J. Phys. Chem. B 2010, 114, 2549−2564.
(35) Valeev, E. F.; Coropceanu, V.; da Silva Filho, D. A.; Salman, S.;
Bredas, J.-L. J. Am. Chem. Soc. 2006, 128, 9882−9886.
́
(36) Baumeier, B.; Kirkpatrick, J.; Andrienko, D. Phys. Chem. Chem.
Phys. 2010, 12, 11103.
(37) Blom, P.; Martens, H.; Huiberts, J. Synth. Met. 2001, 121, 1621−
1624.
(38) Tsang, S. W.; So, S. K.; Xu, J. B. J. Appl. Phys. 2006, 99,
013706−013706-7.
(39) Mulliken, R. S. J. Chem. Phys. 1955, 23, 1833.
(40) Novikov, S. V.; Vannikov, A. V. J. Phys. Chem. 1995, 99, 14573−
14576.
(41) Novikov, S. V.; Dunlap, D. H.; Kenkre, V. M.; Parris, P. E.;
Vannikov, A. V. Phys. Rev. Lett. 1998, 81, 4472.
13822
dx.doi.org/10.1021/ja305310r | J. Am. Chem. Soc. 2012, 134, 13818−13822