Rational In Silico Design of an Organic Semiconductor with Improved Electron Mobility

Article abstract of DOI:10.1002/adma.201703505

Organic semiconductors find a wide range of applications, such as in organic light emitting diodes, organic solar cells, and organic field effect transistors. One of their most striking disadvantages in comparison to crystalline inorganic semiconductors is their low charge-carrier mobility, which manifests itself in major device constraints such as limited photoactive layer thicknesses. Trial-and-error attempts to increase charge-carrier mobility are impeded by the complex interplay of the molecular and electronic structure of the material with its morphology. Here, the viability of a multiscale simulation approach to rationally design materials with improved electron mobility is demonstrated. Starting from one of the most widely used electron conducting materials (Alq₃), novel organic semiconductors with tailored electronic properties are designed for which an improvement of the electron mobility by three orders of magnitude is predicted and experimentally confirmed.

Full text of DOI:10.1002/adma.201703505

CommuniCation
Computational Materials Design
www.advmat.de
Rational In Silico Design of an Organic Semiconductor with
Improved Electron Mobility
Pascal Friederich, Verónica Gómez, Christian Sprau, Velimir Meded, Timo Strunk,
Michael Jenne, Andrea Magri, Franz Symalla, Alexander Colsmann,
Mario Ruben,* and Wolfgang Wenzel*
as its synthetic accessibility. Yet, its compa-
−10
Organic semiconductors find a wide range of applications, such as in organic
light emitting diodes, organic solar cells, and organic field effect transis-
tors. One of their most striking disadvantages in comparison to crystalline
inorganic semiconductors is their low charge-carrier mobility, which mani-
fests itself in major device constraints such as limited photoactive layer
thicknesses. Trial-and-error attempts to increase charge-carrier mobility are
impeded by the complex interplay of the molecular and electronic structure of
the material with its morphology. Here, the viability of a multiscale simulation
approach to rationally design materials with improved electron mobility is
demonstrated. Starting from one of the most widely used electron conducting
rably low electron mobility of 1.1 × 10
–
−7
2
−1 −1 [3]
3.7 × 10 cm V s , which is at least
three orders of magnitude lower than the
hole mobility of typical hole transporting
[3]
materials such as α-NPD, have moti-
vated intense but costly and protracted
experimental efforts to develop better ETL
[4,5]
materials.
So far, however, these mate-
rials often show lower electron affinities
[4]
compared to Alq3, which leads to unbal-
anced charge carrier injection and hence
reduced device efficiencies. Development
of better materials, thus, requires system-
atic improvement of their intrinsic charge
carrier mobility under various additional
constraints, such as energy level compat-
ibility with optimized OLED architectures,
materials (Alq ), novel organic semiconductors with tailored electronic
3
properties are designed for which an improvement of the electron mobility by
three orders of magnitude is predicted and experimentally confirmed.
Improvement of the charge carrier mobility of functional mate-
rials in organic light emitting diodes[1] (OLEDs) is directly con-
nected to their efficiency, as higher mobility leads to lower
resistivity, improved better charge balance, and lower power
consumption for a given applied electric field. Although today
a wide range of hole transport materials has become available,
Tris-(8-hydroxyquinoline)aluminum (Alq3)[ has been used
as a reference material for electron transport layers (ETLs) in
OLEDs because of its high electron affinity, its stability as well
processing, and charge balancing requirements. Presently,
some of the most prominent solutions to solve these problems
[6]
comprise a spectrum of doping techniques, which entails the
use of materials mixtures in the device. In order to tackle this
challenge, many materials aspects must be addressed in a scale-
bridging approach that incorporates transport models in amor-
2]
[7,8]
phous organic semiconductors (see Bässler and co-workers,
[9]
[10]
extensions of these models, KMC techniques, and multiscale
[11,12]
modeling approaches
).
In this work, we use a multiscale modeling approach to calcu-
late the electron mobility of new aluminum complexes derived
[
+]
Dr. P. Friederich, Dr. V. Gómez, Dr. V. Meded, Dr. T. Strunk,
[
++]
Dr. A. Magri, Dr. F. Symalla, Prof. M. Ruben, Prof. W. Wenzel
Karlsruhe Institute of Technology
Hermann-von-Helmholtz-Platz 1
6344 Eggenstein-Leopoldshafen, Germany
E-mail: mario.ruben@kit.edu; wolfgang.wenzel@kit.edu
from the well-known Alq . Systematic analysis of the com-
3
pounds as shown in Figure 1a–d enable us to deduce design
rules for enhanced electron mobility under the constraint of
energy level preservation. We apply these rules to develop the
materials shown in Figure 1e with predicted electron mobility
outperforming the electron mobility of Alq3 by up to three
orders of magnitude. We identify and synthesize the most prom-
7
C. Sprau, M. Jenne, Dr. A. Colsmann
Karlsruhe Institute of Technology
Kaiserstraße 12, 76131 Karlsruhe
Prof. M. Ruben
ising candidate Alq MHept and confirm the theoretical findings
2
CNRS-Université de Strasbourg
3, rue du Loess, BP 43, 67034 Strasbourg cedex 2, France
using space charge limited current (SCLC) experiments.
In the multiscale simulation protocol, we first generate a
morphology model of the material based on molecule-specific
2
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adma.201703505.
[13]
force fields, with molecular geometries and partial charges
parameterized by B3-LYP/def2-SV(P) density functional theory
[+]Present address: Palacký University Olomouc, K rˇ ížkovského 8, 771 47
Olomouc, Czech Republic
[++]Present address: MIXER Spa, Via Chiara 6/C, 48012 Villa Prati di
Bagnacavallo, Italy
[14]
(DFT) calculations.
Internal dihedral potentials are param-
eterized using the semiempirical PM6 functional as imple-
[15]
mented in MOPAC. We use a molecular simulation approach
[13]
DOI: 10.1002/adma.201703505
mimicking vapor deposition to generate atomistic thin film
Adv. Mater. 2017, 1703505
1703505 (1 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 1. a–e) The aluminum complexes studied in this work are presented. a,b) Molecules where at least one of the three quinoline ligands of Alq3
was interchanged by ligands with larger or smaller conjugated π-systems, respectively. c) The systematic exchange of quinoline ligands of Alq by
3
Acac ligands. d) Alq X molecules where the interchanged ligand X was chosen to minimize the dipole moment and e) Alq X molecules where X was
2
2
chosen to minimize the electrostatic potential created by the molecule. f) The calculated bulk IP and EA energies for all tested materials. The dashed
horizontal lines indicate the IP/EA gap of the reference material Alq . The materials Alq to AlAcac (see (c)) are marked in red as an illustration of the
3
3
3
systematic increase of the EA (see text). Alq MHept with the largest simulated electron mobility is marked in green. Bulk HOMO and LUMO energies
2
can be found in Figure S2 (Supporting Information).
Adv. Mater. 2017, 1703505
1703505 (2 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
morphologies containing 1000 molecules per system, which are
periodically extended in x- and y-direction. A detailed analysis
of the atomistic systems can be found in Figure S1 of the Sup-
porting Information. Based on these morphologies, we com-
pute distributions of the polaron levels based on bulk ionization
moment alone is insufficient to fully rationalize the mobility
as a function of composition. Several outliers (marked in red
and illustrated in Figure 1d) show a high energy disorder while
having a comparably low molecular dipole moment. While the
interchange of a single ligand reduces the overall molecular
dipole moment in these materials, the ligand fails to screen the
polar core of the molecule causing substantial intermolecular
electrostatic interaction, which results in high energy disorder.
This is illustrated by the analysis of the mean electrostatic poten-
tial in close to medium proximity to the molecule. As shown in
Figure 2b, energy disorder σ and electron mobility μ are both
strongly correlated with the mean interaction potential. The
strongest correlation with a correlation coefficient of r = 0.80
is achieved by averaging the electrostatic potential of the mol-
ecule on its accessible surface areas with distances between 3.2
and 4.6 Å to the atom positions. We note that the mean electro-
static potential can be fine-tuned by adjusting the interchanged
ligand, while keeping the electron affinity essentially constant.
A decrease of the mean electrostatic potential diminishes the
energy disorder of a given material and thus boosts the corre-
sponding charge carrier mobility orders of magnitude. This can
be formulated into a second design rule: high charge mobility
requires molecules with a low mean electrostatic potential.
Using the second design rule, we systematically varied the
potentials (IP) and electron affinities (EA) using the Quantum
Patch method,[
16,17]
which accounts for polarization effects due to
additional electrons and holes self-consistently at the DFT level.
Electronic couplings are calculated using the Löwdin orthogo-
nalization[ in electrostatically self-consistent, neutral systems.
From the latter, the dipole moments of molecules embedded
in the morphology are extracted. Reorganization energies are
18]
[19]
estimated using Nelsens four-point-procedure. The zero-field
electron mobility is then computed for each material in the gen-
eralized effective medium model approximation.[
20]
In addition to the mobility, the on-site energy of charge car-
riers is a key parameter determining charge carrier injection
and driving voltage of organic semiconductor devices. For ETL,
the electron affinity has to be closely aligned with the work
function of the electron injecting electrode as any significant
energetic misalignment would lead to either large energy bar-
riers or substantial energy losses. Calculated IPs/EAs as well
as highest occupied/lowest unoccupied molecular orbital
(HOMO/LUMO) energies of amorphous thin-films are shown
in Figure 1f. Polarization effects due to the local and global
environment modify IP and EA levels and, to a smaller extent,
even HOMO/LUMO energies. In order to accurately determine
these levels and their distribution, a detailed quantum-mechan-
ical treatment of the polarization in the bulk is required.
We observe a relation between the size of the largest conju-
gated π-system of a molecule and its energy levels (see Figure 1f
and Figure S2 as well as the discussion in the Supporting
Information). All materials having at least one quinoline-ligand
as largest conjugated electronic system show comparable IPs
and LUMO levels as the reference molecule, whose IP/EA gap
is indicated as dashed lines. Observed stronger delocalization of
molecular orbitals on, e.g., the Op ligand compared to quino-
line leads to larger electron affinities and lower LUMO energies.
The opposite effect can be observed in the very small conju-
gated system of, e.g., the Acac ligand, leading to the low electron
third ligand of Alq to decrease the mean interaction potential
3
from 140 mV for Alq to 67 mV for Alqm MHept (Figure 2b).
3
2
The resulting compounds are displayed in Figure 1e. We
have computed the electron mobility of all of these molecules
−
4
2
−1 −1
using the same approach and find μ = 1.4 × 10 cm V
s
for the most promising candidate (Alq MHept) (see Figure 2c).
2
The additional methyl group at the central carbon atom of the
interchanged ligand decreases the dipole moment while the
elongation of the alkyl chain from five to seven carbon atoms
improves the shielding of the molecular core. The low electron
affinity of AlAcac disobeys the first design rule, which is why
3
we discard it as an improved ETL candidate.
We have confirmed the theoretical prediction by synthesis of
Alq MHept followed by preparation of spin-coated devices (see
2
Figure 3a and Figure S3 in the Supporting Information) and
SCLC measurements (see Figures 2c and 3b and Figure S4 in
the Supporting Information). The electron mobility μ is derived
affinity of AlAcac and a shallow LUMO level. This behavior is
3
2
systematically analyzed upon the stepwise interchange of quino-
line by Acac (Alq Alq Acac, AlqAcac and AlAcac , indicated in
from the quadratic regime (I ∼U ) of the SCLC measurements
using Equation (1)
3,
2
2,
3
red in Figure 1f). Upon interchange of a single quinoline ligand,
the EA shift is smaller than 0.2 eV, whereas the step from
2
9
U
d
I = Aε
0
εµ
(1)
3
AlqAcac to AlAcac decreases the electron affinity by more than
2
3
8
0.5 eV. As a first design rule for new organic semiconductors,
2
we want to keep IP and EA levels constant by exploiting charge
localization on the largest π-conjugated molecular subsystems.
IP and EA preserving single ligand substitution can, thus, be
used to improve the charge carrier mobility by specific design of
the interchanged ligand.
where I is the measured current, A = 0.035 cm is the device
area, ε is the vacuum permittivity, ε ≈ 3 is the relative per-
0
[
21]
mittivity, U is the applied voltage, and d is the layer thick-
ness obtained from surface profilometer measurements
(d = 150 nm for Alq , d = 185 nm for Alq Acac, and d = 250 nm
3
2
The performance of a device intrinsically depends on the
charge carrier mobility, which exponentially depends on the
for Alq MHept). Equation (1) yields an electron mobility of
2
−
8
2
−1 −1
−5
2
−1 −1
μ = 6.5 × 10 cm V
s
for Alq , μ = 1.1 × 10 cm V
s
3
[10,20]
−4
2
−1 −1
square of the energy disorder of the constituting materials.
for Alq Acac, and μ = 3.3 × 10 cm V
s
for Alq MHept.
2
2
It is generally accepted that apart from internal distortions of
the molecules in the amorphous phase, one of the main sources
These experimental values confirm the theoretical predictions
−8
2
−1 −1
−6
2
−1 −1
(μ = 7.4·10 cm V
Alq Acac, and μ = 1.4 × 10 cm V
show an improvement of almost four orders of magnitude
s
for Alq , μ = 6.2 × 10 cm V
s
for
3
[7,12]
−4
2
−1 −1
of energy disorder are the molecular dipole moments.
As Figure 2a demonstrates, consideration of the dipole
s
for Alq MHept) and
2
2
Adv. Mater. 2017, 1703505
1703505 (3 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 2. a,b) The energy disorder σ and electron mobility μ of the compounds, compared to their dipole moments and their mean electrostatic
potentials. While comparison of dipole moments with mobility and energy disorder (see (a)) leads to a group of outlies marked in red, the mean
electrostatic potential (see (b)) correlates better with mobility and energy disorder. c) The calculated electron mobilities of all studied materials. The
reference material Alq is highlighted in red, Alq MHept as a promising high electron mobility material is marked in green. Its calculated (green bar)
3
2
and measured (yellow bar) electron mobility is more than 3 orders of magnitude higher than that of Alq . Due to its low electron affinity, AlAcac is
3
3
discarded as an ETL candidate. A detailed analysis of the charge carrier mobility is shown in Figure S3 (Supporting Information).
between Alq and Alq MHept. This result demonstrates that
going from λ = 523 nm for Alq to λ = 511 nm for Alq Acac,
3 2
3
2
detailed understanding of structure-property relations of mate-
rials enables targeted in silico molecule design. To demonstrate
the general applicability of the approach to other classes of
materials, the same design principles were applied to a state-
of-the-art electron conducting material (P4PyMPM) with an
electron mobility about three orders of magnitude higher than
and λ = 512 nm for Alq MHept.
2
To confirm the theoretically predicted stability of the IP of
Alq , Alq Acac, and Alq MHept, with calculated IP of 5.34, 5.46,
3
2
2
and 5.31 eV, respectively, we performed photoelectron yield
spectroscopy in air measurements (PESA). We find an ioniza-
tion potential of IP = 5.85 eV of Alq and ionization potentials
3
that of Alq , leading to a similar boost in electron mobility as
of IP = 5.88 eV and IP = 5.89 eV for Alq Acac and Alq MHept,
3
2
2
Alq2MHept. The theoretical results are shown in Figure S6
respectively. The IP differences between the materials are
within the range of their energy disorder (up to σ = 0.19 eV),
while deviations between predicted and measured absolute IPs
(Supporting Information).
In order to substitute Alq by a higher mobility material in
3
[
22]
ETLs, there should be little change in the energy levels to main-
tain optimal charge injection. The energy level shifts can be
directly observed in, e.g., optical measurements or photoelec-
tron spectroscopy. Figure 3c–e shows experimental absorption
and photoluminescence spectra of Alq , Alq Acac, and Alq M-
can be attributed to well-known DFT limitations. Given the
stability of the IP and the band gap, we infer stability of the EA.
In conclusion, we demonstrate the predictive power of an
in silico multiscale model for key parameters of amorphous
organic semiconductors, including charge carrier mobility and
energy levels, thereby advancing an emerging trend for vir-
3
2
2
Hept. The absorption maximum of Alq3 is at a wavelength
[
23]
of λ = 390 nm while Alq Acac and Alq MHept show maxima
tual materials design. Starting from one of the most widely
used ETL materials, our approach allows a systematic improve-
ment of the electron mobility of organic materials, poten-
tially boosting power efficiency by decreasing recombination
2
2
at λ = 384 nm and λ = 386 nm, respectively (see Figure 3d).
This confirms the theoretical findings, where we also observe a
slightly blue-shifted IP-EA gap upon substitution of quinoline
with smaller π-conjugated ligands. Similar, but slightly larger,
blue-shifts can be observed in the PL peaks shown in Figure 3e,
[
24]
and driving voltage of contemporary OLEDs. We show that
energy disorder and charge mobility strongly depend on the
Adv. Mater. 2017, 1703505
1703505 (4 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Figure 3. a) Spin-coat layers of Alq and Alq MHept as well as the device architecture used for SCLC measurements. b) The current voltage character-
3
2
istics of Alq , Alq Acac, and Alq MHept. A transition from the linear to the quadratic SCLC regime is visible. The vertical dotted lines indicate the range
3
2
2
2
of data points used for extracting the charge carrier mobility. Additional dotted lines with I ∼ U and I ∼ U are shown to guide the eye. The dashed/
dashed-dotted lines are fits to the experimental data in the linear/quadratic regime. c) Absorbance (solid lines) and photoluminescence (dashed line)
spectra of Alq , Alq Acac, and Alq MHept are depicted. Absorbance peaks in the λ = 380–390 nm range correspond to the quinolone ligands and the
absorbance peak in the λ = 300–320 nm range correspond to the Acac/MHept ligands. Photoluminescence maxima are found in the λ = 510–530 nm
range. d,e) The normalized luminescence and absorbance maxima.
3
2
2
DFT-optimized molecular conformations and partial charges (B3-LYP/
near-field electrostatic potential of a molecule, which is not fully
captured by the dipole moment. We furthermore demonstrate
def2-SV(P)). Energy disorder and HOMO/LUMO levels as well as IPs
[
16]
and EA were calculated using the Quantum Patch method.
a computational protocol to control the energy levels of the
molecule. Using these design criteria, we develop a new elec-
tron conducting material with an electron mobility four orders
of magnitude higher than that of the reference material. As
our observations apply to amorphous organic semiconductors
in general, the design rules established in this work may be
helpful to design new materials with improved charge carrier
mobility relevant to many applications of organic electronics.
1
13
Characterization: H and C NMR spectra were recorded using
a Bruker FT-NMR Avance III spectrometer with working frequency
5
00 MHz. The residual solvent peak was used as an internal reference.
−
1
Infrared spectra were recorded in the 4000–400 cm range with a
PerkinElmer Spectrum 6× FT-IR spectrophotometer. Elemental analyses
of C, H, and N were carried out on a Vario Micro Cube. Electrospray
ionization (ESI)-time-of-flight (TOF) mass spectrometric analytical
data was acquired on a Bruker microOTOF-Q II equipped with a
nanospray source. Single crystal X-ray diffraction measurements were
made at 180.15 K on a STOE IPDS2T diffractometer with graphite
monochromated Mo Kα radiation (λ = 0.71073 Å). The structure was
solved with the SHELXS program using Patterson Method and refined
with the SHELXL program[ using Least Squares minimization. CCDC-
1535127 contains the crystallographic data for this paper.
Experimental Section
28]
Simulations: DFT calculations were carried using the Turbomole
[
14]
[25]
Package. All calculations were performed using the hybrid B3-LYP
functional. Reorganization energies were calculated using the def2-
Synthesis: Al(CH ) , Al(NO ) , 8-hydroxyquinoline, 2,4-pentanedione,
3
3
3 3
piperidine, and 4-methyl-3,5-heptanedione were purchased and used
without further purification. Toluene was distilled from sodium beads
under argon.
[
26]
TZVP basis-set while for energy levels, energy disorder, and electronic
couplings, the def2-SV(P) basis-set was used. Atomistically resolved
morphologies were generated using the Metropolis Monte Carlo
[27]
Alq : It was synthesized following the procedure reported in the
3
[13]
[29]
based simulated annealing method DEPOSIT. This method required
literature.
Adv. Mater. 2017, 1703505
1703505 (5 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
Alq Acac: It was synthesized slightly modifying a procedure reported
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
2
[
30]
in the literature:
in 20 mL ethanol and heated to 60 °C. A solution of 2,4-pentanedione
0.31 mL, 3 mmol) and piperidine (0.29 mL, 3 mmol) in 10 mL ethanol
was added dropwise into the above solution and the reaction was stirred
for 30 min. Then, a second solution of 8-hydroxyquinoline (0.870 g, Supporting Information
Al(NO ) ∙9H O (1.125 g, 3 mmol) was dissolved
3 3 2
(
6
mmol) and piperidine (0.59 mL, 6 mmol) in 10 mL ethanol was
Supporting Information is available from the Wiley Online Library or
from the author.
added dropwise and the mixture was stirred for 4 h. The product
was collected by filtration and washed with ethanol and diethyl ether.
Yield 1.01 g (81%).
Alq MHept: Al(CH ) (2 m toluene, 0.48 mL, 1 mmol) was added
2
3 3
to a 100 mL three-necked flask with 40 mL freshly distilled toluene at
room temperature under argon. 4-Methyl-3,5-heptanedione (0.15 mL,
Acknowledgements
1
mmol) in 10 mL freshly distilled toluene was added dropwise. The
solution was stirred for 30 min. Then, 8-hydroxyquinoline (0.290 g,
mmol) in 10 mL freshly distilled toluene was added dropwise and the
The authors acknowledge funding by the Helmholtz program “Science
and Technology of Nanosystems” (STN), the EXTMOS EU project
(grant number: 646176), the Transregio “3Met” Project C5 and the Carl-
Zeiss Foundation for funding the project “Multiskalen Modellierung
elektronischer Eigenschaften von Materialien in der organischen
Elektronik”. The calculations were performed on the computational
resources bwUniCluster and ForHLR Phase I funded by the Ministry of
Science, Research and the Arts Baden-Württemberg and DFG (“Deutsche
Forschungsgemeinschaft”). UV–vis–NIR absorption spectrophotometry
and PESA were made available by the Federal Ministry of Education and
Research (BMBF) under contract 03EK3504 (project TAURUS).
2
mixture was stirred overnight. The solid, which is Alq , was removed by
filtration and the solution was concentrated under reduced pressure.
Successive recrystallizations of the crude from diethyl ether lead to
the desired product. Yield: 0.050 g (11%). H-NMR (500 MHz, CD Cl ,
2
3
1
2
2
5 °C, δ (ppm)): 8.59 (d, J = 4.2 Hz, 2H), 8.17 (d, J = 8.1 Hz 2H),
.49 (t, J = 8.0 Hz 2H), 7.28 (dd, J = 4.2 and 8.1 Hz, 2H), 7.03 (2xd,
7
J = 8.0 Hz, 4H), 2.34 (q, J = 7. Hz 4H), 1.88 (s, 3H), 0.78 (t, J = 7.0 Hz,
6
1
1
3
H); C-NMR (500 MHz, CD Cl , 25 °C, δ (ppm)): 194.5, 190.2, 159.5,
2 2
44.4, 139.0, 133.2, 131.0, 129.5, 121.9, 111.7, 111.6, 31.4, 13.4, 9.3; IR
−
1
(KBr pellet, cm ): 3446m, 1589s, 1579s, 1499vs, 1472vs, 1383s, 1286s,
1
116s, 1029m, 826s,752s, 650m, 547m; elemental analysis (calcd, found
−
1
for C H AlN O , MW = 456.48 g mol ): C (68.41, 68.2), H (5.52, 5.5), Conflict of Interest
2
6
25
2
4
N (6.14, 6.0); ESI-mass spectrocopy (MS) (CH Cl , m/z, rel. intensity,
2
+
2
The authors declare no conflict of interest.
assigned structure): 457.1617 (100%, [M+H] , calcd. = 457.1639),
3
+
12.1329 (68.30%, Alq(MHept) , calcd. = 312.1175).
Mobility Measurements: For SCLC measurements, electron-
Keywords
only devices were built according to the architecture depicted in
Figure 3a (glass substrate/indium tin oxide (ITO) (125 nm)/ZnO +
polyethyleneimine, PEI) (15 nm)/AlqX / LiF (1 nm)/Al (200 nm), with
charge mobility, computational material design, multiscale modeling,
organic electronics, organic semiconductors
−
1
AlqX = Alq or Alq MHept. ITO coated glass substrates (R ≈ 15 Ω sq )
3
2
sq
were structured in hydrochloric acid and sequentially cleaned
in acetone and isopropanol in an ultrasonic bath (10 min). ZnO
nanoparticles (Nanograde N-10) were spin cast (4000 rpm, 30 s) from
isopropanol dispersion (0.55 wt%) in nitrogen atmosphere, followed by
thermal annealing (80 °C, 10 min), resulting in a thickness of 15 nm.
For an efficient electron injection, a layer of PEI was spin coated atop
Received: June 23, 2017
Revised: August 28, 2017
Published online:
−
1
(4000 rpm, 30 s) from 2-methoxyethanol solution (3.9 g L ) under
[
1] a) T.-H. Lai, J. W. Lee, J. R. Manders, F. So, Sci. Rep. 2014, 4,
946; b) J. Zou, K. Zhang, J. Li, Y. Zhao, Y. Wang, S. K. R. Pillai,
H. V. Demir, X. Sun, M. B. Chan-Park, Q. Zhang, Sci. Rep. 2015, 5,
nitrogen atmosphere. The substrates were then annealed in ambient
atmosphere (100 °C, 10 min) and rinsed with ethanol (30 s) to remove
excess PEI. The aluminum-quinoline layer was prepared by dissolving
5
−
1
−1
Alq in chloroform (20 g L ), Alq Acac in dichloromethane (20 g L ),
11755.
3
2
−
1
and Alq MHept in dichloromethane (20 g L ), respectively, and spin
[2] C. W. Tang, S. VanSlyke, Appl. Phys. Lett. 1987, 51, 913.
[3] a) S. Naka, H. Okada, H. Onnagawa, Y. Yamaguchi, T. Tsutsui,
Synth. Met. 2000, 111, 331; b) S. Tse, K. Kwok, S. So, Appl. Phys.
Lett. 2006, 89, 262102.
4] a) S. Naka, H. Okada, H. Onnagawa, T. Tsutsui, Appl. Phys.
Lett. 2000, 76, 197; b) P. Kathirgamanathan, S. Surendrakumar,
R. R. Vanga, S. Ravichandran, J. Antipan-Lara, S. Ganeshamurugan,
M. Kumaraverl, G. Paramaswara, V. Arkley, Org. Electron. 2011, 12,
2
casting (800 rpm, 40 s) in ambient atmosphere. The substrates were
−
6
then transferred to a vacuum chamber (10 mbar) to deposit the
counter electrode of LiF (1 nm) and Al (200 nm) by thermal evaporation.
Current density–voltage (J–V) characteristics were recorded using a
source measure unit (Keithley 238) at room temperature under nitrogen
atmosphere. The thickness of the layers was determined using a tactile
stylus profiler (Dektak XT, Bruker).
[
Absorbance, Photoluminescence and Photoelectron Yield: For
measurements of absorbance, photoluminescence, and photoelectron
yield, aluminum-quinoline thin films were spin cast on glass substrates
according to the procedure described above. Absorbance spectra
were measured on an UV–vis–NIR spectrophotometer (Cary5000,
Agilent Technologies). Photoluminescence spectra were recorded on a
fluorescence spectrophotometer (Cary Eclipse, Agilent Technologies),
with a chosen excitation wavelength of 385 nm. The IPs were determined
by photo-electron yield spectroscopy in air (PESA) (AC-2E, Rieken Keiki).
Atomic Force Microscopy (AFM): AFM (Dimension ICON, Bruker)
images were recorded on thin films on glass substrates in tapping mode.
CCDC-1535127 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
6
66.
[
[
5] D. Yokoyama, H. Sasabe, Y. Furukawa, C. Adachi, J. Kido, Adv.
Funct. Mater. 2011, 21, 1375.
6] M. Pfeiffer, K. Leo, X. Zhou, J. Huang, M. Hofmann, A. Werner,
J. Blochwitz-Nimoth, Org. Electron. 2003, 4, 89.
7] H. Bässler, Phys. Status Solidi B 1981, 107, 9.
8] A. Dieckmann, H. Bässler, P. Borsenberger, J. Chem. Phys. 1993, 99,
[
[
8
136.
[9] a) W. F. Pasveer, J. Cottaar, C. Tanase, R. Coehoorn, P. A. Bobbert,
P. W. M. Blom, D. M. De Leeuw, M. A. J. Michels, Phys. Rev. Lett.
2005, 94, 206601; b) J. Cottaar, L. Koster, R. Coehoorn, P. Bobbert,
Phys. Rev. Lett. 2011, 107, 136601; c) R. Coehoorn, W. Pasveer,
Adv. Mater. 2017, 1703505
1703505 (6 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.advmat.de
P. Bobbert, M. Michels, Phys. Rev. B 2005, 72, 155206; d) A. Massé,
R. Coehoorn, P. Bobbert, Phys. Rev. Lett. 2014, 113, 116604;
e) I. Fishchuk, A. Kadashchuk, S. Hoffmann, S. Athanasopoulos,
J. Genoe, H. Bässler, A. Köhler, Phys. Rev. B 2013, 88, 125202.
[20] V. Rodin, F. Symalla, P. Friederich, V. Meded, D. Danilov,
A. Poschlad, G. Nelles, F. von Wrochem, W. Wenzel, Phys. Rev. B
2015, 91, 155203.
[21] R. L. Martin, J. D. Kress, I. H. Campbell, D. L. Smith, Phys. Rev. B
2000, 61, 15804.
[
10] H. Bässler, Phys. Status Solidi B 1993, 175, 15.
[
11] a) J. J. Kwiatkowski, J. Nelson, H. Li, J. L. Bredas, W. Wenzel,
C. Lennartz, Phys. Chem. Chem. Phys. 2008, 10, 1852; b) P. Kordt,
J. J. van der Holst, M. Al Helwi, W. Kowalsky, F. May, A. Badinski,
C. Lennartz, D. Andrienko, Adv. Funct. Mater. 2015, 25, 1955.
12] P. Friederich, V. Meded, A. Poschlad, T. Neumann, V. Rodin,
V. Stehr, F. Symalla, D. Danilov, G. Lüdemann, R. Fink, I. Kondov,
F. von Wrochem, W. Wenzel, Adv. Funct. Mater. 2016, 26, 5757.
13] T. Neumann, D. Danilov, C. Lennartz, W. Wenzel, J. Comput. Chem.
[22] X. Blase, C. Attaccalite, V. Olevano, Phys. Rev. B 2011, 83,
115103.
[23] a) M. Schwarze, W. Tress, B. Beyer, F. Gao, R. Scholz, C. Poelking,
K. Ortstein, A. A. Günther, D. Kasemann, D. Andrienko, Science
2016, 352, 1446; b) R. Gómez-Bombarelli, J. Aguilera-Iparraguirre,
T. D. Hirzel, D. Duvenaud, D. Maclaurin, M. A. Blood-Forsythe,
H. S. Chae, M. Einzinger, D.-G. Ha, T. Wu, Nat. Mater. 2016, 15,
1120.
[24] H. Sasabe, E. Gonmori, T. Chiba, Y.-J. Li, D. Tanaka, S.-J. Su,
T. Takeda, Y.-J. Pu, K.-i. Nakayama, J. Kido, Chem. Mater. 2008, 20,
5951.
[
[
[
2
013, 34, 2716.
14] R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kölmel, Chem. Phys. Lett.
989, 162, 165.
1
[
[
15] J. J. P. Stewart, J. Comput.-Aided Mol. Des. 1990, 4, 1.
16] P. Friederich, F. Symalla, V. Meded, T. Neumann, W. Wenzel,
J. Chem. Theory Comput. 2014, 10, 3720.
[25] A. D. Becke, J. Chem. Phys. 1993, 98, 1372.
[26] A. Schafer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829.
[27] A. Schafer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571.
[28] G. M. Sheldrick, Acta Crystallogr., Sect. C: Struct. Chem. 2015,
71, 3.
[29] H. Li, F. Zhang, Y. Wang, D. Zheng, J. Mater. Sci. Eng. B 2003, 100,
40.
[30] B. Xu, L. Chen, X. Liu, H. Zhou, H. Xu, X. Fang, Y. Wang, Appl. Phys.
Lett. 2008, 92, 103305.
[
[
[
17] P. Friederich, V. Meded, F. Symalla, M. Elstner, W. Wenzel, J. Chem.
Theory Comput. 2015, 11, 560.
18] V. Stehr, J. Pfister, R. F. Fink, B. Engels, C. Deibel, Phys. Rev. B 2011,
8
3, 155208.
19] S. F. Nelsen, S. C. Blackstock, Y. Kim, J. Am. Chem. Soc. 1987, 109,
77.
6
Adv. Mater. 2017, 1703505
1703505 (7 of 7)
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim